Nature outlook stroke by Javier Ortega

OUTLOOK

STROKE

26 June 2014 / Vol 510 / Issue No 7506

OUTLOOK STROKE

Produced with support from:

The race to save the brain

Cover art: Nik Spencer

Editorial Herb Brody, Michelle Grayson, Brian Owens, Anna York Art & Design Wes Fernandes, Mohamed Ashour, Alisdair Macdonald, Andrea Duffy Production Karl Smart, Susan Gray, Ian Pope, Christopher Clough Sponsorship Janice Stevenson, Yvette Smith Marketing Hannah Phipps Project Manager Anastasia Panoutsou Art Director Kelly Buckheit Krause Publisher Richard Hughes Magazine Editor Rosie Mestel Editor-in-Chief Philip Campbell

very year, strokes kill millions of people and leave millions more permanently disabled (page S2). They are terrifying ordeals that usually occur without warning — even though the causes are known (S4 and S12) — and rob people of their independence through impaired speech and movement. Researchers are trying to work out why the blood suddenly stops flowing smoothly to the brain, and how to limit and repair the resulting damage, thereby helping survivors to put their lives back together. Treating stroke is a race against time: starved of oxygen, the brain cells die in the millions every minute. Doctors are therefore experimenting with mobile units that bring the hospital to the patient, drastically reducing the time it takes to deliver drugs that can limit the damage (S5). Once a stroke has left its mark, the focus shifts to repair. One promising approach involves the use of hydrogels — polymers that can deliver treatments directly to the injured brain and stimulate and nurture the growth and differentiation of neural stem cells (S6). Patients face a long road to recovery. But the months or years of physical therapy can be made less gruelling by incorporating the use of robotic helpers and other mechanical devices (S8). For many people, however, physical recovery is the easy part. Strokes often result in depression, and it is becoming clear that this is not just because they are traumatic life events, but because they actually make the brain more susceptible to mood disorders. A topic of some controversy, therefore, is whether every patient should automatically receive antidepressants (S10). We are pleased to acknowledge the financial support of Lundbeck in producing this Outlook. As always, Nature retains sole responsibility for all editorial content. Brian Owens Contributing Editor

Nature Outlooks are sponsored supplements that aim to stimulate interest and debate around a subject of interest to the sponsor, while satisfying the editorial values of Nature and our readers’ expectations. The boundaries of sponsor involvement are clearly delineated in the Nature Outlook Editorial guidelines available at go.nature.com/e4dwzw CITING THE OUTLOOK Cite as a supplement to Nature, for example, Nature Vol. XXX, No. XXXX Suppl., Sxx–Sxx (2014). VISIT THE OUTLOOK ONLINE The Nature Outlook Stroke supplement can be found at http://www. nature.com/nature/outlook/stroke It features all newly commissioned content as well as a selection of relevant previously published material.

CONTENTS S2 STATISTICS

A growing global burden How stroke affects different populations

S4 PERSPECTIVE

Time to tackle blood pressure A simple, but effective, control measure

S5 FIRST RESPONSE

Race against time Mobile stroke units can save lives

S6 DRUG DELIVERY

Brain food Delivery systems that minimize damage

S8 REHABILITATION

Machine recovery Robots aid motor-skills recovery

S10 MENTAL HEALTH

Ups and downs The controversy over depression

S12 PERSPECTIVE

Silent, but preventable, perils Covert strokes major cause of dementia

COLLECTION

S13 The worldwide burden of stroke — a blurred photograph B. Fuentes & E. Díez Tejedor

S15 Capillary pericytes regulate cerebral blood flow in health and disease C. N. Hall et al.

All featured articles will be freely available for 6 months. SUBSCRIPTIONS AND CUSTOMER SERVICES For UK/Europe (excluding Japan): Nature Publishing Group, Subscriptions, Brunel Road, Basingstoke, Hants, RG21 6XS, UK. Tel: +44 (0) 1256 329242. Subscriptions and customer services for Americas — including Canada, Latin America and the Caribbean: Nature Publishing Group, 75 Varick St, 9th floor, New York, NY 10013-1917, USA. Tel: +1 866 363 7860 (US/Canada) or +1 212 726 9223 (outside US/Canada). Japan/China/Korea: Nature Publishing Group — Asia-Pacific, Chiyoda Building 5-6th Floor, 2-37 Ichigaya Tamachi, Shinjuku-ku, Tokyo, 162-0843, Japan. Tel: +81 3 3267 8751. CUSTOMER SERVICES Feedback@nature.com Copyright © 2014 Nature Publishing Group

S21 Social interaction plays a critical role in neurogenesis and recovery after stroke V. R. Venna et al.

S30 Cerebral ischemic stroke: is gender important? C. L. Gibson

S37 A large-sample assessment of possible association between ischaemic stroke and rs12188950 in the PDE4D gene H. Lövkvist et al.

2 6 J U N E 2 0 1 4 | VO L 5 1 0 | NAT U R E | S 1

OUTLOOK STROKE

A GROWING GLOBAL BURDEN

Stroke is a major public-health problem that tends to affect poorer countries more and leaves richer countries with ballooning medical costs. Yet it is often preventable. By Zoë Corbyn. ROOTS OF STROKE A stroke occurs when blood flow to the brain is interrupted, either by a blockage or a burst vessel. Without oxygen-rich blood, the brain cells die and damage starts to show in the parts of the body controlled by those cells.

Frontal lobe Controls movement, personality, behaviour, language, decision-making and short-term memory. A stroke in the right hemisphere affects movement in the left side of the body, and vice versa.

Ischaemic stroke Occurs when an artery feeding the brain becomes blocked by a blood clot or a build-up of fatty deposits. It is the most common type of stroke, but is less likely to be fatal. Risk factors are varied and treatment is available. Blockage

Parietal lobe Controls speech and sensation (touch and pressure).

Haemorrhagic stroke Occurs when a blood vessel bursts in the brain. It is less common than ischaemic stroke but more likely to be fatal. It often results from uncontrolled hypertension, and the only treatment is surgery to stop the bleeding, relieve pressure and bypass the ruptured vessel.

Haemorrhage Occipital lobe Controls vision.

Temporal lobe

Transient ischaemic attack TIA or ‘mini-stroke’ occurs when blood flow to the brain is interrupted only briefly. Most TIAs last less than 5 minutes, after which symptoms disappear. A TIA is often warning of a full-blown stroke.

Controls hearing, speech and long-term memory.

Cerebellum Controls balance and coordination.

Temporary blood clot

Brain stem Controls vital functions such as breathing and heartbeat. Although rare, strokes here are the most life-threatening and can leave people in a vegetative state or with severe impairments.

33 M people are living with the effects of stroke.

OUT OF CONTROL

LIFESTYLE CHOICES

Haemorrhagic stroke presents less of a burden in high-income countries because they tend to have better control over high blood pressure, a dominant risk factor for this type of stroke.

Some risk factors cannot be controlled, such as age, sex and ethnicity. However, ten factors are collectively associated with nearly 90% of the burden, and all can be mitigated.

200

Rate per 100,000 people

Incidence

High-income countries Low- and middle-income countries

High blood pressure Lack of physical activity Obesity

150

High cholesterol Smoking Poor diet

100

Stress and depression

Interventions that reduce blood pressure and smoking, and that promote physical activity and a healthy diet, could substantially reduce the burden of stroke.

Cardiac disorders 50

Mortality

Diabetes High alcohol intake

0 Ischaemic

Percentage of strokes attributable to risk factor

Haemorrhagic

S 2 | NAT U R E | VO L 5 1 0 | 2 6 J U N E 2 0 1 4

STROKE OUTLOOK

DEADLY TRENDS Stroke is responsible for roughly one-tenth of deaths, making it the second most common cause after heart disease worldwide. Death rates are highest in Eastern Europe, Russia and southeast Asia, regions that tend to lack specialized health services for treatment.

Insufficient data <36.7 36.7–51.9 52.0–85.1 85.2–103.0 103.1–136.7 >136.7 Age-adjusted deaths in 2010 per 100,000 people. Data presented by region.

Stroke is the fourth leading cause of death in the United States.

County data show that mortality is concentrated in the south, in what is known as a ‘stroke belt’.

Insufficient data (21) <72.8 (651) 72.8–81.2 (642) 81.3–88.9 (635) 89.0–100.1 (636) >100.1 (640)

Age-adjusted deaths in 2010 per 100,000 people (number of counties)

Although Afghanistan is in a region of moderate mortality, it has the highest in the world, at 264 per 100,000 people.

69% of strokes and 71% of deaths occur in low- and middleincome countries.

15M

Some of the excess stroke risk may be accounted for by high rates of smoking and obesity, but much remains unexplained.

Alaska

AGE CONCERN

FINANCIAL TIMES

Both incidence and mortality are falling in high-income countries. But prevalence is rising, particularly in high-income nations.

High-income countries Low- and middle-income countries

Incidence

200 Mortality

100

1990

2005

2010

200 Treatment cost (US$ billions)

Prevalence per 100,000 people

Rate per 100,000 people

Direct medical costs are expected to rise substantially in the United States because prevalence will increase as the population ages.

800

400

5M people die.

Hawaii

300

people are left permanently disabled.

people have a stroke every year.

600

400

200

1990

2005

2010

150

100

If prevalence rises by 20%, as predicted, then costs are projected to more than double by 2030.

2015

2020

2025

2030

Sources: Out of control: R. V. Krishnamurthi et al. Lancet Glob. Health 1, e259–e281 (2013). Lifestyle choices: M. J. O'Donnell, et al. Lancet 376, 112–123 (2010). Deadly trends and Age concern: V. L. Feigin et al. Lancet 383, 245–255 (2014); World Health Organization; US Centers for Disease Control and Prevention. Financial times: B. Ovbiagele et al. Stroke 44, 2361–2375 (2013).

2 6 J U N E 2 0 1 4 | VO L 5 1 0 | NAT U R E | S 3

NIH

OUTLOOK STROKE

PERSPECTIVE Time to tackle blood pressure

Simply lowering blood pressure would reduce the risk of both stroke and age-related cognitive impairment, says Walter J. Koroshetz.

troke has obvious, permanent, devastating consequences: it affects physical and mental abilities that strike at the very nature of what makes us human. The good news is that we know how to keep strokes from happening. Controlling factors such as high blood pressure, obesity, smoking and sedentary lifestyle is effective in preventing strokes. The number of people dying from strokes in the United States has fallen by as much as 70% since the 1960s (ref. 1) — rivalling some of the advances in controlling infectious disease. The greatest risk factor for stroke is high blood pressure. The data tell us that the likelihood of stroke decreases as blood pressure is lowered2. And by getting to grips with blood pressure, we can prevent not just the death and physical disability that so often result from stroke, but also age-related cognitive impairment and dementia3. Most people who die with a diagnosis of dementia have a mixture of Alzheimer’s and vascular disease4. A fall in stroke rates would therefore be expected to translate into a decreased risk of dementia — and a study last year from the United Kingdom suggests that this is indeed happening5. On the bright side, most strokes can be prevented using medical information that we already have — we know a great deal about how to control blood pressure. We are probably sitting on a set of very effective antidementia therapies, but this message has not made it to the public or to policy-makers. In fact, the blood-pressure front seems to be backsliding. Earlier this year, influential guidelines published in the Journal of the American Medical Association (JAMA) recommended that physicians become less strict about their patients’ blood pressure6: people over 60 years old can have systolic readings of up to 150 mmHg — a significant increase from today’s threshold of 140 mmHg. This is a perilous step. The risk of stroke increases dramatically with each decade after the age of 60 — and stroke rates in this age group decrease with every incremental drop in blood pressure. It is not clear where that effect plateaus. For instance, even when it failed to decrease the risk of major cardiovascular events overall, targeting a systolic blood pressure of 120 mmHg rather than 140 mmHg did lower the incidence of stroke in people with type 2 diabetes7. And in a trial aimed at preventing second strokes, aggressive blood-pressure control — along with reduced lipid intake and other lifestyle modifications — almost halved the incidence of second strokes compared with the less-strict approach taken in another trial in a similar population8. If practitioners adopt the liberal targets, they will almost certainly be inviting an increase in the stroke rate. Perhaps the most effective way to promote brain health — lifestyle modification — is inexpensive but sociologically difficult. Exercise is thought to improve brain health by lowering blood pressure, decreasing obesity and preventing stroke and heart disease. However, using medications to control blood pressure is the true low-hanging fruit. And it is underused. In the United States, 29% of adults have high blood

pressure — but only half have it under control9. That is still a lot better than in China, where the prevalence of high blood pressure is almost identical to that in the United States but only 9.3% of people have their condition under control10. It is no coincidence that stroke is China’s leading cause of death and the risk of dying from stroke there is twice as high as in the United States. Why doesn’t the United States control blood pressure more effectively? This is of particular concern in the African American population, which is especially vulnerable to the tissue damage linked to high blood pressure — such as white-matter stroke, kidney disease and heart disease — at a relatively young age. There are still things that researchers do not know, such as what the optimal blood pressure is. The Systolic Blood Pressure Intervention Trial funded by the US National Institutes of Health is testing the effects of keeping systolic blood pressure readings below 120 mmHg in 9,000 people at risk of heart or kidney disease, and a sub-study called SPRINT-MIND is using magnetic resonance imaging and cognitive testing to find out whether lowering blood pressure is good for the brain for reasons other than preventing stroke. Quality-of-life measures are important in such studies to learn how to balance the risk of stroke against the adverse effects of aggressive blood-pressure lowering, such as lightheadedness and fainting. Some studies, including the JAMA guidelines, treat stroke and falls from dizziness as equally bad. But this is misguided: quality of life can return to a good level even in the case of a hip fracture, but it rarely does after a major stroke. Data amassed from hundreds of thousands of people in trials of drugs to treat high blood pressure show that stroke is exquisitely sensitive to reductions in blood pressure. The trick lies in getting this message out in a way that will reduce the burden of stroke. Most diseases are difficult to prevent because treatment is expensive or risky or because screening tests are either inadequate or too expensive to scale up for a population. But treatment for high blood pressure is not expensive. Screening is not difficult, expensive or invasive; the tool is a simple blood-pressure cuff. To protect the brain, pump it up! ■

MOST

STROKES CAN BE PREVENTED USING MEDICAL INFORMATION THAT WE

ALREADY HAVE.

Walter J. Koroshetz is deputy director of the US National Institute of Neurological Disorders and Stroke in Bethesda, Maryland. e-mail: koroshetzw@ninds.nih.gov 1. Gillespie, C. D., Wigington, C. & Hong, Y. MMWR Surveill. Summ. 62, 157–160 (2013). 2. Lv, J. et al. PLoS Med. 9, e1001293 (2012). 3. Gorelick, P. B. et al. Stroke 42, 2672–2713 (2011). 4. Cholerton, B. et al. J. Alzheimers Dis. 36, 699–709 (2013). 5. Matthews, F. E. et al. Lancet 382, 1405–1412 (2013). 6. James, P. A. et al. J. Am. Med. Assoc. 311, 507–520 (2014). 7. ACCORD Study Group N. Engl. J. Med. 362, 1575–1585 (2010). 8. Derdeyn, C. P. Lancet 383, 333–341 (2014). 9. Egan, B. M., Zhao, Y. & Axon, R. N. J. Am. Med. Assoc. 303, 2043–2050 (2010). 10. Wang, J., Zhang, L., Wang, F., Liu, L. & Wang, H. Am. J. Hypertens. http://dx.doi. org/10.1093/ajh/hpu053 (2014).

S 4 | NAT U R E | VO L 5 1 0 | 2 6 J U N E 2 0 1 4

R. KOOP, UNIV. CLINIC SAARLAND

STROKE OUTLOOK

Mobile stroke units, such as this one run by Saarland University in Germany, contain all the tools needed to diagnose and start treating a stroke.

FIRST RESPO NSE

Race against time Mobile stroke units can save lives by treating people before any damage starts to take hold. BY ED YONG

omewhere in your bloodstream, a clot loosens. It travels to your brain and blocks one of the blood vessels there, depriving the surrounding neurons of oxygen and glucose. Your face suddenly drops on one side, one of your arms weakens and your speech becomes slurred. You are having a stroke. The hourglass turns — your time is running out. Neurologists have a saying: ‘time is brain’. The brain guzzles so much energy, and is so dependent on oxygen and glucose, that it is extremely vulnerable to any shortfall in these substances. For every minute a stroke goes untreated, a patient loses 1.9 million neurons and 14 billion nerve-cell junctions, known as synapses. For every hour without treatment, the brain effectively ages by 3.6 years1. Doctors can use an enzyme called tissue plasminogen activator (tPA) to dissolve a clot, but patients are most likely to avoid disability if they receive the drug within a few hours of their first symptoms. Ideally, they would be treated within the first ‘golden’ hour. This rarely happens: fewer than 5% of patients in the developed world receive tPA, because most arrive at the hospital too late. Neurologist James Grotta from the University of Texas Health Science Center at Houston has a solution: rather than waiting for patients to arrive at hospital, he is bringing the hospital to them. Earlier this year, he unveiled the first

mobile stroke unit in the United States — an ambulance fitted with a computed tomography (CT) scanner and a telemedicine system that sends data from the scanner to the hospital. When an emergency call fits the profile, Grotta’s team rushes to the scene in the ambulance alongside the usual emergency services. If the CT scan reveals a clot, the team injects tPA right away. The scanner also tells the crew whether the stroke is ischaemic, caused by a clot, or haemorrhagic, caused by a burst blood vessel. About 85% of strokes are ischaemic, but the distinction is crucial — tPA can break up a clot, but make things worse for a bleed.

NO TIME FOR DELAY

Mobile units have a long history. In the First World War, Marie Curie fitted 20 vans with X-ray scanners, then learned to drive so that she could take them to the front line herself. In 2003, Klaus Fassbender at Saarland University in Saarbrücken, Germany, resurrected the idea for people having a stroke2. The first ambulance rolled out in 2008, and dealt with more than 300 patients before an accident put it out of commission 18 months ago (a replacement vehicle will start running on 1 July). In 2012, the team showed that the unit could halve the median time from emergency call to treatment, to just 35 minutes3. Grotta was paying attention. “Our door-toneedle times were around 60 minutes, and no matter how hard we tried, they remained there,”

he says. “I read Fassbender’s paper, visited him and thought: we should do that here.” Two years later, and with US$1.1 million in funding from local businesses and philanthropists, the unit saw its first patient on 4 June. A Norwegian team has adopted a similar approach by adding CT scanners to its air ambulances. The fleet of light aircraft and helicopters can reach most of the population within 30 minutes. Mobile units are an expensive and labourintensive approach. In Houston, the ambulance and scanner cost around $600,000, and supplies, drugs and personnel further inflate the costs. But Grotta says that every case of stroke costs the US health-care system around $140,000, largely because of lengthy hospital stays and long-lasting disability. “We’re talking about people who are facing paralysis, but who return to normal if the treatment works,” he says. “The pay-off is immediate and dramatic.” If the mobile unit can deliver tPA to just four patients early enough to avoid long-term consequences, the health-care savings downstream would pay for the initial financial outlay. But the biggest lingering question is whether early treatment actually saves lives — a crucial issue that Fassbender’s trial3 was not designed to address. It is hard to say, says Grotta. Shortening the time to treatment from 3 hours to around 90 minutes is known to improve the chances of complete recovery4, but no one knows what an even smaller delay could accomplish, because so few patients are treated within that window. “That’s what we intend to find out,” he says. Fassbender is optimistic. “Every solution is better than the current situation,” he says. “We can only win.” ■ Ed Yong is a freelance science writer in London. 1. 2. 3. 4.

Saver, J. L. Stroke 37, 263–266 (2006). Fassbender, K. et al. Stroke 34, e44 (2003). Silke, W. et al. Lancet Neurol. 11, 397–404 (2012). Fassbender, K. et al. Lancet Neurol. 12, 585–596 (2013).

2 6 J U N E 2 0 1 4 | VO L 5 1 0 | NAT U R E | S 5

LIAM MALONEY

OUTLOOK STROKE

Molly Shoichet has developed a polymer than can deliver drugs to the brain with the minimum of damage.

DRUG DELIVERY

Brain food The key to stroke recovery is to coax the brain cells to heal without creating more damage in the process. A clever delivery system may just do the trick. BY HANNAH HOAG

olly Shoichet searches her desk for a sandwich-sized zip-up bag. Stowed in the bag is a syringe containing roughly 2 millilitres of a rich-turquoise fluid. She pops the cap off the needle and pushes the plunger gently until a small bead of the material emerges. She tilts the syringe upside down, but the tiny bulb stays put. Almost instantly, the liquid had transformed into a gel — one that Shoichet hopes will be able to deliver drugs directly to the brain. The material in the syringe is a three-dimensional network of polymers called a hydrogel. It is a viscous fluid when flowing down a narrowgauged needle, but becomes a gel once it has been injected into the body. “It’s important that it gels quickly, so that it doesn’t disperse too widely,” says Shoichet, a polymer chemist and biomedical engineer at the University of Toronto in Canada. Shoichet designed the hydrogel to aid the

repair of brain tissue after a stroke deprives the brain of oxygen. The brain has self-renewing cells — called neural stem cells — that can differentiate into multiple cell types1. After a stroke, these cells move towards the damage to repair it2, but as a general rule, not enough cells mature into functioning tissue3, so lasting signs of damage remain. To coax the brain to heal, researchers face two main challenges: how to keep the cells alive long enough for them to mature and how to integrate them into the neural circuitry so that they can repair the tissue and restore function. The hydrogel strategy aims to solve both problems. Suspended within the gel are two cellular growth factors, both neatly encapsulated in polymer shells that control the timing of their release. One factor causes the stem cells to multiply and the other prompts NATURE.COM them to differentiate For more on the potential into the neurons and of hydrogels, visit: glial cells that will go.nature.com/pkyrj1

S 6 | NAT U R E | VO L 5 1 0 | 2 6 J U N E 2 0 1 4

become the new brain tissue. Like a gardener who prepares for planting by mixing the best soil and adding the right amount of fertilizer, Shoichet is creating flexible dressings that feed and cultivate immature brain cells. Shoichet began working on stroke in 2008. Her lab had been collaborating with Charles Tator, a neurosurgeon at the University of Toronto, and Dimpy Gupta, his graduate student, to design a hydrogel that could be used to treat spinal-cord injury4. The material had to be liquid enough to flow through an ultrafine needle, but solid enough to stay in place after being injected into the space between the spinal cord and the thick membrane that surrounds it. The substance she developed contains hyaluronan to keep it fluid inside the syringe, and methylcellulose to allow it to gel once it emerges. Cindi Morshead, a stem-cell biologist at the University of Toronto who was working on another part of the spinal-cord injury project, suggested that they try a similar approach for the stroke-injured brain.

STROKE OUTLOOK In 2006, Morshead had used infusions of two growth factors to coax the brains of stroke-injured rats to heal5. After letting the stroke take hold for 4 days, she used a hollow tube called a cannula to infuse the brain with epidermal growth factor for 7 days and prompt the small number of stem cells in the brain to proliferate and migrate to the site of the injury. She then infused erythropoietin for 7 days to encourage the cells to differentiate into neurons (see ‘Right on target’). The combination worked, increasing the number of neurons at the site of the injury and reducing the size of the injured area.

AVOID HEALTHY TISSUE

The trouble was that the technique required surgery that damaged healthy brain tissue. “It was too invasive,” says Morshead. She wondered whether Shoichet’s hydrogel would eliminate the need for the cannula. Shoichet’s first challenge was to find a way to allow for both sequential and sustained release of the growth factors: one needed to be released 4 days after the stroke, the second 7 days later, and each had to be released over a week. One way to control the delivery is to coat the drugs with a polymer that takes time to break down in the body. This was one of Shoichet’s areas of expertise. She had earned her undergraduate degree at the Massachusetts Institute of Technology in Cambridge, where she says she “got really excited about polymers and their use in medicine and pharmaceutical chemistry”. After completing her PhD in polymer science and engineering at the University of Massachusetts Amherst, she joined a start-up called Cytotherapeutics in Providence, Rhode Island, where she designed polymers to encapsulate cells for the treatment of Parkinson’s disease, diabetes and chronic pain. It was there that her materials expertise fell under the influence of life science. “I was surrounded by really smart biologists,” she says. “Rather than take something off the shelf and see if it works, I could design a material to produce the response they wanted.” For the hydrogel, Shoichet began by wrapping nanometre-sized droplets of each of the two growth factors in polymers. She then added a layer of polysebacic acid to the erythropoietin bundle so that the drug would not be released at the same time as the endothelial growth factor. Then she suspended the beads in the hydrogel6. With the challenge of developing a gel behind them, she moved on to test it in a mouse model of stroke. Shoichet, Morshead and their collaborators fixed a small disk with an opening over a hole in the skull and dripped the hydrogel directly onto the brains of the stroke-injured mice. It worked: the brains had smaller areas of dead tissue, more living neurons in the damaged area and less inflammation compared with mice that received the growth factors through the cannula6. And because the 4-day

R I G H T O N TA R G E T A hydrogel polymer may allow targeted, time-sensitive delivery of brain-repairing drugs. Hydrogel is dripped directly onto brain

Injury site

Epidermal growth factor is released immediately to stimulate regrowth of neurons

Growth-nurturing erythropoietin is coated in polysebacic acid to delay its release Gel contains hyaluronan to keep it liquid and methylcellulose to let it gel once on the brain

Release profile Stroke Day 0

Inject composite Day 4

delay meant that very few cells were still alive, the living neurons the researchers saw in the brains were likely to be new cells rather than repaired ones. Although Shoichet and Morshead are most interested in using the brain’s existing stem cells to repair itself after stroke, they have also experimented with stem-cell transplantation. In unpublished work, they have mixed the hydrogel with neural stem cells from a healthy adult mouse, then injected these directly into the brain of a stroke-injured mouse. The cells lived longer when introduced through the hydrogel than when transplanted on their own, but the results still fall short of Shoichet’s aims. “There aren’t as many neurons as we’d like,” she says. She hopes that the number of neurons can be boosted by allowing the stem cells to differentiate before transplanting them.

YOUNG AT HEART

Shoichet and Morshead remain optimistic about their hydrogel approach. It avoids the complications faced by intravenous administration of stem cells or drugs, an approach that is being tested in early clinical trials. The main problem with intravenous administration is that the blood–brain barrier limits the passage of growth factors into the brain, so higher drug doses are often needed, leading to unwanted side effects. It also avoids the damage to healthy tissue caused by infusing the drugs through a cannula. Shoichet and Morshead are now tweaking the hydrogel’s components to improve its performance. They have added ciclosporin, a drug that promotes the survival of neural precursor cells7. They are also looking for other molecules and strategies that can help to repair the damaged site, such as breaking down the scar of glial tissue that typically forms after a stroke. Another goal is to see whether the hydrogel therapy works in older mammals because “that is the most prominent population that will suffer from these deficits, and the older

Epidermal growth factor Day 11

Erythropoietin Day 18

body doesn’t work as well as the younger one”, says Morshead. Other researchers are also developing biomaterials to repair the stroke-injured brain. The combination of substances such as polymers, growth factors and cytokines (small proteins that act as cell signals) varies among the research groups, as does the method of application. None replicates Shoichet and Morshead’s less-invasive hydrogel approach, but some are following a similar path. For example, a group at the Nagoya City University Graduate School of Medical Sciences in Japan has injected the brains of mice with gelatin microspheres laced with a growth factor. Insulin growth factor 1 increased the number of neurons in a part of the brain known as the subventricular zone; and hepatocyte growth factor increased the number of neurons migrating from this region to the damaged site8. For scientists focused on regenerating nerve tissue, getting beyond the brain’s protective barriers can be a challenge. But Shoichet and Morshead think that they are getting close to finding a way to help the brain heal itself. “Celltransplant strategies are costly and complex,” says Shoichet. “It would be great if we could take advantage of what was already there.” ■ Hannah Hoag is a freelance science writer in Toronto, Canada. 1. Reynolds, B. A. & Weiss, S. Science 255, 1707–1710 (1992). 2. Jin, K. et al. Proc. Natl Acad. Sci. USA 103, 13198–13202 (2006). 3. Abrous, D. N., Koehl, M. & Le Moal, M. Physiol. Rev. 85, 523–569 (2005). 4. Gupta, D., Tator, C. H. & Shoichet, M. S. Biomaterials 27, 2370–2379 (2006). 5. Kolb, B. et al. J. Cereb. Blood Flow Metab. 27, 983–997 (2007). 6. Wang, Y., Cooke, M. J., Sachewsky, N., Morshead, C. M. & Shoichet, M. S. J. Control. Release 172, 1–11 (2013). 7. Caicco, M. J. et al. J. Control. Release 166, 197–202 (2013). 8. Nakaguchi, K. et al. Stem Cells Int. http://dx.doi. org/10.1155/2012/915160 (2012). 2 6 J U N E 2 0 1 4 | VO L 5 1 0 | NAT U R E | S 7

H. I. KREBS, NEWMAN LAB, MIT

OUTLOOK STROKE

The MIT-Manus, developed at the Massachusetts Institute of Technology, helps patients to improve movement in their arm after a stroke.

RE H ABILITATIO N

Machine recovery Interactive devices are helping people who have had a stroke to regain their motor function. B Y M O H E B C O S TA N D I

tienne Burdet leans over a glass table, places his elbow in a spring-loaded support clamped to its edge and uses both hands to cut a virtual tomato into thin slices. The tomato is one of several manipulable objects displayed on a touchscreen fitted with force and motion sensors, motors and lightemitting diodes, all of which are connected by Bluetooth to an Android tablet. Burdet finishes slicing the tomato and moves on to unscrewing the lid of a virtual jar. Burdet, who directs the Human Robotics Group at Imperial College London, is demonstrating a prototype of the iTable, a device that helps people who have had strokes and other neurological injuries to regain use of their hands. Each task draws on reaching, grasping and coordination skills. Sensors detect and analyse the user’s movements, then relay the information back to the tablet1. “Strictly speaking, this isn’t a robot,” Burdet says — it does not have moving parts that are powered by electric motors, or perform tasks automatically. “It’s an interactive object that tells the patient what to do and how to do it.” The

iTable is a ‘haptic’ device — patients interact with it using their sense of touch — but unlike other such devices, the user interface and visual display are one and the same. The patients interact directly with the virtual objects rather than watching them on a separate monitor, “so whatever they learn is directly transferable to the real world”, Burdet says. Rehab devices are in huge demand. Stroke is the world’s leading cause of severe disability. In the United Kingdom alone, more than 150,000 people have a stroke in any one year, and about 1.1 million people are living with the consequences, more than half of whom need help to carry out their everyday activities. And because the world’s population is ageing, the number of people who have strokes will only increase.

RISE OF THE MACHINES

Stroke usually causes paralysis or weakness on one side of the body, but the injuries can often be treated through intensive training that involves repetitive movements of the affected limbs. The process is time consuming and laborious for both patient and therapist. “The more intense the practice, the better the outcome,”

S 8 | NAT U R E | VO L 5 1 0 | 2 6 J U N E 2 0 1 4

says Diane Playford, a consultant neurologist at University College London. Evidence from rats2 suggests that each movement needs to be repeated between 400 and 600 times a day to recover function. “We’re essentially trying to re-educate the brain,” says James Patton, a bioengineer at the Rehabilitation Institute of Chicago in Illinois, “and that means that patients have to do a lot of homework.” In the United Kingdom, people who have a stroke typically spend only a few days in a hospital stroke unit. Most of their recovery happens after discharge: they need to do up to two hours of exercises a day for many months to regain use of their limbs. Many people lack the patience. And a shortage of therapists and caregivers, at least in Europe, exacerbates the situation. Researchers hope that combining robotics with conventional physiotherapy might help people to stick to their daily exercise regimes. Machines are already invaluable in rehabilitation — they measure patients’ progress with great accuracy and are widely used to compare the outcomes and identify the people most likely to benefit from therapies. Some researchers think that machines will transform the way

HOCOMA

STROKE OUTLOOK in which treatment is delivered. As the cost of technology drops, that of manual labour rises; researchers hope that robots will not only lend help to deliver the intensive therapy needed for stroke recovery, but also do it more costeffectively. “It’s a very exciting time,” says Hermano Igo Krebs, a mechanical engineer at the Massachusetts Institute of Technology (MIT) in Cambridge. In 1998, Krebs, along with his colleague Neville Hogan, founded Interactive Motion Technologies to commercialize MIT-Manus, an interactive arm device they had developed. The company has sold about 250 devices to clinics around the world. “I believe that robots will change the rehab landscape within the next five years,” says Krebs. Research interest in rehabilitation devices has grown over the past 20 years. A survey published in January3 lists more than 120 devices for arm rehabilitation alone, ranging from passive braces for the shoulder to complex exoskeletons that actively assist users in performing certain movements. David Reinkensmeyer, an engineer at the University of California, Irvine, has been involved in the development of more than a dozen devices, several of which emerged from the Machines Assisting Recovery from Stroke centre at the Rehabilitation Institute of Chicago, which he co-directs with Patton. “In neurological rehabilitation, it’s critical that the patient exerts as much effort as possible,” he says. “An active robot can actually discourage patients from being active.”

SPRING INTO ACTION

That is the idea behind the ArmeoSpring, an arm exoskeleton that Reinkensmeyer developed with Tariq Rahman, an engineer at the University of Delaware in Newark. The device “has springs that relieve the weight of the paralysed or weak arm”, says Reinkensmeyer. “It’s a mechanically passive device that demands activity from the patient — it won’t move unless the person is active.” The ArmeoSpring has been commercialized by Hocoma in Volketswil, Switzerland, which has sold more than 500 units to hospitals around the world and makes a similar device for the hand, the ManovoSpring. One way to get people moving is to make the task fun and engaging. Reinkensmeyer’s latest invention is the Finger-individuating Grasp Exercise Robot (FINGER), a hand exo skeleton containing two motorized rod-andlever mechanisms, one each for the index and middle fingers. The system, co-developed with Eric Wolbrecht of the University of Idaho in Moscow, helps users to move their fingers individually in a natural curling motion. The device captures their movements and monitors their effort and progress while they play a computer game similar to Guitar Hero, in which players use a guitar-shaped controller to simulate playing the real instrument4. Other researchers are developing low-cost,

home-based systems that use virtual-reality and gaming technology to make exercise routines more enjoyable, and relay information to therapists who can then monitor patients’ progress remotely. The Limbs Alive project, founded in 2010 by neuroscientist Janet Eyre and occupational therapist Janice Pearse in partnership with Newcastle University, UK, brought together game developers and mathematicians to produce Circus Challenge, a set of games that can be played

36-week course of robot-assisted therapy cost several thousand dollars less than the same duration of conventional therapy, even when the cost of the devices was taken into account. To answer the economics question more definitively, UK researchers have begun a fouryear, multi-centre trial to evaluate the costeffectiveness of robot-assisted rehabilitation in the National Health Service. The study is led by neurologist Helen Rodgers of Newcastle University, and is the largest of its kind, with plans to enrol up to 800 patients. Another large trial, led by Olivier Rémy-Néris, a psychiatrist at the University of Western Brittany in Brest, is evaluating the ArmeoSpring at 21 hospitals in France.

