Medicine & Science in Sports & Exercise:
Clinical Sciences: Clinically Relevant
Reduced Amount and Disrupted Temporal Pattern of Spontaneous Exercise in Diabetic Rats
VAN LUNTEREN, ERIK1,2; MOYER, MICHELLE1; POLLARINE, JENNIFER1
1Department of Medicine (Pulmonary), Cleveland Department of Veterans Affairs Medical Center and Case Western Reserve University, and 2Department of Neurosciences, Case Western Reserve University, Cleveland, OH
Address for correspondence: Erik van Lunteren, M.D., Pulmonary 111J(W), Cleveland Department of Veterans Affairs Medical Center, 10701 East Boulevard, Cleveland, OH 44106; E-mail: email@example.com.
Submitted for publication November 2003.
Accepted for publication June 2004.
This study was supported in part by the Department of Veterans Affairs Medical Research Service and by NIH grant HL-70697.
Introduction/Purpose: The beneficial effects of exercise for subjects with diabetes or prediabetic states are well established. However, the converse, that is, the effect of diabetes on spontaneous exercise performance, is not as well defined. Mice with mdx muscular dystrophy not only reduce total spontaneous running distance, but also decrease the duration of periods during which they are active, suggesting a defect in endurance. Studies tested the hypothesis that Type I diabetes causes similar changes in spontaneous exercise performance.
Methods: Wistar rats received streptozotocin to produce a model of Type I diabetes or buffer alone, and had access to running wheels for the next 8 wk.
Results: Diabetic rats had elevated serum glucose levels (689 ± 85 vs 270 ± 21 mg·dL−1, P = 0.0003) but normal serum bicarbonate levels. After 8 wk, diabetic rats were running for considerably lower distances than normal animals (daily distance 182 ± 58 vs 4981 ± 1373 m, P = 0.006). Furthermore, the average consecutive running time was much shorter in diabetic than normal rats (16 ± 1 vs 40 ± 6 min, P = 0.004). Differences in running behavior between diabetic and normal mice were absent early after injection of streptozotocin, but were fully established by week 4 for both total distance and consecutive running times.
Conclusion: Severe untreated Type I diabetes in rats reduces spontaneous exercise in a manner similar to that seen in mdx mouse muscular dystrophy, with reduced running distance and consecutive running times.
Exercise is an important factor in preventing Type II diabetes (27). In addition, it exerts profound beneficial effects on glycemic control in already established diabetic subjects (1) and, more importantly, reduces both all-cause and cardiovascular disease-specific mortality in humans with diabetes (7). The converse is not as well delineated; that is, does diabetes affect exercise capacity and spontaneous exercise performance? If diabetes reduces exercise performance, then the potential beneficial effects of exercise will be attenuated by the low levels of exercise participation. It does appear that humans with both Type I diabetes (24) and Type II diabetes (12,18) have reduced maximal exercise capacity during formal testing. However, the effect of diabetes on spontaneous exercise performance is a more relevant measure for population-wide disease prevention programs, as imposing directly supervised exercise programs on large numbers of subjects with diabetes is neither practical nor affordable. Assessing effects of disease on spontaneous exercise performance is difficult in human populations, due to the many psychosocial factors that modulate exercise behavior in people.
Rodent models have provided valuable information about spontaneous exercise performance in disease states. Mice with mdx muscular dystrophy have the analogous genotype as humans with Duchenne muscular dystrophy, although they have a milder phenotype, with muscle weakness being prominent in the diaphragm but minimal in the limbs. Studies of spontaneous wheel running in mdx mice have found not only reductions in total running distance (4,9) but also alterations in the temporal pattern of running. Specifically, mdx dystrophic mice demonstrate much more intermittent running than normal (a “stop and go” pattern), which is believed to be due to an abnormality in muscle endurance (9). Whether diabetes alters spontaneous running in rodents is not as clear. One study found unchanged spontaneous running in a Type II diabetic model (3), another found unchanged running distance but modestly reduced running speed in a Type I diabetic model (28), and two studies found reduced running distance in two different Type I diabetic models (13,19). Of note is that one of the latter studies in a Type I diabetic model examined running for only 5 h·d−1 (presumably during the daytime, whereas most rodents are nocturnal) (13), providing an incomplete view of the 24-h running pattern. Furthermore, none of the studies reported data in sufficient detail to determine whether diabetic animals have the same “stop and go” pattern of running as do mdx mice.