MONEY MAGNETS

The ManovoSpring trains users in grasping skills.

on televisions, desktop computers, laptops and tablets. Players use motion-sensitive controllers to take charge of various circus performers, using increasingly complex arm and hand movements as they progress through the game. The controllers that move the on-screen avatars capture the speed, accuracy and smoothness of players’ movements, allowing therapists to perform validated clinical assessments, which would otherwise take “I believe that a whole day to do, in robots will less than 20 minutes of game play, says change the rehab landscape Eyre. Few devices have within the next been tested in randfive years.” omized clinical trials. Of those that are commercially available, the MIT-Manus has been assessed the most thoroughly, in several large, multi-centre trials and in more than 800 patients. The results have shown that robot-assisted therapy works about as well as conventional therapy, but can modestly improve outcomes when delivered at the same time5–7. On the basis of such results, the American Stroke Association now endorses the use of robotics for stroke rehabilitation. Playford, who runs the rehabilitation unit at the National Hospital for Neurology and Neurosurgery in London, has come up with similar preliminary findings. She is testing the HapticKnob, a simple robot developed in Burdet’s lab that trains patients in the movements needed to turn a doorknob8. “If you use robots as an adjunct to usual practice you get slightly better results,” Playford says, “but if you use them instead there’s no difference.” The real improvements may lie not in effectiveness but in economic efficiency. A 2011 study 9 of the MIT-Manus showed that a

At present, cost remains the biggest barrier to the widespread use of rehab robots. “Most commercially available robots are still very expensive,” says Playford. “They take time to set up and need an expert, so at this stage it probably isn’t cheaper to have a robot in most settings.” Ultimately, Playford says, cost savings will come only when portable robots are available that are inexpensive to make and can be delivered to people’s homes for a course of therapy. In line with that thinking, Burdet’s team is now focusing on developing cheap, compact machines. Most of the creations coming out of his lab are either passive, sensor-based devices such as the iTable, or simple robots such as the HapticKnob. Krebs, for his part, sees a time when both large and small robots are an integral part of the therapeutic mix. “Large machines are more suitable for clinics, which could have different machines to train different movements,” he says. “Then you might move to simpler devices that go to the home, based on the individual’s needs.” Krebs predicts that home-based machines will soon become commonplace. He and his team are therefore developing devices for use in what they call social spaces, or virtual environments in which rehab exercises are built into multiplayer games. In the not-too-distant future, then, patients may motivate each other to do their daily exercises by virtually beating each other up, even though they are thousands of kilometres apart. ■ Moheb Costandi is a freelance science writer in London. 1. Jarrassé, N. et al. IEEE Int. Conf. Rehabil. Robot. http://dx.doi.org/10.1109/ICORR.2013.6650379 (2013). 2. Kleim, J. A., Barbay, S. & Nudo, R. J. J. Neurophysiol. 80, 3321–3325 (1998). 3. Maciejasz, P., Eschweiler, J., Gerlach-Hahn, K., Jansen-Troy, A. & Leonhardt, S. J. Neuroeng. Rehabil. 11, 3 (2014). 4. Taheri, H. et al. J. Neuroeng. Rehabil. 11, 10 (2014). 5. Lo, A. C. et al. N. Engl. J. Med. 362, 1772–1783 (2010). 6. Burgar, C. G. et al. J. Rehabil. Res. Dev. 48, 445–458 (2011). 7. Volpe, B. T. et al. Neurology 54, 1938–1944 (2000). 8. Lambercy, O. et al. J. Neuroeng. Rehabil. 8, 63 (2011). 9. Wagner, T. H. et al. Stroke 42, 2630–2632 (2011). 2 6 J U N E 2 0 1 4 | VO L 5 1 0 | NAT U R E | S 9

Physical rehabilation was the easy part of Lynn Betts’s recovery; her depression was harder to beat.

ME NTA L HEALTH

Ups and downs Strokes can shatter a person’s identity and make it difficult to find the light. But there are ways to help patients cope. B Y S U J ATA G U P TA

er troubles started with a headache. Lynn Betts powered her way through and stayed up late to complete her paperwork for her job as a speech and lan guage pathologist. Early the next morning, she was awakened by something touching her face, but when she went to brush it off, she realized she couldn’t move her left arm. She

woke up her husband and told him: “I think I’m having a stroke.” It was April 2009, and Betts had indeed had a stroke. She was rushed by ambulance to the University of Virginia hospital in Charlottes ville, where she spent two weeks before being moved back to her home hospital in Lynch burg for another six weeks. There, she began work to regain the basic physical skills she had lost — including walking and sitting upright.

S 1 0 | NAT U R E | VO L 5 1 0 | 2 6 J U N E 2 0 1 4

At night, Betts says that she could hear other patients crying. She sometimes felt like cry ing, too. “Why do I feel down in the dumps?” she would chide herself. “I’ve got my second chance to live. I’ve got my loving family. My church, my friends.” But when Betts’s minister came to visit her, she confided to him: “It feels like my soul died with my stroke.” The sadness intensified when Betts got home. Just 61 years old, she found herself unable to continue working at the job she loved, and reluctantly decided to retire early. She required constant care. She could no longer drive, and missed her independence. At her next doctor’s visit — four months after the stroke — she confessed that she had been feeling blue. The doctor suspected that the stroke had damaged parts of her brain that control emotions, and prescribed an anti depressant. Gradually some of her old exuber ance returned, albeit with lingering physical disabilities. “I used to run up to people and give them a hug,” she says. “Now I hobble to people and give them a one-armed hug.” Betts’s situation is not uncommon. Almost one-third of people who have a stroke in the United States experience major depression (compared with just under 7% of people in the general population). People younger than 65, women and those with a history of depression are at greater risk. Left unchecked, depression makes it harder for patients to face the rigours of rehabilitation or to stay con nected to loved ones. And a person experi encing depression is three times more likely to die in the ten years after their stroke than is a patient without depression1. The good news is that if people are given antidepressants in the six months after their stroke, their chances of survival improve2. What is more, antidepressants can help patients undergoing physiotherapy to regain motor control, even if they aren’t depressed3. Could it be that the drugs prevent the onset of depression, which dampens motivation in rehab? Or do they somehow help the brain to rewire? Some researchers now advocate an aggres sive approach to managing depression after stroke. Evidence is mounting that “every patient who doesn’t have a contraindication to the use of an antidepressant should receive an antidepressant after stroke”, says Robert Robinson, a psychiatrist at the University of Iowa in Iowa City.

A COMMON PROBLEM

The link between stroke and depression was first proposed a century ago. But the gloomy mood was generally believed to be the logi cal consequence of suffering a debilitating injury. The attitude, says Alan Carson, a neuropsychiatrist at the University of Edin burgh, UK, was largely: “Why wouldn’t stroke patients be depressed? They just had a stroke!” But in the mid-1970s, Robinson, fresh out

BILL BETTS

OUTLOOK STROKE

STROKE OUTLOOK of medical school, began inducing strokes in rats by severing one of the arteries that sup ply blood to the brain. He then measured how the procedure affected the concentration of catecholamines, hormones responsible for the body’s fight-or-flight response. He found that the stroke rats initially had lower levels of these hormones than the control rats4. The stroke rats were also more sluggish and less inclined to run maniacally on exercise wheels. That work soon shifted from rats to humans. In 1977, researchers at Johns Hop kins University in Baltimore, Maryland, compared the mental health of two groups of people who no longer had fully functional legs as a result of either stroke or orthopae dic injuries, such as hip fractures or arthritis. The researchers found that almost half of those “You need to who had had a stroke were continually depressed compared with re-evaluate just one-fifth of those with orthopaedic injuries5. The for these study triggered a marked problems.” change in how neurolo gists viewed psychological complications after a brain injury6 and research into depres sion following a stroke “went from an intel lectual backwater too dull for neurologists to even bother seeing”, says Carson, to being a hot topic. A few years ago, for instance, Nada El Husseini, a neurologist at Duke University in Durham, North Carolina, compared rates of depression in two types of stroke: ‘classic’ strokes, which tend to result in long-lasting functional impairment; and transient ischae mic attacks, which have similar symptoms but are fleeting and do not even show up on a brain scans. A year on, El Husseini found, those with transient ischaemic attacks were just as likely to be depressed as those with full-blown strokes 7. Apparently, there is something about how strokes affect the brain that triggers depression.

STIGMA ISSUES

Despite such research, the attitude persists that people who have had a stroke should be able to simply power through their depression. When Betts was recuperating in the Lynchburg hos pital, she was automatically allocated speech, occupational and physical therapists. But she had to ask to see a mental-health therapist. And although depression after stroke is starting to garner more attention, identify ing and treating these patients remains prob lematic. In 2012, the Joint Commission, a non-profit group that accredits and certifies health-care programmes across the United States, mandated that all comprehensive stroke centres screen patients for depression before they are discharged. Yet when Stephanie Casal, a nurse at the Stanford Stroke Center in Califor nia, evaluated how well current screening

methods worked, she was taken aback. Of the 423 patients admitted to hospital for a stroke between September 2012 and June 2013, she had to weed out those with language or cogni tive impairments because the screen was not validated for those conditions. That left her with just 10% of her initial pool. Moreover, the screen asks how people have been feel ing over 7–14 days (typically asking ques tions such as “In the past few weeks have you been bothered by: little interest in pleasure or doing things? feeling down, depressed or hopeless? sleep problems?”), yet most patients are discharged within five days. With those limitations, Casal identified just two people with depression — and both were depressed before their strokes. Given that one-third of people who have a stroke experience depres sion, Casal concluded that new screening tools are clearly needed8. Because most people who have a stroke ini tially visit a hospital, screening there makes logistical — although perhaps not medical — sense. Ideally, says Nathan Herrmann, a psychiatrist at Sunnybrook Health Sciences Centre in Toronto, Canada, patients should be screened every time they visit a new health professional. If this were the case, then Betts would have been screened not just when she finally confessed to her doctor, but also dur ing her earlier two-week stay in Charlottes ville and six-week stay in Lynchburg. “You need to continually re-evaluate for these problems,” Herrmann says. But other research suggests that screening might have little value. In 2008, Robinson selected 176 people who had had a stroke but showed no signs of depression and divided them into three groups. The first group was given the antidepressant escitalopram, a second received a psychological interven tion known as problem-solving therapy and the third was given a placebo pill. The patients were then evaluated for depression over 12 months9. The results were startling. Of the participants who completed the trial (27 dropped out before it started), those who received no intervention were twice as likely to develop depression as were those going through problem-solving therapy, which entailed having patients visit psychologists for 12 sessions and work through problems using a seven-step model. Furthermore, patients in the control group were more than four times more likely to develop depression than those on the antidepressant. That suggests that patients should consider therapy even if their early depression screen comes back negative. Further support for this idea came from a 2011 study3 by French researchers, who dis covered that when they gave the antidepressant fluoxetine to people who did not have depres sion but did have weakness or paralysis on one side of the body after a stroke, the patients’ motor recovery soared. The researchers specu lated that the drug enhanced the brain’s ability

to heal itself, but another possibility is that the antidepressants kept depression at bay and helped patients to push through rehab. The French trial involved just over 100 participants and closed after only three months, so several labs are now working to replicate those results by assessing thousands of individuals over longer time frames.

GENERAL PRESCRIPTION

If antidepressants really do protect against depression and enhance motor recovery, rea sons Robinson, why not prescribe them to all people who have a stroke? Waimei Tai, a neu rologist at the Stanford University School of Medicine, says that she prescribes the drugs to anyone who has had a stroke and has depres sion or motor problems, unless their language is impaired. Like any drug, antidepressants do not work for everyone and do have side effects, ranging from dry mouth and diges tion issues to a slightly increased risk of future stroke. “These drugs are not benign,” Tai says. And, says Allan House, a psychiatrist at the Leeds Institute of Health Sciences, UK, the recovery process looks “more like grief than a mental illness”. That means that unless the use of antidepressants to improve motor recovery is validated, prescribing them to every patient is akin to putting every grieving widow on drugs straight after their spouse’s death. Some mourners will rebound on their own; others may need additional help. Nor are medications a fix-all. A stroke is an earthquake that shatters routines and relation ships. Betts says that her husband went from a loving equal to being her caregiver. Her career came to an abrupt end. She could no longer hold her grandchildren for fear of dropping them. She would have loved regular visits with a skilled mental-health therapist. “Just a pill doesn’t cut it,” she says. But Betts is grateful for the antidepressant, which she continues to take. And now, five years later, she can speak more positively about the experience that has left her unable to walk unaided. She now has time, she says, to just “sit and listen to my grandchildren”. ■ Sujata Gupta is freelance journalist living in Burlington, Vermont. 1. Morris, P. L., Robinson, R. G., Andrzejewski, P., Samuels, J. & Price, T. R. Am. J. Psychiatry 150, 124–129 (1993). 2. Jorge, R. E., Robinson, R. G., Arndt, S. & Starkstein, S. Am. J. Psychiatry 160, 1823–1829 (2003). 3. Chollet, F. et al. Lancet Neurol. 10, 123–130 (2011). 4. Robinson, R. G., Shoemaker, W. J., Schlumpf, M., Valk, T. & Bloom, F. E. Nature 255, 332–334 (1975). 5. Folstein, M. F., Maiberger, R. & McHugh, P. R. J. Neurol. Neurosurg. Psychiatry 40, 1018–1020 (1977). 6. Carson, A. J. J. Neurol. Neurosurg. Psychiatry 83, 859 (2012). 7. El Husseini, N. et al. Stroke 43, 1609–1616 (2012). 8. Casal, S. L., Finley-Caufield, A. & Tai, W. A. Stroke 45, ANS3 (2014). 9. Robinson, R. G. et al. J. Am. Med. Assoc. 299, 2391–2400 (2008). 2 6 J U N E 2 0 1 4 | VO L 5 1 0 | NAT U R E | S 1 1

OTTAWA HOSPITAL

OUTLOOK STROKE

PERSPECTIVE Silent, but preventable, perils

‘Covert’ strokes are a leading cause of dementia — and their incidence will rise in step with that of vascular risk factors, says Antoine M. Hakim.

“I

can’t really explain it, but I have not been right in the head for a few months. My mind is in a fog.” That is how my 38-year-old patient with a history of high blood pressure and obesity described his symptoms. He had not experienced any of the symptoms of a ‘mini-stroke’, or transient ischaemic attack, which can result in temporary motor, sensory or speech deficits. Other than his high blood pressure and obesity, his physical and neurological examination showed nothing out of the ordinary. But something was clearly wrong: a standard test revealed a deficit in his cognitive function. A magnetic resonance imaging (MRI) scan showed the probable cause: four ‘hyperintensities’ — bright signals that indicate damage — in the white matter, which coordinates communication between parts of the brain. Most neurologists would probably agree that, left untreated, this patient’s mind would have continued to deteriorate, leading eventually to dementia. His cognitive difficulties were probably a result of this damage, yet he had few symptoms. We are learning that these covert strokes are much more widespread than we thought, and that they are a major cause of dementia. Most overt strokes are the result of a blockage in one of the arteries in the grey matter, which controls a lot of the brain’s functions. But because the tissue is supplied by parallel arteries, blood can pass through the other vessels if one is blocked, thereby mitigating — although not eliminating — the impact. By contrast, the deeper insulated fibres in the white matter affected by a covert stroke are supplied by single, smaller vessels called arterioles. As a result, a blockage there results in tissue death and the loss of insulation. Although they are typically not noticed by the patient, the events do disrupt pathways that are essential for normal cognitive function. These strokes have been shown to result in cognitive deficits in rats1, and to atrophy in the human hippocampus — a structure crucial for memory2. Although they have the same risk factors as overt strokes, the causes and effects are much more complex3. The most troubling aspect of silent strokes is how common they are. Anywhere from 8% to 28% of the population have had a covert stroke, and their incidence rises with age4. In 1998, roughly 9 million people in the United States had a covert stroke, compared with 770,000 who had an overt event4. The silent nature of these events, plus the fact that accurate diagnosis requires an MRI, makes it likely that small-vessel disease in the white matter is under-appreciated as a cause of dementia. In fact, cerebrovascular disease contributes to three-quarters of dementia cases in North America5. Several factors suggest that without firm action taken relatively early in life, an epidemic of dementia caused by vascular disease is likely. Dementia is a disease of ageing, and Statistics Canada, a government agency, estimates that by 2031 almost one-quarter of the country’s population will be aged 65 years or older6. And there is a

growing crisis in the prevalence of cardiovascular disease risk factors across North America — particularly in younger people — with high blood pressure, diabetes and obesity having risen dramatically over the past decades. The brain is more sensitive than the heart to three factors: high blood pressure, obesity and a sedentary lifestyle7. The incidence of covert strokes is thought to increase as resting systolic blood pressure rises above 130 mmHg (ref. 8). Despite this, formal recommendations allow higher pressure in individuals who have no other risk factors, and many health-care providers are reluctant to treat high blood pressure aggressively, particularly in elderly people. These factors may be combining to increase the possibility of cognitive decline, and should induce us to reconsider both our bloodpressure targets and our management of vascular risk factors, not least because evidence suggests that reducing systolic blood pressure gradually to 130 mmHg is quite feasible9. If there is a silver lining in this story, it is the implication that we have significant control over how well individuals age and over the prevalence of dementia. But it will not be easy to do so. Small-vessel disease in the brain seems to result from prolonged exposure to risk factors, so preventing it will require early action. We must therefore urgently institute social measures that begin in schools to encourage and potentially reward healthy and active lifestyles, promoting the consumption of low-salt and low-sugar foods that are less processed, and stimulating the commitment and drive towards normal weights and blood pressures. Such changes can have dramatic, positive results. By changing his diet from daily fast food to plant-based, Mediterranean-style meals and walking 45 minutes a day at a brisk pace, the patient described in the opening statement lost 18 kilograms and normalized both his blood pressure and cognitive function. The fog is lifting. ■

WE HAVE SIGNIFICANT

CONTROL

OVER HOW WELL INDIVIDUALS

AGE

AND OVER THE PREVALENCE OF

DEMENTIA.

Antoine M. Hakim is chief executive and scientific director of the Canadian Stroke Network and director of the Neuroscience Research Program at the Ottawa Hospital Research Institute and the University of Ottawa in Canada. e-mail: ahakim@ohri.ca 1. 2. 3. 4. 5. 6. 7. 8. 9.

Shih, A. et al. Nature Neurosci. 16, 55–63 (2013). Neltner, J. H. et al. Brain 137, 255–267 (2014). Iadecola, C. Neuron 80, 844–866 (2013). Leary, M. C. & Saver, J. L. Cerebrovasc. Dis. 16, 280–285 (2003). Agüero-Torres, H., Kivipelto, M. & von Strauss, E. Dement. Geriatr. Cogn. Disord. 22, 244–249 (2006). Statistics Canada Population Projections: Canada, the Provinces and Territories (2010). O’Donnell, M. J. et al. Lancet 376, 112–123 (2010). SPS3 Study Group Lancet 382, 507–515 (2013). Pergola, P. E. et al. Am. J. Hypertens. http://dx.doi.org/10.1093/ajh/hpu027 (2014).

S 1 2 | NAT U R E | VO L 5 1 0 | 2 6 J U N E 2 0 1 4

ARTICLE

A large-sample assessment of possible association between ischaemic stroke and rs12188950 in the PDE4D gene Ha˚kan Lo¨vkvist*,1,6, Sandra Olsson2, Peter Ho¨glund3,6, Olle Melander4,6, Christina Jern2, Marketa Sjo¨gren4,6, Gunnar Engstro¨m4,6, J Gustav Smith4,6, Bo Hedblad4,6, Gunnar Andsberg1,6, Hossein Delavaran1,6, Katarina Jood2, Ulf Kristoffersson5,6, Holger Luthman4,6, Bo Norrving1,6 and Arne Lindgren1,6 Previous reports have shown ambiguous findings regarding the possible associations between ischaemic stroke (IS) and single nucleotide polymorphisms (SNPs) in the phosphodiesterase 4D (PDE4D) gene region. The SNP rs12188950 (or SNP45) has often been studied in this context. We performed a multi-centre study involving a large sample of 2599 IS patients and 2093 control subjects from the south and west regions of Sweden to replicate previous studies regarding IS risk and rs12188950. Subjects from Lund Stroke Register (LSR), Malmo¨ Diet and Cancer Study (MDC) and Sahlgrenska Academy Study on Ischemic Stroke (SAHLSIS) were enroled. Subgroups of participants with hypertension and participants o55 years of age, as well as the TOAST subgroups large vessel disease, small vessel disease and cardioembolism, were also assessed. Univariate odds ratios (ORs) and ORs controlling for hypertension, diabetes and current smoking were calculated. We additionally performed a meta-analysis including 10 500 patients and 10 102 control subjects from 17 publications (including the present study). When assessing pooled data from LSR, MDC and SAHLSIS we obtained no association between IS and rs12188950 for all participants (OR¼0.93; 95% confidence interval (CI): 0.83–1.05). Significant associations were not found for hypertensive participants or participants with age o55, or when separately evaluating patients from the three different TOAST subgroups. The meta-analysis showed no significant overall estimate (OR¼0.96; 95% CI: 0.89–1.04) with significant heterogeneity for random effect (P¼0.042). No effect from rs12188950 on IS was found from either our pooled multi-centre data or the performed meta-analysis. We did not find any association between the examined subgroups and rs12188950 either. European Journal of Human Genetics (2012) 20, 783–789; doi:10.1038/ejhg.2012.4; published online 25 January 2012 Keywords: stroke; polymorphism; PDE4D; rs12188950; meta-analysis

INTRODUCTION Stroke is a major cause of death and functional disability in industrialised countries. The clinical presentation of stroke includes intracranial bleedings as well as cerebral diseases with a cardiovascular aetiology. A genetic contribution to stroke risk, also for common ischaemic stroke (IS) subtypes, is considered likely, but results have been equivocal. Two previous reports, one from an Icelandic study in 2003 and one reported by us in 2008, suggested that the minor allele in the single nucleotide polymorphism (SNP) in rs12188950 (also denoted SNP45), located at the 5¢ end of the phosphodiesterase 4D (PDE4D) gene in chromosome 5, might have an inhibiting effect on IS risk.1,2 Our previous study also showed a significant inhibiting effect of rs12188950 minor (T) allele on IS among subjects with hypertension.2 However, a meta-analysis in our previous publication in 2008 showed ambiguous results concerning the possible association between IS risk and variants in rs12188950.3–12

We were aware of three replicating studies, that showed significant allelic association with IS by detecting and examining haplotype blocks located at the 5¢ end region of the PDE4D gene.4,13,14 Moreover, a meta-analysis including four studies indicated significantly increased IS risk when considering an at-risk haplotype, including rs12188950.15 These findings suggest that rs12188950, as a polymorphism located in the centre of the 5¢ end region of PDE4D, might be particularly interesting for further study to confirm or refute a relation to IS. In the present multi-centre study involving three case–control groups from the west and south of Sweden, we intended to test for replication of previous findings regarding rs12188950 by assessing a large number of individuals. We also updated our previous meta-analysis by including new publications and our new results regarding the subject of the possible association between rs12188950 and IS. We focused on possible discrepancies between included populations and subpopulations because of sex, nationality or ethnicity, rather than discrepancies between the publications themselves in this updated meta-analysis.

1Department of Clinical Sciences, Neurology, Lund University, Lund, Sweden; 2Institute of Neuroscience and Physiology, Department of Clinical Neuroscience and Rehabilitation, The Sahlgrenska Academy at University of Gothenburg, Gothenburg, Sweden; 3Department of Laboratory Medicine, Clinical Pharmacology, Lund University, Lund, Sweden; 4Department of Clinical Sciences, Lund University, Malmo¨, Sweden; 5Department of Clinical Genetics, Lund University, University and Regional Laboratories, Lund, Sweden; 6Ska˚ne University Hospital, Lund, Sweden *Correspondence: H Lo¨vkvist, Competence Centre for Clinical Research, Ska˚ne University Hospital, Lund SE-221 85, Sweden. Tel: +46 46 177932; Fax: +46 46 176085; E-mail: hakan.lovkvist@skane.se Received 28 September 2011; revised 12 December 2011; accepted 4 January 2012; published online 25 January 2012

Ischaemic stroke and rs12188950 in PDE4D H Lo¨vkvist et al 784

MATERIAL AND METHODS Study population and sample characteristics Phenotypic and genotypic data for subjects from Lund Stroke Register (LSR), the Malmo¨ Diet and Cancer study (MDC) and the Sahlgrenska Academy Study on Ischemic Stroke (SAHLSIS) were included in our analyses. Originally, the total number of individuals consisted of 2682 IS patients and 2139 control subjects. After exclusion of 129 subjects owing to genotyping failure for rs12188950, the remaining 2599 patients and 2093 control subjects were included in our multi-centre study. Complete characteristics of the data used are described elsewhere.2,16–21 Briefly, LSR is a prospective, consecutive epidemiological study from which we have included 920 patients with first-ever stroke from the local catchment area of Ska˚ne University Hospital in Lund, and 566 control subjects from the same geographical area, selected randomly after stratification by age and sex to match the patient group (when including intracranial bleedings as well as cerebral infarctions).2,17 The LSR data was collected between March 2001 and August 2007. In a previous study by us from 2008 we assessed 932 IS patients and 396 control subjects from LSR.2 In order to perform a replication being independent of assessments made previously, we did not include any of these participants in the current study. However, the results from 2008 were included as a separate study/population in our meta-analysis. MDC is a prospective population-based study that was initiated in the 1990s as a cohort study at Ska˚ne University Hospital in Malmo¨, Sweden.16,21 MDC initially included a total of 28 499 randomly selected men, born between 1923 and 1945, and women, born between 1923 and 1950. Cases with a first stroke after the baseline examination (n¼874) were matched for age, sex and month of the baseline examination with control subjects (n¼881) free from stroke at the time of the stroke event of the corresponding patient. SAHLSIS is a consecutive study comprising patients from four stroke units situated in the west of Sweden. SAHLSIS involves patients presenting with firstever or recurrent acute IS before reaching the age of 70 years.18–20 A total of 805 IS patients and 646 controls, recruited between 1998 and 2008 within the SAHLSIS project, were included in this study. All three above mentioned studies are approved by the ethics committee of the University of Gothenburg (SAHLSIS data) and by the ethics committee of Lund University (LSR and MDC data). All participants or, when necessary, their next of kin provided informed consent before enrolment.

Definition of IS The diagnosis of IS was determined in accordance with World Health Organization criteria.22 All patients in the study had computer tomography or magnetic resonance scan, or autopsy findings, compatible with IS. In addition, a classification of IS according to the Trial of Org 10172 in Acute Stroke Treatment (TOAST) nomenclature was made for the LSR and SAHLSIS populations. Three main TOAST subgroups of IS were defined: (i) Large vessel disease/large artery atherosclerosis (LVD); (ii) small vessel disease/small artery occlusion (SVD) and iii) cardioembolism or significant aortic plaques (CE).23 Three patients diagnosed with aortic plaques were included in the LVD group in SAHLSIS.

Definition of intermediate phenotypes Three risk factors were ascertained for all participants in the study. (1) Hypertension was defined as medical treatment for hypertension, or blood pressure 160/90 mm Hg or higher at the time of discharge or after at least 1 week of hospitalisation. (2) Diabetes mellitus was defined as dietary or medical treatment for diabetes mellitus or measurements at discharge or at least 1 week of hospitalisation as follows: blood glucose 6.1 mmol/l or higher at two occasions, plasma glucose 7.0 mmol/l or higher at two occasions or alternatively more than 11 mmol/l at one occasion with concurrent symptoms suggestive of diabetes mellitus (measurements of hypertension and diabetes mellitus among SAHLSIS cases were performed at 3 months follow-up).18 (3) Current smoking was defined through self-report (current smokers versus previous smokers and never smokers). For LSR and SAHLSIS, age was defined at first-time stroke onset for patients and at inclusion for control subjects. A nested case–control design was used in MDC. Age at the time of the stroke event was used both for MDC stroke cases European Journal of Human Genetics

and their corresponding control subjects (age was omitted for 63 control subjects of the MDC owing to the absence of matching patients).

Genotyping Genotyping of data from the three included populations was performed by matrix-assisted, laser desorption ionisation time-of-flight mass spectrometry on a MASSarray platform with iPLEX genotyping technology (Sequenom, San Diego, CA, USA).20 All laboratory analyses were performed by the SWEGENE Resource Center for Profiling Polygenic Disease (Ska˚ne University Hospital, Malmo¨, Sweden), with no access to case/control status or personal identities of the samples. Data were stored at the Bioinformatics Unit at the same facility.

Data analysis and statistical methods Allelic association tests were used to compare patients and control subjects regarding possible allelic association between variants in rs12188950 and IS. Odds ratio (OR) estimates of the minor (alternative) T allele over the major (reference) C allele against cases and controls are presented with 95% confidence intervals (CI). OR estimates for each subpopulation (LSR, MDC and SAHLSIS) were calculated as well as an overall pooled OR estimate for the total of these subpopulations were joined together. We correspondingly assessed a specific subgroup of 1727 patients and 1056 control subjects with hypertension. Similarly, we defined a subgroup of 382 patients and 321 control subjects below 55 years of age for a separate assessment. We also evaluated a subsample of patients with either SVD, CE or LVD according to TOAST classification criteria compared with control subjects for LSR and SAHLSIS data joined together.23 Furthermore, we carried out unweighted and weighted simple logistic regression (LR) as well as weighted multiple LR for individuals genotyped using an additive model counting 0, 1 or 2 minor alleles per individual. Regression weights were created by sex and age (categorized into three age groups: participants below 55 years; participants between 55 and 65 years; and participants 65 years or more). Patients and control subjects of LSR, MDC and SAHLSIS (ie six strata referring to cases/controls in three geographical domains of study) were thus adjusted within each stratum to a weight variable referring to sex and age distribution in uniformity with the age and sex distribution of the entire sample of data. We performed a goodness-of-fit test (a w2 test using one degree of freedom) for possible departure from Hardy–Weinberg equilibrium in control subject data. Statistical computations were performed by using PASW (SPSS) Statistics (version 18) statistical software (PASW/SPSS software, IBM Corporation, Armonk, NY, USA). Power tests were carried out by using the PS Power and Sample Size Calculations software (version 3.0.14; Department of Biostatistics, Vanderbilt University, Nashville, TN, USA).

Meta-analysis We searched the electronic PubMed database (using http://www.ncbi.nlm.nih. gov/pubmed/) for studies including the search keywords ‘STRK1’, ‘PDE4D’ or ‘Phosphodiesterase 4D’ in combination with keywords ‘Stroke’ or ‘Cerebrovascular’. All keywords were searched in titles or abstracts of publications issued between January 2003 and August 2011. A total of 58 publications were found and thereafter manually reviewed. We excluded 42 publications owing to the following reasons: 17 publications were studies comprising reviews or metaanalyses; 3 publications were editorials or correspondence articles; 5 publications were association studies excluding rs12188950 because the SNP was found to be monomorphic after being genotyped; 8 were association studies that did not consider rs12188950 at all; whereas 9 publications were not allelic association studies assessing genotypes versus IS risk. After this process, 16 publications, comprising 19 distinguishable populations suitable for the meta-analysis, remained.1–13,24–26 The meta-analysis thus involved a total of 20 populations from 17 publications when our current study was included. The 42 publications that remained after the above selection for the metaanalyses were either studies comprising reviews or meta-analyses (11 publications); or editorials or correspondence articles (3 publications); or association studies where rs12188950 was excluded because of non-polymorphic traits

Ischaemic stroke and rs12188950 in PDE4D H Lo¨vkvist et al 785 (5 publications), or studies that were not including rs12188950 or in other respects were irrelevant for our meta-analysis (23 publications). We performed a random effects DerSimonian–Laird meta-analysis, and our intention was to examine the effect of rs12188950 on IS in a large population providing genetic information, as well as comparing different population subgroups regarding the possible association between rs12188950 allelic variation and IS risk. R version 2.12.0 (The R Foundation for Statistical Computing c/o Institute for Statistics and Mathematics, Vienna, Austria) was used for the meta-analysis calculations.

RESULTS The 2599 IS patients were older than the 2030 control subjects with non-missing values of age (see Table 1). The LSR subpopulation showed a higher rate of men in control subjects than in patients, whereas the SAHLSIS subpopulation conversely showed a higher rate of women. As expected, there were higher prevalences of the vascular risk factors hypertension, diabetes mellitus and current smokers among patients compared with control subjects.

The genotyping success rate for rs12188950 was 98% when considering all data. The genotype distribution in LSR, MDC and SAHLSIS control subject data joined together conformed to Hardy–Weinberg equilibrium (P¼0.072). According to a power test, using the parameters alpha¼0.05; 1-beta¼0.8, we found the entire sample size of our data to be sufficient to detect significant OR estimates when true OR values were below 0.84 or above 1.18. For a subgroup of 382 patients and 321 control subjects (covering the number of participants below 55 years of age), a true OR o0.61 or 41.52 would be required for a significant finding. Univariate association tests For the entire study data, we did not detect any statistically significant association between IS risk and allelic variants in rs12188950. Moreover, we did not obtain statistically significant ORs when assessing LSR, MDC and SAHLSIS data separately. Separate assessments of hypertensive participants or participants younger than 55 years of age did not show any results with such significance either. When examin-

Table 1 Phenotype characteristics of patients and control subjects included in Lund Stroke Register (LSR), Malmo¨ Diet and Cancer study (MDC) and the Sahlgrenska Academy Study on Ischemic Stroke (SAHLSIS) Patients LSR (N¼1486) Age: mean (±SD)

Control subjects

P-value a

N¼920

N¼566

74.8 (±12.3) 76.7, 18.2, 101.7

73.9 (±12.4) 76.3, 18.0, 96.4

Age: non-missing N Men: frequency/non-missing N (%)

920 463/920 (50.3)

566 324/566 (57.2)

0.010

Hypertension: frequency/non-missing N (%) Diabetes mellitus: frequency/non-missing N (%)

618/896 (69.0) 256/875 (29.3)

319/566 (56.4) 45/562 (8.0)

o0.001 o0.001

Current smokers: frequency/non-missing N (%)

171/907 (18.9)

60/565 (10.6)

o0.001

N¼874 69.3 (±7.0)

N¼881 69.4 (±7.0)

Age: median, min, max Age: non-missing N

69.9, 48.4, 82.7 874

70.1, 48.4, 82.7 818d

Men: frequency/non-missing N (%) Hypertension: frequency/non-missing N (%)

481/874 (55.0) 645/873 (73.9)

476/881 (54.0) 513/880 (58.3)

0.701 o0.001

Diabetes mellitus: frequency/non-missing N (%) Current smokers: frequency/non-missing N (%)

86/874 (9.8) 283/854 (33.1)

25/881 (2.8) 185/869 (21.3)

o0.001 o0.001

Age: median, min, max

MDC (N¼1755) Age: mean (±SD)

SAHLSIS (N¼1451) Age: mean (±SD)

0.163b 0.207c

0.680b 0.630c

N¼805

N¼646

Age: median, min, max

56.2 (±10.6) 58.8, 18.1, 69.8

56.0 (±10.4) 58.4, 18.6, 70.4

Age: non-missing N Men: frequency/non-missing N (%)

805 527/805 (65.5)

646 376/646 (58.2)

0.005

Hypertension: frequency/non-missing N (%) Diabetes mellitus: frequency/non-missing N (%)

464/794 (58.4) 150/805 (18.6)

224/645 (34.7) 31/644 (4.8)

o0.001 o0.001

Current smokers: frequency/non-missing N (%)

307/801 (38.3)

127/646 (19.7)

o0.001

All subjects in study (N¼4692) Age: mean (±SD)

0.698b 0.471c

N¼2599

N¼2093

Age: median, min, max

67.2 (±12.8) 67.6, 18.1, 101.7

66.4 (±12.7) 67.4, 18.0, 96.4

Age: non-missing N Men: frequency/non-missing N (%)

2599 1471/2599 (56.6)

2030d 1176/2093 (56.2)

0.790

Hypertension: frequency/non-missing N (%) Diabetes mellitus: frequency/non-missing N (%)

1727/2563 (67.4) 492/2554 (19.3)

1056/2091 (50.5) 101/2087 (4.8)

o0.001 o0.001

Current smokers: frequency/non-missing N (%)

761/2562 (29.7)

372/2080 (17.9)

o0.001

0.033b 0.045c

aP-values

are based on Fisher’s exact test if not footnoted otherwise. t-test (with equal variances assumed). non-parametric test. dAge of ischaemic stroke patient-matched control subjects of whom 63 such subjects from MDC data did not match. bStudent’s

cMann–Whitney

European Journal of Human Genetics

Ischaemic stroke and rs12188950 in PDE4D H Lo¨vkvist et al 786

Table 2 Association between rs12188950 (SNP45) and prevalence of ischaemic stroke for Lund Stroke Register (LSR), Malmo¨ Diet and Cancer study (MDC) and the Sahlgrenska Academy Study on Ischemic Stroke (SAHLSIS) Population/ Subpopulation

Number of non-missing patients/controls

Number of minor alleles in patients/controlsa

Allelic association test: OR; 95% CI; P-valueb

920/566

257/177

0.88; 0.71–1.08; 0.219

874/881 805/646

244/248 192/168

0.99; 0.82–1.20; 0.923 0.91; 0.73–1.13; 0.396

2599/2093

693/593

0.93; 0.83–1.05; 0.251

LSR MDC

618/319 645/513

174/94 176/145

0.95; 0.72–1.24; 0.728 0.96; 0.76–1.22; 0.762

SAHLSIS

464/224

116/59

0.94; 0.67–1.32; 0.730

1727/1056

466/298

0.95; 0.81–1.11; 0.521

58/44

16/13

0.92; 0.42–2.04; 1.000

31/29

10/12

0.74; 0.29–1.87; 0.638

SAHLSIS

293/248

53/63

0.68; 0.46–1.01; 0.061

Total (LSR, MDC, SAHLSIS)

382/321

79/88

0.73; 0.53–1.00; 0.057

161/1212

40/345

0.85; 0.60–1.22; 0.395

339/1212 410/1212

85/345 123/345

0.86; 0.67–1.11; 0.285 1.06; 0.85–1.33; 0.605

All patients and controls LSR MDC SAHLSIS Total (LSR, MDC, SAHLSIS) Patients and controls with hypertension

Total (LSR, MDC, SAHLSIS) Age below 55 years LSR MDC

IS sub-groups due to TOAST classificationc LVD SVD CE

Abbreviations: CE, cardiembolism; CI, confidence interval; LVD, large vessel disease; OR, odds ratio; SVD,small vessel disease. aMinor alleles T, major alleles C. bCIs computed using Wald approximation of variances, P-values computed using Fisher’s exact test. A small divergence may cause conflicting null-hypothesis conclusions. cTOAST classification of 1678 cases and1786 control subjects from LSR and SAHLSIS was performed, MDC is not included.

ing cases defined as IS patients with either LVD, SVD or CE subgroup, we did not find any significant association with rs12188950 (Table 2).23 LR analyses of additive rs12188950 data From the weighted simple LR analyses we obtained results for LSR, MDC and SAHLSIS after adjustments for heterogeneity caused by dissimilar distributions on sex and age. No significant associations were found (Table 3). These results were approximately similar to the corresponding univariate OR estimates (Table 2). The weighted multiple LRs, controlling for the dichotomous phenotype risk factors hypertension, diabetes mellitus and current smoking, did not change the results. All of these risk factor covariates were significant with Po0.001, but no significant contribution from rs12188950 could be seen. Meta-analysis The outcome of the meta-analysis of 10 500 IS patients and 10 102 subjects from 17 published studies (including our previous study) and this study, containing 20 distinguishable populations is visualised in Figure 1. The overall OR of 0.96 (95% CI: 0.89–1.04) indicates no significant overall effect in these 20 populations. The heterogeneity test (P¼0.042) shows a slightly significant difference between the ORs of the 20 populations, indicating that the OR estimates from these populations may be caused by other reasons than random variations. European Journal of Human Genetics

Only two populations showed ORs that significantly suggest allelic association between variants in rs12188950 and IS risk.1,2 When carrying out a meta-analysis that only comprised the Icelandic and the Swedish populations (five studies including the present study), the OR remained insignificant (OR¼0.88; 95% CI: 0.76–1.01; P-value from heterogeneity test¼0.047).1,2,5,9 DISCUSSION In this large multi-centre study we could not detect an association between rs12188950 and IS. This might be in conflict with the results of our previous study and the Icelandic study, although the upper CI limit of 0.91 and 0.92 from these studies, respectively, are close to the estimated OR (0.93) for rs12188950 in this present study.1,2 Considerations about IS subgroups None of the TOAST-classified subgroups of IS met the significance level in our analyses. A previous Swedish study by Kostulas et al5 in 2007 showed a relative risk (RR) of 1.43 (to be compared with 1.07 when including all IS causes and transient ischaemic attack) when testing the major (G) allele of rs12188950 against LVD risk. Another study by Kuhlenba¨umer et al 6 in 2006 found an OR estimate of 1.06 when testing major alleles of rs12188950 against LVD risk among a German population (OR¼1.05 when assessing all IS causes). The Icelandic study by Gretarsdottir et al1 showed RR¼1.77 when assessing patients with combined carotid and cardiogenic causes of stroke (all IS causes provided an RR estimate of 1.31). From these findings, only the Icelandic RR estimates were significant. However, the tendency

920/566

0.88; 0.70–1.10; 0.253

Controlling for IS risk factorsb OR; 95% CI; P-value

805/646

874/818

0.86; 0.67–1.09; 0.215

4.73; 2.86–7.82; o0.001 2.45; 1.95–3.07; o0.001

0.98; 0.80–1.21; 0.873 2.02; 1.63–2.50; o0.001

Diabetes Current smoker

rs12188950 Hypertension

0.93; 0.83–1.05; 0.240

2599/2093

0.90; 0.80–1.02; 0.095

2599/2030

4.40; 3.50–5.53; o0.001 2.64; 2.28–3.07; o0.001

0.91; 0.80–1.04; 0.159 2.02; 1.77–2.29; o0.001

3.37; 2.56–4.42; o0.001

0.91; 0.73–1.14; 0.419

0.93; 0.77–1.13; 0.478

Current smoker

805/646

874/881

2.52; 2.00–3.16; o0.001 3.86; 2.64–5.66; o0.001

0.90; 0.72–1.13; 0.372

0.99; 0.82–1.20; 0.920

Hypertension Diabetes

rs12188950

Diabetes Current smoker

rs12188950 Hypertension

Abbreviations: CI, confidence interval; OR, odds ratio. aWeight obtained by age and sex standardization of included populations. bDichotomous risk factors for ischaemic stroke: hypertension, diabetes mellitus and current smoking.

Pooled data

SAHLSIS

MDC

0.86; 0.70–1.06; 0.154

Number of non-missing cases/controls

2.61; 1.92–3.55; o0.001

920/566

OR; 95% CI; P-value

Current smoker

0.88; 0.72–1.08; 0.214

Number of non-missing cases/controls

2498/2002

793/645

849/795

856/562

Number of non-missing cases/controls

Weighted Multiple LR a

1.73; 1.37–2.18; o0.001 4.18; 2.94–5.96; o0.001

rs12188950

LSR

OR; 95% CI; P-value

Weighted Simple LR a

Hypertension Diabetes

Covariate

Population

Simple LR

Table 3 Simple and multiple logistic regression (LR) performing an additive approach for assessing possible association between rs12188950 and ischaemic stroke risk for Lund Stroke Register (LSR), Malmo¨ Diet and Cancer study (MDC) and the Sahlgrenska Academy Study on Ischemic Stroke (SAHLSIS)

Ischaemic stroke and rs12188950 in PDE4D H Lo¨vkvist et al 787

European Journal of Human Genetics

Ischaemic stroke and rs12188950 in PDE4D H Lo¨vkvist et al

Study Reference

788

AUSTRALIAN (Staton 2006) BRITISH (Bevan 2005) DUTCH (van Rijn 2005) GERMAN (Lohmussaar 2005) GERMAN (Kuhlenbäumer 2006) GREEK (Fidani 2007) ICELANDIC (Gretarsdottir 2003) ITALIAN (Quarta 2010) PORTUGESE (Domingues-Montanari 2010) SPANISH (Domingues-Montanari 2010) SWEDISH (Nilsson-Ardnor 2005) SWEDISH (Kostulas 2007) SWEDISH (Lövkvist 2008) SWEDISH (Lövkvist 2011) US WHITE (Meschia 2005) US NON-WHITE (Meschia 2005) US WHITE FEMALE >65yrs (Brophy 2006) US WHITE FEMALE 15-49yrs (Song 2006) US AFRO-AM FEMALE 15-49yrs (Song 2006) US WHITE MALE (Zee 2006) Summary

0.40

0.50

0.63

0.79

1.00

1.26

1.58

2.00

2.51

Odds Ratio

Figure 1 Random effects (DerSimonian–Laird) meta-analysis of 16 previously published studies and the present study, comprising 20 distinguishable populations assessing the association between the minor (T) allele of rs12188950 and IS in 10 500 IS patients and 10 102 controls. Overall OR: 0.96 (95% CI: 0.89–1.04), test for heterogeneity w2¼30.9 (P¼0.0416).

towards a protecting effect of minor alleles against IS in the three studies as a whole might be an interesting subject of further studies of LVD subgroups. Discrepancy between present and previously reported LSR data Our previously reported study showed a protecting effect against IS that could not be found in our current study, despite similar LSR data collection routines.2 The dataset in the previous study published in 2008 (LSR-I) did not differ considerably from the present LSR data (LSR-II) regarding mean age and proportion of current smokers. Nevertheless, LSR-I comprises a lower frequency of hypertensive controls (33.7%) than LSR-II (56.4%). Patients and controls with diabetes mellitus were also more frequent in LSR-II (29.1% and 8.1%, respectively) compared with that in LSR-I (21.0% and 6.3%, respectively). LSR-I was collected between 2001 and 2004, and the major part of LSR-II was collected between 2004 and 2008. Meta-analysis findings Our meta-analysis that included populations from Europe, North America and Australia did not find any overall tendency of IS being affected by rs12188950 minor alleles in neither protective nor augmenting direction. A previously published meta-analysis, with similarities to ours regarding selection of publication, presented a significant overall OR¼1.29 (P¼0.05) after removal of publications causing heterogeneity.27 The slightly significant heterogeneity test (P¼0.042) suggests a nonrandom effect explaining the difference between the ORs among the 20 compared populations in our meta-analysis. Figure 1 visualises this ambiguity regarding the OR estimates. Dissimilar ethnicities in the studied populations might be one explaining factor. However, other demographic attributes (such as age, sex or occupational status of participants) may also contribute to such an effect. Besides, conclusions regarding possible ethnical differences might be connected with various problems regarding adequate definitions of a particular group of study.28 European Journal of Human Genetics

We found five publications in our PubMed search that presented monomorphic allele representation regarding rs12188950.14,29–32 Moreover, we found a Chinese meta-analysis on Asian populations that had excluded rs12188950 from five studies regarding PDE4D gene data because of total absence of individuals with heterozygote allele pairs within that SNP.33 A number of published studies examined SNPs in the PDE4D gene without including rs12188950. Uncommented exclusion of a SNP might cause selection bias in meta-analyses. Multi-centre data considerations One main issue when assessing multi-centre data is how to cope with a non-homogenous group. When performing a heterogeneity test of the three subpopulations (LSR, MDC and SAHLSIS) using the algorithm applied in the meta-analysis, we obtained a non-significant P-value of 0.6735. However, although the subpopulations were heterogeneous because of age distribution, we fitted weighted LR models using age- and sex-standardizing stratum weights. We thereby obtained OR estimates that were more similar than those obtained when not using such stratum weights (Table 3). Another source of heterogeneity between LSR, MDC and SAHLSIS is the registration of vascular risk factors (hypertension, diabetes and current smoking). Data for these phenotypes was collected at stroke onset for LSR and SAHLSIS, but at the baseline time of screening for MDC. The MDC data collection regarding these risk factors was therefore performed at an estimated mean time of about 5.5 years before stroke onset. CONCLUSION Our assessment in a large study including 2599 patients and 2093 control subjects did not show any association between rs12188950 and IS. A meta-analysis including 10 500 patients and 10 102 control subjects provided the same result. No association was detected in any of the subgroups. However, a hypothetical protecting effect (although it is non-significant with a P-value of 0.057 in this study)

Ischaemic stroke and rs12188950 in PDE4D H Lo¨vkvist et al 789

of the rs12188950 minor allele on IS among younger individuals might be subject for further studies. CONFLICT OF INTEREST The authors declare no conflict of interest. ACKNOWLEDGEMENTS We thank all the subjects as well as the attendant nursing staff, technical advisers and other co-workers involved for their participation in this study. This study was supported by grants from the Swedish Research Council (K2008-65X-14605-06-03, K2011-65X-14605-09-6, K2010-61X-20378-04-3), the Swedish State (ALFGBG-148861), the Swedish Heart and Lung Foundation (20100256), the Yngve Land Foundation for Neurological Research, the Crafoord Foundation, Region Ska˚ne, the Freemasons Lodge of Instruction EOS in Lund, King Gustav V and Queen Victoria’s Foundation, Lund University, the Department of Neurology Lund, the Swedish Stroke Association, the Tore Nilsson Foundation and the Emelle Foundation.

1 Gretarsdottir S, Thorleifsson G, Reynisdottir ST et al: The gene encoding phosphodiesterase 4D confers risk of ischemic stroke. Nat Genet 2003; 35: 131–138. 2 Lo¨vkvist H, Smith JG, Luthman H et al: Ischaemic stroke in hypertensive patients is associated with variations in the PDE4D genome region. Eur J Hum Genet 2008; 16: 1117–1125. 3 Bevan S, Porteous L, Sitzer M, Markus HS: Phosphodiesterase 4D gene, ischemic stroke, and asymptomatic carotid atherosclerosis. Stroke 2005; 36: 949–953. 4 Brophy VH, Ro SK, Rhees BK et al: Association of phosphodiesterase 4D polymorphisms with ischemic stroke in a US population stratified by hypertension status. Stroke 2006; 37: 1385–1390. 5 Kostulas K, Gretarsdottir S, Kostulas V et al: PDE4D and ALOX5AP genetic variants and risk for Ischemic Cerebrovascular Disease in Sweden. J Neurol Sci 2007; 263: 113–117. 6 Kuhlenba¨umer G, Berger K, Huge A et al: Evaluation of single nucleotide polymorphisms in the phosphodiesterase 4D gene (PDE4D) and their association with ischaemic stroke in a large German cohort. J Neurol Neurosurg Psychiatry 2006; 77: 521–524. 7 Lo˜hmussaar E, Gschwendtner A, Mueller JC et al: ALOX5AP gene and the PDE4D gene in a central European population of stroke patients. Stroke 2005; 36: 731–736. 8 Meschia JF, Brott TG, Brown RD et al: Phosphodiesterase 4D and 5-lipoxygenase activating protein in ischemic stroke. Ann Neurol 2005; 58: 351–361. 9 Nilsson-Ardnor S, Wiklund PG, Lindgren P et al: Linkage of ischemic stroke to the PDE4D region on 5q in a Swedish population. Stroke 2005; 36: 1666–1671. 10 Song Q, Cole JW, O’Connell JR et al: Phosphodiesterase 4D polymorphisms and the risk of cerebral infarction in a biracial population: the Stroke Prevention in Young Women Study. Hum Mol Genet 2006; 15: 2468–2478. 11 Staton JM, Sayer MS, Hankey GJ et al: Association between phosphodiesterase 4D gene and ischaemic stroke. J Neurol Neurosurg Psychiatry 2006; 77: 1067–1069. 12 van Rijn MJE, Slooter AJC, Schut AFC et al: Familial aggregation, the PDE4D gene, and ischemic stroke in a genetically isolated population. Neurology 2005; 65: 1203–1209.

13 Fidani L, Clarimon J, Goulas A et al: Association of phosphodiesterase 4D gene G0 haplotype and ischaemic stroke in a Greek population. Eur J Neurol 2007; 14: 745–749. 14 Nakayama T, Asai S, Sato N, Soma M: Genotype and haplotype association study of the STRK1 region on 5q12 among Japanese – A case-control study. Stroke 2006; 37: 69–76. 15 Bevan S, Dichgans M, Gschwendtner A, Kuhlenbaeumer G, Ringelstein EB, Markus HS: Variation in the PDE4D gene and ischemic stroke risk – A systematic review and meta-analysis on 5200 cases and 6600 controls. Stroke 2008; 39: 1966–1971. 16 Dahlberg J, Smith G, Norrving B et al: Genetic variants in serum and glucocortocoid regulated kinase 1, a regulator of the epithelial sodium channel, are associated with ischaemic stroke. J Hypertens 2011; 29: 884–889. 17 Hallstro¨m B, Jo¨nsson AC, Nerbrand C, Petersen B, Norrving B, Lindgren A: Lund Stroke Register: hospitalization pattern and yield of different screening methods for first-ever stroke. Acta Neurol Scand 2007; 115: 49–54. 18 Jood K, Ladenvall C, Rosengren A, Blomstrand C, Jern C: Family history in ischemic stroke before 70 years of age: the Sahlgrenska Academy Study on Ischemic Stroke. Stroke 2005; 36: 1383–1387. 19 Olsson S, Jood K, Melander O et al: Lack of association between genetic variations in the KALRN region and ischemic stroke. Clin Biochem 2011; 44: 1018–1020. 20 Olsson S, Melander O, Jood K et al: Genetic variant on chromosome 12p13 does not show association to ischemic stroke in 3 Swedish case-control studies. Stroke 2011; 42: 214–216. 21 Smith JG, Melander O, Lo¨vkvist H et al: Common genetic variants on chromosome 9p21 confers risk of ischemic stroke: a large-scale genetic association study. Circ Cardiovasc Genet 2009; 2: 159–164. 22 WHO MONICA Project Principal Investigators: The World Health Organization MONICA Project (monitoring trends and determinants in cardiovascular disease). A major international collaboration. J Clinical Epidemiol 1988; 41: 105–114. 23 Adams Jr HP, Bendixen BH, Kappelle LJ et al: Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment. Stroke 1993; 24: 35–41. 24 Domingues-Montanari S, Fernandez-Cadenas I, del Rio-Espinola A et al: Association of a genetic variant in the ALOX5AP with higher risk of ischemic stroke: A Case-Control, Meta-Analysis and Functional Study. Cerebrovasc Dis 2010; 29: 528–537. 25 Quarta G, Stanzione R, Evangelista A et al: Phosphodiesterase 4D and 5-lipoxygenase activating protein genes and risk of ischemic stroke in Sardinians. Eur J Hum Genet 2009; 17: 1448–1453. 26 Zee RYL, Brophy VH, Cheng S, Hegener HH, Erlich HA, Ridker PM: Polymorphisms of the phosphodiesterase 4D, cAMP-specific (PDE4D) gene and risk of ischemic stroke – a prospective, nested case-control evaluation. Stroke 2006; 37: 2012–2017. 27 Yoon D, Park SK, Kang D, Park T, Park JW: Meta-analysis of homogeneous subgroups reveals association between PDE4D gene variants and ischemic stroke. Neuroepidemiology 2011; 36: 213–222. 28 Census, race and science. Nat Genet 2000; 24: 97–98. 29 Lin HF, Liao YC, Liou CW, Liu CK, Juo SH: The phosphodiesterase 4D gene for early onset ischemic stroke among normotensive patients. J Thromb Haemost 2007; 5: 436–438. 30 Matsushita T, Kubo M, Yonemoto K et al: Lack of association between variations of PDE4D and ischemic stroke in the Japanese population. Stroke 2009; 40: 1245–1251. 31 Sun Y, Huang YY, Chen X et al: Association between the PDE4D gene and ischaemic stroke in the Chinese Han population. Clin Sci(Lond) 2009; 117: 265–272. 32 Woo D, Kaushal R, Kissela B et al: Association of phosphodiesterase 4D with ischemic stroke – a population-based case-control study. Stroke 2006; 37: 371–376. 33 Xu XW, Li X, Li JJ, Ou R, Sheng WL: Meta-analysis of association between variation in the PDE4D gene and ischemic cerebral infarction risk in Asian populations. Neurogenetics 2010; 11: 327–333.

European Journal of Human Genetics

REVIEW ARTICLE

Cerebral ischemic stroke: is gender important? Claire L Gibson Cerebral stroke continues to be a major cause of death and the leading cause of long-term disability in developed countries. Evidence reviewed here suggests that gender influences various aspects of the clinical spectrum of ischemic stroke, in terms of influencing how a patients present with ischemic stroke through to how they respond to treatment. In addition, this review focuses on discussing the various pathologic mechanisms of ischemic stroke that may differ according to gender and compares how intrinsic and hormonal mechanisms may account for such gender differences. All clinical trials to date investigating putative neuroprotective treatments for ischemic stroke have failed, and it may be that our understanding of the injury cascade initiated after ischemic injury is incomplete. Revealing aspects of the pathophysiological consequences of ischemic stroke that are gender specific may enable gender relevant and effective neuroprotective strategies to be identified. Thus, it is possible to conclude that gender does, in fact, have an important role in ischemic stroke and must be factored into experimental and clinical investigations of ischemic stroke. Journal of Cerebral Blood Flow & Metabolism (2013) 33, 1355–1361; doi:10.1038/jcbfm.2013.102; published online 12 June 2013 Keywords: acute stroke; brain ischemia; gender; neuroprotection; steroids

INTRODUCTION Cerebral stroke is a leading cause of death and the main cause of adult long-term disability in developed countries. The majority (B85%) of cerebral strokes are ischemic in nature and result from the occlusion of a major cerebral artery by a thrombus or an embolism, which leads to loss of blood flow and subsequent tissue death in the affected region. There is an increasing recognition of differences between men and women in relation to stroke, not only in terms of stroke risk but also in terms of influencing the etiology, symptoms, and outcomes. Women appear to have a higher overall lifetime risk of stroke in addition to higher rates of poststroke mortality, disability, depression, and dementia, compared with men. Such gender differences have largely been attributed to the longer life expectancy of women, consistent with the fact that age is the strongest independent risk factor for stroke1 and also a negative predictor for clinical outcome.2 Such a detrimental impact of age upon stroke risk and outcome, initially observed in white populations, has been consistently reported across studies involving participants of varying race and ethnicity.3 In the general population, men have been found to experience more ischemic strokes4–6 whereas women tend to have more infarctions involving the anterior circulation and experience more subarachnoid hemorrhages.5,7 In terms of stroke onset, women tend to be, on average, approximately 4 years older than men at the age of ischemic stroke onset. A recent meta-analysis on data from 2,566 patients revealed that the mean age of onset of first ischemic stroke was 66.6 years in men compared with 70.0 years in women.8 As age is a significant risk factor and predictor of outcome after ischemic stroke, it is reported that elderly women carry the major burden associated with ischemic stroke.9 However, it is reported that in the premenopausal years, women benefit from a native cerebrovascular protection provided by reproductive hormones. For example, The Framingham study reported that the incidence of ischemic stroke is lower in women than men within the 45- to 54-year-old age cohort, composed

mainly of premenopausal and preimenopausal women, but is equalized in the 55- to 64-year-old cohort.10 However, nonhormonal factors also contribute to the sexual dimorphic risk of ischemic stroke as gender differences not only extend well beyond the menopausal years but are also clearly present in neonatal and childhood populations where circulating hormones are equivalent between the sexes.11,12 Interestingly, although some studies do report a higher incidence of stroke recurrence in females compared with males13 it seems that actually, after comparing for age, medical history, and other relevant risk factors, no gender differences are observed in stroke recurrence.14 Gender may also influence both the mechanisms of injury and outcome after ischemic stroke and thus, gender should be considered in both experimental and clinical stroke studies. There is an urgent need for the development of effective pharmacological therapies to counter the deleterious effects of ischemic stroke. The only current treatment available is thrombolysis with tissue plasminogen activator, but because of its narrow therapeutic window (o4.5 hours) and safety concerns, less than 5% of stroke patients receive tissue plasminogen activator and the majority of stroke patients only receive supportive care.15 The development of safe and effective treatments is, therefore, a major challenge to experimental and clinical stroke research. This review discusses the various aspects of ischemic stroke (risk, symptoms, outcome, etc.), which are beginning to emerge as being sexually dimorphic. In addition, it is discussed how knowledge of the influence of gender on the pathologic mechanisms after ischemic stroke may reveal potential treatment targets. GENDER INFLUENCES THE CLINICAL CHARACTERISTICS ASSOCIATED WITH ISCHEMIC STROKE Important modifiable risk factors for ischemic stroke include hypertension, high blood cholesterol, diabetes, cigarette smoking, obesity, and lack of physical activity.16–19 Data on sex-specific risk

School of Psychology, University of Leicester, Leicester, UK. Correspondence: Dr CL Gibson, School of Psychology, University of Leicester, Henry Wellcome Building, Lancaster Road, Leicester LE1 9HN, UK. E-mail: cg95@le.ac.uk Received 8 April 2013; revised 23 May 2013; accepted 24 May 2013; published online 12 June 2013

Cerebral ischemic stroke CL Gibson

1356 factors in ischemic stroke have initially revealed that men report a higher prevalence of current smoking, history of smoking, coronary artery disease, and peripheral artery disease.20–22 Such variations in risk factors, according to gender, were also confirmed by others23 who reported that although women diagnosed with ischemic stroke were older than men, the history of hyperlipidemia, smoking, and coronary heart disease were less frequent. However, a systematic review and meta-analysis revealed that women with stroke are more likely than men to have a parental history of stroke, which is accounted for by an excess maternal history of stroke.24 Such a finding could be explained by sex-specific genetic, epigenetic, or non-genetic mechanisms. It is also important to consider whether gender differences occur in the clinical presentation of a patient experiencing an ischemic stroke, as this may impact upon the prompt recognition, diagnostic testing, and receipt of appropriate treatment. In general terms, the symptoms of ischemic stroke can be separated into classic symptoms, which include hemibody numbness, hemiparesis, aphasia, dysarthria, visual disturbance, diplopia, facial weakness, discoordination/ataxia, and vertigo, and non-specific to stroke symptoms including pain, light headedness, mental status change, headache, and other neurologic or non-neurologic symptoms. Data on gender differences in acute stroke presentation are somewhat limited and few population-based studies exist. However, some studies report that women present with more non-specific stroke symptoms such as pain or altered mental status8,25–27 whereas others report that women are as likely as men to report weakness and clumsiness, as well as both classic and non-specific symptoms that include pain, disorientation, and altered consciousness.28 In one of the few population-based studies, men were found to be more likely to present with classic focal neurologic symptoms such as sensory abnormalities, ataxia, and diplopia whereas women tended to present with diffuse symptoms such as disorientation, generalized weakness, fatigue, and mental status change.29 However, it is reported that women delay up to three times longer than men in seeking care for stroke symptoms,30 which of course may contribute to variations in the presentation of symptoms. If women do clinically present differently with ischemic stroke than men, then there may be significant public health implications. Recognition of potential gender differences in stroke symptoms through education aimed at both the public and health care professionals could result in decreased out-of-hospital and in-hospital delays, thus increasing access to acute stroke therapy in women. Many public health campaigns, for example ‘Give me 5’in the United States and ‘FAST’ in the United Kingdom, are aimed primarily at targeting the classic symptoms of ischemic stroke.31 It may be that public health campaigns and stroke intervention efforts should be modified to address the possibility that women, experiencing an ischemic stroke, may experience a number of non-specific symptoms. However, in the first instance, data from a large multicenter trial could determine whether gender differences in presentation of symptoms occur across a representative population of stroke cases and importantly, whether differences in symptom presentation have any clinical consequences. GENDER EFFECTS ON OUTCOME AFTER ISCHEMIC STROKE Numerous studies report gender differences in both mortality and morbidity after ischemic stroke. Typically, female stroke patients present at an older age than male stroke patients thus, older female patients are likely to comprise the majority of strokerelated mortality and disability.3,32 Data on gender differences in mortality rates after ischemic stroke are conflicting with some reporting higher in-hospital mortality rates in women,33 whereas others report no gender differences in mortality rates both inhospital5 and at 3 months after ischemic stroke.4 Journal of Cerebral Blood Flow & Metabolism (2013), 1355 – 1361

However, the general consensus seems to be that women have poorer functional outcomes, than men, after ischemic stroke.5,33,34 Such a gender difference appears to be sustained even after making adjustments for age and other sex differences in medical history and presentation. In fact, one study34 found that at 6 months poststroke, female sex is still an independent predictor of poor prognosis even when adjusting for other predictors of functional outcome. Others have reported that at 3 months poststroke, women are more likely to have a poorer functional outcome4 and, in addition, women show significantly worse locomotor function than men at both 1- and 5-year-follow-up after ischemic stroke.14 After ischemic stroke, women are less likely to be discharged home35 and more likely to have impairments and activity limitations on followup.5 It is reported that poststroke women experience more mental impairment,35 depression,36 fatigue,37 and have a lower overall quality of life38–40 than men. It may be though that if women are more likely to delay in seeking care for stroke symptoms,30 this could result in treatment delays, which would contribute to worse outcomes. However, it needs also to be considered whether gender influences the treatment a patient receives both during and after the diagnosis of ischemic stroke. GENDER EFFECTS ON THE TREATMENT RECEIVED AND RESPONSE TO TREATMENT During diagnosis of ischemic stroke, several studies report gender differences in the use of echocardiography, carotid imaging, and endarterectomy.2,41,42 In addition, data from different national registers indicate that women are less likely to be seen by a stroke team or have their lipids investigated.5,43 In terms of in-hospital treatment, the use of antiplatelet drugs and statins is reported to be lower among women than men and some report that women are less likely to receive thrombolytic treatment.6,43 However, it must be noted that any differences in treatment may be as a direct consequence of sex differences in symptoms at presentation. In addition to influencing the treatment a patient may actually receive, gender probably also influences the response to certain pharmacological treatments. Such a notion is largely overlooked in experimental stroke studies (discussed below) and only a few clinical trials to date have specifically investigated whether sexspecific responses to treatment exist. The Physicians Health Study found that aspirin substantially reduced the risk of cardiovascular disease, but not stroke, in men.44 By contrast, in the Women’s Health Study, aspirin reduced the risk of stroke by 24% but had no effect on cardiovascular disease in women.45,46 A subsequent sexspecific meta-analysis confirmed the differential effects of aspirin in men and women with respect to stroke and coronary artery disease risk.44 Oral anticoagulation with warfarin also seems to be affected by gender as shown in a prospective study of atrial fibrillation patients.47 In that study, men and women were anticoagulated to equal degrees but anticoagulated women had a relative risk of 2.0 for ischemic stroke versus anticoagulated men, even after age correction. Biologic sex may also affect the response to thrombolytic therapy as, after stroke, women have worse functional outcomes compared with men yet several studies have demonstrated an absence of sex differences in functional outcome after acute thrombolytic therapy.48–50 This normalization of baseline differences in functional outcome suggests that women may experience a greater benefit from tissue plasminogen activator than men.49 Thus, gender differences do seem to exist across many clinical aspects of ischemic stroke from affecting how the patient may present in the clinic through to how they may respond to a treatment. Although the current American Heart Association/American Stroke Association guidelines for acute stroke treatment51 make no distinction between genders in terms of presentation of symptoms or possible therapeutic approaches. However, the influence of & 2013 ISCBFM

Cerebral ischemic stroke CL Gibson

1357 gender is often overlooked in the design of experimental stroke studies and the investigation of the pathologic mechanisms influenced by gender may inform the design of potential treatment strategies. THE IMPACT OF GENDER—INTRINSIC VERSUS HORMONAL MECHANISMS Although it would appear that many aspects of ischemic stroke are sexually dimorphic, it is relevant to consider the reasons behind such dimorphism. Differences according to biologic sex may be due to intrinsic factors, i.e., the sex chromosomes, which are consistent throughout the lifespan, or caused by sex steroid hormones, which vary between the genders and fluctuate at various stages of the lifespan. The influence of biologic sex has both organizational, i.e., occurs at a particular period, for example during development, and persists even after the hormonal exposure is removed, and activational effects, which are reversible and persist only while the influencing sex steroid hormone is present. As discussed below, gender differences in the outcome after ischemic stroke can be explained by both hormoneindependent and hormone-dependent mechanisms (reviewed also elsewhere52,53). INTRINSIC MECHANISMS Aside from adult males experiencing a higher incidence of ischemic stroke, this ‘male sensitivity’ to ischemic stroke can also be observed in childhood populations (for example, see references 11,12). As levels of steroid hormones are low in childhood populations, this suggests that other factors, i.e,. non-hormonal, have a role in mediating this sensitivity to ischemic stroke in males. Emerging evidence suggests that the sexually dimorphic response seen to ischemic stroke is partly due to the genetic complement of cells. Sex-specific cultures, in terms of both hippocampal neurons54,55 and astrocytes56,57 have shown that male-derived cells are more sensitive to ischemic injury than female-derived cells. Such cells are derived from neonatal populations and cultured in the absence of hormonal influences thus, confirming that intrinsic cell differences exist between males and females. In addition, several molecular mechanisms of injury, not linked to sexual development or function, have been shown to act dimorphically under ischemic conditions including nitric oxide synthase,58 poly-ADP ribose polymerase (PARP),59 and caspases.60,61 However, the role of sexspecific hormones must also be considered, as revealing their role in determining the effects of ischemic stroke has also identified potential neuroprotective candidates for ischemic stroke treatment. BIOLOGIC MECHANISMS Experimental studies, using both in vitro and in vivo approaches, support the notion that gender differences influence the outcome after induction of ischemic conditions.6 The increased occurrence of ischemic stroke in males has been identified in studies using spontaneously hypertensive, genetically stroke-prone rats.62 It is also well established that female rodents sustain less injury than males after experimental stroke.63–66 The overall neuroprotective effect of female gender on ischemic stroke injury is evident even in the presence of specific stroke risk factors such as diabetes67 and hypertension.63 Such a neuroprotective effect of female gender is abolished by ovariectomy and reproductive senescence, which both cause a decline in the circulating levels of the sex steroid hormones, estrogen and progesterone.63,68 In fact, experimental work has revealed that both estrogens and progesterone are potential neuroprotective factors after ischemic stroke (reviewed elsewhere69,70). However, clinical trials examining the utility of estrogens as an intervention treatment after ischemic stroke have failed to show positive results71,72 and & 2013 ISCBFM

no clinical trials, to date, have been designed to specifically examine the role of progesterone after ischemic stroke. SOCIOCULTURAL FACTORS In recent years, a considerable amount of attention has been focused on the importance of considering social and cultural factors when considering differences in stroke epidemiology. For example, although it is often stated that men are at overall higher risk of ischemic stroke than females, a recent meta-analysis reported that there was a significantly larger male predominance in studies from Australasia and the Americas compared with studies from Europe.29 However, that same review highlighted the fact that the distribution of contributing studies to such metaanalyses is uneven with regions such as Africa, Asia, and South America being underrepresented. Racial and ethnic disparities in stroke incidence and outcome are well documented with, for example, age-adjusted stroke hospitalization rates for blacks being over three times higher than for whites.73 In terms of social factors, a number of studies have shown that the incidence of ischemic stroke increases with lower socioeconomic status in both men and women.74,75 Although, interestingly, adjustment for a variety of classic lifestyle and psychosocial risk factors did not alter the impact of socioeconomic status on stroke risk in men whereas some attenuation of socioeconomic differential was seen after adjustment for risk factors in women.76 MECHANISMS OF INJURY MAY DIFFER ACCORDING TO GENDER Cerebral ischemia triggers a cascade of pathologic events including excitotoxicity, cell necrosis, apoptosis, inflammation, blood–brain barrier breakdown etc., which ultimately culminate in cellular dysfunction and death. Although much progress has been made in dissecting the molecular pathways of pathologic processes such as excitotoxicity, oxidative stress, inflammation, and apoptosis after ischemic stroke, this has largely failed to translate into effective therapies. However, emerging evidence suggests that gender differences occur in certain aspects of the pathologic cascade, induced after cerebral ischemia, including molecular pathways activated, which culminate in cell death and various aspects of the peripheral and brain inflammatory response (reviewed elsewhere77) although gender (and age) are largely overlooked in the design of experimental studies. Ischemic cell death is triggered by an influx of calcium, with subsequent oxidative damage and mitochondrial dysfunction activating several distinct cell death pathways.78 The immediate energy failure seen in the ischemic core leads to swelling of the mitochondria and loss of membrane integrity, typical of necrotic cell death. Reducing necrotic cell death is extremely difficult and likely requires reperfusion during the ischemic insult. However, a delayed apoptotic cell death also occurs in the penumbral region of the ischemic brain. Over the past decade, both caspasedependent and caspase-independent cell-death pathways have been recognized, adding considerable complexity to studies of ischemic cell death.79 Although cell death and apoptosis occur after ischemic injury, it is relevant to consider that the mechanism of injury between the genders could differ. Previous studies have shown that ischemic cell death pathways are different in the male and female brains, females often showing caspase-mediated cell death of individual neurons, whereas males are more sensitive to caspase-independent cell death.77 Apoptotic pathology involves at least two signaling cascades, one initiated through intrinsic, mitochondria-mediated mechanisms involving cytochrome C release, apoptosome assembly, and caspase cleavage, and hence is known as a ‘caspase-dependent’ pathway. An alternative caspase-independent pathway is triggered by postischemic DNA damage and involves the activation of Journal of Cerebral Blood Flow & Metabolism (2013), 1355 – 1361

Cerebral ischemic stroke CL Gibson

1358 Table 1.

Molecular components of cell death pathways, which are proposed to show sexually dimorphic patterns of expression and/or function

Molecule 80

AIF Caspase 360,80 PARP55 Nitric oxide synthase81 Glutathione82 Akt83 Astrocytic aromatase57 GFAP84 Angiotensin II type 2 receptor85 P450 enzyme86

Biologic function in relation to cell death/survival Triggers DNA fragmentation and chromatin condensation Has a central role in the execution of apoptosis after ischemia DNA repair enzyme May produce toxic levels of nitric oxide under ischemic conditions Involved in DNA synthesis and repair Phosphorylation of a variety of proteins (e.g., BAD) associated with apoptosis Induced after ischemic injury and contributes to tissue damage Intermediate filament protein, which maintains astrocyte integrity Mediates neuronal apoptosis Under ischemic conditions, can produce reactive oxygen species resulting in oxidative stress and cell death

AIF, apoptosis-inducing factors; GFAP, glial fibrillary acidic protein; PARP, poly-ADP ribose polymerase.