A number of studies have found that diabetes impairs respiratory and limb muscle contractile performance (5,8,10,14–17,23,25,26). We therefore hypothesized that Type I diabetes not only reduces total spontaneous exercise but also fragments the temporal pattern of running, similar to that described in mdx dystrophic mice (9). Studies used an acquired Type I diabetes model (streptozotocin-induced diabetes) rather than a genetic diabetic model to make certain that diabetic and normal animals would not differ genetically in ways unrelated to diabetes that could affect spontaneous running behavior. In addition, this model ensured that both control and diabetic animals had normal and comparable antecedent growth and development before the onset of the diabetes, eliminating potential interactive effects of diabetes and prior maturational processes. In addition, with a Type II diabetes model, there is both obesity and diabetes, and it would be more difficult to determine how much of the altered running behavior is due to diabetes per se as opposed to solely the higher weight of the obese-diabetic animals.
All experiments were performed in accordance with the animal care and welfare guidelines of the American College of Sports Medicine and the National Institutes of Health, were approved by the Institutional Animal Care and Use Committee, and conducted in an institution with an AAALAC-accredited animal facility. Male Wistar rats were obtained from Charles River Laboratories (Wilmington, MA) and housed in a 12:12 light:dark cycle. The animals were divided randomly into two groups. Control animals were injected intraperitoneally with 20-mM sodium citrate buffer (pH 4.5, N = 8), and diabetes was induced in the other animals using intraperitoneal streptozotocin in a dose of 60 mg·kg−1 dissolved in citrate buffer (N = 7) (10,21). The streptozotocin-diabetic rat is a model of Type I diabetes. Rats were housed singly in standard size polycarbonate living chambers, which were modified to allow access to an immediately adjacent running wheel (22). The running wheel was equipped with a magnetic counter, the output of which was sent to a PC-type computer, allowing quantification of the number of revolutions (Lafayette Instruments, Lafayette, IN). All animals were given ample access to food and water 24 h·d−1. Frequent bedding changes were required due to the excess drinking and urination of the diabetic animals, and these were always done during the daytime, during which time running is minimal to absent in nocturnal animals, including the rats of the present study. The animals entered the study at an age of 8 wk, having previously been housed in conventional caging, and spent the subsequent 8 wk in running-wheel–equipped housing. At the end of 8 wk in the running wheel cages, animals were anesthetized with intraperitoneal urethane, and a blood sample was obtained for analysis of glucose and electrolytes. The animals were not fasted before blood sampling. Rat weight was monitored after induction of diabetes to detect any rapid weight loss that might indicate that a diabetic rat was overtly ill and would adversely affect the results of the study, which did not occur in any animal. None of the animals were part of previous studies from our lab examining diabetes (21) or wheel running exercise training (22).
The number of rotations of each running wheel was recorded over 10-min intervals 24 h·d−1 for the entire 8-wk period and stored on a PC-based computer. Data was initially analyzed with respect to the total distance run per day. Active running times were then identified modified from the approach of Hara et al. (9), defined as any 10-min interval during which the animals ran enough to produce at least one rotation of the wheel, although higher thresholds for defining active running times were also examined, in particular five rotations (see Results). Average consecutive running times were calculated from the number of 10-min periods in a row during which the animals ran. Thus, the consecutive running time indicates the total duration of time during which the animals ran enough to produce at least the predefined minimum number of wheel rotations during each 10-min interval, and does not imply that the animals ran without stopping the entire time. In a previous study, we found that normal Sprague-Dawley rats gradually increased their running activity over the course of 4–5 wk and then maintained a plateau for the remainder of the 3–4 wk examined (22). We therefore focused the data analysis on the eighth week of running as indicative of a steady state, but also examined earlier time points to determine the time course over which differential running patterns developed. Statistical analysis comparing two data sets was performed with the unpaired t-test. Statistical analysis of multiple data sets was performed with analysis of variance, followed by the Bonferroni method in the event of a significant analysis of variance result. A P value of < 0.05 (two-tailed) was chosen to indicate significance.