PARP, release of apoptosis-inducing factor (AIF) from the mitochondria, and translocation of AIF to the nucleus to induce chromatin condensation and large scale DNA fragmentation. Both pathways of apoptosis involve multiple molecular signals, and there are numerous examples of such molecules that have sexually dimorphic roles in cerebral ischemic injury (see Table 1) including; AIF,80 caspase 3,60,80 PARP,55 nitric oxide synthase,81 glutathione,82 Akt,83 astrocytic aromatase,57 glial fibrillary acidic protein,84 angiotensin II type 2 receptor,85 and the P450 enzyme, soluble epoxide hydrolase.86 The DNA repair enzyme PARP-1 induces caspase-independent apoptotic cell death via neuronal nitric oxide synthase (nNOS), PARP-1, and AIF, which all appear to have a key role in regulating cell death in the male brain. For example, although neuronal nitric oxide synthase deletion and inhibition results in neuroprotection in adult male mice, such genetic and pharmacological manipulations of neuronal nitric oxide synthase activity actually exacerbate stroke damage in adult female mice.58 Such a sexually dimorphic response is not affected by hormone exposure and/or replacement suggesting it is either hormone-independent (a sex-specific effect) or secondary to organizational or epigenetic factors induced by early gonadal steroid exposure. In addition, PARP-1 gene deletion is protective in male pups exposed to hypoxic injury but has no effect in female pups.59 In adults, females sustain greater damage when PARP-1 is absent through genetic deletion or inhibited pharmacologically suggesting that activational effects or again female gender itself amplify these sex differences. Although Yuan et al87 demonstrated that, in fact, equivalent activation of the PARP pathway occurs in both sexes after ischemia, the detrimental effects of this pathway seem to be only present in males. Importantly, it appears that AIF translocation and PARP formation do not mediate ischemic injury in the female brain, therefore agents designed to reduce PARP-1 activation, such as minocycline, are unlikely to benefit females.88 A novel cell-death pathway, autophagy, has also been recently implicated in the pathology occurring after ischemic stroke and appears to differ according to gender. Autophagy involves engulfing segments of the cell’s cytoplasm into an autophagosome destined for degradation by the lysosome allowing for recycling of cellular contents.89 Autophagic cell death occurs after neonatal hypoxic injury in male animals as shown by increases in autophagosome formation and lysosome activity. Inhibitors of autophagy are protective in models of neonatal hypoxic injury but seem to preferentially protect male-derived neuronal populations rather than female-derived ones.90 It is widely accepted that inflammation has a key role in contributing to the ensuing damage after ischemic stroke. It has been reported that young female rodents exhibit a decreased inflammatory response to ischemic injury compared with males and a considerable body of evidence strongly implicates estrogens as a significant regulator of inflammatory pathways.4,91 Journal of Cerebral Blood Flow & Metabolism (2013), 1355 – 1361

Gene expression changes in the blood are a reflection of the immune response and a consequence of both environmental and genetic factors. Interestingly, Tian et al92 found gender-associated changes of gene expression in whole blood of humans after ischemic stroke at the whole-genome level. Although this does not inform as to whether these changes represent a cause or an effect of the stroke, it does provide evidence that gender-specific changes in gene expression occur after ischemic stroke. Such gender differences in ischemic stroke may reflect differences in the immune system, inflammatory responses, and cell death between the different genders after ischemic stroke. In addition, because women with Turner syndrome experience an almost three-fold higher risk of cerebrovascular diseases than XX women, it has been suggested that the dosage of X chromosome may be influential in ischemic stroke.93 However, experimental studies report that the number of X chromosomes does not have a significant impact on infarct damage after cerebral ischemia.94 In terms of the pathology produced after cerebral ischemia, attention has traditionally focused on damage to the gray matter, although cerebral white matter is highly vulnerable to the effects of focal ischemia.95 Pathologic changes in oligodendrocytes and myelinated axons occur early after the onset of ischemia and seems to be concomitant with, but independent of, neuronal perikaryal injury.96 It is relevant to consider how white matter function may alter, according to gender, after ischemic injury. Although reports of sexual dimorphism in white matter have been reported, these have largely been restricted to gross anatomic differences and numbers of myelinated axons. Previous studies have reported that significant differences occur in the development of white matter between the genders.97 In the mature nervous system, the average axonal diameter is larger and myelin sheaths are thicker in males than in females.98 However, a study by Cerghet et al99 found that the density of oligodendrocytes in three different white matter tracts (corpus callosum, fornix, and spinal cord) were on average 30% greater in males than females and interestingly, the rate of both oligodendrocyte proliferation and cell death were increased in males compared with females. However, whether gender influences the rate and distribution of oligodendrocyte cell death under pathologic conditions, such as cerebral ischemia, remains to be determined. In addition, although the rate of oligodendrocyte cell death has been shown to differ between the sexes, it remains to be determined whether gender affects the mechanism of cell death for oligodendrocytes, as it seems to for neuronal populations. HORMONES AND CEREBRAL ISCHEMIA Numerous experimental studies have identified the contribution of hormonal influences on the outcome after ischemic stroke. For example, ovariectomized female mice have increased stroke volume compared with females with intact ovarian function.63 & 2013 ISCBFM

Cerebral ischemic stroke CL Gibson

1359 Such observations have led researchers to investigate the potential protective role of steroid hormones as a candidate treatment after ischemic stroke. In experimental studies, treatment with estrogens is neuroprotective in most animal models, and estrogen reduces infarct size when acutely administered to males, ovariectomized females, and reproductively senescent females (reviewed elsewhere100,101). Although estrogen confers protection in experimental stroke, clinical translation of estrogen’s protective effect has been unsuccessful to date.102 Along with estrogens, progesterone levels also dramatically reduce after the menopause and thus, may contribute to the endogenous cardiovascular protection women experience during their premenopausal years. A large number of experimental studies (reviewed elsewhere70) have investigated the neuroprotective potential of progesterone after cerebral ischemia. The majority of these studies have reported a positive effect of progesterone treatment after ischemic stroke and this appears to be consistent across numerous studies including both rats and mice, animals of both genders, and using both transient and permanent models of ischemia. Although as highlighted in a recent systematic review,103 further studies assessing the effectiveness of progesterone need to examine dose responses, time window of treatment effects, longer-term outcomes, and effects of gender and hormonal status of animals used. Because of increased incidence of stroke in males, and worse outcome in male animals after experimental stroke, it has been suggested that androgens may be detrimental to the consequence of ischemic stroke. Recent investigations have focused on the role of testosterone in this male ischemic sensitivity largely because of the clinical observation that elevated endogenous testosterone levels correlate with increased stroke risk in young boys but not in young girls.104 However, in adults it seems that low circulating testosterone levels are associated with increased stroke incidence and worse outcomes after stroke in men.105,106 In animals, low doses of testosterone administered to castrated animals decreases the amount of infarcted tissue after ischemic stroke, whereas infarct damage is exacerbated with higher doses of testosterone.107 The androgen receptor antagonist, flutamide, blocked these effects, suggesting a key role of the androgen receptor mediating the ischemic response to testosterone.108 Such results suggest that while testosterone may contribute to the ‘ischemia-sensitive’ phenotype seen in males such effects are both complex and dose dependent in vivo.109 It is now well established that tissues and organs outside of the reproductive system, i.e., the brain, can also synthesize steroid hormones and this extragonadal production of hormones may also influence stroke sensitivity. In female animals, the absence of aromatase, which converts androgens to estrogens, results in increased infarction area after ischemic stroke.108 In vitro, femalederived astrocytes are protected from oxygen and glucose deprivation compared with male-derived astrocytes, which is abolished by pharmacological inhibition of aromatase56 suggesting that gender differences in estradiol production, by aromatase, may also contribute to the sex differences in sensitivity to cell death after ischemic insult. CEREBRAL ISCHEMIC STROKE—THE IMPORTANCE OF AGE AND GENDER The total incidence of stroke is projected to rise substantially over the next 20 years, because of the rising age of the population. Age is the most important independent risk factor with stroke rates doubling every decade after the age of 55.1 In addition, age is a significant predictor of outcome independent of stroke severity, etiology, efficacy of thrombolysis, gender, and other vascular risk factors.110 In terms of outcome, older patients tend to have higher in-hospital mortality rates as well as poorer functional outcomes. Despite the importance of age and the knowledge that stroke is & 2013 ISCBFM

sexually dimorphic, experimental studies are largely conducted in healthy, young adult males thus, overlooking the impact of both age and gender. However, it appears that in experimental studies, the impact of both age and gender are not straightforward as the consistency of results in experimental stroke studies using aged animals is dependent upon the gender of the animals used. In young rodents, the consensus seems to be that males sustain more histologic damage after experimental stroke than females. However, in aging animals, infarct size and functional outcomes can be worse in females than males.111 Studies involving aging female animals consistently report worse stroke outcome, in terms of increased lesion volume and poorer functional ability, than in younger females regardless of strain and stroke model used (reviewed elsewhere52). However, experimental studies using aged males have produced inconsistent results in terms of determining whether aged males have worse outcomes compared with young males. Some studies report that aged males experience greater ischemic damage, some report smaller ischemic lesions in aged males, and others report that aged males exhibit similar ischemic damage compared with young males.52 Although the aging process itself is associated with a greater risk of mortality and poorer long-term functional outcomes,52,112,113 these detrimental effects, in terms of mortality and longer-term functional ability, seem to be direct consequences of the aging process per se. If the infarct volume is reduced in aged females, by hormone supplementation, to a similar size to that seen in young females, greater functional disability and increased mortality remain in aged females.52 Possible explanations for worse outcome in aged females include the fact that female rodents demonstrate an age-related impairment in astrocyte function, which is not present in males.114 This could directly contribute to the infarct severity by inefﬁcient glutamate clearance and enhanced cytokine production. Thus, normal aging and female gender may both be associated with an increased inﬂammatory response.115

CONCLUSIONS Evidence exists to suggest that gender inﬂuences many aspects of ischemic stroke including stroke risk/incidence, diagnosis, symptoms, treatment and outcomes. It is well documented from epidemiologic studies that during the premenopausal years, female gender is associated with a reduced risk of ischemic stroke and hormonal factors have been investigated as potential protective treatments. However, as discussed in this review, sex differences in ischemic stroke probably result from a combination of factors, including elements intrinsic to the sex chromosomes as well as the effects of sex hormone exposure throughout the lifespan, although we cannot discount the inﬂuence also of cultural and social factors. Research investigating the sexual dimorphism of stroke is only beginning to emerge but several studies suggest that different cell pathways are activated in males and females after ischemic stroke. Stroke researchers must consider the translational relevance of such sex differences as many are unaware of the potential confounding factors of sex differences. The majority of experimental stroke studies continue to focus on using young male animals despite the Stroke Therapy Academy Industry Roundtable (STAIR) recommendations that neuroprotective studies be performed in both male and female rodents.116 A greater understanding of the mechanisms underlying sex differences in stroke and responsiveness to neuroprotection will lead to more appropriate treatment strategies for patients of both sexes.

DISCLOSURE/CONFLICT OF INTEREST The author declares no conﬂict of interest.

Journal of Cerebral Blood Flow & Metabolism (2013), 1355 – 1361

Cerebral ischemic stroke CL Gibson

1360 REFERENCES 1 Rojas JI, Zurru MC, Romano M, Patrucco L, Cristiano E. Acute ischemic stroke and transient ischemic attack in the very old—risk factor profile and stroke subtype between patients older than 80 years and patients aged less than 80 years. Eur J Neurol 2007; 14: 895–899. 2 Seshadri S, Beiser A, Kelly-Hayes M, Kase CS, Au R, Kannel WB et al. The lifetime risk of stroke: estimates from the Framingham Study. Stroke 2006; 37: 345–350. 3 Reeves MJ, Bushnell CD, Howard G, Gargano JW, Duncan PW, Lynch G et al. Sex differences in stroke: epidemiology, clinical presentation, medical care, and outcomes. Lancet Neurol 2008; 7: 915–926. 4 Di Carlo A, Lamassa M, Baldereschi M, Pracucci G, Basile AM, Wolfe CD et al. Sex differences in the clinical presentation, resource use, and 3-month outcome of acute stroke in Europe: data from a multicenter multinational hospital-based registry. Stroke 2003; 34: 1114–1119. 5 Kapral MK, Fang J, Hill MD, Silver F, Richards J, Jaigobin C et al. Sex differences in stroke care and outcomes: results from the Registry of the Canadian Stroke Network. Stroke 2005; 36: 809–814. 6 Simpson CR, Wilson C, Hannaford PC, Williams D. Evidence for age and sex differences in the secondary prevention of stroke in Scottish primary care. Stroke 2005; 36: 1771–1775. 7 Niewada M, Kobayashi A, Sandercock PA, Kaminski B, Czlonkowska A. Influence of gender on baseline features and clinical outcomes among 17,370 patients with confirmed ischaemic stroke in the International Stroke Trial. Neuroepidemiology 2005; 24: 453–460. 8 Gargano JW, Wehner S, Reeves MJ. Delays among patients with acute stroke. do presenting symptoms explain sex differences in emergency department? Stroke 2009; 40: 1114–1120. 9 Rosamond W, Flegal K, Furie K, Go A, Greenlund K, Haase N et al. Heart disease and stroke statistics—2008 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation 2008; 117: e25–146. 10 Anderson KM, Odell PM, Wilson PW, Kannel WB. Cardiovascular disease risk profiles. Am Heart J 1991; 121: 293–298. 11 Fullerton HJ, Wu YW, Zhao S, Johnston SC. Risk of stroke in children. Neurol 2003; 61: 189–194. 12 Golomb MR, Fullerton HJ, Nowak-Gottl U, Deveber G. Male predominance in childhood ischemic stroke: findings from the international pediatric stroke study. Stroke 2009; 40: 52–57. 13 Wang Z, Li J, Wang C, Yao X, Zhao X, Zhao X et al. Gender differences in 1-year clinical characteristics and outcomes after stroke: results from the China national stroke registry. PLoS ONE 2013; 8: e56459. 14 Fukuda M, Kanda T, Kamide N, Akutsu T, Sakai F. Gender differences in long-term functional outcome after first-ever ischemic stroke. Inter Med 2009; 48: 967–973. 15 Elijovich L, Chong JY. Current and future use of intravenous thrombolysis for acute ischaemic stroke. Curr Atheroscler Rep 2010; 12: 316–321. 16 Allen CL, Bayraktutan U. Risk factors for ischaemic stroke. Int J Stroke 2008; 3: 105–116. 17 Hankey GJ. Smoking and risk of stroke. J Cardiovasc Risk 1999; 6: 207–211. 18 Kirshner HS. Differentiating ischemic stroke subtypes: risk factors and secondary prevention. J Neurol Sci 2009; 279: 1–8. 19 Whisnant JP. Effectiveness versus efficacy of treatment of hypertension for stroke prevention. Neurology 1996; 46: 301–307. 20 Arboix A, Oliveres M, Garcia´-Eroles L, Maragall C, Massons J, Targa C. Acute cerebrovascular disease in women. Eur Neurol 2001; 45: 199–205. 21 Palm F, Urbanek C, Olf J, Buggle F, Kleeman T, Hennerici MG et al. Etiology, risk factors and sex differences in ischemic stroke in the Ludwigshafen stroke study, a population-based stroke registry. Cerebrovasc Dis 2012; 33: 69–75. 22 Roquer J, Campello AR, Gomis M. Sex differences in first-ever acute stroke. Stroke 2003; 34: 1581–1585. 23 Fo¨rster A, Gass A, Kern R, Wolf ME, Ottomeyer C, Zohsel K et al. Gender differences in acute ischemic stroke. Etiology, stroke patterns and response to thrombolysis. Stroke 2009; 40: 2428–2432. 24 Touze´ E, Rothwell PM. Sex differences in heritability of ischemic stroke. A systematic review and meta-analysis. Stroke 2008; 39: 16–23. 25 Heim LJ, Brunsell SC. Heart disease in women. Prim Care 2000; 27: 741–746. 26 Lisabeth LD, Brown DL, Hughes R, Majersik JJ, Morgenstern LB. Acute stroke symptoms, comparing women and men. Stroke 2009; 40: 2031–2036. 27 Stuart-Shor E, Wellenius GA, Dellolacono DM, Mittleman MA. Gender differences in presenting and prodromal stroke symptoms. Stroke 2009; 40: 1121–1126. 28 Labiche L, Chan W, Saldin K, Morgenstern L. Sex and acute stroke presentation. Ann Emerg Med 2002; 40: 453–460. 29 Jerath NU, Reddy C, Freeman D, Jerath AU, Brown RD. Gender differences in presenting signs and symptoms of acute ischemic stroke: a population-based study. Gend Med 2011; 8: 312–319. 30 Mandelzweig L, Goldbourt U, Bokyo V, Tanne D. Perceptual, social, and behavioural factors associated with delays in seeking medical care in patient with symptoms of acute stroke. Stroke 2006; 37: 1248–1253.

Journal of Cerebral Blood Flow & Metabolism (2013), 1355 – 1361

31 Kleindorfer DO, Miller R, Moomaw CJ, Alwell K, Broderick JP, Khoury J et al. Designing a message for public education regarding stroke: does FAST capture enough stroke? Stroke 2007; 38: 2864–2868. 32 Appelros P, Stegmayr B, Terent A. Sex differences in stroke epidemiology: a systematic review. Stroke 2009; 40: 1082–1090. 33 Park SJ, Shin SD, Ro YS, Son KJ, Oh J. Gender differences in emergency stroke care and hospital outcome in acute ischemic stroke: a multicenter observational study. Am J Emerg Med 2013; 31: 178–184. 34 Silva GS, Lima FO, Camargo ECS, Smith WS, Lev MH, Harris GJ et al. Gender differences in outcomes after ischemic stroke: role of ischemic lesions volume and intracranial large-artery occlusion. Cerebrovasc Dis 2010; 30: 470–475. 35 Glader EL, Stegmayr B, Norrving B, Terent A, Hulter-Asberg K, Wester PO et al. Sex differences in management and outcome after stroke: a Swedish national perspective. Stroke 2003; 34: 1970–1975. 36 Eriksson M, Asplund K, Glader EL, Norrving B, Stegmayr B, Terent A et al. Selfreported depression and use of antidepressants after stroke. Stroke 2004; 35: 936–941. 37 Glader EL, Stegmayr B, Asplund K. Poststroke fatigue: a 2-year follow-up study of stroke patients in Sweden. Stroke 2002; 33: 1327–1333. 38 Franzen-Dahlin A, Laska AC. Gender differences in quality of life after stroke and TIA: a cross-sectional survey of out-patients. J Clin Nurs 2012; 21: 2386–2391. 39 Gokkaya NK, Aras MD, Cakci A. Health-related quality of life of Turkish stroke survivors. Int J Rehabi Res 2005; 28: 229–235. 40 Sturm JW, Donnan GA, Dewey HM, Macdonell RA, Gilligan AK, Srikanth V et al. Quality of life after stroke: the North East Melbourne Stroke Incidence Study (NEMESIS). Stroke 2004; 35: 2340–2345. 41 Ramani S, Byrne-Logan S, Freund KM, Ash A, Yu W, Moskowitz MA. Gender differences in the treatment of cerebrovascular disease. J Am Geriatr Soc 2000; 48: 741–745. 42 Smith MA, Lisabeth LD, Brown DL, Morgenstern LB. Gender comparisons of diagnostic evaluation for ischemic stroke patients. Neurology 2005; 65: 855–858. 43 Gargano JW, Wehner S, Reeves M. Sex differences in acute stroke care in a statewide stroke registry. Stroke 2008; 39: 24–29. 44 Group SCotPHSR. Final report on the aspirin component of the ongoing physicians’ health study. Steering committee of the Physicians’ Health Study Research Group. N Engl J Med 1989; 321: 129–135. 45 Ridker PM, Cook NR, Lee IM, Gordon D, Gaziano JM, Manson JE et al. A randomized trial of low-dose aspirin in the primary prevention of cardiovascular disease in women. N Engl J Med 2005; 352: 1293–1304. 46 Berger JS, Roncaglioni MC, Avanzini F, Pangrazzi I, Tognoni G, Brown DL. Aspirin for the prevention of cardiovascular events in women and men: a sex-specific meta-analysis of randomized controlled trials. JAMA 2006; 295: 306–313. 47 Poli D, Antonucci E, Grifoni E, Abbate R, Gensini GF, Prisco D. Gender differences in stroke risk of atrial fibrillation patients on oral anticoagulant treatment. Thromb Haemost 2009; 101: 938–942. 48 Hill MD, Kent DM, Hinchey J, Rowley H, Buchan AM, Wechsler LR et al. Sex-based differences in the effect of intra-arterial treatment of stroke: analysis of PROACT-2 study. Stroke 2006; 37: 2322–2325. 49 Kent DM, Buchan AM, Hill MD. The gender effect in stroke thrombolysis: of CASES, controls, and treatment-effect modification. Neurology 2008; 71: 1080–1083. 50 Shah SH, Liebeskind DS, Saver JL, Starkman S, Vinuela F, Duckwiler G et al. Influence of gender on outcomes after intra-arterial thrombolysis for acute ischemic stroke. Neurology 2006; 66: 1745–1746. 51 Jauch EC, Saver JL, Adams HP, Bruno A, Connors JJ, Demaerschalk BM et al. Guidelines for the early management of patients with acute ischemic stroke: a guideline for healthcare professionals from the American Heart Association/ American Stroke Association. Stroke 2013; 44: 870–947. 52 Liu M, Dziennis S, Hurn PD, Alkayed NJ. Mechanisms of gender-linked brain injury. Restor Neurol Neurosci 2009; 27: 163–179. 53 Vagnerova K, Koerner IP, Hurn PD. Gender and the injured brain. Anesthesia and Analgesia 2008; 107: 201–204. 54 Heyer A, Hasselblatt M, von Ahsen N, Hafner H, Siren AL, Ehrenreich H. In vitro gender differences in neuronal survival on hypoxia in 17b-estradiol-mediated neuroprotection. J Cereb Blood Flow Metab 2005; 25: 317–321. 55 Li H, Pin S, Zeng Z, Wang MM, Andreasson KA, McCullough LD. Sex differences in cell death. Ann Neurol 2005; 58: 317–321. 56 Liu M, Hurn PD, Roselli CE, Alkayed NJ. Role of P450 aromatase in sex-specific astrocytic cell death. J Cereb Blood Flow Metab 2007; 27: 135–141. 57 Liu M, Oyarzabal EA, Yang R, Murphy SJ, Hurn PD. A novel method for assessing sex-specific and genotype-specific response to injury in astrocyte culture. J Neurosci Methods 2008; 171: 214–217. 58 McCullough LD, Zeng Z, Blizzard KK, Debchoudhury I, Hurn PD. Ischemic nitric oxide and ploy (ADP-ribose) polymerase-1 in cerebral ischemia: male toxicity, female protection. J Cereb Blood Flow Metab 2005; 25: 502–512.

& 2013 ISCBFM

Cerebral ischemic stroke CL Gibson

1361 59 Hagberg H, Wilson MA, Matsushita H, Zhu C, Lange M, Gustavsson M et al. PARP1 gene disruption in mice preferentially protects males from perinatal brain injury. J Neurochem 2004; 90: 1068–1075. 60 Liu F, Li Z, Li J, Siegel C, Yuan R, McCullough LD. Sex differences in caspase activation after stroke. Stroke 2009; 40: 1842–1848. 61 Renolleau S, Fau S, Goyenvalle C, Jloy LM, Chauvier D, Jacotot E et al. Specific caspase inhibitor Q-VD-OPh prevents neonatal stroke in P7 rat: a role for gender. J Neurochem 2007; 100: 1062–1071. 62 Yamori Y, Horie R, Handa H, Sato M, Fukase M. Pathogenetic similarity of strokes in stroke-prone spontaneously hypertensive rats and humans. Stroke 1976; 7: 46–53. 63 Alkayed NJ, Harukuni I, Kimes AS, London ED, Traystman RJ, Hurn PD. Genderlinked brain injury in experimental stroke. Stroke 1998; 29: 159–166. 64 Hurn PD, Aardelt AA, Alkayed NJ, Crain BJ, Hu W, Kearney ML et al. In: Krieglsein J (eds) Pharmacology of Cerebral Ischemia. Medpharm Scientific Publishers: Stuttgart, 2002 P17–P24. 65 Wise PM, Dubal DB, Wilson ME, Rau SW, Bottner M, Rosewell KL. Estradiol is a neuroprotective factor in the adult and aging bran: understanding of mechanisms derived from in vivo and in vitro studies. Brain Res Brain Res Rev 2001; 37: 313–319. 66 Cheng J, Hurn PD. Sex shapes experimental ischemic brain injury. Steroids 2009; 75: 754–759. 67 Toung TJ, Hurn PD, Traystman RJ, Sieber FE. Estrogen decreases infarct size after temporary focal ischemia in a genetic model of type I diabetes mellitus. Stroke 2000; 31: 2701–2706. 68 Alkayed NJ, Murphy SJ, Traystman RJ, Hurn PD. Neuroprotective effect of female gonadal steroids in reproductively senescent female rats. Stroke 2000; 31: 161–168. 69 Gibson CL, Gray LJ, Murphy SP, Bath PMW. Estrogens and experimental ischemic stroke: a systematic review. J Cereb Blood Flow Metab 2006; 26: 1103–1113. 70 Gibson CL, Coomber B, Rathbone J. Is progesterone a candidate neuroprotective factor for treatment following ischemic stroke? Neuroscientist 2009; 15: 324–332. 71 Viscoli CM, Brass LM, Kernan WN, Sarrel PM, Makuch R, Horwitz RJ. A clinical trial of estrogen-replacement therapy after ischemic stroke. N Engl J Med 2001; 345: 1243–1249. 72 Wassertheil-Smoller S, Hendrix SL, Limacher M, Heiss G, Kooperberg C, Baird A et al. Effect of estrogen plus progestin on stroke in post-menopausal women—the Women’s Health Initiative: a randomized trial. JAMA 2003; 289: 2673–2674. 73 Pathak EB, Sloan MA. Recent racial/ethnic disparities in stroke hospitalizations and outcomes for young adults in Florida, 2001-2006. Neuroepidemiology 2009; 32: 302–311. 74 Avendano M, Kawachi I, Van Lenthe F, Boshuizen HC, Mackenbach JP, Van den Bos GAM et al. Socioeconomic status and stroke incidence in the US elderly: the role of risk factors in the EPESE study. Stroke 2006; 37: 1368–1373. 75 Cox AM, McKevitt C, Rudd AG, Wolfe CDA. Socioeconomic status and stroke. Lancet Neurol 2006; 5: 181–188. 76 McFadden E, Luben R, Wareham N, Bingham S, Khaw K-T. Social class, risk factors, and stroke incidence in men and women. A prospective study in the European prospective investigation into cancer in Norfolk cohort. Stroke 2009; 40: 1070–1077. 77 Siegel C, Turtzo C, McCullough LD. Sex differences in cerebral ischemia: possible molecular mechanisms. J Neurosci Res 2010; 88: 2765–2774. 78 Mehta SL, Manhas N, Rahubir R. Molecular targets in cerebral ischemia for developing novel therapeutics. Brain Res Brain Res Rev 2007; 54: 34–66. 79 Yu SW, Wang H, Poitras MF, Coombs C, Bowers WJ, Federoff HJ et al. Mediation of poly(ADP-ribose) polymerase-1-dependent cell death by apoptosis-inducing factor. Science 2002; 297: 259–263. 80 Zhu C, Xu F, Wang X, Shibata M, Uchiyama Y, Blomgren K et al. Different apoptotic mechanisms are activated in male and female brains after neonatal hypoxia-ischemia. J Neurochem 2006; 96: 1016–1027. 81 Park EM, Cho S, Frys KA, Glickstein SB, Zhou P, Anrather J et al. Inducible nitric oxide synthase contributes to gender differences in ischemic brain injury. J Cereb Blood Flow Metab 2006; 26: 392–401. 82 Du L, Bayir H, Lai Y, Zhang X, Kockanek PM, Watkins SC et al. Innate genderbased proclivity in response to cytotoxicity and programmed cell death. J Biol Chem 2004; 279: 38563–38570. 83 Kitano H, Young JM, Cheng J, Wang L, Hurn PD, Murphy SJ. Gender-specific response to isoflurane preconditioning in focal cerebral ischemia. J Cereb Blood Flow Metab 2007; 27: 1377–1386. 84 Cordeau Jr P, Lalancette-Herbert M, Weng YC, Kriz J. Live imaging of neuroinflammation reveals sex and estrogen effects on astrocyte response to ischemic injury. Stroke 2008; 39: 935–942. 85 Sakata A, Mogi M, Iwanami J, Tsukada K, Min L-J, Fujita T et al. Sex-different effect of angiotensin II type 2 receptor on ischemic brain injury and cognitive function. Brain Res 2009; 1300: 14–23. 86 Zhang W, Iliff JJ, Campbell CJ, Wang RK, Hurn PD, Alkayed NJ. Role of soluble epoxide hydrolase in the sex-specific vascular response to cerebral ischemia. J Cereb Blood Flow Metab 2009; 29: 1475–1481.

& 2013 ISCBFM

87 Yuan M, Siegel C, Zeng Z, Li J, Liu F, McCullough LD. Sex differences in the response to activation of the poly (ADP-ribose) polymerase pathway after experimental stroke. Exp Neurol 2009; 217: 210–218. 88 Li J, McCullough LD. Sex differences in minocycline-induced neuroprotection after experimental stroke. J Cereb Blood Flow Metab 2009; 29: 670–674. 89 Shintani T, Klionsky DJ. Autophagy in health and disease: a double-edged sword. Science 2004; 306: 990–995. 90 Puyal J, Vaslin A, Mottier V, Clarke PG. Postischemic treatment of neonatal cerebral ischemia should target autophagy. Ann Neurol 2009; 66: 378–389. 91 Maggi A, Ciana P, Belcredito S, Vegeto E. Estrogens in the nervous system: mechanisms and nonreproductive functions. Annu Rev Physiol 2004; 66: 291–313. 92 Tian Y, Stamova B, Jickling GC, Liu D, Ander BP, Bushnell C et al. Effects of gender on gene expression in the blood of ischemic stroke patients. J Cereb Blood Flow Metab 2012; 32: 780–791. 93 Schoemaker MJ, Swerdlow AJ, Higgins CD, Wright AF, Jacobs PA. Mortality in women with turner syndrome in Great Britain: a national cohort study. J Clin Endocrinol Metab 2008; 93: 4735–4742. 94 Turtzo LC, McCullough LD. Sex differences in stroke. Cerebrovasc Dis 2008; 26: 462–474. 95 Pantoni L, Garcia JH, Gutierrez JA. Cerebral white matter is highly vulnerable to ischemia. Stroke 1996; 27: 1641–1647. 96 Po C, Kalthoff D, Kim YB, Nelles M, Hoehn M. White matter reorganization and functional response after focal cerebral ischemia in the rat. PLOS One 2012; 7: e45629. 97 Clayden JD, Jentschke S, Munoz M, Cooper JM, Chadwick JM, Banks T et al. Normative development of white matter tracts: similarities and differences in relation to age, gender, and intelligence. Cereb Cortex 2012; 22: 1738–1747. 98 Yang S, Li C, Zhang W, Wang W, Tang Y. Sex differences in the white matter and myelinated nerve ﬁbers of Long-Evans rats. Brain Res 2008; 1216: 16–23. 99 Cerghet M, Skoff RP, Bessert D, Zhang Z, Mullins C, Ghandour MS. Proliferation and death of oligodendrocytes and myelin proteins are differentially regulated in male and female rodents. J Neurosci 2006; 26: 1439–1447. 100 Liu R, Yang SH. Window of opportunity: estrogen as a treatment for ischemic stroke. Brain Res 2013; 1511: 83–90. 101 Suzuki S, Brown CM, Wise PM. Neuroprotective effects of estrogens following ischemic stroke. Front Neuroendocrinol 2009; 30: 201–211. 102 Henderson VW, Lobo RA. Hormone therapy and the risk of stroke: perspectives 10 years after the Women’s Health Initiative trials. Climacteric 2012; 15: 229–234. 103 Gibson CL, Gray LJ, Bath PMW, Murphy SP. Progesterone for the treatment of experimental brain injury: a systematic review. Brain 2008; 131: 318–328. 104 Vannucci SJ, Hurn PD. Gender differences in pediatric stroke: is elevated testosterone a risk factor for boys? Ann Neurol 2009; 66: 713–714. 105 Jeppesen LL, Jorgensen HS, Nakayam H, Raaschou HO, Olsen TS, Winther K. Decreased serum testosterone in men with acute ischemic stroke. Arterioscler Thromb Vasc Biol 1996; 16: 749–754. 106 Yeap BB, Hyde Z, Almeida OP, Norman PE, Chubb SA, Jamrozik K et al. Lower testosterone levels predict incident stroke and transient ischemic attack in older men. J Clin Endocrinol Metab 2009; 94: 2353–2359. 107 Cheng J, Hu W, Toung TJ, Zhang Z, Parker SM, Roselli CE et al. Age-dependent effects of testosterone in experimental stroke. J Cereb Blood Flow Metab 2009; 29: 486–494. 108 Uchida M, Palmateer JM, Herson PS, DeVries AC, Cheng J, Hurn PD. Dosedependent effects of androgens on outcome after focal cerebral ischemia in adult male mice. J Cereb Blood Flow Metab 2009; 29: 1454–1462. 109 McCullough LD, Blizzard K, Simpson ER, Oz OK, Hurn PD. Aromatase cytochrome P450 and extragonadal estrogen play a role in ischemic neuroprotection. J Neurosci 2003; 23: 8701–8705. 110 Moulin T, Tatu L, Vuiller F, Berger E, Chvot D, Rumbach L. Role of a stroke data bank in evaluating cerebral infarction subtypes: patterns and outcomes of 1,776 consecutive patients from the Besancon stroke registry. Cerebrovasc Dis 2000; 10: 261–271. 111 Liu F, Yuan R, Benashski SE, McCullough LD. Changes in experimental stroke outcome across the life span. J Cereb Blood Flow Metab 2009; 29: 792–802. 112 Shapira S, Sapir M, Wengier A, Grauer E, Kadar T. Aging has a complex effect on a rat model of ischemic stroke. Brain Res 2002; 925: 148–158. 113 Wang RY, Wang PS, Yang YR. Effect of age in rats following middle cerebral artery occlusion. Gerontology 2003; 49: 27–32. 114 Lewis DK, Thomas KT, Selvamani A, Sohrabji F. Age-related severity of focal ischemia in female rats is associated with impaired astrocyte function. Neurobiol Aging 2012; 33: 1123.e1–1123.e16. 115 Mouton PR, Long JM, De-Liang L, Howard V, Jucker M, Calhoun ME et al. Age and gender effects on microglia and astrocyte numbers in brains of mice. Brain Res 2002; 956: 30–35. 116 Fisher M, Feuerstain G, Howells DW, Hurn PD, Kent TA, Savitz SI et al. Update of the stroke therapy academic industry roundtable preclinical recommendations. Stroke 2009; 40: 2244–2250.

Journal of Cerebral Blood Flow & Metabolism (2013), 1355 – 1361

ARTICLE

doi:10.1038/nature13165

Capillary pericytes regulate cerebral blood flow in health and disease Catherine N. Hall1*, Clare Reynell1*, Bodil Gesslein2*, Nicola B. Hamilton1*, Anusha Mishra1*, Brad A. Sutherland3, Fergus M. O’Farrell1, Alastair M. Buchan3, Martin Lauritzen2,4 & David Attwell1

Increases in brain blood flow, evoked by neuronal activity, power neural computation and form the basis of BOLD (bloodoxygen-level-dependent) functional imaging. Whether blood flow is controlled solely by arteriole smooth muscle, or also by capillary pericytes, is controversial. We demonstrate that neuronal activity and the neurotransmitter glutamate evoke the release of messengers that dilate capillaries by actively relaxing pericytes. Dilation is mediated by prostaglandin E2, but requires nitric oxide release to suppress vasoconstricting 20-HETE synthesis. In vivo, when sensory input increases blood flow, capillaries dilate before arterioles and are estimated to produce 84% of the blood flow increase. In pathology, ischaemia evokes capillary constriction by pericytes. We show that this is followed by pericyte death in rigor, which may irreversibly constrict capillaries and damage the blood–brain barrier. Thus, pericytes are major regulators of cerebral blood flow and initiators of functional imaging signals. Prevention of pericyte constriction and death may reduce the long-lasting blood flow decrease that damages neurons after stroke.

Pericytes are isolated contractile cells on capillaries that may regulate cerebral blood flow1,2, in addition to stabilizing newly formed capillaries3, maintaining the blood–brain barrier4–6, contributing to the ‘glial scar’ in pathology7, and having stem cell properties8. Pericytes can be constricted and dilated by neurotransmitters in vitro1,9, through poorly understood signalling pathways, and capillary blood flow heterogeneity might reflect differences in pericyte tone10–12. Pericytes can constrict in vivo, but it was suggested2 that they do not relax actively to increase blood flow13. Similar controversy surrounds the effect of pericytes on blood flow in pathology14,15. We have now characterized the responses of pericytes in the neocortex and cerebellum to neuronal activity and ischaemia. We demonstrate that, unexpectedly, pericytes are the first vascular elements to dilate during neuronal activity, making them the initiators of functional imaging signals. Furthermore, they die readily in ischaemia, and this is expected to promote brain damage.

Signals that regulate pericyte dilation We assessed the signalling systems that dilate molecular layer capillaries in cerebellar slices1, using 95% or 20% O2 in the superfusate to produce a supra-normal or a physiological O2 concentration, [O2], in the slice16,17. Capillaries were defined as vessels of less than 10 mm diameter and lacking a continuous layer of smooth muscle. Their diameter was larger (P 5 0.00023) at the lower [O2] (5.36 6 0.30 mm (mean 6 s.e.m., n 5 59) in 20%, and 4.07 6 0.14 mm (n 5 154) in 95% O2). Capillary pericytes can be identified by labelling for the chondroitin sulphate proteoglycan NG2 or the PDGFRb receptor (Fig. 1a), or using mice expressing DsRed under control of the NG2 promoter18 (Fig. 1b, c), and are a different cell class from Iba1-expressing perivascular microglia or macrophages19 (Fig. 1c). For neocortical capillaries in postnatal day 21 (P21) rats the pericyte soma density was 2.2 6 0.2 per 100 mm of capillary length (950 mm of capillary analysed in each of 11 confocal stacks). Pericytes extend processes along and around vessels (Fig. 1a–c), which regulate capillary diameter.

The capillary diameter was extremely stable over the course of 60 min if no drugs were applied (see below). Application of noradrenaline (2 mM), to mimic its release from the locus coeruleus in vivo, produced a sustained constriction mediated by pericytes1 (Fig. 1d, e), that was not affected by [O2] (Fig. 1f). Superimposing glutamate (500 mM), to mimic neuronal glutamate release13, dilated capillaries at pericyte locations1 (Fig. 1d, g and Supplementary Video 1). As a percentage of the diameter without drugs, this dilation was twice as large with 20% as with 95% O2 (Fig. 1h), possibly due to less production of vasoconstricting 20-HETE (20-hydroxyeicosatetraenoic acid) in low [O2] (see below and Extended Data Fig. 1). Independent of [O2], most pericytes (72% in 95% O2; 70% in 20% O2) constricted .5% to noradrenaline, and dilated .5% to glutamate (63% in 95% O2 (n 5 154); 66% in 20% O2 (n 5 59)), and most pericytes that constricted to noradrenaline also dilated to glutamate (71% in 95% O2; 78% in 20% O2). Glutamate also dilated some capillaries in the absence of noradrenaline: in 20% O2, 29% (7/24) of pericytes dilated .5%, which is less (P 5 0.005, chi-squared test) than the 66% of pericytes that dilated .5% after applying noradrenaline. In subsequent experiments noradrenaline was used to pre-constrict capillaries, to aid analysis of the signalling underlying glutamate-evoked dilation. These experiments do not establish which cells the noradrenaline and glutamate act on (they may be neurons, astrocytes or pericytes themselves13) to release the downstream messengers that control pericyte tone. However, we can rule out the idea that noradrenaline generates 20-HETE to constrict pericytes, because blocking 20-HETE synthesis with 1 mM HET0016 did not affect the noradrenaline-evoked constriction (Extended Data Fig. 2b, g). Glutamate releases nitric oxide (NO), a vasodilator, by activating NMDA (N-methyl-D-aspartate) receptors13 and an NO donor (DETA-NONOate, 100 mM) evoked capillary dilation (Extended Data Fig. 2m). Blocking NO synthase with L-NG-nitroarginine (L-NNA, 100 mM) reduced the glutamate-evoked dilation (Fig. 1i, l; P 5 0.002 from analysis of

Department of Neuroscience, Physiology and Pharmacology, University College London, Gower Street, London, WC1E 6BT, UK. 2Department of Neuroscience and Pharmacology and Center for Healthy Aging, University of Copenhagen, DK-2200 Copenhagen N, Denmark. 3Acute Stroke Programme, Radcliffe Department of Medicine, University of Oxford, Oxford OX3 9DU, UK. 4Department of Clinical Neurophysiology, Glostrup University Hospital, DK-2600 Glostrup, Denmark. *These authors contributed equally to this work. 3 A P R I L 2 0 1 4 | VO L 5 0 8 | N AT U R E | 5 5

RESEARCH ARTICLE a

d 10 μm

20 μm

Control

NA + Glut 20 μm

IB4 IB4 NG2 PDGFRβ 10 μm NG2-DsRed

NG2-DsRed Iba1

Merge Merge

2 0

5 Time (min)

L-NNA NA

60 40 20 59

154

P = 0.02

Δ Diameter in glut (%)

l 20

131 30

–10

o Diameter (μm)

Control L-NNA

P = 0.04

95 O2 (%)

L161,982 NA Glut

4 3 2 1 0

5 10 15 Time (min)

2 4 Time (min)

Control ODQ

131 28

P = 0.05

154

20 95 O2 (%)

Glut

10 15 20 25 Time (min)

HET HET + L-NNA

30 20 10

q Control L161,982

10 19

95 O2 (%)

P = 0.01 30

O2 (%)

HET0016 L-NNA NA

5 10 15 20 Time (min)

Glut

Δ Diameter in glut (%)

Diameter (μm)

Glut

ODQ NA

Glut

20 95 O2 (%)

Diameter (μm)

P = 0.01

Δ Diameter in glut (%)

P = 0.68

Δ Diameter in glut (%)

Diameter in NA (%)

Diameter (μm)

NA 6

100

Diameter (μm)

Δ Diameter in glut (%)

10 μm

P = 0.004

4 35

–4

O2 (%)

Figure 1 | Signalling pathways that control capillary diameter. a, Capillaries in the molecular layer of rat cerebellum labelled using isolectin B4; pericytes (arrows) labelled with NG2 or PDGFRb antibodies. b, Cerebellar capillaries in NG2-DsRed mouse; pericytes are red. c, Neocortical capillaries in NG2-DsRed mouse; microglia labelled for Iba1. d, Rat cerebellar capillary response to 2 mM noradrenaline (NA) and superimposed 500 mM glutamate (Glut). Line shows lumen diameter. e, NA-evoked prolonged constriction (95% O2; a large constriction is shown for clarity). f, Diameter in NA was not affected by [O2] (graphs show percentage of baseline diameter before drugs). g, Glutamate dilates capillaries (20% O2; a large dilation is shown for clarity). h, Dilation was larger in low [O2]. i, NOS blocker L-NNA (100 mM) inhibits Glut-evoked dilation (20% O2). j, Guanylyl cyclase blocker ODQ (10 mM) does not block dilation (20% O2). k, Blocking 20-HETE production (HET0016, 1 mM) abolishes the inhibitory effect of L-NNA (20% O2). l, L-NNA reduces glutamate-evoked dilations at high and low [O2] (P 5 0.002, ANOVA on all data; P values at each [O2] from post-hoc t-tests). m ODQ does not affect glutamate-evoked dilation. n, HET0016 abolishes effect of L-NNA. o, p, Blocking EP4 receptors (L161,982, 1 mM) inhibits glutamate-evoked dilation (o, 20% O2) at high and low [O2] (p). Data in d–p are from rat cerebellar capillaries. q, EP4 block abolishes glutamate-evoked dilation in rat neocortical capillaries (20% O2). Drug effects on baseline diameter are in Extended Data Fig. 2. Data shown as mean 6 s.e.m.

variance (ANOVA); L-NNA and other blockers used did not inhibit the noradrenaline-evoked constriction, see Extended Data Fig. 2). Unexpectedly, the dilation was not affected by blocking guanylyl cyclase with ODQ (1H-[1,2,4]Oxadiazolo[4,3-a]quinoxalin-1-one, 10 mM, P 5 1, ANOVA), so NO does not act by raising [cGMP] in the pericyte (Fig. 1j, m

and Extended Data Fig. 2e). However, when production of vasoconstricting 20-HETE was blocked using HET0016 (1mM), L-NNA no longer inhibited the dilation (Fig. 1k, n; P 5 0.0005 from ANOVA on black bars in l and n), implying that NO promotes dilation by preventing 20-HETE formation. As a robust dilation occurs with both NO and 20-HETE synthesis blocked (Fig. 1k, n), another messenger must be active. Blocking synthesis of epoxy derivatives of arachidonic acid with the inhibitor MS-PPOH (10 mM) did not affect the dilation (Extended Data Fig. 2i, ANOVA P 5 0.92), but blocking EP4 receptors for prostaglandin E2 (with 1 mM L-161,982) greatly reduced it (Fig. 1o–p, P 5 0.001, ANOVA). A similar inhibition of capillary dilation by blocking EP4 receptors was seen in neocortical pericytes (Fig. 1q, P 5 0.004). Application of prostaglandin E2 (PGE2) itself dilated cerebellar capillaries (Extended Data Fig. 2n). We therefore identify the messenger that dilates capillaries in response to glutamate as prostaglandin E2 (or a related species active at EP4 receptors), but this dilation requires NO release to suppress 20HETE formation (Extended Data Fig. 1b). Glutamate (500 mM) or NMDA (100 mM) (Fig. 2a, b, d), or parallel fibre stimulation (Fig. 2c, d), evoked an outward membrane current in pericytes patch-clamped at 255 to 275 mV (sometimes preceded by a small inward current; Extended Data Fig. 3). The stimulation-evoked outward current was inhibited (P 5 0.005, paired t-test) by blocking action potentials with tetrodotoxin (TTX) (Fig. 2c, reduced to 15 6 12% (n 5 4) of its amplitude without TTX, not significantly different from zero, P 5 0.28), whereas the NMDA-evoked current was unaffected by TTX (reduced by 12 6 19%, n 5 5, P 5 0.56), consistent with stimulation evoking the outward current by generating action potentials that release glutamate. The stimulation-evoked current of approximately 30 pA is expected to hyperpolarize the cells by approximately 9 mV (see Methods) and decrease voltage-gated Ca21 entry, causing active relaxation9 (although other Ca21 sources, and cyclic nucleotides, may also regulate contractile tone9,20). Indeed, parallel fibre stimulation produced a dilation of 14.9 6 3.1% in 21 capillaries in 20% O2 (Fig. 2e–h and Supplementary Video 2), which (unlike the noradrenaline-evoked constriction) was blocked by TTX and by blocking EP4 receptors (Fig. 2g, h). An outward current at negative potentials is not consistent with activation of glutamatergic ionotropic receptors, but is consistent with K1 current activation. These data suggest that glutamate release generates PGE2, which dilates the capillaries by activating an outward K1 current in pericytes. PGE2 similarly activates an outward K1 current in aortic smooth muscle21, and relaxes kidney pericytes22.

Pericytes increase blood flow in vivo To assess whether pericyte relaxation regulates blood flow in vivo, we electrically stimulated the whisker pad (at 3 Hz) and used two-photon imaging of the vasculature (labelled with FITC-dextran) in somatosensory cortex to monitor dilations of penetrating arterioles (entering the cortex from the pia) and capillaries, in anaesthetized mice expressing DsRed in pericytes (Fig. 3a). The capillary diameter in vivo was 4.4 6 0.1 mm in 633 capillary regions (Extended Data Table 1). Pericytes were visualized up to 200 mm deep in the cortex (layer 2/3). Brief whisker pad stimulation (2 s) evoked vessel dilations that peaked just after the end of the stimulation (,2.5 s, Fig. 3b). Longer (15 s) stimulation produced dilations that initially followed the same time course, then dilated further throughout the stimulation (Fig. 3b and Extended Data Fig. 4). Most imaging employed 15 s stimuli, which increased the response magnitude and measurement accuracy. Repeated stimulation gave reproducible responses (Extended Data Fig. 5a, b). We segmented the vasculature by the branching order of the vessels, zero being the penetrating arteriole, one being the primary capillary branching off the arteriole, and so on (Extended Data Fig. 1; see Extended Data Table 1 for resting diameters, dilations and numbers of each vessel order). Whiskerpad stimulation dilated vessels of all orders (Fig. 3c and Extended Data Table 1). The fraction of vessels responding (that is, with a dilation .5%) was similar in penetrating arterioles and in first-order capillaries, whereas

5 6 | N AT U R E | VO L 5 0 8 | 3 A P R I L 2 0 1 4

ARTICLE RESEARCH

–30

10 0 12 Hz

–10

12 Hz

10 Time (min)

40 20

Stim

Glut

NMDA

100

First-order capillary

Zero-order penetrating arteriole

0 0

Control TTX L161,982

Diameter in NA (%)

100

8 7 6 0

5 10 Time (min)

Diameter (% of peak)

10 μm

12 Hz

12 10 15 10

–1

4 s 1 vs 2 vs 3 vs ≥ 2 3 0v 1 Branching orders compared

All vessels

First-order capillary

Zero-order arteriole

0 0

Zero-order Zero order arteriole arteriole

Paired vessels

Time (s)

Second order Second-order capillary capillary

All vessels P = 0.039

P = 0.040

Time (s)

P = 0.80

13 52 89 0

0 1 2 Branching order

Time (s)

10 5 7

Second-order capillary

P = 0.017

P = 0.01

First-order First order capillary

100

15 s stim

Zero order Zero-order penetrating penetrating arteriole arteriole

Time (s)

–5

Figure 2 | Pericyte membrane current and capillary dilation in cerebellar slices. a, DsRed-labelled patch-clamped pericyte in molecular layer of mouse cerebellum. Lucifer yellow from pipette overlaps with DsRed. b, Glutamate (Glut, 500mM) and NMDA (100mM) evoked outward current (255mV). c, Parallel fibre stimulation (Stim, 12 Hz)-evoked outward current (274mV) is blocked by 1 mM TTX. d Mean outward currents evoked by stimulation, glutamate and NMDA (b–d are in 95% O2). e, f, Parallel fibre stimulation (in NA 1 mM) in rat cerebellar slice evokes capillary dilation (20% O2). g, h, Constriction by NA (g) was unaffected by TTX or EP4 block (L161,982), which abolished the stimulation-evoked dilation (h). P values from one-way ANOVA with Dunnett’s post-hoc tests. NA, noradrenaline. Data shown as mean 6 s.e.m.

the frequency of capillary responses decreased with increasing order (Fig. 3c). To establish where vasodilation is initiated, we imaged different orders of vessel simultaneously. Strikingly, first-order capillaries usually dilated before penetrating arterioles (Fig. 3d, e and Supplementary Video 3), with vasodilation onset (assessed as the time to 10% of the maximum dilation) in the capillary being 1.38 6 0.38 s earlier than for the penetrating arteriole (Fig. 3e, f, P 5 0.015). Further along the vascular tree there was no significant difference in the time to dilation of simultaneously imaged capillaries of adjacent order (Fig. 3f and Extended Data Fig. 5c). Thus, capillaries dilate before the penetrating arteriole feeding them. Averaging over all vessels of the same order (not just those imaged simultaneously) showed a similar faster dilation of capillaries than of penetrating arterioles (Fig. 3g), with the time to 10% of the maximum dilation for penetrating arterioles (3.7 6 0.3 s) being significantly longer

P = 7.5 × 10–11 Response frequency (%)

Δ Diameter to stim. (%)

NA + Stim

Diameter (μm)

First-order 1st order capillaries capillaries

Diameter (% of peak)

464 168

yte icyte ric Pe -per n No

Cumulative probability (%)

TTX

0 1 2 3 ≥4 Branching order

Paired vessels

100

80 Outward current (pA)

Current (pA)

e 10 μm First order Zero order

Diameter (% of peak)

10 Time (min)

Time to 10% of peak (s)

20 μm

10 20 Time (s)

205

164 164

10 μm

NG2-DsRed

P = 0.83

50 μm

First order

P = 0.70

Zero order

100

Second order Second order

2 s stim

P = 0.015

15 s stim

FITC-dextran

P = 0.98

Percentage responding

Time difference to 10% of max (s)

NMDA

100

Non-pericyte P = 3.2 × 10–5

75 50

Pericyte

25 0 –20

0 20 40 Δ Diameter (% of initial diameter)

Δ Diameter on stimulation (% of initial diameter)

Lucifer yellow

Glut

120

Δ Diameter on stimulation (% of initial diameter)

NG2-DsRed

Current (pA)

P = 8.0 × 10–13 8

464 168 yte yte ric eric e P n-p No 0

Figure 3 | Active dilation of capillaries by pericytes in vivo in mouse cerebral cortex. a, Confocal stack (90 mm thick, maximum intensity projection) of FITC-dextran-filled vessels in vivo in somatosensory cortex of NG2-DsRed mouse (pericytes are red). Enlargement (single image) shows a penetrating arteriole (0th order) giving off a capillary (first order) which splits into second-order branches. b, Response of 45 capillary regions to 2-s and 15-s whisker-pad stimulation. c, Percentage of vessel regions of different orders (number studied on bars) showing .5% dilation to stimulation. d, Simultaneous imaging (top, lines show measurement loci) of penetrating arteriole and first-order capillary: capillary dilates 3 s before arteriole (bottom, smoothing in d–g is explained in Methods and Extended Data Fig. 4). e, Dilation time course in simultaneously imaged penetrating arterioles and first-order capillaries. f, Time to 10% of peak dilation for (j–1)th order (third order for j $ 4) vessel minus that of jth-order vessel (j is an integer $ 1). Capillaries dilate faster than arterioles. g Dilation time course in all responding (.5%) penetrating arterioles and first- and second-order capillaries. Inset expands initial response. h, Time to 10% of peak dilation in all zero- to second-order responding vessels. i, Percentage of capillary locations with or without pericytes showing .5% dilations. j, Cumulative probability of capillary diameter changes (including ‘non-responding’ capillaries with ,5% dilations) in 464 pericyte and 168 non-pericyte locations. Diameter changes ,0% (constrictions) represent random changes and measurement error. k, Mean responses for distributions in j (P values from Mann–Whitney U-tests). Data shown as mean 6 s.e.m.

than the values (,2.7 s) obtained for first- and second-order capillaries (P 5 0.040 and 0.039 respectively, Fig. 3h and Extended Data Fig. 5d). As expected, the time course of the blood flow increase in capillaries, 3 A P R I L 2 0 1 4 | VO L 5 0 8 | N AT U R E | 5 7

RESEARCH ARTICLE a Control

Before

14 min

18 min

22 min

PI (60 min)

Ischaemia

2 P

EC 4

b Diameter (%)

120

Control

100

4 2

3 Ischaemia from t = 0

40 20 0 0

20 Time (min)

120

Ischaemia from t = 0

1 20

Control

Dead pericytes per 100 μm

Control

100 Diameter (%)

0 0

20 Time (min)

Figure 4 | In ischaemia, pericytes constrict capillaries and then die in rigor. a, Top, capillary in a P21 rat cortical slice in normal solution. Bottom, capillary exposed to simulated ischaemia. Right, propidium iodide (PI, 7.5 mM) labelling after one hour, showing dead pericytes (P) and an endothelial cell (EC). b, Vessel diameters at regions indicated in a, over time (red lines are for diameter near pericytes that die). c, Mean diameter and number of dead pericytes per 100 mm from nine capillaries in ischaemia (measured at 18 locations) and six capillaries in normal solution (13 locations). Diameter in ischaemia was reduced from control (P 5 1.3310215 and 7.7 3 10217 at 30 and 60 min). Pericyte death was higher in ischaemia at 40 and 60 min (P 5 0.041 and 0.021, Mann–Whitney U-test). A few pericytes also died on capillaries in non-ischaemic slices, but did not constrict capillaries (1.3 6 1.5% diameter decrease at three dead pericytes on two vessels). Data shown as mean 6 s.e.m.

assessed from the speed of red blood cell movement with line-scanning23, was similar to that of the capillary dilation (Extended Data Fig. 5f). The faster dilation in capillaries compared to arterioles indicates that capillary dilation is not a passive response to a pressure increase produced by arteriole dilation. To assess whether pericytes generate this dilation, we measured the diameter changes of capillaries at locations where DsRedlabelled pericytes were present (either somata or processes, responses did not differ significantly at these locations; Extended Data Fig. 5e) or where no pericyte was visible. The resting diameter of capillaries was larger where pericyte somata or processes were present (4.62 6 0.09 mm, n 5 464) than in pericyte-free zones (3.72 6 0.08 mm, n 5 168, Mann– Whitney P 5 2.731027), suggesting that pericytes induce an increase of capillary diameter. Dilations over 5% were much more frequent at pericyte locations (Fig. 3i; P 5 7.53 10211, chi-squared test), where the responses were also larger (P 5 3.23 1025, Kolmogorov–Smirnov test; Fig. 3j, k). These data confirm that pericytes actively relax to generate the capillary dilation.

Pericytes constrict and die in ischaemia A key question is whether pericyte control of capillary diameter also has a role in pathology. Pericytes constrict some retinal capillaries in ischaemia1, perhaps because pericyte [Ca21]i rises when ion pumping is inhibited by ATP depletion. Cortical capillaries also constrict following middle cerebral artery occlusion14 (MCAO) in vivo. Clinically, reperfusion of thrombus-occluded arteries using tissue plasminogen activator can be achieved in 1 to 6 h24,25 but, even when arterial flow is restored, a long-lasting reduction of cerebral blood flow can ensue26–29. This may reflect pericyte constriction outlasting ischaemia14, but it is unclear why the constriction is so prolonged. We examined the effect of ischaemia on pericyte health using propidium iodide to mark cell death. Live imaging of cerebral cortical slices exposed to simulated ischaemia (oxygen-glucose deprivation with ATP synthesis by glycolysis and oxidative phosphorylation inhibited with iodoacetate and antimycin) revealed that, within approximately 15 min, grey matter capillaries constricted at spatially restricted regions near pericytes (Fig. 4a–c). In contrast, the diameter of capillaries not exposed to ischaemia was stable over 60 min (reduced by 3.2 6 3.0%, P 5 0.31, n 5 13; Fig. 4c). Although capillaries that were not exposed to ischaemia showed little pericyte death even after 1 hour (as assessed by propidium iodide labelling), ischaemia led to most pericytes on capillaries dying after approximately 40 min, usually at locations at which the earlier constriction had occurred. All capillaries exposed to ischaemia that we examined showed a consistent response, in which the pericytes first constricted the capillaries, and then died (Fig. 4c). Death of pericytes in rigor, after they have been constricted by a loss of energy supply, should produce a long-lasting increase in the resistance of the capillary bed. To sample more pericytes than is possible while live imaging the capillary diameter, and to examine mechanisms contributing to their death, we acquired confocal stacks of brain slices exposed to simulated ischaemia, which were then fixed and labelled for NG2 and/or isolectin B4. Ischaemia led to pericytes apposed to capillaries dying rapidly in both the white matter of the cerebellum (Fig. 5a, b) and the grey matter of the cerebral cortex (Fig. 5c, d). For pericytes exposed to simulated ischaemia as above, approximately 90% of pericytes died within 1 hour (Fig. 5b). This was unaffected by blocking action potentials (with TTX, 1mM) but was halved by blocking AMPA(a-amino-3-hydroxy-5-methyl4-isoxazole propionic acid)/kainate (25mM NBQX) or NMDA receptors (50mM D-AP5, 50mM MK-801 and 100mM 7-chlorokynurenate), implying an excitotoxic contribution to pericyte death. When oxygenglucose deprivation (OGD) alone was used (without antimycin and iodoacetate), to allow ATP generation when oxygen and glucose were restored, approximately 40% of pericytes died after 1 hour, but OGDevoked death increased 1.5-fold during 1 h of reperfusion (Fig. 5c, d). Ionotropic glutamate receptor block or removal of external Ca21 again significantly reduced the death (Fig. 5d), while blocking NO production had a small protective effect and lowering free radical levels by scav? enging O2 2 had no significant effect (Fig. 5d and Extended Data Fig. 6a). Blocking metabotropic glutamate receptors or 20-HETE production, which might prevent [Ca21]i rises, or blocking mitochondrial calcium uptake, also had no effect (Extended Data Fig. 6b). Pericytes apposed to capillaries, unlike endothelial cells, also died in vivo after 90 min of MCAO (followed by 22.5 h recovery: Fig. 5e, f, and Extended Data Fig. 6c). In contrast, a sham operation occluding only the internal carotid artery (see Methods) produced less pericyte death, and a sham operation with no artery occlusion (which did not affect blood flow; see Methods) induced no more death than in naive untreated animals. Thus, pericyte death is a rapid response of the cerebral vascular bed to ischaemia, both in brain slices and in vivo.

Discussion Understanding what initiates the blood-flow increase in response to neuronal activity is crucial for understanding both how information processing is powered and how functional imaging signals are generated30. Most neurons are closer to capillaries (,8.4 mm distant, in hippocampus31)

5 8 | N AT U R E | VO L 5 0 8 | 3 A P R I L 2 0 1 4

ARTICLE RESEARCH a

Control

P = 0.003 P = 0.001

10 μm

Control 20 μm

Pericytes (% dead)

P = 0.77

NG2 PI Ischaemia

100 75

Pericytes (% dead)

iGluR blockers

IB4 PI NG2 Neocortical grey matter

0 [Ca2+]0

Scavengers of ROS

L-NNA

13 13

6 6

18 18 18 18

-re

f 50 μm

IB4 PI

8 8

Cells (% dead)

Control hemisphere

-r No

Lesioned hemisphere

ol ia X ck k ntr em TT blo bloc Co scha R R I PA DA AM NM

CON OGD

OGD

Cerebellar white matter

P = 4.5 × 10–8

-re

11 11 11 11

-re No

11 13

-re

Control hemisphere Lesioned hemisphere MCAO Naive Sham Sham without ICA with ICA occlusion occlusion

3 3 25

3 3

6 6

3 3

6 6

s s s l l s l l yte helia icyte helia icyte helia icyte helia ric er er ot ot er ot ot d Pe P P d d d P En En En En

Figure 5 | Pericyte death in ischaemia. a, Rat cerebellar slice white matter capillaries labelled for NG2 and propidium iodide (PI) after 1 h of control (white arrow, living pericyte) or ischaemia solution containing antimycin and iodoacetate (red arrow, dead pericyte). b, Percentage of pericytes dead in control or after 1 h ischaemia (as in a) alone or with block of action potentials (TTX, 1 mM), AMPA/kainate receptors (25 mM NBQX), or NMDA receptors (50 mM D-AP5, 50 mM MK-801, 100 mM 7-chlorokynurenate); P values from one-way ANOVA with Dunnett’s post-hoc tests. c, Rat neocortical slice grey matter capillaries labelled for IB4, NG2 and PI after 1 h of control solution or oxygen and glucose deprivation (OGD). d, Percentage of pericytes (as in c) dead after 1 h OGD (No-reoxy) or OGD followed by 1 h of control solution (Reoxy) with no drugs, or with iGluR block (NBQX (25 mM), AP5 (50 mM) and 7CK (100 mM)), zero [Ca21]0, NOS block (100 mM L-NNA), or free radical scavenging (150 mM MnTBAP or 100 mM PBN, pooled data from Extended Data Fig. 5a) throughout. OGD killed pericytes (P 5 10213, ANOVA) and death increased during reperfusion (P 5 3.3 310213). iGluR block or zero [Ca21]0 reduced death (P 5 2.7 3 1024 and 6.0 3 1027, ANOVA with Dunnett’s post-hoc test). Blocking NOS had a small protective effect (P 5 0.026); reactive oxygen species (ROS) scavenging did not (P 5 0.99). e, f, Confocal images of striatal capillaries labelled with IB4 and PI (e) and percentage of striatal pericytes and endothelial cells that are dead (f) from the control and lesioned hemisphere of in vivo MCAO-treated rats (90 min, assessed 22.5 h later), sham-operated rats (with or without filament being inserted into the internal carotid artery (ICA)), and naive control animals. More pericytes die than endothelial cells (P 5 1026, repeated measures ANOVA). For pericytes, but not endothelial cells, cell death is greater in lesioned hemisphere (main effect of hemisphere, P 5 0.004; hemisphere-cell type interaction P 5 0.003) and cell death is greater in MCAO-lesioned animals than in naive or sham animals without ICA occlusion (P 5 0.005 and 0.01, Tukey post-hoc tests). See Extended Data Fig. 5 for data from cortex. Data shown as mean 6 s.e.m.

than to arterioles (70 mm distant31), suggesting that neurons might adjust their energy supply by initially signalling to pericytes (Extended Data Fig. 1a). Our data support this concept: neuronal activity leads to a release of messengers that activate an outward membrane current in pericytes (Fig. 2) and dilate capillaries before arterioles (Fig. 3). Capillary dilation

implies that there is a resting tone set by the pericytes, perhaps as a result of noradrenaline release from locus coeruleus axons, two-thirds of the perivascular terminals of which end near capillaries rather than arterioles32. Vascular diameter responses can propagate between adjacent pericytes1,9, but it is not yet known whether arterioles receive a signal to dilate from pericytes, or from vasoactive messengers which reach arterioles later than they reach pericytes. We have identified the cause of pericyte dilation as being EP4 receptor activation, by PGE2 or a related compound, although NO production is also needed to suppress synthesis of vasocontricting 20HETE (Fig. 1 and Extended Data Fig. 1b). These mechanisms resemble those controlling arteriole dilation, and may reflect glutamate release activating the production of arachidonic acid and its derivatives in astrocytes or neurons13. The more frequent occurrence of capillary dilation at pericyte locations (Fig. 3i–k), and the faster onset of capillary dilation than of arteriole dilation in vivo (Fig. 3d–h) suggest that pericytes actively relax to dilate capillaries. Our data suggest that capillaries have two conceptually separate roles in regulating cerebral blood flow. First, being closer to neurons, they detect neuronal activity earlier than arterioles can, and may pass a hyperpolarizing vasodilatory signal back to arterioles (through gap junctions between pericytes or endothelial cells)1,9. Second, capillary vasodilation itself contributes significantly to increasing blood flow. To assess the extent to which capillary dilations increase blood flow, we used data on the vascular tree in mouse cortex33 (see Methods). For a mean arteriole dilation of 5.9% during prolonged (15 s) stimulation (Extended Data Table 1), a capillary dilation of 6.7% (averaged over all capillary orders, Extended Data Table 1) and ignoring venule dilation, the steady state blood flow was predicted to increase by 19%. Omitting the capillary dilation predicted a flow increase of only 3%. Thus, capillary dilation is estimated to generate 1003 (19 2 3)/19 5 84% of the steady-state increase in blood flow evoked by neuronal activity, and capillaries dilate approximately 1 s before penetrating arterioles (Fig. 3d–h). These results suggest that BOLD functional imaging signals largely reflect capillary dilation by pericytes. For ischaemia, our data support, but modify, the suggestion that pericyte constriction1,14 may be a cause of the long-lasting decrease of cerebral blood flow that occurs even when a blocked artery is opened up after stroke26–29. It was previously envisaged that constriction of capillaries was by healthy pericytes and could be reversed by suppressing oxidative stress14, whereas we find that, after ischaemia has constricted them (Fig. 4), pericytes die readily (Fig. 5). This death is mediated in part by glutamate, but is not reduced by free radical scavenging, suggesting that the constriction and death differ at least partly in their causes. Pericyte death in rigor will produce a long-lasting decrease of capillary blood flow26–29, as well as a breakdown of the (pericyte-maintained4–6) blood–brain barrier. Both of these will contribute to ongoing neuronal damage, highlighting the potential importance of preventing pericyte death as a therapeutic strategy after stroke, particularly in the penumbra of an affected region. To develop this approach, it will be necessary to develop small molecule inhibitors of pericyte death that could be administered, perhaps with tissue plasminogen activator, soon after a stroke has occurred.

METHODS SUMMARY Brain slices from rats or NG2-DsRed mice were made and capillaries were imaged as described previously1. The fluorescence of DsRed-labelled capillary pericytes was used to identify pericytes for patch-clamping. Two-photon imaging of FITCdextran-labelled cortical vessels and DsRed-labelled pericytes in vivo was performed in mice as previously described for cerebellum34. Online Content Any additional Methods, Extended Data display items and Source Data are available in the online version of the paper; references unique to these sections appear only in the online paper. Received 29 July 2013; accepted 19 February 2014. Published online 26 March 2014. 1.

Peppiatt, C. M., Howarth, C., Mobbs, P. & Attwell, D. Bidirectional control of CNS capillary diameter by pericytes. Nature 443, 700–704 (2006). 3 A P R I L 2 0 1 4 | VO L 5 0 8 | N AT U R E | 5 9

RESEARCH ARTICLE 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.