Diabetic rats had substantially higher nonfasting serum glucose levels than did control animals (Table 1), measured 8 wk after injection of streptozotocin or buffer. They also had polydipsia and polyuria during most of the 8 wk, based on gross observation of the amount of water consumed and the need to change the bedding, respectively (although neither was quantified). Values for serum sodium, potassium, and bicarbonate did not differ significantly for diabetic and control animals, whereas chloride levels were diminished modestly (Table 1). Diabetic and control rats had similar weights at the onset of the 8-wk period. However, the diabetic rats subsequently gained weight at a reduced rate (Fig. 1), and hence weighed less than the control rats at the end of 8 wk. Other than being slightly thinner, the physical appearance and the gross behavior of the diabetic rats did not differ from that of the controls. Furthermore, none of the animals were identified by the animal facility technicians or veterinarian staff (or the investigators) as needing medical attention or other interventions.
Spontaneous running performance 8 wk after streptozotocin injection.
Diabetic rats ran considerably less per day than the control animals (P = 0.006), as quantified during the last 2 d of the eighth week in Figure 2. The normal animals had a wide variation in daily running distances, ranging from 476 to 10,260 m, with a coefficient of variation [SD/mean × 100%] of 87%. The running distances of the diabetic animals were spread over a smaller range in absolute terms (32–492 m), but due to the much lower average among the group, the coefficient of variation was almost as great as that of the normal animals (84%). Daily running distances did not correlate with blood glucose values for either diabetic or normal animals.
FIGURE 2Average dail...Image Tools
Examples of running activity during the course of 48 h are depicted for two normal and two diabetic rats in Figure 3. All diabetic and control rats ran predominantly at night. Control animals often ran for multiple consecutive 10-min periods, in some cases up to 10–12 in a row, but also ran for short periods of time (single 10-min periods). In contrast, diabetic rats ran for only a few consecutive 10-min periods, and rarely more than three or four in a row. The average consecutive running times for all animals, depicted in Figure 4, were significantly shorter for diabetic compared with normal rats (P = 0.004). Changing the threshold for defining the consecutive running time from at least one revolution to a higher value did not change the finding of shorter consecutive running times in diabetic compared with control animals, as depicted in Figure 4 for a threshold of 5 m per 10-min interval. The average consecutive running times did not correlate with blood glucose values for either diabetic or normal animals.
Time course over which changes in spontaneous running occurred.
To determine the time course over which diabetic and normal animals diverged in their running performance, data were quantified at three earlier time points after the injection of streptozotocin: day 3, day 6, and the end of week 4. The daily running distance was similar for the diabetic and normal rats during day 3 (Fig. 5). By day 6, the two groups were starting to diverge, although there were still two (of seven) diabetic rats that ran more than 1000 m·d−1 and three (of eight) normal rats that ran less than 1000 m·d−1. By the end of 4 wk, the normal rats as a group had increased their daily running distance approximately threefold compared with values during the first week, and this high distance persisted at the end of week 8. In contrast, the diabetic rats had a gradual decline in running, so that values at 4 and 8 wk were significantly lower than those of normal animals.
FIGURE 5Changes in a...Image Tools
The average consecutive running times were similar (albeit slightly lower) in the diabetic compared with the normal rats during the third day after streptozotocin injection (Fig. 6). By day 6, average consecutive running times were diverging for the two groups of animals, and they were significantly different during weeks 4 and 8. The divergence between diabetic and normal rats appeared to be due predominantly to a halving of the consecutive running in the diabetic rats, as the continuous running times of the normal animals was not markedly different at the end of 8 wk compared with day 3.