Ferna´ndez-Klett, F., Offenhauser, N., Dirnagl, U., Priller, J. & Lindauer, U. Pericytes in capillaries are contractile in vivo, but arterioles mediate functional hyperemia in the mouse brain. Proc. Natl. Acad. Sci. USA 107, 22290–22295 (2010). Chan-Ling, T. et al. Desmin ensheathment ratio as an indicator of vessel stability: evidence in normal development and in retinopathy of prematurity. Am. J. Pathol. 165, 1301–1313 (2004). Bell, R. D. et al. Pericytes control key neurovascular functions and neuronal phenotype in the adult brain and during brain aging. Neuron 68, 409–427 (2010). Armulik, A. et al. Pericytes regulate the blood–brain barrier. Nature 468, 557–561 (2010). Daneman, R., Zhou, L., Kebede, A. A. & Barres, B. A. Pericytes are required for blood-brain barrier integrity during embryogenesis. Nature 468, 562–566 (2010). Go¨ritz, C. et al. A pericyte origin of spinal cord scar tissue. Science 333, 238–242 (2011). Dore-Duffy, P., Katychev, A., Wang, X. & Van Buren, E. CNS microvascular pericytes exhibit multipotential stem cell activity. J. Cereb. Blood Flow Metab. 26, 613–624 (2006). Puro, D. G. Physiology and pathobiology of the pericyte-containing retinal microvasculature: new developments. Microcirculation 14, 1–10 (2007). Villringer, A., Them, A., Lindauer, U., Einha¨upl, K. & Dirnagl, U. Capillary perfusion of the rat brain cortex. An in vivo confocal microscopy study. Circ. Res. 75, 55–62 (1994). Schulte, M. L., Wood, J. D. & Hudetz, A. G. Cortical electrical stimulation alters erythrocyte perfusion pattern in the cerebral capillary network of the rat. Brain Res. 963, 81–92 (2003). Jespersen, S. N. & Østergaard, L. The roles of cerebral blood flow, capillary transit time heterogeneity, and oxygen tension in brain oxygenation and metabolism. J. Cereb. Blood Flow Metab. 32, 264–277 (2012). Attwell, D. et al. Glial and neuronal control of blood flow. Nature 468, 232–243 (2010). Yemisci, M. et al. Pericyte contraction induced by oxidative-nitrative stress impairs capillary reflow despite successful opening of an occluded cerebral artery. Nature Med. 15, 1031–1037 (2009). Vates, G. E., Takano, T., Zlokovic, B. & Nedergaard, M. Pericyte constriction after stroke: the jury is still out. Nature Med. 16, 959 (2010). Hall, C. N., Klein-Flu¨gge, M. C., Howarth, C. & Attwell, D. Oxidative phosphorylation, not glycolysis, powers presynaptic and postsynaptic mechanisms underlying brain information processing. J. Neurosci. 32, 8940–8951 (2012). Hall, C. N. & Attwell, D. Assessing the physiological concentration and targets of nitric oxide in brain tissue. J. Physiol. (Lond.) 586, 3597–3615 (2008). Zhu, X., Bergles, D. E. & Nishiyama, A. NG2 cells generate both oligodendrocytes and gray matter astrocytes. Development 135, 145–157 (2008). Krueger, M. & Bechmann, I. CNS pericytes: concepts, misconceptions, and a way out. Glia 58, 1–10 (2010). Hamilton, N. B., Attwell, D. & Hall, C. N. Pericyte-mediated regulation of capillary diameter: a component of neurovascular coupling in health and disease. Front. Neuroenergetics 2, 5 (2010). Serebryakov, V., Zakharenko, S., Snetkov, V. & Takeda, K. Effects of prostaglandins E1 and E2 on cultured smooth muscle cells and strips of rat aorta. Prostaglandins 47, 353–365 (1994). Crawford, C. et al. An intact kidney slice model to investigate vasa recta properties and function in situ. Nephron Physiol. 120, 17–31 (2012).

23. Kleinfeld, D., Mitra, P. P., Helmchen, F. & Denk, W. Fluctuations and stimulusinduced changes in blood flow observed in individual capillaries in layers 2 through 4 of rat neocortex. Proc. Natl. Acad. Sci. USA 95, 15741–15746 (1998). 24. Khatri, P. et al. Good clinical outcome after ischemic stroke with successful revascularization is time-dependent. Neurology 73, 1066–1072 (2009). 25. Gupta, E. et al. Higher volume endovascular stroke centers have faster times to treatment, higher reperfusion rates and higher rates of good clinical outcomes. J. Neurointerv. Surg. 5, 294–297 (2013). 26. Hauck, E. F., Apostel, S., Hoffmann, J. F., Heimann, A. & Kempski, O. Capillary flow and diameter changes during reperfusion after global cerebral ischemia studied by intravital video microscopy. J. Cereb. Blood Flow Metab. 24, 383–391 (2004). 27. Leffler, C. W., Beasley, D. G. & Busija, D. W. Cerebral ischemia alters cerebral microvascular reactivity in newborn pigs. Am. J. Physiol. 257, H266–H271 (1989). 28. Nelson, C. W., Wei, E. P., Povlishock, J. T., Kontos, H. A. & Moskowitz, M. A. Oxygen radicals in cerebral ischemia. Am. J. Physiol. 263, H1356–H1362 (1992). 29. Baird, A. E. et al. Reperfusion after thrombolytic therapy in ischemic stroke measured by single-photon emission computed tomography. Stroke 25, 79–85 (1994). 30. Attwell, D. & Iadecola, C. The neural basis of functional imaging signals. Trends Neurosci. 25, 621–625 (2002). 31. Lovick, T. A., Brown, L. A. & Kay, B. J. Neurovascular relationships in hippocampal slices: physiological and anatomical studies of mechanisms underlying flow-metabolism coupling in intraparenchymal microvessels. Neuroscience 92, 47–60 (1999). 32. Cohen, Z., Molinatti, G. & Hamel, E. Astroglial and vascular interactions of noradrenaline terminals in the rat cerebral cortex. J. Cereb. Blood Flow Metab. 17, 894–904 (1997). 33. Blinder, P. et al. The cortical angiome: a 3-D interconnected vascular network with noncolumnar patterns of blood flow. Nature Neurosci. 16, 889–897 (2013). 34. Mathiesen, C., Brazhe, A., Thomsen, K. & Lauritzen, M. Spontaneous calcium waves in Bergmann glia increase with age and hypoxia and may reduce tissue oxygen. J. Cereb. Blood Flow Metab. 33, 161–169 (2013). Supplementary Information is available in the online version of the paper. Acknowledgements We thank B. Clark, A. Gibb, A. Gourine, C. Howarth, R. Jolivet, C. Madry, P. Mobbs, B. Richardson and A. Silver for comments on the manuscript. This work was supported by the Fondation Leducq, European Research Council, Wellcome Trust, UK Medical Research Council, Rosetrees Trust, Nordea Foundation via the Center for Healthy Aging, the Lundbeck Foundation, NOVO-Nordisk Foundation and Danish Medical Research Council. Author Contributions All authors (C.N.H., C.R., B.G., N.B.H., A.M., B.A.S., F.M.O., A.M.B., M.L. and D.A.) contributed to designing the study, doing the experiments, analysing the results and writing the paper. Author Information Reprints and permissions information is available at www.nature.com/reprints. The authors declare no competing financial interests. Readers are welcome to comment on the online version of the paper. Correspondence and requests for materials should be addressed to D.A. (d.attwell@ucl.ac.uk) or M.L. (mlauritz@sund.ku.dk).

6 0 | N AT U R E | VO L 5 0 8 | 3 A P R I L 2 0 1 4

ARTICLE RESEARCH METHODS Animals. Experiments used Sprague-Dawley or Wistar rats and NG2-DsRed C57BL/ 6J mice of either sex. Animal procedures were carried out in accordance with the guidelines of the UK Animals (Scientific Procedures) Act 1986, the Danish National Ethics Committee and European Directive 2010/63/EU. Each experiment was conducted on tissue from at least three animals on at least three different experimental days. Brain slice preparation. Slices (200–300 mm thick) were prepared1,35 on a vibratome in ice-cold oxygenated (95% O2, 5% CO2) solution. This solution was usually artificial cerebrospinal fluid (aCSF) containing (in mM) 124 NaCl, 2.5 KCl, 26 NaHCO3, 1 MgCl2, 1 NaH2PO4, 10 glucose, 0.1–1 Na ascorbate, 2 CaCl2 (to which 1 kynurenic acid was added to block glutamate receptors), and slices were incubated at room temperature (21–23 uC) in the same solution until used in experiments. For Fig. 2e–h and 4, the slicing solution contained (mM) 93 N-methyl-D-glucamine chloride, 2.5 KCl, 30 NaHCO3, 10 MgCl2, 1.2 NaH2PO4, 25 glucose, 0.5 CaCl2, 20 HEPES, 5 Na ascorbate, 3 Na pyruvate, 1 kynurenic acid, and the slices were incubated at 34uC in the same solution for 10 min, and then incubated at room temperature (21–23 uC) until used in experiments in a similar solution with the NMDG-Cl, MgCl2, CaCl2 and Na ascorbate replaced by (mM) 92 NaCl, 1 MgCl2 2 CaCl2 and 1 Na ascorbate. Immunohistochemical labelling of pericytes. Isolectin B4 binds to a-D-galactose residues in the basement membrane secreted by endothelial cells36,37, which surrounds pericytes. Cerebellar slices were incubated in FITC-conjugated 10 mg ml21 isolectin B4 (Sigma) for 1 h then fixed for 20 min in 4% PFA, and incubated for 4 to 6 h in 0.05% Triton X-100, 10% goat serum in phosphate-buffered saline at 21uC, then with primary antibody at 21uC overnight with agitation, and then for 4 to 8 h at 21uC with secondary antibody. Primary antibodies were: guinea pig NG2 (from W.B. Stallcup, 1:100), rabbit NG2 (Millipore AB5320, 1:300) and rabbit PDGFRb (Santa Cruz sc432, 1:200). Secondary antibodies (goat) were: anti-rabbit (Molecular Probes, 1:200) and anti-guinea-pig (Jackson Laboratory, 1:100). Pericytes, with a bump-on-a-log morphology on capillaries, or located at the junction of capillaries, labelled for NG2 and PDGFRb (Fig. 1). As perivascular immune cells are sometimes confused with pericytes19, we labelled the former with antibody to Iba1 (rabbit Iba1, Synaptic Systems 234003, 1:500) in seven cortical and eight cerebellar slices from five NG2-DsRed animals, and found that none of 135 (cortex) and 212 (cerebellum) NG2-DsRed labelled perivascular cells co-labelled for Iba1 (although Iba1 labelled 110 cells in the cortical slices and 136 cells in the cerebellar slices, Fig. 1c), implying that pericytes defined by an on-capillary location and NG2 expression differ from perivascular microglia or macrophages. Imaging of capillaries in brain slices. Slices were perfused with bicarbonatebuffered aCSF, as described above, but without the kynurenic acid, at 31 to 35uC. In experiments using 20% oxygen, the perfusion solution was bicarbonate-buffered aCSF, gassed with 20% O2, 5% CO2 and 75% N2. For bright-field recording of capillary diameter, sagittal cerebellar slices were prepared from postnatal day 10 (P10) to P21 Sprague-Dawley rats or coronal cortical slices were prepared from P12 rats. On average 1.3 capillary regions were imaged per slice. Capillaries were imaged1 at approximately 30 mm depth within the molecular layer of cerebellar slices or the grey matter of slices containing the somatosensory and motor cortices, using a 340 water immersion objective, a Coolsnap HQ2 CCD camera, and ImagePro Plus or Metafluor acquisition software. Images were acquired every 1 to 5 s, with an exposure time of 5 ms. The pixel size was 160 or 300 nm. Vessel internal diameters were measured by manually placing a measurement line (perpendicular to the vessel, Fig. 1d) on the image (at locations near visible pericytes which constricted when noradrenaline was applied), using ImagePro Analyzer, Metamorph or ImageJ software, with the measurer blinded as to the timing of drug applications. The end of the measurement line was placed at locations representing the measurer’s best estimate of where the rate of change of intensity was greatest across pixels under the vessel edge, and the diameter was estimated to a precision of one pixel. Where necessary, images were aligned by manually tracking drift, or by using the Image Pro ‘Align Global Images’ macro. Experiments in which changes in focus occurred were excluded from further analysis. Data in the presence of blockers of signalling pathways were compared with interleaved data obtained without the blockers. For experiments in which the parallel fibres were stimulated in the molecular layer, coronal slices were used to preserve the parallel fibres, and stimuli of 60 to 100 ms duration, at 50 to 90 V and 12 Hz, were applied for 25 s using a patch-pipette electrode placed approximately 100 mm away from the imaged vessel. To check that parallel fibres were being successfully activated, the field potential was monitored in the molecular layer using a 4 MV patch pipette filled with aCSF. To ensure that pericytes were healthy we excluded capillaries that did not constrict to 1 mM noradrenaline. Stimulation evoked dilation (Fig. 2e, f), except in two capillaries that constricted, presumably owing to direct depolarization of a pericyte by the stimulus

since when TTX was applied (to one of these vessels) a stimulation-evoked constriction was still seen in TTX: these two vessels were excluded from the analysis. Patch-clamp recordings of pericytes. Coronal slices of cerebellum were prepared35 from P10 to P17 NG2-DsRed C57BL/6J mice. Slices were superfused with bicarbonate-buffered solution containing (in mM) 124 NaCl, 26 NaHCO3, 1 NaH2PO4, 2.5 KCl, 1 MgCl2, 2.5 CaCl2, 10 glucose, bubbled with 95% O2/5% CO2, pH 7.3, at 21 to 23 uC or 33 to 36 uC (stimulation-evoked currents were not significantly different at the two temperatures and were pooled). Pericytes were identified as DsRed-expressing cells located on capillaries (oligodendrocyte precursor cells also express NG2-DsRed but these can be distinguished from pericytes morphologically and by their position in the parenchyma). Pericytes were whole-cell clamped between 255 and 275 mV with pipettes containing solution comprising (in mM) 130 K-gluconate, 4 NaCl, 0.5 CaCl2, 10 HEPES, 10 BAPTA, 2 Na2ATP, 2 MgCl2, 0.5 Na2GTP, 0.05 Alexa Fluor 488, pH set to 7.3 with KOH. Electrode junction potentials were compensated. Patch-clamped cells were morphologically confirmed to be pericytes by dye filling. Series resistance was 20 to 40 MV. In 17 cells the mean resting potential was 247.6 6 2.1 mV, and the mean input resistance at the resting potential was 292 6 27 MV. A 30-pA outward current evoked by neuronal activity (see main text) is thus expected to hyperpolarize pericytes by approximately 9mV. Imaging of vessels in vivo: animal preparation. Sixteen adult NG2 DsRed C57BL/ 6J mice (18–37 g, of either sex) were prepared for experiments by cannulation of the trachea for mechanical ventilation (SAR-830; CWE). Catheters were placed into the left femoral artery and vein and perfused with physiological saline. The end-expiratory CO2 (microCapstar End-tidal CO2 Monitor, CWE) and blood pressure (Pressure Monitor BP-1; World Precision Instruments) were monitored continuously in combination with blood gases in arterial blood samples (pO2 115– 130 mm Hg; pCO2 35–40 mm Hg; pH 7.35–7.45; ABL 700 Series; Radiometer Medical) to ensure the animals were kept under physiological conditions. The temperature was measured and maintained at 37uC during the experiment with a rectal thermometer regulated heating pad (TC-1000 Temperature Controller; CWE). The animals were anaesthetized with xylazine (10 mg kg21 intraperitoneal) and ketamine (60 mg kg21 intraperitoneal) during surgery, and then switched to alphachloralose (50 mg kg21h21 intravenous) during the experiment. The skull was glued to a metal plate using cyanoacrylate gel (Loctite Adhesives) and the plate was fixed in the experimental setup. A craniotomy was drilled with a diameter of approximately 4 mm with the centre 0.5 mm behind and 3 mm to the right of the bregma over the sensory barrel cortex region. The dura was removed and the preparation covered with 0.75% agarose (type III-A, low EEO; Sigma-Aldrich) and moistened with artificial cerebrospinal fluid (in mM: 120 NaCl, 2.8 KCl, 22 NaHCO3, 1.45 CaCl2, 1 Na2HPO4, 0.9 MgCl2 and 2.6 glucose; pH 7.4) at 37 uC and bubbled with 95% air, 5% CO2. The craniotomy was covered with a glass coverslip. When experiments were complete, mice were euthanized by intravenous injection of anaesthesia (pentobarbital, 200 mg ml21 and lidocaine hydrochloride, 20 mg ml21) followed by decapitation. Imaging of vessels in vivo: whisker-pad stimulation. The mouse sensory barrel cortex was activated by stimulation of the contralateral ramus infraorbitalis of the trigeminal nerve using a set of custom-made bipolar electrodes inserted percutaneously. The cathode was positioned at the hiatus infraorbitalis, and the anode was inserted into the masticatory muscles38. Thalamocortical infraorbitalis stimulation was performed at an intensity of 1.5 mA (ISO-flex; AMPI) and lasting 1 ms, in trains of 2 s or 15 s at 3 Hz. The stimulation was controlled by a sequencer file running within Spike2 software (version 7.02; Cambridge Electronic Design). Imaging of vessels in vivo: cortical response imaging. For each animal, the haemodynamic response to stimulation was detected using intrinsic optical imaging (IOS) and used to identify the region of brain activated by whisker-pad stimulation, for further vascular imaging. Two-photon imaging of blood vessels was then conducted near the centre of this activated region. The IOS was recorded on a Leica microscope with 35 magnification that included the entire preparation in the field of view. The light source consisted of LEDs with green light filters and a fast camera (QuantEM 512SC; Photometrics) sampled 29 images per s before and during 15 s of 3 Hz stimulation. As haemoglobin strongly absorbs green light, the captured light intensity decreases as the total haemoglobin concentration increases during changes in cerebral blood volume and flow39. Images during stimulation were subtracted from control images40, allowing the area of brain where blood flow increases during whisker stimulation to be revealed. Imaging of vessels in vivo: two-photon imaging. Fluorescein isothiocyanatedextran (FITC-dextran, 2% w/v, MW 70,000, 50 ml, Sigma-Aldrich) was administered into the femoral vein to label the blood plasma34. In vivo imaging of blood vessel diameter and pericyte location was performed using a commercial twophoton microscope (SP5, Leica), a MaiTai HP Ti:Sapphire laser (Millennia Pro; Spectra Physics, mean output power 10 mW), and a 20 3 1.0 N.A. water-immersion objective (Leica). Tissue was excited at 900 nm wavelength, and the emitted light was filtered to collect red and green light from DsRed (pericytes) and FITC-dextran

RESEARCH ARTICLE (vessel lumens). Penetrating arterioles were identified unequivocally in vivo in two ways, first by tracing their connections back to pial arterioles with obvious smooth muscle around them, and second by observing the direction of flow of the red blood cells from the arterioles into the parenchymal capillaries. Z-stack images were taken of the area of interest. XY-time series were taken to image pericytes and blood vessels during stimulation, with a frame size 512 3 300 pixels (170 ms per frame; pixel size was 93 to 201 nm depending on the magnification used, with a mean value of 155 nm; pixel dwell time was 1.1 ms). Imaging of vessels in vivo: image analysis. The presence of red blood cells (RBCs) leads to apparent holes in the images of the capillary and (to a lesser extent) the arteriole (Extended Data Fig. 4a, b). This image noise was reduced by smoothing images with a maximum intensity 10 frame (1.7 s) running summation of the green channel showing the FITC-labelled vascular lumen (see Extended Data Fig. 4a–d for sample images, and the effect this would have on the apparent timing of a step increase of diameter and on the dilations shown in Fig. 3d). This channel was then processed to extract blood vessel diameters. Lines were placed across the vessel (perpendicular to the vessel wall) at a spacing of approximately 20 mm (for example, Fig. 3d). The edges of the vessel were located using the ImagePro caliper tool, which finds the greatest gradient in light intensity along the line (where d2intensity/dx2 5 0). As for brain slice imaging above, interpolation of the image intensity across pixels allows, in principle, the position of the edge to be estimated to change by less than one pixel, but in practice we measured the diameter with a precision of one pixel. We considered whether the smoothing applied could artefactually preferentially accelerate the dilation seen for capillaries. In fact this is not possible, because the smoothing uses a maximum intensity sum of all the 10 frames: consequently, a black area produced by a passing RBC will not add to the summed image, and any increase in diameter that has occurred will only be detected in a later frame. Thus, passage of an RBC can only delay the apparent time at which dilation is detected, this effect should be larger in capillaries in which the RBCs are moving more slowly, and yet the capillaries dilate before the arterioles (Fig. 3). We conclude that the smoothing cannot cause a preferential acceleration of the time course of dilation in the capillary data. For Fig. 3e, g, the time courses of the measured diameter were also smoothed with a five-point FFT procedure that removes frequencies over 1.16 Hz (OriginLab software), the effect of which is shown in Extended Data Fig. 4e. Responding capillaries were defined as those showing a change upon stimulation of more than 5% of the initial vessel diameter (since 4.99% was twice the standard deviation of the baseline diameter averaged over all vessels studied). Where multiple regions responded on a single vessel, their response times were averaged for comparison between paired vessels (Fig. 3f and Extended Data Fig. 5c). Blood flow in capillaries was assessed from the velocity of RBCs, which appear as dark patches inside FITC-dextran labelled vessels, using line-scan imaging23. Repetitive line-scans (0.358 ms per line of 512 pixels) along the axis of the vessel before, during and after whisker stimulation (3Hz, 15 s) were used to form a space-time image in which moving RBCs produce streaks with a slope that is equal to the inverse of the speed. The slope was calculated using an automated image-processing algorithm41. The baseline speed of RBCs averaged over all capillaries studied was 1.73 6 0.20 mm s21 (n 5 49). Cell-death experiments: chemical ischaemia. For chemical ischaemia experiments (Figs 4 and 5a, b), coronal cortical slices from P21 (Fig. 4) or sagittal cerebellar slices from P7 (Fig. 5a, b) Sprague-Dawley rats were incubated at 37uC in an ischaemic solution in which glucose was replaced with 7 mM sucrose and oxygen was removed by equilibrating solutions with 5% CO2 and 95% N2. In addition, 2 mM iodoacetate and 25 mM antimycin were added to the ischaemic solution to block ATP generation by glycolysis and oxidative phosphorylation, respectively42. Control slices were incubated in aCSF, gassed as usual with 5% CO2, 95% O2. Propidium iodide (PI, 37 mM for Fig. 5a-d, but 7.5 mM for Fig. 4) was added to both solutions to label dead cells. For Fig. 5, after 60 min incubation in this ischaemic solution, slices were fixed for 20 min in 4% paraformaldehyde and immunohistochemistry for NG2 was performed, as described above. Cell-death experiments: oxygen-glucose deprivation. For the oxygen-glucose deprivation experiments shown in Fig. 5c, d and Extended Data Fig. 6a, b, coronal forebrain slices were prepared from P21 Sprague Dawley rats then incubated in aCSF in which glucose was replaced with 7 mM sucrose and oxygen was removed by equilibrating solutions with 5% CO2 and 95% N2. Control slices were incubated in aCSF, gassed as usual with 5% CO2, 95% O2. After 60 min, some slices were immediately fixed in 4% paraformaldehyde, while others were placed in control aCSF, to reoxygenate for a further 60 min. All solutions also contained 37 mM PI and 10 mg ml21 FITC-conjugated isolectin B4 to label dead cells and blood vessels, respectively. Slices were swiftly washed in aCSF before fixation for 20 min in 4% paraformaldehyde, washed three times in PBS and mounted on microscope slides in Dako hard set mounting medium. Slides were then imaged using a Zeiss LSM 700 or 710 confocal microscope. PI-positive dead vascular and parenchymal cells were counted using ImageJ software, by an experimenter who was blind to their condition.

Cells in the 20 mm closest to the slice surface were excluded from analysis to prevent confounds from slicing-induced damage. Dead or alive pericytes were identified by their ‘bump-on-a-log’ morphology on vessels surrounded by isolectin B4 labelling. To check that pericytes could be identified by isolectin B4 labelling alone, in parallel experiments we labelled slices for isolectin B4 and NG2: the great majority (93% of 718 cells assessed) of pericytes identified this way were found to be positive for NG2. Although some pericytes may slightly move away from capillaries after hypoxia or brain injury43,44, we only counted pericytes apposed to capillaries in this study. Cell-death experiments: middle cerebral artery occlusion. Male Wistar rats (Harlan) weighing 253 to 312g, housed on a 12-h light–dark cycle with ad libitum access to food and water, underwent transient middle cerebral artery occlusion (MCAO) as described previously45. In brief, animals were anaesthetized with 4% isoflurane and maintained in 1.5 to 2% isoflurane carried in 70% N2O and 30% O2. A midline incision was made in the neck, the right external carotid artery was cauterised and cut, and the right common carotid and internal carotid arteries were temporarily ligated. Through a small arteriotomy in the external carotid artery stump, a 4-0 nylon filament coated with silicone at the tip (Doccol) was advanced up the internal carotid artery to occlude the right middle cerebral artery at its origin. For sham animals, the entire procedure was followed except that either: the filament was only advanced up the beginning of the internal carotid artery (ICA) before being withdrawn after 3 min (sham with ICA occlusion); or the external carotid artery was permanently ligated (which had no effect on cerebral blood flow) and the common and internal carotid arteries were exposed but not ligated, and the animals remained under anaesthesia for the same length of time as sham animals (sham without ICA occlusion). Core temperature was maintained at 37uC by a rectal thermister probe attached to a heating pad. Cerebral blood flow was continuously monitored by placing a laser Doppler probe (Oxford Optronix) over a thinned skull of the MCA territory approximately 4 mm lateral and 1.5 mm caudal to bregma. In MCAO animals, averaged over the period of occlusion, cerebral blood flow on the treated side fell to 34.9 6 7.1% of baseline (n 5 6) for 90 min, while in sham animals with ICA occlusion it dropped significantly less to 67.9 6 11.0% of baseline (P 5 0.035, t-test, n 5 3) for 16 min (this smaller drop occurs because of occlusion of the ICA), and in sham animals without ICA occlusion blood flow was unaffected (104.0 6 3.9% of baseline, n 5 3). Following 90 min of MCAO, the filament was retracted, and the common carotid artery ligation was released to allow maximal reperfusion. Anaesthesia was then removed, and at 22.5 h of reperfusion, neurological deficit was assessed by investigating limb symmetry, motor function, activity and response to sensory stimulation (modified from ref. 46 by scoring spontaneous movement, symmetry in limb movement and forepaw outstretching on a scale of 0–3 with 0 5 normal and 3 5 most severely affected, and scoring climbing, body proprioception and response to vibrissae touch on a 0 (normal) to 2 scale). A maximum score of 15 equates to severe neurological deficit, while a minimum score of 0 implies no neurological deficit. MCAO animals had a mean score of 7.5 6 1.7 (n 5 6), which was significantly greater than that of sham animals with ICA occlusion (0.3 6 0.3, n 5 3, P 5 0.027) and sham animals without ICA occlusion (0 6 0, n 5 3 P 5 0.039, corrected for multiple comparisons). Animals (including an additional three naive control animals) were then anaesthetized, decapitated, and 200 mm forebrain slices were prepared on a vibratome and labelled with PI (37 mM) and FITC-isolectin B4 (10 mg ml21) in aCSF for 60 min, then washed, fixed, mounted and cortical and striatal images were captured as described above. Live and dead pericytes were counted in both regions as above except that dead endothelial cells were also counted (identified by their elongated nuclei). More pericyte death was seen in brain slices made from naive control animals in this in vivo series of experiments (Fig. 5f) than in experiments studying pericyte death in slices (Fig. 5b, d); this may be because, for the adult rats used for the in vivo experiments, it takes longer to kill the animal and remove its brain, than for the younger animals used for slice experiments. The total number of endothelial cells present, and therefore the percentage of dead endothelial cells, was estimated from the total number of pericytes, assuming a 1:3 ratio of pericytes to endothelial cells47. Statistics. All data are presented as mean 6 s.e.m. P values are from ANOVA (univariate, unless otherwise stated) and post-hoc Dunnett’s or Student’s t-tests, chi-squared tests, Kolmogorov–Smirnov tests or Mann–Whitney U-tests (for nonnormally distributed data), as appropriate. Two-tailed tests were used. P values quoted in the text are from independent samples t-tests unless otherwise stated. For multiple comparisons, P values are corrected using a procedure equivalent to the Holm-Bonferroni method (for N comparisons, the most significant P value is multiplied by N, the second-most significant by N-1, the third-most significant by N-2 and so on; corrected P values are considered significant if they are less than 0.05). Normality of data was assessed using Kolmogorov–Smirnov tests. All statistical analysis was conducted using IBM SPSS21, OriginLab or Excel software. Contribution of capillary dilation to cerebral blood-flow increases. To assess how capillary dilations increase blood flow in the steady state, we used data from

ARTICLE RESEARCH a recent analysis of the vascular tree in mouse cortex33. For blood flow from the cortical surface (where for simplicity we assume blood pressure to be constant) through a penetrating arteriole to layer 4 of the cortex, through the array of interconnected capillaries, and back to the cortical surface through a penetrating venule, that analysis33 concluded that the resistances (at baseline diameter) of the arteriole, capillary and venule segments of this path were, respectively, 0.1, 0.4 (for a path from an arteriole to a venule separated by approximately 200 mm, Figs 2g and 5c of ref. 33) and 0.2 poise per mm3, so that capillaries provide 57% of the total resistance. For a mean neuronal activity evoked arteriole dilation of 5.9% during prolonged (15 s) stimulation (Extended Data Table 1), a capillary dilation of 6.7% (averaged over all capillary orders, Extended Data Table 1) and ignoring venule dilation, then for resistance inversely proportional to the fourth power of diameter (Poiseuille’s law) the blood flow should increase by 19% in the steady state. Omitting the capillary dilation predicts a flow increase of only 3%, while omitting the arteriole dilation (so that only capillaries dilate) predicts a flow increase of 15%. Deviations from Poiseuille’s law in the capillaries33 make only a small correction to these values. Thus, capillary dilation is predicted to generate 84% of the steady state increase in blood flow evoked by prolonged neuronal activity. This figure would be reduced somewhat if pial arteriole dilation48 significantly contributes to the flow increase. For example, if pial arterioles are assumed to dilate by the same 5.9% as penetrating arterioles, and to have the same resistance as penetrating arterioles, then the capillary contribution to the blood flow increase is predicted to be 73% (however the larger diameter of pial vessels and their anastomoses49 suggest that their contribution to the total resistance and blood flow control will be much less than that of the penetrating arterioles). Relationship to earlier work. A previous failure to observe active capillary dilation2 may reflect two factors. First, in that study2, unlike ours, the authors were unable to reproducibly evoke blood flow increases by whisker-pad stimulation and so used bicuculline to excite neurons: conceivably this induces seizure-like activity which may generate a non-physiological release of vasoconstricting 20-HETE, as well as of dilating messengers. Second, the use of thiopental anaesthetic in that study2 may have reduced blood flow increases, since the same group reported that the closely related anaesthetic thiobutabarbital suppresses neuronally evoked blood flow increases by 40% compared to those seen using a-chloralose (the anaesthetic that we use)50. In addition, the vessels they define as precapillary arterioles2 (their Fig. 4h), which did show active dilations, seem to have isolated pericytes on them (rather than continuous smooth muscle) and would be called first-order capillaries in our nomenclature. Duration of rigor after pericyte death. Rigor occurs because when the ATP level falls, muscle cells like pericytes cannot detach their myosin heads from actin to relax. At the whole body level rigor mortis following death (the stiffness resulting from skeletal muscles dying in rigor) lasts approximately 18 to 36 h, depending on temperature. We would therefore expect pericytes to stay in rigor for hours (that

is, the period after ischaemia for which blood flow is known to be reduced26), unless removed by microglia. 35. Marcaggi, P. & Attwell, D. Endocannabinoid signaling depends on the spatial pattern of synapse activation. Nature Neurosci. 8, 776–781 (2005). 36. Peters, B. P. & Goldstein, I. J. The use of fluorescein-conjugated Bandeiraea simplicifolia B4-isolectin as a histochemical reagent for the detection of alpha-D-galactopyranosyl groups. Their occurrence in basement membranes. Exp. Cell Res. 120, 321–334 (1979). 37. Laitinen, L. Griffonia simplicifolia lectins bind specifically to endothelial cells and some epithelial cells in mouse tissues. Histochem. J. 19, 225–234 (1987). 38. Norup Nielsen, A. & Lauritzen, M. Coupling and uncoupling of activity-dependent increases of neuronal activity and blood flow in rat somatosensory cortex. J. Physiol. (Lond.) 533, 773–785 (2001). 39. Frostig, R. D., Lieke, E. E., Ts’o, D. Y. & Grinvald, A. Cortical functional architecture and local coupling between neuronal activity and the microcirculation revealed by in vivo high-resolution optical imaging of intrinsic signals. Proc. Natl. Acad. Sci. USA 87, 6082–6086 (1990). 40. Harrison, T. C., Sigler, A. & Murphy, T. H. Simple and cost-effective hardware and software for functional brain mapping using intrinsic optical signal imaging. J. Neurosci. Methods 182, 211–218 (2009). 41. Schaffer, C. B. et al. Two-photon imaging of cortical surface microvessels reveals a robust redistribution in blood flow after vascular occlusion. PLoS Biol. 4, e22 (2006). 42. Allen, N. J., Ka´rado´ttir, R. & Attwell, D. A preferential role for glycolysis in preventing the anoxic depolarization of rat hippocampal area CA1 pyramidal cells. J. Neurosci. 25, 848–859 (2005). 43. Dore-Duffy, P. et al. Pericyte migration from the vascular wall in response to traumatic brain injury. Microvasc. Res. 60, 55–69 (2000). 44. Gonul, E. et al. Early pericyte response to brain hypoxia in cats: an ultrastructural study. Microvasc. Res. 64, 116–119 (2002). 45. Nagel, S. et al. Neuroprotection by dimethyloxalylglycine following permanent and transient focal cerebral ischemia in rats. J. Cereb. Blood Flow Metab. 31, 132–143 (2011). 46. Garcia, J. H., Wagner, S., Liu, K. F. & Hu, X. J. Neurological deficit and extent of neuronal necrosis attributable to middle cerebral artery occlusion in rats. Statistical validation. Stroke 26, 627–635 (1995). 47. Pardridge, W. M. Blood-brain barrier biology and methodology. J. Neurovirol. 5, 556–569 (1999). 48. Iadecola, C., Yang, G., Ebner, T. J. & Chen, G. Local and propagated responses evoked by focal synaptic activity in cerebellar cortex. J. Neurophysiol. 78, 651–659 (1997). 49. Blinder, P., Shih, A. Y., Rafie, C. & Kleinfeld, D. Topological basis for the robust distribution of blood to rodent neocortex. Proc. Natl. Acad. Sci. USA 107, 12670–12675 (2010). 50. Lindauer, U., Villringer, A. & Dirnagl, U. Characterization of CBF response to somatosensory stimulation: model and influence of anesthetics. Am. J. Physiol. 264, H1223–H1228 (1993).

RESEARCH ARTICLE a PIAL SURFACE

smooth muscle

3. Pericyte dilation may spread to arteriole 1st order capillary

2nd order capillary

1. Active cells

2nd order capillary

0th order penetrating arteriole

Key neuron astrocyte pericyte

3rd order capillary 3rd order capillary

b AA

O 2 Km 10µM 55µM

O2 K m 350µM

2. Dilators released by active neurons and astrocytes reach pericytes before arterioles

PgE2

EP4R

dilation

20-HETE

constriction

NO blocks production of constrictor 20-HETE

[Ca2+]i

2. Pericyte constricts, blocking capillary

1st order capillary

smooth muscle

PIAL SURFACE

1. Low [ATP] raises pericyte [Ca2+]i

0th order penetrating arteriole

Ischaemia

2nd order capillary

3. Pericyte dies in rigor, producing long-lasting decrease in blood flow

Extended Data Figure 1 | Signalling to capillary pericytes in health and disease. a, As vasodilators released from active neurons and their associated astrocytes (1) diffuse through the brain, they encounter pericytes before arteriole smooth muscle (2) because neurons are closer to capillaries than to arterioles31. This may partly explain why capillaries dilate before arterioles (Fig. 3). Pericyte dilation may spread to arteriole smooth muscle by current flow or [Ca21]i changes passing through gap junctions (3). b, Oxygen-dependent signalling pathways regulating vessel diameter (adapted from ref. 13). Neuronal activity leads to the generation of nitric oxide (NO) and arachidonic acid. Arachidonic acid (AA) is converted into PGE2, which dilates vessels via

EP4 receptors, but also into the vasoconstrictor 20-HETE. Production of 20-HETE is inhibited by NO. Together these pathways regulate capillary diameter (Fig. 1). Larger dilations to glutamate in low [O2] may reflect less production of 20-HETE from arachidonic acid. c, In ischaemia, the decrease of ATP concentration leads to a rise of [Ca21]i in pericytes. This results in some of them contracting and constricting capillaries1, which will prevent the passage of white and red blood cells. Most pericytes then die (Figs 4 and 5). Death of pericytes in rigor will produce a long-lasting decrease of cerebral blood flow, and reduce the ability of the microvasculature to increase blood flow in response to neuronal activity.

ARTICLE RESEARCH

46 95

50 40 23 132 30 20 95

O2 (%)

40 20 0

O2 (%)

f Diameter in HET (%)

p=4x10-3

Control ODQ

28 16 23 10 50 20 95 p=0.76

10 10 20

23 16 95

O2 (%) Δ Diameter (%)

Diameter (μm)

NA NO

4 2 0

8 20

22 95

Time (min)

Extended Data Figure 2 | Drug effects on capillary baseline diameter and constriction to noradrenaline. a, Rat capillaries are constricted by 100 mM L-NG-nitroarginine (L-NNA), suggesting that there is some tonic release of NO in the slice, with a greater constriction occurring at high [O2] (bars in a–d, f–n show percentage of initial drug-free diameter). b, However, L-NNA does not affect the diameter of vessels preconstricted with noradrenaline (NA; 2 mM, P 5 0.81, ANOVA). c, The guanylyl cyclase blocker ODQ (10 mM) does not affect baseline capillary diameter, suggesting that the constricting effect of tonic NO release seen in a is not through the cGMP pathway but is via suppression of 20-HETE release. d, ODQ slightly enhances the constriction achieved with NA (P 5 0.003, ANOVA). P values on the graph are from post hoc t-tests. e, As expected, ODQ blocks cGMP production by guanylyl cyclase, as assessed by radioimmunoassay. f, Inhibition of 20-HETE formation with 1 mM HET0016 (HET) does not affect baseline capillary diameter (white bars, at 20% O2, P 5 0.78; at 95% O2, P 5 0.49), presumably because the tonic NO release (in a) is sufficient to suppress tonic 20-HETE release, and there is no significant difference in baseline diameter in the presence of HET between the two O2 concentrations (P 5 0.57, t-test). Unlike application of L-NNA alone

19 19 20

Control L161,982

75 23 50

19 8 20

22 95

NA PgE2

4 2 0

p=0.96

100

50 n Co

O2 (%) 6

O2 (%)

100

n p=6x10-5

23 16

Control L161,982

O2 (%)

m 6

Diameter (μm)

19 19 20

16 10 95

O2 (%)

p=0.7 p=0.96

100

Control MSPPOH

100

Diameter in NA (%)

Diameter in L161,982 (%)

∆ Diameter in glutamate (%)

Control MS-PPOH

O2 (%)

100

50 40 31 132 28 20 95

HET HET+L-NNA

O2 (%)

28 95

p=0.04

O2 (%)

Control L-NNA

31 20

p=0.31

Δ Diameter (%)

23 20

Diameter in NA (%)

Control ODQ

100

Diameter in NA (%)

100

p=0.81 p=0.96 p=0.74

Diameter in NA (%)

p=2x10-3

Control L-NNA

Diameter in ODQ (%)

Diameter in NA (%)

Diameter in L-NNA (%)

100 p=0.03

cGMP (pmol/mg protein)

p=0.02

Diameter in NA (%)

Time (min)

l 2 tro ,98 61 L1

p=4x10-4

PgE2

(see a), application of L-NNA and HET together (black bars) does not significantly change capillary diameter (P 5 0.59, ANOVA compared to HET alone). Indeed, comparing vessel diameters in HET plus L-NNA with those in L-NNA alone (black bars in f versus panel a) reveals that HET significantly relieves the constriction produced by L-NNA (P 5 0.03, ANOVA). g, HET does not affect either the constriction to NA (white bars versus white bars in b, P 5 0.51, ANOVA) or the diameter of vessels in L-NNA and NA (black bars versus black bars in b, P 5 0.26, ANOVA). h, i, Blocking synthesis of epoxy-derivatives of arachidonic acid with MS-PPOH 10 mM does not affect the degree of constriction to NA (h, P 5 0.92, ANOVA) or the dilation to glutamate (i; P 5 0.92, ANOVA). j–l, Blocking EP4 receptors with 1 mM L161,982 had no effect on baseline diameter in cerebellum (j) or cortex (97.3 6 1.4% of baseline in 20% O2, P 5 0.07, n 5 34) and also did not affect the constriction to NA in either area (cerebellum: k, P 5 0.90, ANOVA over both [O2]; l, cortex in 20% O2). m, n, Applying DETA-NONOate (m, 100 mM) or prostaglandin E2 (n, 1 mM) dilated cerebellar capillaries preconstricted with noradrenaline (in 20% O2). Data shown as mean 6 s.e.m.

RESEARCH ARTICLE

Current (pA)

Inward Outward

Glut.

NMDA

Stim -40

Extended Data Figure 3 | Pericyte current responses in NG2-DsRed mouse cerebellar slices. Mean initial inward and later outward currents (as in Fig. 2a窶電, 95% O2) evoked in pericytes by stimulation of the parallel fibres, and by superfusion of 500 mM glutamate or 100 mM NMDA. Numbers of cells apply to both black and white bars. Error bars show s.e.m.

ARTICLE RESEARCH a

10 μm

Effect of 10 frame smoothing on the diameter trace 1

Normalised diameter

RBC “holes” in FITC-dextran Unprocessed image lumen labelling (at time t)

Running 10 frame maximum intensity projection (starting at time t)

Step change in dilation, occurring between frames 0 and 1.

Forward-looking time window for averaging frames for signal at time T

Dilation change appears 10 frames earlier in smoothed images.

0 -20

0 20 Frame number

Removing the time advance caused by the 10 frame smoothing Diameter profiles from smoothed images

1st order capillary

Effect of FFT smoothing on the diameter profiles

0 traces from smoothed images as here imply original time courses as here

1 Time course corrected by shifting profile backwards by 10 frames (1.72s)

0 0

Time (s)

Extended Data Figure 4 | Smoothing of the in vivo diameter changes. a, Specimen single frame from the image sequence for Fig. 3d (penetrating arteriole on the right, capillary on the left). The presence of red blood cells (RBC) leads to apparent holes in the image of the capillary diameter. b, RBC movement results in the holes being removed when averaging over 10 frames. c, For a step increase in vessel diameter (top), the effect on the measured diameter time course (bottom) of a running maximum intensity average being calculated over 10 frames starting at each time being considered: the maximum intensity summation results in the largest diameter at any time dominating

Unfiltered data

Normalised diameter

0th order arteriole

1.0

0.5 Data smoothed using a FFT filter to remove frequencies over 1.16 Hz

0 0

Time (s)

the smaller diameters at other times, and so the diameter increase is brought forward by 10 frames. If this were not corrected for, the diameter would appear to increase 10 frames (1.72 s) before it actually does. To correct for this, the time axis needs to be advanced by 1.72 s. d, Correction of the 10-frame averaged time courses of the data used for Fig. 3d for the time shift introduced by the averaging. e, Effect of the five-point FFT procedure (which removes frequencies over 1.16 Hz) applied to the averaged time courses in d. The smooth dashed lines are the traces plotted in Fig. 3d.

15s stim at 3Hz

∆ Diameter on stimulation (% of initial diameter)

∆ Diameter on stimulation (%)

RESEARCH ARTICLE b

6 1st stim 2nd stim 4 2 0

p=0.54

p=0.82

1st stim 2nd stim

0 5

Time (s) 3

Percent of peak reached:

All vessels

10% 20% 50% 80%

Time (s)

Paired vessels

p=0.015 p=0.012

Time difference (s)

Branching order 0 1 2 3 ≥4

4 2

-1 -2

0 vs. 1

1 vs. 2

2 vs. 3

3 vs. ≥4

Branching orders compared

Vessel diameter or RBC velocity (% of peak)

100

Processes Soma

25 0 -20

∆ Diameter on stimulation (% of initial diameter)

Extended Data Figure 5 | Responses of capillaries in somatosensory cortex in vivo. a, b, Reproducibility of response to whisker-pad stimulation at 101 capillary locations in NG2-DsRed mice. Mean capillary response time courses are the same on repeated stimulation (a). Responses at 5 and 15 s into 15 s whisker-pad stimulation did not differ significantly between the first and second stimulation (b). c, Time to reach a certain percentage of the maximum dilation in j–1th order vessel minus that in jth order vessel ( j is an integer $ 1) imaged simultaneously. The time to 10% and 20% of the peak is faster in first-order capillaries than in zero-order penetrating arterioles (see main text and Fig. 3f), although there are no significant differences between vessels of adjacent orders for any of the other bars shown (P 5 0.33–1, t-tests; n for each

All vessels 1st order capillary

100

Percent of peak reached (%)

e Cumulative probability (%)

Time (s)

RBC velocity

0th order arteriole

Time (s)

comparison was as in Fig. 3f). d, Time course of responses in all responding (.5%) vessels of different order. Arterioles were significantly slower to reach 10% of their peak response than first- and second-order vessels (see main text). e, The response distributions of capillaries do not differ near pericyte somata or processes (P 5 0.24, Kolmogorov–Smirnov test; 172 somata locations, 292 process locations). f, Comparison of time course of dilation of penetrating arterioles and first-order capillaries with that of the blood-flow increase in capillaries (n 5 49, all orders averaged) assessed by line-scanning (normalized to the average value at the peak from 11.7 to 13.2 s). Data shown as mean 6 s.e.m.

ARTICLE RESEARCH a

b MnTBAP

aCSF OGD

18 18

HET0016

50 6 6 18 18

18 18

MCPG

OGD

50 18 18

Ru360

aCSF Pericytes (% dead)

Pericytes (% dead)

PBN

tr on

ol R

x eo

y. C

tr on

ol R

x eo

y. C

tr on

ol R

x eo

ntr

CONTROL HEMISPHERE

LESIONED HEMISPHERE

Cells (% dead)

Naive

Sham without ICA occlusion

Sham with ICA occlusion

MCAO

ri Pe

tes E

o nd

the

lia

ri Pe

tes E

o nd

the

Extended Data Figure 6 | Ischaemia-evoked pericyte death in cortical slices and in vivo. a, Rat pericyte death is not affected by either of the free radical scavengers MnTBAP (150 mM) or PBN (100 mM; P 5 0.78 and P 5 1, respectively, ANOVA with Dunnett’s post hoc test versus no drug control). The amount of pericyte death did not differ between the two different scavengers (P 5 0.88, ANOVA) so the data from the two scavengers were combined for the analysis in the main text (Fig. 4d). b, In addition to the drugs discussed in the main text, none of the following drugs affected pericyte death following OGD and reoxygenation (ANOVA with Dunnett’s post hoc test versus no drug control): an inhibitor of mitochondrial calcium uptake, Ru360 (50 mM, P 5 1), the metabotropic glutamate receptor antagonist MCPG (500 mM, P 5 0.93) or the 20-HETE synthesis blocker HET0016 (HET, 1 mM, P 5 1). c, The percentage of dead cerebral cortical pericytes and endothelial cells after 24 h in the control and treated hemispheres of MCAO-treated rats, sham-operated animals where a filament was inserted into the ICA but was not advanced far enough to completely occlude the vessel (see Methods), sham

lia

ri Pe

tes E

o nd

the

lia

l ri Pe

lia tes the do En

animals without ICA occlusion, and naive animals that did not experience any surgery before being killed. These data were analysed together with the striatal data in the main Fig. 5f. There was no difference in cell death between cortex and striatum (P 5 0.55, repeated measures ANOVA). A much greater proportion of pericytes died than endothelial cells (P 5 2.131027, repeated measures ANOVA), and pericyte death, but not endothelial cell death, was greater in the lesioned hemisphere (effect of hemisphere, P 5 0.008, repeated measures ANOVA; interaction between hemisphere and cell type, P 5 0.007). As expected, most pericyte death occurred in the MCAO treated animals, whereas sham-operated animals with ICA occlusion showed intermediate levels of death between MCAO and naive (or sham with no ICA occlusion) animals (MCAO versus naive animals, P 5 0.004; MCAO versus sham without ICA occlusion, P 5 0.01; MCAO versus sham with ICA occlusion, P 5 0.20; sham with ICA occlusion versus naive, P 5 0.13; Tukey post-hoc tests). Data shown as mean 6 s.e.m.

RESEARCH ARTICLE Extended Data Table 1 | Properties of vessels of different branching orders in vivo Penetrating arteriole

Order

Capillaries of different order

All capillaries

24 (24)

100 (59)

205 (88)

164 (79)

164 (72)

633 (298)

257

Baseline diameter (µm)

12.4±0.9

5.5±0.2

4.6±0.1

4.1±0.1

3.7±0.1

4.4±0.1

Dilation of all regions (%)

5.9±1.3

7.7±1.0

8.3±0.1

5.4±0.1

5.3±0.1

6.7±0.4

Dilation of responders (%)

10.3±1.3

13.8±1.1

17.9±1.1

13.9±1.1

14.0±1.0

15.3±0.6

Number of regions (on N vessels) Number responding >5%

Data shown as mean 6 s.e.m.

NEWS & VIEWS STROKE

The worldwide burden of stroke—a blurred photograph Blanca Fuentes and Exuperio Díez Tejedor

Assessing the burden of stroke is essential to improve stroke care. Two recent systematic reviews provide a picture of the worldwide stroke burden. However, poor-quality data sources in low-income and middle-income countries make this a blurred photograph, and set a challenge to expand stroke registries. Fuentes, B. & Tejedor, E. D. Nat. Rev. Neurol. 10, 127–128 (2014); published online 11 February 2014; corrected online 25 March 2014; doi:10.1038/nrneurol.2014.17

countries, as well as the limited quality of the patient registries. Two accompanying editorials in The Lancet and The Lancet Global Health have highlighted the worldwide coverage as one of the major strength of these studies. 3,4 Data from registries in 21 high-income, middle-income and low-income regions were included in both systematic reviews. However, this strength is also the Achilles heel of both these studies, as data sources from low-income and medium-income countries are limited and of questionable quality, as acknowledged by the authors of both the systematic reviews and the editorials. Surprisingly, data from the 191 low-income and middle-income countries included in the analysis was based on 61 articles, only 14 (23%) of which were rated as high-quality studies according to guidelines in the STROBE statement for assessing the quality of observational studies.5,6 Moreover, the majority (88.3%) of the studies selected for analysis in these two systematic reviews were carried out in 1990–2004, and only 14 were carried out in 2005–2010 (11.7%). Data for 2010, therefore, seem to be based on less than 12% of the total included studies. The small and nonsignificant increase in the prevalence of stroke in low-income and middle-income countries could also be explained by the limitations of the quality of the stroke registries, as has been previously suggested.3 The increases in the incidence of stroke and stroke-related mortality in these countries over the past two decades could be due to so-called epidemiological transition, caused by an increase in the incidence of vascular risk factors. Other possibilities that have so far been ignored, however, are increasing recognition of stroke symptoms

NATURE REVIEWS | NEUROLOGY

in these countries, and the possible development of improved registries that, on the surface, might contribute to the increased incidence of stroke. The incidence of stroke in low-income and middle-income countries is still underestimated, possibly owing to the competition with other diseases with a higher burden of disease than stroke, such as infectious diseases.7 Moreover, stroke is classifixed among noncommunicable diseases by the WHO. This classification is another potential reason for the underestimation of stroke incidence: it is not feasible to obtain data on the incidence of stroke or stroke-related mortality from official registries, as they do not provide enough data on noncommunicable diseases. The development of improved stroke registries should become a priority for future actions of the World Stroke Organization. The data from both of the new studies1,2 are of great importance in terms of setting a challenge to improve stroke registries in low-income and middle-income countries. Previous studies such as the WHO MONICA Project show that multinational comparisons of stroke incidence present considerable problems, the most important

Thinkstock

Two recently published systematic reviews have raised great interest in the global burden of stroke, which has increased over the past two decades. The first article focuses on the global and regional burdens of stroke, the main findings of which are increases in the absolute numbers of people with stroke and stroke-related deaths, as well as an increase in the global burden of stroke in terms of disability-adjusted life years (DALYs) in middle-income and low-income countries.1 The second study, which addresses the global and regional burden of first-ever ischaemic or haemorrhagic stroke, shows the burden of both of these types of stroke has increased in the past two decades in terms of the absolute number of people with stroke, stroke-related deaths, and DALYs, despite a decrease in age-s tandardized mortality. 2 Compared with ischaemic stroke, haemorrhagic stroke has a lower incidence, is responsible for a higher burden of disease, and is the cause of more stroke-related deaths and DALYs in middle-income and low-income countries.2 The good news from these studies is a decrease in age-standardized mortality over the past two decades in high-income countries, which could reflect an improvement in acute stroke management.1,2 The decrease in age-standardized incidence of stroke, also in high-income countries, probably relates to the success of preventative health-care policies. The reduction in agestandardized mortality over the past 20 years is less in low-income and middle-income countries than in high-income countries. This small reduction is probably due to the greater heterogeneity in reporting and case ascertainment among low-income to middleincome countries than among high-income

NEWS & VIEWS being incomplete case ascertainment.8 To facilitate the standardization of the datacollection process and improve the quality of stroke registers, the WHO, in collaboration with the World Stroke Organization, developed a manual of basic methods to establish a stepwise approach to set up stroke registries (STEPS Stroke) that has become a useful tool in low-income and middle-income countries.9 Moreover, it has been suggested that the involvement of national stroke associations in collaboration with the World Stroke Organization or regional institutions is crucial to ensuring the quality of stroke registers in low-income and middle-income countries. 9 In this context, we would like to highlight the Safe Implementation of Treatments in Stroke from the Iberoamerican Cerebrovascular Diseases Society as an example of a noninterventional monitored register for patients with acute stroke. This register was developed in the Iberoamerican region—in collaboration with the Karolinska Institute, Stockholm, Sweden—and could improve the knowledge of stroke burden in this g eographical area.10 In conclusion, these two systematic reviews are welcome as they report for the first time the global burden of stroke in terms of incidence, prevalence, mortality, DALYs lost, and age-stratified mortality to stroke incidence ratios. The global and regional data coverage shows differences in stroke burden between high-income and low-income or middle-income countries. Moreover, the studies report for the first time the differences in the global burden of two types of stroke—ischaemic and haemor rhagic. It is clear, however, that clinicians and researchers should make an effort to improve the quality of stroke registries, particularly in low-income and middle-income countries. Such an effort could ensure a clearer picture of the worldwide stroke burden, and could also aid promotion of acute care and stroke prevention measures in these countries. Department of Neurology and Stroke Centre, La Paz University Hospital, IdiPAZ Health Research Institute, Autonomous University of Madrid, Paseo de la Castellana 261, 28046 Madrid, Spain (B.F., E.D.T.). Correspondence to: E.D.T. exuperio.diez@idipaz.es Competing interests The authors declare no competing interests. 1.

Feigin, V. L. et al. Global and regional burden of stroke during 1990–2010: findings from the Global Burden of Disease Study 2010. Lancet 383, 245–255 (2014). Krishnamurthi, R. V. et al. Global and regional burden of first-ever ischaemic and

128 | MARCH 2014 | VOLUME 10

haemorrhagic stroke during 1990–2010: findings from the Global Burden of Disease Study 2010. Lancet Glob. Heal. 1, e259–e281 (2013). Hankey, G. J. The global and regional burden of stroke. Lancet Glob. Heal. 1, e239–e240 (2013). Giroud, M., Jacquin, A. & Béjot, Y. The worldwide landscape of stroke in the 21st century. Lancet 383 195–197 (2013). Von Elm, E. et al. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting of observational studies. Internist (Berl.) 49, 688–693 (2008). Global Burden of Disease Expert Group & Bennett, D. A. Methodology of the global and regional burden of stroke study. Neuroepidemiology 38, 30–40 (2012). The global burden of disease 2004 update. World Health Organization [online], http://

www.who.int/healthinfo/global_burden_ disease/GBD_report_2004update_full.pdf (2004). 8. Asplund, K. et al. Multinational comparisons of stroke epidemiology. Evaluation of case ascertainment in the WHO MONICA Stroke Study. World Health Organization monitoring trends and determinants in cardiovascular disease. Stroke 26, 355–360 (1995). 9. Truelsen, T. et al. Standard method for developing stroke registers in low-income and middle-income countries: experiences from a feasibility study of a stepwise approach to stroke surveillance (STEPS Stroke). Lancet Neurol. 6, 134–139 (2007). 10. SIECV SITS Iberoamerican Stroke register to be implemented in 2009. Safe Implementation of Treatments in Stroke [online], https:// sitsinternational.org/news/siecv-sitsiberoamerican-stroke-register-to-beimplemented-in-2009 (2009).

PARKINSON DISEASE

Intestinal levodopa infusion in PD —the first randomized trial Regina Katzenschlager and Werner Poewe

The first randomized, placebo-controlled trial of intrajejunal levodopa provides support for the benefits of continuous dopaminergic drug delivery. Motor fluctuations in patients with advanced Parkinson disease were markedly improved. In clinical practice, the symptomatic benefits of this treatment need to be weighed carefully against its adverse effect profile. Katzenschlager, R. & Poewe, W. Nat. Rev. Neurol. 10, 128–129 (2014); published online 18 February 2014; doi:1038/nrneurol.2014.26

Dopaminergic replacement therapies significantly improve motor function in most patients with Parkinson disease (PD); however, long-term levodopa therapy is frequently associated with troublesome motor fluctuations and drug-induced involuntary movements (dyskinesias). Results from the first placebo-controlled study assessing the benefits of continuous levodopa delivery support previous evidence for continuous delivery of dopaminergic drugs in PD:1 the daily ‘off ’-time in the continuous delivery group was almost 2 h shorter than in the immediate-release group. Although our understanding of the mechanisms that underlie the motor complications in PD remains incomplete, discontinuous drug delivery is thought to be involved. Under normal physiological conditions, striatal dopamine release is relatively stable. Pharmacotherapy with short-half-life drugs, such as levodopa, is associated with oscillating plasma levels of the drug and subsequent fluctuations in synaptic dopamine concentrations. Over time, this variability is

believed to induce maladaptive changes in basal ganglia motor circuits.2,3 Amelioration of motor response oscillations and, to some extent, dyskinesias in patients receiving continuous subcutaneous infusions of the dopamine agonist apomorphine or intra intestinal infusions of a gel formulation of levodopa has been repeatedly observed in uncontrolled, open-label case series. The trial by Olanow and colleagues, published recently in The Lancet Neurology,1 is the first ever randomized, placebocontrolled study comparing continuous pump delivery and oral immediate-release dopaminergic pharmacotherapy. In this 12-week multicentre study, all 71 patients underwent initial gastrostomy and were randomly assigned to either intrajejunal or oral immediate-release levodopa, receiving the alternative drug as a dummy, with all dose changes performed in both modalities during the titration phase.1 The mean baseline duration of daily ‘off ’ time was typical for patients enrolled into studies of advanced PD. The degree of improvement with www.nature.com/nrneurol

ADDENDUM The worldwide burden of stroke—a blurred photograph Blanca Fuentes and Exuperio Díez Tejedor Nat. Rev. Neurol. 10, 127–128 (2014); doi:10.1038/nrneurol.2014.17 The correspondence details for this article have changed: the authors wish to update the email address to exuperio.diez@idipaz.es. The details have been amended in the HTML and PDF versions of the article.

NATURE REVIEWS | NEUROLOGY

OPEN

ORIGINAL ARTICLE

Social interaction plays a critical role in neurogenesis and recovery after stroke VR Venna1, Y Xu1, SJ Doran1, A Patrizz1 and LD McCullough1,2,3 Stroke survivors often experience social isolation. Social interaction improves quality of life and decreases mortality after stroke. Male mice (20–25 g; C57BL/6N), all initially pair housed, were subjected to middle cerebral artery occlusion (MCAO). Mice were subsequently assigned into one of three housing conditions: (1) Isolated (SI); (2) Paired with their original cage mate who was also subjected to stroke (stroke partner (PH-SP)); or (3) Paired with their original cage mate who underwent sham surgery (healthy partner (PH-HP)). Infarct analysis was performed 72 h after stroke and chronic survival was assessed at day 30. Immediate poststroke isolation led to a significant increase in infarct size and mortality. Interestingly, mice paired with a healthy partner had significantly lower mortality than mice paired with a stroke partner, despite equivalent infarct damage. To control for changes in infarct size induced by immediate post-stroke isolation, additional cohorts were assessed that remained pair housed for three days after stroke prior to randomization. Levels of brain-derived neurotrophic factor (BDNF) were assessed at 90 days and cell proliferation (in cohorts injected with 5-bromo-2′-deoxyuridine, BrdU) was evaluated at 8 and 90 days after stroke. All mice in the delayed housing protocol had equivalent infarct volumes (SI, PH-HP and PH-SP). Mice paired with a healthy partner showed enhanced behavioral recovery compared with either isolated mice or mice paired with a stroke partner. Behavioral improvements paralleled changes in BDNF levels and neurogenesis. These findings suggest that the social environment has an important role in recovery after ischemic brain injury. Translational Psychiatry (2014) 4, e351; doi:10.1038/tp.2013.128; published online 28 January 2014 Keywords: brain-derived neurotrophic factor; infarct; middle cerebral artery occlusion; neurogenesis; post-stroke recovery; social isolation

INTRODUCTION It is increasingly accepted that social interactions are critical for maintaining physical, mental and social well-being. Social isolation has been associated with increased blood pressure, elevated cortisol levels,1 enhanced inflammatory and metabolic responses to stress,2,3 and modifications in transcriptional pathways linked with glucocorticoid and inflammatory signaling.4–7 Compared with individuals with social cohesion, isolated individuals have enhanced susceptibility for the development of cardiovascular disorders,8 infectious diseases,9 cognitive decline10 and increased risk of cerebrovascular disorders including ischemic stroke.11 Stroke is now the fourth leading cause of death in the United States.12 Stroke-related mortality has dropped over the past 10 years, primarily due to improvements in medical care and acute stroke treatment, but this has not been accompanied by a decrease in the incidence of ischemic stroke. This is alarming, as stroke is the leading cause of chronic disability in adults, and increases in the number of stroke survivors will further escalate health care costs over the next several decades as our population ages.12,13 Therefore, any potential opportunity to reduce stroke-related disability and improve functional outcome would have tremendous public health impact. Data from the US census show an increase in the number of people living alone since 1980s; a trend that is expected to

continue.14 Individuals who report lack of social support or isolation have an increased incidence of recurrent stroke, poorer recovery, greater functional decline and higher mortality following a stroke compared with individuals with strong social support.15–18 As not all strokes can be prevented, more attention is needed to develop strategies to improve function after a stroke has occurred. Attesting to the importance of social factors in stroke outcome is that these same effects can be reproducibly demonstrated in animals; social interaction has been shown to reduce histological damage from experimental stroke, whereas pre-stroke social isolation (SI) enhances injury.7,19,20 To date, previous studies have focused on isolation prior to, or at the time of, induction of injury.21,22 However, as most patients are not identiﬁed as ‘isolated’ until after their stroke, the examination of the effects of post-stroke isolation is important. In addition, as isolation increases infarct size, prior studies have been confounded by the larger infarcts seen in isolated animals. We hypothesized that social interaction will enhance functional recovery independently of infarct size and that the partner’s health status might have a role. To test this hypothesis, we utilized a novel delayed poststroke housing paradigm to normalize infarct size between groups. The effect of post-stroke environmental manipulation on histological outcomes, IL-6 levels, chronic behavioral recovery and neurogenesis was assessed.

1 Department of Neuroscience, University of Connecticut Health Center, Farmington, CT, USA; 2Department of Neurology, University of Connecticut Health Center, Farmington, CT, USA and 3The Stroke Center at Hartford Hospital, Hartford, CT, USA. Correspondence: Dr LD McCullough, Department of Neuroscience, University of Connecticut Health Center, 263 Farmington Avenue, Farmington, CT 06030, USA. E-mail: lmccullough@uchc.edu Received 13 September 2013; revised 25 November 2013; accepted 7 December 2013

Post-stroke social interaction hastens recovery VR Venna et al

2 MATERIALS AND METHODS Experimental animals Six-week-old C57Bl/6 mice weighing 20–22 g were purchased from Charles River laboratories (Wilmington, MA, USA) and allowed to acclimate in groups of four for two weeks prior to any housing manipulation. All mice were maintained in a temperature- and humidity-controlled vivarium, with ad libitum access to food and water.

with 100 μl of saline subcutaneously for 5 days after surgery. This prevents weight loss and reduces mortality, both stroke and sham mice were administered injections to control for any potential stress of handling or repeated injection. Quantiﬁcation of infarct volume and all behavioral tests were performed and recorded by an investigator blinded to housing conditions. All behavior tests were performed between 0800 and 1100 h, immediately at the beginning of the light cycle after acclimatizing mice to the testing room lighting conditions. Testing chambers were cleaned with 70% ethanol between mice.

Housing conditions and experimental design All experimental animals were screened for baseline locomotor activity in the open field. Mice were housed in standard mouse cages (11"L, 6"W, 6"H) with a 12-h light/dark schedule (lights on at 0700 h). Body weights were recorded and mice were pair housed (all mice were housed two per cage with an original cage mate) for 2 weeks before subjecting them to right middle cerebral artery occlusion (MCAO) or sham surgery. Animals were then randomly assigned to one of the three housing conditions—either housed isolated (SI), pair housed with a stroke partner (PH-SP) or pair housed with a healthy partner (PH-HP) either immediately after stroke (cohort-1) or three days after MCAO (cohort-2). In the first cohort, two endpoints were assessed; a three-day survival group was used for neurological deficit scores (NDS), infarct analysis and changes in serum interleukin-6 levels; a 30-day survival group that was used to assess mortality rates (see Figure 1a). A second cohort (cohort-2) was used to study delayed post-stroke housing effects with two different endpoints at 8 and 90 days. Animals were tested in the cylinder, corner and on the tail suspension test (TST) (see Figure 2a for timeline of behavioral testing) and sacrificed at day 90, for BDNF analysis. In the 8-day endpoint, mice were injected with BrdU from day 3–7 and analysis was performed 24 h after the last injection (day 8), an additional sub-group of mice was allowed to survive for 90 days to confirm neuronal maturation. All animals were given wet mash for the initial 72 h after stroke or sham surgery. As mortality is higher when chronic endpoints are assessed, all animals that had planned survival endpoints over 3 days were injected

Exclusion criteria If any mouse did not gain at least one gram of weight during the 2 weeks of pre-stroke pair housing, lost weight during the two weeks of PH, or exhibited signs of ﬁght wounds due to dominant/aggressive behavior of the partner, both mice were excluded from the study as social aversion/ stress can also inﬂuence stroke outcome.4,23,24 If one mouse of the pair dies at any point during the experiment, that pair is excluded from further study.

Middle cerebral artery occlusion model Focal transient cerebral ischemia was induced by right MCAO followed by reperfusion as described previously.7 Core body temperature was monitored with a rectal probe connected to a temperature control system (Fine science tools, North Vancouver, BC, Canada). Temperature was maintained with an automatic heating pad at ∼37 °C during surgery and ischemia. Cerebral blood flow measurements by laser Doppler flowmetry confirmed ischemia during MCAO (to o80% of baseline) and restoration of flow during reperfusion in all cohorts.

Infarct analysis Following 72 h of reperfusion, all animals in cohort 1 were euthanized by pentobarbital overdose, brains were removed and cut into ﬁve 2-mm

d0 (stroke)

d30

Pair housed for 2 weeks Assigned to one of the three housing condition

Mortality recorded

PH-SP

PH-HP

100

PH-HP

40 30 20

2 1 0 PH-SP PH-HP

PH-SP PH-HP

1.5 h

72 h

250 200

25-50%

* NS

50-75%

pg/ml

% of Infarct

PH-SP

80 70

Neurological deficit score

TTC staining

150

100 50

0 Cortex

Striatum

Total

PH-SP

PH-HP

Figure 1. Effects of post-stroke social isolation on stroke outcome: (a) Schematic of experimental design. Following a 60-min MCAO, mice were either single housed or pair housed with a partner mouse also subjected to stroke (PH-SP) or who was a healthy sham (PH-HP) on day 0. (b) Seventy-two hours after stroke (d3), brains were taken for histological analysis using TTC staining. (c) Quantification of infarcts at 72 h after stroke (d3) showed that SI mice had significantly larger ischemic damage compared with PH mice (*Po0.05); however, no significant differences were seen in mice that were pair housed with either a stroke partner or a healthy partner; n = 10 per group. (d) Neurological deficit scores prior to sacrifice showed a lack of recovery in SI cohort relative to intra-ischemic score, while both pair-housed groups recovered significantly (there were no significant differences between PH-SP and PH-HP groups); Friedman’s test was used for analysis, *Po0.05; n = 10 per group. (e) Post-stroke, SI mice had significantly enhanced serum levels of IL-6 compared with PH-SP and PH-HP mice; n = 6 per group; *Po 0.05 compared with PH groups. Error bars denote s.e.m. HP, healthy partner; PH, pair housed; MCAO, middle cerebral artery occlusion; SI, social isolation; SP, stroke partner; TTC, 2,3,5-triphenyltetrazolium chloride. Translational Psychiatry (2014), 1 – 9

Post-stroke social interaction hastens recovery VR Venna et al

BrDU 75 mg/kg/day

Stroke d0

Pair housed for 2 weeks

d14

d21

d90

d28

Cylinder & Corner testing Cell proliferation assessed on day 8

(Social Isolated)

Assigned to one of the three housing condition (SI/ paired with stroke partner / paired with healthy partner)

c 1.0

0.6 0.5

0.9

0.4

0.3

0.8

R/(R+L)

Laterality Index

Tail Suspension Test BDNF and neurogenesis analysis

0.2 0.1

14d

21d

-0.2

* **

0.6

21d

28d

0.5

0 -0.1

0.7

7d Stroke+SI Sham+SI

Stroke+PH-SP Sham+PH-SP

28d Stroke+PH-HP Sham+PH-HP

3d 0.4 0.3

7d Stroke+SI Stroke+PH-HP Sham+PH-SP

14d

Stroke+PH-SP Sham+SI Sham+PH-HP

Figure 2. Beneficial effects of delayed post-stroke pair housing on stroke outcome were independent of infarct volume. (a) Schematic of the experimental design used to assess effects of delayed housing on functional recovery. (b) Mice housed with a healthy partner (PH-HP) show enhanced recovery in the cylinder test by day 7 compared with PH-SP and SI. PH-HP cohorts almost fully recovered by day 21. Conversely, SI mice demonstrated significant forelimb impairment even at day 28 compared with shams and both pair-housed stroke groups. (c) Significant delay in recovery was also observed in SI mice in the corner test compared with PH-SP mice at day 14, whereas a significantly faster recovery was achieved in PH-HP mice by day 7 compared with PH-SP. *Significant difference (Po 0.05) compared with SI group; **Significant difference in PH-HP compared with PH-SP. Error bars denote s.e.m. HP, healthy partner; PH, pair housed; SI, social isolation; SP, stroke partner.

coronal sections and stained with 1.5% 2,3,5-triphenyltetrazolium chloride (TTC) for 8 min at 38 °C. Slices were formalin-ﬁxed (4%) and then digitalized and infarct volumes analyzed (Sigma Scan Pro, San Jose, CA, USA) as previously described.7 The ﬁnal infarct volumes are presented as a percentage (percentage of contralateral structures with correction for edema) as in Venna et al.7 For histology and immunohistochemistry, brains were collected from additional chronic survival cohorts. Animals were deeply anesthetized and perfused transcardially with ice-cold sodium phosphate-buffered saline followed by 4% paraformaldehyde. Brains were cryoprotected in 30% sucrose.