FIGURE 6Changes in a...Image Tools
Several studies have examined total spontaneous running distance of diabetic rodent models. On the one hand, Dubuc et al. (3) found no difference in total running distance between ob/ob (obese) mice with mean glucose values of 320 mg·dL−1 mice and lean mice with mean blood glucose values of 183 mg·dL−1. Woodiwiss et al. (28) found that diabetic rats with blood glucose values of 20.4 mM (368 mg·dL−1) ran the same distance, albeit at a modestly slower average speed, than normal rats with glucose levels of 5.6 mM (101 mg·dL−1). On the other hand, Keller et al. (13) found that nonobese diabetic mice with blood glucose values of ∼25 mM (∼283 mg·dL−1) ran less than half that of normal mice with glucose values of ∼6 mM (∼108 mg·dL−1). In addition, Rowland and Caputo (19) found that streptozotocin-diabetic hamsters with blood glucose values of 398 mg·dL−1 exercised half to two-thirds that of normal animals with glucose values of 116 mg·dL−1 when maintained in a light-dark cycle (14:10), and that the extent of exercise in diabetic animals was reduced even further to ∼25% of control when animals were housed in continuous light. The results of the previous studies along with the present study suggest that mild to moderate degrees of diabetes may or may not diminish exercise activity, whereas animals with more severe hyperglycemia as in the present study exercise less than normal.
Less information is available about the temporal pattern of exercise in diabetes. Dubuc et al. (3) and Woodiwiss et al. (28) did not provide information along these lines. Keller et al. (13) allowed mice access to exercise wheels for only 5 h·d−1, 5 d·wk−1, and noted that both diabetic and normal animals exercised continuously with few interruptions. Rowland and Caputo (19) commented that the diabetic hamsters generally ran less during each hour than the normal animals, but further temporal details were not provided. The present study thus extends previous observations, by demonstrating a marked reduction in the consecutive running time in Type I diabetic compared with normal animals, similar to that described in mdx dystrophic mice (9).
One possible factor that could have contributed to the reduced spontaneous exercise behavior of the diabetic rats in the present study is that they did not eat enough food to maintain a normal rate of weight increase, so that lack of exercise might have been due to relative malnutrition. Arguing against this are two rodent studies indicating that reductions in caloric intake are associated with increases rather than reductions in spontaneous wheel running. Jones et al. (11) found that normal rats that were food restricted (caloric intake one-third to one-half of ad libitum-fed animals) ran 2–4 times as much as normally fed animals (with the exact magnitude of the increase depending on gender). Dubuc et al. (3) found that ob/ob mice that were food restricted (by pair-feeding them to the caloric intake of lean mice) ran almost twice as much as normal.
Several other factors also seem unlikely to have contributed to the altered exercise behavior of diabetic rats. Blood bicarbonate levels were almost identical in diabetic and normal animals, arguing against diabetic ketoacidosis or other types of metabolic acidosis (e.g., due to renal failure) having played a role. Alterations in extracellular sodium and potassium can affect both skeletal and cardiac muscle contractile performance, but blood levels of these ions were similar in the two sets of animals. Blood chloride levels were reduced modestly in the diabetic animals, although probably not sufficiently to impair muscle contractility.
Another consideration for the cause of the reduced spontaneous exercise in the diabetic animals is the general malaise and feelings of tiredness that are often associated with chronic disease states. Diabetic humans report a lower quality of life than normal, albeit higher than that of other major chronic diseases (2,6,20). The importance of this factor in determining, or even modulating the present findings, can not be determined definitively from the outcome measures used in the present study. Of note, however, is that the gross behavior of the diabetic animals in terms of moving around the cage, eating and drinking appeared normal, which according to standard animal care guidelines are generally accepted as measures indicating the absence of a substantial amount of distress or discomfort. However, lesser degrees of malaise may have been present, which could have been a factor.
Several physiological factors are more likely to have played a role in the diminished spontaneous running and the reduced duration of running per episode in the diabetic rodents. First, based on the degree of hyperglycemia and polyuria, it is possible that the animals were volume depleted, despite having been provided with ample drinking water, which may adversely affect exercise performance. Serum Na+ values roughly the same for diabetic than normal rats (142 vs 144 mmol·L−1), arguing against a major degree of volume depletion, however.