Corner test Following the cylinder test animals were tested in the corner test as described in (Venna et al.26). The mouse was placed between two pieces of cardboard, each with a dimension of 30 × 20 × 1 cm. The two boards were gradually moved to enclose the mouse from both sides to encourage the mouse to enter into a corner of 30° angle. After vibrissae stimulation, the mouse rears forward and upward, then turns back to face the opening. Twenty trials were performed for each mouse and the percentage of right turns was calculated.

Open field Neurological deficit scores Neurological deficit scores were obtained in the intra-ischemic period and at 72 h post-stroke in short-term survival cohort-1. The scoring system was as follows: 0, no deficit; 1, forelimb weakness and torso turning to one side when held by tail; 2, circling to affected side; 3, unable to bear weight on affected side and circling immediately when placed on a bench; and 4, no spontaneous locomotor activity or barrel rolling as described previously as in Venna et al.7

IL-6 ELISA Serum was collected from the mice sacriﬁced at 72 h post-stroke in mice subjected to the three housing conditions (SI, PH-SP or PH-HP) immediately after stroke. Interleukin-6 (IL-6) levels were examined utilizing an enzyme-linked immune absorbent (ELISA) assay (eBiosciences, San Diego, CA, USA) per manufacturer’s instructions.

Mice were placed in an open ﬁeld chamber (15" × 15") equipped with 16 infrared beam emitting LEDs. The total number of beam breaks were automatically registered by a computer-operated PAS Open Field system (San Diego Instruments, San Diego, CA, USA). The total number of beam breaks during a 20 min session was analyzed as a measure of spontaneous locomotor activity.

Tail suspension test Animals were returned to their home cages after open field analysis and allowed free access to food and water at least for 30 min before testing them in TST. TST was performed as in Venna et al.26 Mice were suspended horizontally by the tail to a metal bar using adhesive scotch tape for a total duration of 6 min. The test was analyzed and the times of active escape behavior (mobility) versus passive behavior with no limb or head movements (immobility) was recorded. Latency to first immobile period was also quantified27 and is reported as latency. Behavior was scored manually by an experienced rater blinded to housing condition.

Cylinder test

BrdU Treatment

The cylinder test was used to assess forelimb use after stroke. The mouse was placed in a transparent Plexiglas cylinder 9-cm diameter and 15 cm in height. A mirror was placed behind the cylinder with an angle to enable the rater to record forelimb movements when the mouse was turned away from the rater. A total of 20 forelimb placements were counted for analysis as previously described.7,25

Separate cohorts of mice were injected intraperitoneally with 75 mg kg − 1per day of BrdU once daily during days 3–7 after stroke.28 BrdU (Sigma Aldrich, MO, USA) was dissolved in 0.007 N NaOH in 0.9% NaCl. Mice were either sacriﬁced on day 8 or on day 90 (n = 4) to conﬁrm neuronal maturation. BrdU+ cell counts were performed on stained sections from perfused brains29,30 (n = 8 per group) as detailed below.

Translational Psychiatry (2014), 1 – 9

Post-stroke social interaction hastens recovery VR Venna et al

4 100

** Percent survival

signiﬁcant. Each dyad was treated as one individual sample for statistical analysis. Investigators performing MCAO, behavioral and infarct size analysis were blind to housing conditions.

* 60 Sham

SI PH-SP

PH-HP

0 0

Days

Figure 3. Chronic survival rates in isolated and pair-housed mice: survival rates were recorded in mice that were placed immediately into assigned housing conditions after stroke. Mortality rates assessed through day 30 revealed a signiﬁcantly higher mortality rate in SI mice compared with PH-SP, *Po0.05 and **P o0.01 in PHHP cohorts (n = 18 mice per group initially). Data presented in Kaplan–Meier curves and Z-test for proportions are used for statistical analysis. HP, healthy partner; PH, pair housed; SI, social isolation; SP, stroke partner.

Immunofluorescence Brains were cut in 30 micron sections and mounted on superfrost plus slides (Fisher scientific, Pittsburg, PA, USA), then subjected to microwave irradiation for 5 min in 0.1 M, pH 6 citrate buffer solution followed by incubation in blocking solution.7 Sections were incubated overnight with rat anti-BrdU (1:100, AbD Serotec, Raleigh, NC, USA) and mouse Anti-NeuN (1:200, Millipore, Billerica, MA, USA) to confirm neuronal lineage. Sections were subjected to three five-minute washes with PBS and then incubated for 60 min with fluorescein-conjugated anti-mouse and rhodamineconjugated anti-rat secondary antibodies (Alexafluor, Grand Island, NY). Then brain sections were washed, then dipped in DAPI solution (1:1000) for five minutes and cover slipped. Sections were visualized under an inverted light Zeiss axiovert fluorescence microscope.7 Newborn cells were identified by co-labeling with BrdU+ cells and DAPI. To perform quantification of BrdU+ cells, four coronal brain slices per animal (n = 8 animals per group) were stained from ~1.1 mm bregma, 0.8 mm bregma, 0.5 mm bregma and −0.1 mm bregma. Two 20 × fields per section were analyzed on the infarct side. BrdU+ cells were counted using MacBiophotonics ImageJ software (Hamilton, ON, Canada). The total numbers of cells per field of view were counted by a blinded investigator and the average number of cells per field was used for statistical analysis.29,30

Western blot Brains were rapidly extracted; hippocampi were isolated from brains and were immediately flash frozen. The protein concentration was determined by BCA Protein Assay Kit (Thermo Fisher Scientific Inc., Rockford, lL, USA) to achieve equal loading of 10 μg per well.7 Sample proteins were resolved on SDS electrophoresis gels and transferred to a polyvinylidene diflouride membrane. All blots were blocked with 5% milk and incubated overnight with BDNF (1:200; Millipore) and actin (1:5000; Sigma-loading control) at 4 ° C in TBS containing 4% bovine serum albumin and 0.1% Tween. Secondary antibodies (goat anti-rabbit IgG 1:5000 for BDNF, anti-mouse 1:1000 for actin) were diluted, and ECL detection kit (Amersham Biosciences, Buckinghamshire, UK) was used for signal detection. Densitometry was performed (Adobe). Data is presented as ratio to actin normalized to SI.

Statistics Data were expressed as mean ± s.e.m. except for NDS, which was presented as median (interquartile range). Statistics were performed using SPSS version PASW17. One-way analysis of variance (ANOVA) for analysis involving more than two independent groups, Bonferroni post hoc for multiple comparisons, repeated measures for cylinder and corner tests, a Ztest for proportions was used to analyze survival rates and the Friedman test with Bonferroni correction used for repeated measures ordinal data (i. e., NDS). A probability value, Po0.05 was considered to be statistically Translational Psychiatry (2014), 1 – 9

RESULTS Infarct analysis To examine whether post-stroke housing manipulation influences stroke outcome, all mice were initially pair housed for two weeks prior to a 60-min MCAO and assigned to one of the three ‘immediate post-stroke’ housing conditions: (1) Housed in a cage individually (SI), (2) paired with its original partner also subjected to stroke surgery (PH-SP) (3) paired with its original partner (not subjected to stroke) (PH-HP) (Figure 1a). Regardless of the partner’s health condition, pair-housed mice had significantly smaller infarcts 72 h after stroke compared with SI cohorts in the cortex F(2,26) = 5.12, Po 0.05; striatum F(2,26) = 7.24, Po 0.01 and in the total hemisphere F(2,26) = 13.03, P o0.01. No significant differences were seen in infarct volumes between PH-HP and PHSP cohorts; the mean difference (MD) in cortex = 4.3, P = 0.644; MD striatum = 0.344, P = 0.997 and MD total = 1.56, P = 0.93 (Figure 1c). Neurological deficit scores and serum IL-6 levels The detrimental effects of SI were also reflected in the NDS, Friedman test for repeated measures revealed no recovery in SI mice χ2(1) = 0.0001; P = 1.0, but a significant recovery in both PHSP χ2(1) = 10, P = 0.002 and in PH-HP χ2(1) = 9; P = 0.003 (Figure 1d). SI mice had significantly elevated serum IL-6 compared with PH cohorts after stroke F(2,15) = 16.38, P o 0.05. (Figure 1e), post hoc tests revealed no significant difference between PH-SP and PH-HP groups; P>0.05. Sham mice has no significant changes in serum IL-6 levels F(2,13) = 0.64, P = 0.54. These results revealed a similar pattern to that previously seen in mice exposed to pre-stroke SI with isolated animals sustaining larger infarcts, an enhanced inflammatory response and poorer functional recovery than animals housed together.7,19,20 Exclusion and mortality rates A total of 4% of original mouse pairs were excluded at the beginning of the study due to partner aggression as per pre-set exclusion criteria. In cohort-1 (immediate post-stroke isolation) SI mice had significantly higher mortality (61%) compared with PHSP (39%), P o 0.05. PH-HP mice had only a 17% mortality (P o0.01) (Figure 3) at 30 days. These findings suggest that interactions with a healthy partner might offer additional benefit and improve poststroke survival. These effects appear to be independent of the effects on acute infarct damage, which was equivalent between the two cohorts of PH mice. Survival rates in cohort-2 (delayed post-stroke isolation) were higher as post-stroke care was more aggressive, but SI mice still had significantly higher mortality (20%) compared with PH-SP (10%); (Po 0.05), and in PH-HP (3%); (P o 0.001). Functional recovery is independent of infarct size and associated with interactions and health of the partner One possible explanation for the improved recovery in PH mice is that infarct damage is less due to the acute neuroprotection induced by immediate post-stroke housing. To investigate the specificity of relationship between social interaction, partner’s health and post-stroke recovery, we subjected mice to a ‘delayed post-stroke’ housing paradigm (Figure 2a). As infarct damage is complete in this stroke model by 24 h,31 we allowed the infarct to mature prior to housing in an attempt to normalize infarct size between groups. All mice were kept pair housed for 72 h immediately after stroke. In this way, recovery can be separated from the degree of histological damage, which will be equivalent © 2014 Macmillan Publishers Limited

Post-stroke social interaction hastens recovery VR Venna et al

5 in all cohorts. Mice were then placed back with its previous partner who was either a ‘healthy’ or ‘stroke’ mouse or maintained in isolation (SI). Long-term recovery was assessed with both the cylinder and corner test.25 Repeated measures ANOVA revealed that there were significant asymmetries (reduced contralateral forelimb use) in the cylinder test in stroke cohorts compared with sham animals regardless of housing condition. There was an overall significant effect of surgery F(1,44) = 309.09, P o 0.01 (Figure 2b), an overall significant effect of housing F(2,44) = 18.3, P o0.05. There was also significant surgery-by-housing interaction, F(2,44) = 18.83, P o 0.01 suggesting differences in recovery. Independent analyses at each time point (days 3–28) revealed that these observed effects are due to more rapid and complete forelimb motor recovery in PH-HP mice compared with PH-SP; P o 0.05. SI mice had significantly reduced contralateral limb use compared with PH cohorts even at day 28 (*P o 0.01) (Figure 2b). Sham operated mice had no significant differences between groups. Figure 2c illustrates the results of mean non-impaired side turn percentages using the corner test that showed a consistent pattern of recovery. A significant effect of day-by-surgery F(1,44) = 53.73, P o 0.01 was seen, due to the fact that there was an early reduction in turning preference in PHHP compared with PH-SP; SI mice had a continued turning preference at day 29 compared with PH cohorts. There was also significant surgery F(1,44) = 321.22, P o0.05 and a housing F (2,44) = 8.87; P o 0.01 for the entire testing period. Long-term social isolation reduced mobility in tail suspension test and reduced hippocampal BDNF levels To explore the role of social interaction on depressive phenotypes, mice were subjected to the ‘delayed post-stroke’ housing paradigm detailed above (Figure 2a). When changes in mobility using the TST were evaluated, there was a significant difference between groups F(3,32) = 18.8, P o0.01; Bonferroni post hoc analysis revealed a significantly reduced mobility in SI group

Functional recovery paralleled enhanced neurogenesis Brains were analyzed for histological assessment at 8 days post stroke, no significant differences in infarct size between groups was seen (Figure 5a). Equivalent behavioral deficits were seen at day 3 post stroke prior to housing randomization. Groups (SI vs PH) significantly diverged in their recovery by post-stroke day 7. As ischemia is known to induce neurogenesis, which may be an important factor in functional recovery,32,33 we assessed strokeinduced cell proliferation on day 8 after stroke in the

250

100

200

* 150 100

40 20

0 Shams

PH-SP

Shams

PH-HP

3000 2500 2000 1500 1000 500 0 PH-SP

PH-HP

2 1 0

PH-SP

Actin 4

Shams

BDNF

Relative optical density normalized to actin

Total beam breaks

80 Latency

Duration of mobility (sec.)

compared with both pair-housed conditions P o 0.05, the PH-SP group had significantly less mobility compared with PH-HP (P = 0.04) (Figure 4a). When analyzed for latency to the first bout of immobility, there were significant differences between groups F (3,32) = 23.73, P o 0.01. Bonferroni post hoc analysis revealed significant differences in latencies between SI and PH-SP (P = 0.02) and also a significant difference between PH-HP compared with PH-SP groups (P = 0.03). Pair housing significantly reduced the duration of immobility and prolonged the latency to the first bout of immobility compared with SI group (Figure 4b). Independent analysis showed no differences between sham groups, data are presented as pooled for sham groups (Figure 4). Spontaneous locomotor analysis in open field showed no differences between groups F(3,32) = 0.98; P = 0.42, suggesting that decreased mobility is not due to motor deficits (Figure 4c). Assessment of hippocampal brain-derived neurotrophic factor (BDNF) levels, an important growth factor for neuronal survival and maturation with known antidepressant effects,23,26 showed no differences in sham groups F(2,5) = 0.44, P>0.05; but revealed a significant difference between stroke groups F(3,28) = 20.12, P o0.05. Bonferroni post hoc tests showed that there was a significant reduction of BDNF expression in SI compared with PHSP (P = 0.008) and PH-HP (P o0.001). BDNF levels in PH-HP were higher than PH-SP (P = 0.048) but not significantly different from pooled sham group, P = 0.9 (Figure 4d).

Sham

PH-SP

PH-HP

Figure 4. Housing with a healthy partner significantly improved long-term behavior and enhanced BDNF levels compared with isolated mice or mice housed with a stroke partner. (a) SI mice showed a depression-like phenotype, expressing less mobility when tested at day 90 using the tail suspension test compared with PH-HP and shams. (b) SI mice also expressed a reduced latency to the first bout of immobility compared with PH-HP and shams. (c) All stroke mice expressed similar spontaneous locomotor activity when tested prior to TST. (d) BDNF levels were significantly reduced at day 90 in whole brain homogenates of SI mice. β-actin was used as loading control and densitometry data are presented as normalized ratio. *P o0.05 compared with SI and **P o0.05 compared with PH-SP. Error bars denote s.e.m. HP, healthy partner; PH, pair housed; SI, social isolation; SP, stroke partner. © 2014 Macmillan Publishers Limited

Translational Psychiatry (2014), 1 – 9

Post-stroke social interaction hastens recovery VR Venna et al

100

% Infarct

80 60 NS 40

ii)

iii)

iv)

20 0 SI

PH-SP

PH-HP

PH-SP

PH-HP

Infarct area

SVZ

BrDU cells/focal view in infarct area

BrDU +ve cells/focal view in SVZ

50 40 30 20 10

100 80

40 20 0

0 SI

PH-SP

PH-HP

PH-SP

PH-HP

Figure 5. Functional recovery in pair-housed mice paralleled the increase in BrDU+ cell proliferation. (a) Similar infarct volumes were observed between groups with delayed housing when analyzed at day 8 (n = 8 per group). (b) Representative illustration showing how the brain sections were obtained for cell counts; a total of 32 images per group were counted, 4 images per brain and 8 brains per group; images from red-boxed area were used to count BrDU+ cells in SVZ and images from white-boxed area were used to obtain average BrDU+ cell counts within the infarct. (c) Images showing differences in cell proliferation rates seen in PH mice compared with isolated mice at day 8 suggest a potential implication of this mechanism in recovery; bar = 100 μm. (d) Data represent BrDU+ cells count from 32 images collected from SVZ (n = 8 mice per group). (e) Quantification of BrDU+ cells in infarct area shows significant differences in SI vs PH-SP and in PH-SP vs PH-HP groups. (f) Image showing co-labeling confirms the maturation of newborn cells to neurons, marked with BrDU+NeuN. *P o0.05 compared with SI and **P o0.05 compared with PH-SP. Error bars denote s.e.m. HP, healthy partner; PH, pair housed; SI, social isolation; SP, stroke partner; SVZ, subventricular zone.

subventricular zone in additional cohorts of animals injected with BrdU at days 3–7 after stroke. SI mice had significantly fewer BrdU+ cells compared with either PH cohorts F(2,96) = 60.98, Po0.001 (Figure 5c). There was an additional effect in mice housed with a healthy partner (PH-HP) compared with mice that PH with a stroke partner (PH-SP) at day 8; MD 16.6, Po0.01 (Figure 5c). Quantification revealed that SI mice had significantly fewer BrdU+ cells compared with either PH cohorts F(2,96) = 66.07, Po0.001 in the infarct area. There was an additional effect in mice housed with a healthy partner (PH-HP) compared with mice that PH with a stroke partner (PH-SP) at day 8; MD 22.8, Po0.01 (Figure 5d), these changes are consistent with the enhanced recovery seen in that cohort. When assessed for neuronal maturation on confocal brain Translational Psychiatry (2014), 1 – 9

sections obtained at day 90, we found that a number of BrdU+ cells exhibited markers of mature neurons (NeuN) (Figure 5f). DISCUSSION This study has established that early post-stroke interactions can reduce histological damage after ischemic challenge. These findings are consistent with the beneficial effects of pair housing observed in pre-stroke isolation models, and suggest that this may have important translational relevance for treatment of stroke patients.21 More importantly, this work shows that pair housing, even if delayed for days after stroke, leads to enhanced functional recovery after injury, even in the absence of a significant reduction © 2014 Macmillan Publishers Limited

Post-stroke social interaction hastens recovery VR Venna et al

7 of histological damage. This novel finding suggests that there is considerable brain plasticity and capacity for repair, even days after injury, enhancing the translational significance of this work. Our results also demonstrate that housing stroke mice with a healthy partner significantly hastens functional recovery and reduces mortality compared with mice housed with a stroke partner, whereras isolated mice demonstrated the poorest recovery. These effects parallel increased cell proliferation and higher hippocampal BDNF levels. In previous work, we have demonstrated that pre-stroke social isolation is detrimental.7 However, as stroke is an unpredictable event, manipulating social environments prior to an event in patients at risk is likely to be ineffective from a therapeutic standpoint. Therefore we initially wanted to determine if these same manipulations performed after injury would have similar effects at a time when patients would be presenting to medical care. We found that immediate post-stroke pair housing significantly decreased infarct volume, and this was seen in animals paired with either a healthy partner or a partner that also had been subjected to stroke. We further hypothesized that housing with a sick partner would reduce the benefit of PH while a healthy partner may further ameliorate beneficial effects of PH. When tested, no significant differences were observed between PH-HP and PH-SP in either the amount of histological injury or serum IL-6 levels. We also found a similar improvement in functional outcome, as measured by the NDS, regardless of the health of the cage partner. This suggests that the detrimental effects of acute isolation might also be due to pair dissociation/ separation effects.34 However, an unexpected and interesting pattern emerged when these animals were followed for 30 days, which was the pre-specified behavioral endpoint. Despite having similar infarct sizes, mice housed with a healthy partner (PH-HP) had significantly higher survival rates than mice housed with a partner that had also been subjected to stroke (PH-SP), although both PH cohorts did better than the isolated cohort. This suggested that there were unknown factors that enhanced survival in mice paired with a healthy partner. One limitation of the previously published reports examining social factors and stroke is that all have utilized pre-stroke or immediate post-stroke isolation which invariably leads to an increased infarct size.7,19,20,22 Therefore any increase in inflammation, apoptosis or behavioral deficits in isolated cohorts may simply reflect the amount of ischemic injury. To address this and examine the potential therapeutic window for environmental manipulations, we waited 3 days to place mice in the specified housing environment, as infarct is complete in this model by 72 h after injury.31 This allowed us to control the size of the infarct between the three groups, and may be a more realistic ‘therapeutic time window’ in which to treat an acute stroke patient with environmental manipulations. Despite equivalent histological damage in all groups, isolated mice continued to do poorly, with delayed functional recovery and higher mortality. Mice housed with a healthy partner had significantly better behavioral recovery than those housed with a stroke partner. This suggests that there are factors that contribute to the detrimental effects of isolation7,19,20 such as IL-6, NF-kappaB as well as additional factors that underlie the beneficial effects of affiliative interactions such as grooming, whisker stimulation,35,36 which can be further enhanced based on partner’s health. Although all mice were screened and monitored for aggressive behavior towards the partner and we excluded those dyads from the study, we cannot completely rule out the possibility that stroke induced aggressive behavioral changes by the healthy partner. We feel this is less likely as stroke animals housed with a healthy partner actually did much better both in terms of behavior and mortality than mice housed with a stroke partner. In turn, stroke mice housed with another stroke mouse had better outcomes than stroke mice that were isolated. This suggests that housing was beneficial compared © 2014 Macmillan Publishers Limited

with isolation as the isolated mice are clearly not exposed to any partner, aggressive or otherwise. We hypothesized that the enhanced behavioral recovery seen with PH was secondary to an enhancement in neurogenesis, as has been described with environmental enrichment.37,38 Neurogenesis occurs throughout life in both rodents and humans.39,40 Cerebral ischemia increases cell proliferation in both the subventricular zone and the hippocampus, and these newborn cells can migrate into the infarcted area.30 Enhancements in neurogenesis have been linked to improved functional recovery in many studies,41 but how these cells integrate into the damaged brain and the ultimate potential to enhance repair after stroke is unknown. Both PH-HP and PH-SP mice demonstrated increased cell proliferation rates, which paralleled changes in functional recovery. BDNF is an important mediator of neuronal activity, neurogenesis, brain repair, cell survival and synaptic structure formation.38,41,42 Activation of BDNF and its receptor tropomyosinrelated kinase B (TrkB) enhanced neuronal plasticity and neurogenesis after ischemic injury in mice.43 Recently it has been shown that transplantation of mesenchymal stem cells reduced social deficits in social preference tests mainly by enhancing BDNF levels.44 As we also found a reduction in depression-like behavior in PH mice compared with isolated cohorts, and BDNF signaling has also been implicated in the pathophysiology of depression,45 we examined BDNF levels, which were increased in PH animals compared with SI and further enhanced in mice with a healthy partner. Over ~40% of stroke survivors suffer from depression, and this may be exacerbated by isolation and feelings of loneliness.16,46 Many epidemiological studies have shown that depression is a risk factor for stroke, and is associated with poorer recovery once a stroke has occurred.47,48 The downstream mechanism by which PH prevented depression-like behavior is not known, but likely involves numerous pathways. Nibuya et al.49 showed that antidepressant drugs increase the synthesis of BDNF.49 Traditional antidepressants may act by enhancing BDNF signaling50,51 as BDNF expression increased in depressed patients after treatment.50,52 BDNF activation is essential for manifestation of experience-dependent social aversion and mediating changes in long-term neural and behavioral plasticity in response to social aggression.23 Our work suggests that BDNF also has an important role in the beneficial effects of social interaction after injury. Whether these same effects can be recapitulated with pharmacological treatments is not yet known. In summary, this study provides the first direct experimental evidence to support a role of post-stroke social interaction in stroke recovery independent of infarct damage. Our findings emphasize that although pair housing is beneficial to recovery, the health of the partner is also an important contributing factor to functional recovery. Given that these effects parallel the enhanced cell proliferation and increased hippocampal BDNF levels, more research is needed to understand how the beneficial effects of social interaction could be translated to clinical practice. Recent results from trials designed to provide reciprocal peer support for heart failure patients found that more than half of the participants had no or minimal engagement with the program, and no improvements in outcomes were seen.53 It is possible that social support may be more effective when provided by a healthy individual rather than a disease-affected peer. As a rehabilitative strategy, efforts to encourage interactions among stroke survivors and healthy society members could offer an important opportunity to reduce the substantial burden of stroke in our communities.

CONFLICT OF INTEREST The authors declare no conﬂict of interest.

Translational Psychiatry (2014), 1 – 9

Post-stroke social interaction hastens recovery VR Venna et al

8 ACKNOWLEDGMENTS We thank Dr Lauren Sansing for her help with the statistical analysis. This work was supported by the NIH (NINDS RO1 NSO77769 to LDM) and the AHA (1POST7430045 to VRV).

REFERENCES 1 Grant N, Hamer M, Steptoe A. Social isolation and stress-related cardiovascular, lipid, and cortisol responses. Ann Behav Med 2009; 37: 29–37. 2 Loucks EB, Berkman LF, Gruenewald TL, Seeman TE. Relation of social integration to inflammatory marker concentrations in men and women 70–79 years. Am J Cardiol 2006; 97: 1010–1016. 3 Powell ND, Sloan EK, Bailey MT, Arevalo JM, Miller GE, Chen E et al. Social stress up-regulates inflammatory gene expression in the leukocyte transcriptome via beta-adrenergic induction of myelopoiesis. Proc Natl Acad Sci USA 2013; 110: 16574–16579. 4 Sugo N, Hurn PD, Morahan MB, Hattori K, Traystman RJ, DeVries AC. Social stress exacerbates focal cerebral ischemia in mice. Stroke 2002; 33: 1660–1664. 5 Ploughman M, Windle V, MacLellan CL, White N, Dore JJ, Corbett D. Brain-derived neurotrophic factor contributes to recovery of skilled reaching after focal ischemia in rats. Stroke 2009; 40: 1490–1495. 6 Yang YC, McClintock MK, Kozloski M, Li T. Social isolation and adult mortality: the role of chronic inflammation and sex differences. J Health Soc Behav 2013; 54: 183–203. 7 Venna VR, Weston G, Benashski SE, Tarabishy S, Liu F, Li J et al. Nf-kappab contributes to the detrimental effects of social isolation after experimental stroke. Acta Neuropathol 2012; 124: 425–438. 8 Barth J, Schneider S, von Kanel R. Lack of social support in the etiology and the prognosis of coronary heart disease: a systematic review and meta-analysis. Psychosom Med 2010; 72: 229–238. 9 Cohen S, Doyle WJ, Skoner DP, Rabin BS, Gwaltney JM Jr. Social ties and susceptibility to the common cold. Jama 1997; 277: 1940–1944. 10 Cacioppo JT, Hawkley LC. Perceived social isolation and cognition. Trends Cogn Sci 2009; 13: 447–454. 11 Avendano M, Kawachi I, Van Lenthe F, Boshuizen HC, Mackenbach JP, Van den Bos GA et al. Socioeconomic status and stroke incidence in the us elderly: the role of risk factors in the epese study. Stroke 2006; 37: 1368–1373. 12 Go AS, Mozaffarian D, Roger VL, Benjamin EJ, Berry JD, Borden WB et al. Executive summary: heart disease and stroke statistics--2013 update: a report from the American heart association. Circulation 2013; 127: 143–152. 13 Vaartjes I, O'Flaherty M, Capewell S, Kappelle J, Bots M. Remarkable decline in ischemic stroke mortality is not matched by changes in incidence. Stroke 2013; 44: 591–597. 14 US_Census_Bureau http://www.Census.Gov/hhes/families/data/cps2012.html. 2012. 15 Hawkley LC, Cacioppo JT. Loneliness and pathways to disease. Brain Behav Immun 2003; 17(Suppl 1): S98–105. 16 Boden-Albala B, Litwak E, Elkind MS, Rundek T, Sacco RL. Social isolation and outcomes post stroke. Neurology 2005; 64: 1888–1892. 17 Cacioppo JT, Hughes ME, Waite LJ, Hawkley LC, Thisted RA. Loneliness as a specific risk factor for depressive symptoms: cross-sectional and longitudinal analyses. Psychol Aging 2006; 21: 140–151. 18 Holt-Lunstad J, Smith TB, Layton JB. Social relationships and mortality risk: a metaanalytic review. PLoS Med 2010; 7: e1000316. 19 Craft TK, Glasper ER, McCullough L, Zhang N, Sugo N, Otsuka T et al. Social interaction improves experimental stroke outcome. Stroke 2005; 36: 2006–2011. 20 Karelina K, Norman GJ, Zhang N, Morris JS, Peng H, DeVries AC. Social isolation alters neuroinflammatory response to stroke. Proc Natl Acad Sci USA 2009; 106: 5895–5900. 21 Venna VR, McCullough LD. ‘won't you be my neighbor?’: Deciphering the mechanisms of neuroprotection induced by social interaction. Stroke 2011; 42: 3329–3330. 22 O’Keefe LM, Doran SJ, Mwilambwe-Tshilobo L, Conti LH, Venna VR, McCullough LD. Social isolation after stroke leads to depressive-like behavior and decreased bdnf levels in mice. Behav Brain Res 2014; 260: 162–170. 23 Berton O, McClung CA, Dileone RJ, Krishnan V, Renthal W, Russo SJ et al. Essential role of bdnf in the mesolimbic dopamine pathway in social defeat stress. Science 2006; 311: 864–868. 24 Barik J, Marti F, Morel C, Fernandez SP, Lanteri C, Godeheu G et al. Chronic stress triggers social aversion via glucocorticoid receptor in dopaminoceptive neurons. Science 2013; 339: 332–335. 25 Li X, Blizzard KK, Zeng Z, DeVries AC, Hurn PD, McCullough LD. Chronic behavioral testing after focal ischemia in the mouse: Functional recovery and the effects of gender. Exp Neurol 2004; 187: 94–104.

Translational Psychiatry (2014), 1 – 9

26 Venna VR, Deplanque D, Allet C, Belarbi K, Hamdane M, Bordet R. Pufa induce antidepressant-like effects in parallel to structural and molecular changes in the hippocampus. Psychoneuroendocrinology 2009; 34: 199–211. 27 Castagné V, Moser P, Roux S, Porsolt RD. Rodent models of depression: forced swim and tail suspension behavioral despair tests in rats and mice. Curr Protoc Neurosci 2011; Chapter 8: Unit 8.10A. 28 Liu F, You Y, Li X, Ma T, Nie Y, Wei B et al. Brain injury does not alter the intrinsic differentiation potential of adult neuroblasts. J Neurosci 2009; 29: 5075–5087. 29 Li Y, Yu SP, Mohamad O, Genetta T, Wei L. Sublethal transient global ischemia stimulates migration of neuroblasts and neurogenesis in mice. Transl Stroke Res 2010; 1: 184–196. 30 Jin K, Minami M, Lan JQ, Mao XO, Batteur S, Simon RP et al. Neurogenesis in dentate subgranular zone and rostral subventricular zone after focal cerebral ischemia in the rat. Proc Natl Acad Sci USA 2001; 98: 4710–4715. 31 Liu F, Schafer DP, McCullough LD. TTC, fluoro-jade b and neun staining confirm evolving phases of infarction induced by middle cerebral artery occlusion. J Neurosci Methods 2009; 179: 1–8. 32 Chopp M, Zhang ZG, Jiang Q. Neurogenesis, angiogenesis, and mri indices of functional recovery from stroke. Stroke 2007; 38: 827–831. 33 Li WL, Yu SP, Ogle ME, Ding XS, Wei L. Enhanced neurogenesis and cell migration following focal ischemia and peripheral stimulation in mice. Dev Neurobiol 2008; 68: 1474–1486. 34 Feltenstein MW, Lambdin LC, Webb HE, Warnick JE, Khan SI, Khan IA et al. Corticosterone response in the chick separation-stress paradigm. Physiol Behav 2003; 78: 489–493. 35 Wilby MJ, Hutchinson PJ. The pharmacology of chlormethiazole: a potential neuroprotective agent? CNS Drug Rev 2004; 10: 281–294. 36 Lay CC, Davis MF, Chen-Bee CH, Frostig RD. Mild sensory stimulation reestablishes cortical function during the acute phase of ischemia. J Neurosci 2011; 31: 11495–11504. 37 Nilsson M, Perfilieva E, Johansson U, Orwar O, Eriksson PS. Enriched environment increases neurogenesis in the adult rat dentate gyrus and improves spatial memory. J Neurobiol 1999; 39: 569–578. 38 Rossi C, Angelucci A, Costantin L, Braschi C, Mazzantini M, Babbini F et al. Brainderived neurotrophic factor (bdnf) is required for the enhancement of hippocampal neurogenesis following environmental enrichment. Eur J Neurosci 2006; 24: 1850–1856. 39 Eriksson PS, Perfilieva E, Bjork-Eriksson T, Alborn AM, Nordborg C, Peterson DA et al. Neurogenesis in the adult human hippocampus. Nat Med 1998; 4: 1313–1317. 40 van Praag H, Kempermann G, Gage FH. Running increases cell proliferation and neurogenesis in the adult mouse dentate gyrus. Nat Neurosci 1999; 2: 266–270. 41 Hermann DM, Chopp M. Promoting brain remodelling and plasticity for stroke recovery: therapeutic promise and potential pitfalls of clinical translation. Lancet Neurol 2012; 11: 369–380. 42 Murphy TH, Corbett D. Plasticity during stroke recovery: from synapse to behaviour. Nat Rev Neurosci 2009; 10: 861–872. 43 Han X, Wang W, Xue X, Shao F, Li N. Brief social isolation in early adolescence affects reversal learning and forebrain bdnf expression in adult rats. Brain Res Bull 2011; 86: 173–178. 44 Barzilay R, Ben-Zur T, Sadan O, Bren Z, Taler M, Lev N et al. Intracerebral adult stem cells transplantation increases brain-derived neurotrophic factor levels and protects against phencyclidine-induced social deficit in mice. Transl Psychiatry 2011; 1: e61. 45 Groves JO. Is it time to reassess the bdnf hypothesis of depression? Mol Psychiatry 2007; 12: 1079–1088. 46 Hafner S, Emeny RT, Lacruz ME, Baumert J, Herder C, Koenig W et al. Association between social isolation and inflammatory markers in depressed and nondepressed individuals: results from the monica/kora study. Brain Behav Immun 2011; 25: 1701–1707. 47 Gainotti G, Antonucci G, Marra C, Paolucci S. Relation between depression after stroke, antidepressant therapy, and functional recovery. J Neurol Neurosurg Psychiatry 2001; 71: 258–261. 48 Jonas BS, Mussolino ME. Symptoms of depression as a prospective risk factor for stroke. Psychosom Med 2000; 62: 463–471. 49 Nibuya M, Morinobu S, Duman RS. Regulation of bdnf and trkb mrna in rat brain by chronic electroconvulsive seizure and antidepressant drug treatments. J Neurosci 1995; 15: 7539–7547. 50 Shimizu E, Hashimoto K, Okamura N, Koike K, Komatsu N, Kumakiri C et al. Alterations of serum levels of brain-derived neurotrophic factor (bdnf) in depressed patients with or without antidepressants. Biol Psychiatry 2003; 54: 70–75.

Post-stroke social interaction hastens recovery VR Venna et al

9 51 Mattson MP, Maudsley S, Martin B. Bdnf and 5-ht: A dynamic duo in age-related neuronal plasticity and neurodegenerative disorders. Trends Neurosci 2004; 27: 589–594. 52 Baj G, D’Alessandro V, Musazzi L, Mallei A, Sartori CR, Sciancalepore M et al. Physical exercise and antidepressants enhance bdnf targeting in hippocampal ca3 dendrites: further evidence of a spatial code for bdnf splice variants. Neuropsychopharmacology 2012; 37: 1600–1611.

53 Heisler M, Halasyamani L, Cowen ME, Davis MD, Resnicow K, Strawderman RL et al. Randomized controlled effectiveness trial of reciprocal peer support in heart failure. Circ Heart Fail 2013; 6: 246–253. This work is licensed under a Creative Commons AttributionNonCommercial-ShareAlike 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-sa/3.0/

Translational Psychiatry (2014), 1 – 9