Second, diabetes impairs skeletal muscle contractile performance, even when tested in vitro (5,8,10,14–17,23,25,26). The reported reductions in muscle force are generally relatively mild, but these measurements were made during very brief contractions. Exercise performance is affected not only by absolute force generating capacity but also the extent to which muscle force declines during repetitive contractions. It has been noted previously that mice with mdx muscular dystrophy (who have diaphragm muscle weakness but relatively well preserved limb muscle function) have both reduced running distance and shortened consecutive running times, which was attributed to a disturbance in muscle endurance (9). In the mdx mouse model, abnormal running behavior most likely is due purely to intrinsic muscle factors, as presumably metabolic derangements are absent. Thus, it is likely that in the present study intrinsic muscle contractile abnormalities also limited running endurance, which could have been due to limb muscle dysfunction affecting spontaneous locomotory performance directly, respiratory muscle dysfunction limiting ventilatory responses to exercise and hence reducing locomotory performance indirectly, or both. A number of mechanisms have been implicated in the pathophysiology of diabetes-induced skeletal muscle dysfunction, including mitochondrial dysfunction and alterations in membrane electrophysiological properties. It is possible that exercise may exert its beneficial effects on diabetic patients by reversing these abnormalities. However, the present study was not designed to determine which of these muscle alterations is the most important in limiting exercise behavior.
In the present study, the diabetic animals had markedly elevated glucose levels, consistent with a Type I diabetic model. The control animals had a mild degree of hyperglycemia, the etiology of which was unclear, although of note is that the animals were not fasted before blood analysis. The glucose values in the present study were comparable to a previous study from our lab in nonexercised streptozotocin-diabetic rats (21), in which the controls and diabetic animals had mean values of 252 and 670 mg·dL−1. The lack of correlation between degree of hyperglycemia and amount of exercise in the animals of the present study could have been due to several factors. First, the glucose values reflect only a single point in time, whereas exercise was quantified over the course of several 1- to 2-d periods spread out over the course of 8 wk. Second, there might be several ongoing relationships between glucose levels and exercise, which could cancel each other out. For example, on the one hand high glucose levels could reduce exercise behavior (i.e, if all animals had an equal propensity toward exercise, those with more extreme hyperglycemia might exercise less), whereas on the other hand high exercise levels could reduce glucose levels (i.e., if all animals had equally high blood glucose levels in the absence of an opportunity to exercise, those with an inherent higher tendency toward exercise might have better glycemic control). Third, it may not be the degree of hyperglycemia per se but rather the secondary consequences of diabetes (e.g., alterations in muscle properties) that determine exercise performance.
Types I and II diabetes differ importantly from each other with respect to pathogenesis, as well as associated metabolic disturbances and end-organ responses. The present study used streptozotocin to induce diabetes, which is a model of Type I diabetes. Although the majority of humans with diabetes have Type II rather than Type I diabetes, it is estimated that of the ∼18 million people in the United States affected by diabetes, up to 10% (or 1.8 million) have Type I diabetes. Furthermore, one in every 400–500 children and adolescents in the United States has Type I diabetes. Thus, there is a sizable proportion of the population for which the present data are relevant and important. Future studies are needed in an animal model of Type II diabetes to determine whether they have a similar temporal deficit in running behavior seen in the Type I diabetic rat and mdx mouse models.
1. Boule, N. G., E. Haddad, G. P. Kenny, G. A. Wells, and R. J. Sigal. Effects of exercise on glycemic control and body mass in type 2 diabetes mellitus: a meta-analysis of controlled clinical trials. JAMA
2. Brown, G. C., M. M. Brown, S. Sharma, H. Brown, M. Gozum, and P. Denton. Quality of life associated with diabetes mellitus in an adult population. J. Diabetes Complications
3. Dubuc, P. U., P. J. Cahn, and P. Willis. The effects of exercise and food restriction on obestiy and diabetes in young ob/ob mice. Int J. Obes.
4. Dupont-Versteegden, E. E., R. J. McCarter, and M. S. Katz. Voluntary exercise decreases progression of muscular dystrophy in diaphragm of mdx mice. J. Appl. Physiol.
5. Fahim, M. A., F. el-Sabban, and N. Davidson. Muscle contractility decrement and correlated morphology during the pathogenesis of streptozotocin-diabetic mice. Anat. Rec.
6. Glasgow, R. E., L. Ruggiero, E. G. Eakin, J. Dryfoos, and L. Chobanian. Quality of life and associated characteristics in a large national sample of adults with diabetes. Diabetes Care
7. Gregg, E. W., R. B. Gerzoff, C. J. Caspersen, D. F. Williamson, and K. M. V. Narayan. Relationship of walking to mortality among US adults with diabetes. Arch. Intern. Med.
8. Grossie, J. Contractile and electrical characteristics of extensor muscle from alloxan-diabetic rats: an in vitro study. Diabetes
9. Hara, J., P. M. Nolan, M. O. Scott, M. Bucan, Y. Wakayama, and K. H. Fischbeck. Running endurance abnormality in mdx mice. Muscle Nerve
10. Hida, W., C. Shindoh, J. Satoh, et al. N-Acetylcysteine inhibits loss of diaphragm function in streptozotocin-treated rats. Am. J. Crit. Care Med.
11. Jones, L. C., W. P. Bellingham, and L. C. Ward. Sex differences in voluntary locomotor activity of food-restricted and ad libitum-fed rats: implications for the maintenance of a body weight set-point. Comp. Biochem. Physiol.
12. Katoh, J., Y. Hara, M. Kurusu, J. Miyaji, and K. Narutaki. Cardiorespiratory function as assessed by exercise testing in patients with non-insulin-dependent diabetes mellitus. J. Int. Med. Res.
13. Keller, J. B., W. C. Bevier, L. Jovanovic-Peterson, B. Formby, E. P. Durak, and C. M. Peterson. Voluntary exercise improves glycemia in non-obese diabetic (NOD) mice. Diabetes Res. Clin. Pract.
14. McGuire, M., M. Dumbleton, M. MacDermott, and A. Bradford. Contractile and electrical properties of sternohyoid muscle in streptozotocin diabetic rats. Clin. Exp. Pharmacol. Physiol.
15. McGuire, M., and M. MacDermott. The influence of streptozotocin-induced diabetes and the antihyperglycaemic agent metformin on the contractile characteristics and the membrane potential of the rat diaphragm. Exp. Physiol.
16. McGuire, M., and M. MacDermott. The influence of streptozotocin diabetes and metformin on erythrocyte volume and on the membrane potential and the contractile characteristics of the extensor digitorum longus and soleus muscles in rats. Exp. Physiol.
17. Paulus, S. F., and J. Grossie. Skeletal muscle in alloxan diabetes: a comparison of isometric contractions in fast and slow muscle. Diabetes
18. Regensteiner, J. G., J. Sippel, E. T. McFarling, E. E. Wolfel, and W. R. Hiatt. Effects of non-insulin-dependent diabetes on oxygen consumption during treadmill exercise. Med. Sci. Sports Exerc.
19. Rowland, N., and F. Caputo. Voluntary exercise, food intake, and plasma metabolites in streptozotocin-diabetic Syrian hamsters. Physiol. Behav.
20. Rubin, R. R., and M. Peyrot. Quality of life and diabetes. Diabetes Metab. Res. Rev.
21. van Lunteren, E., and M. Moyer. Streptozotocin-diabetes alters action potentials in rat diaphragm. Respir. Physiol. Neurobiol.
22. van Lunteren, E., and M. Moyer. Wheel running exercise alters rat diaphragm action potentials and their regulation by K+
channels. J. Appl. Physiol.
23. Wanke, T., D. Formanek, M. Auinger, W. Popp, H. Zwick, and K. Irsigler. Inspiratory muscle performance and pulmonary function changes in insulin-dependent diabetes mellitus. Am. Rev. Respir. Dis.
24. Wanke, T., D. Formanek, M. Auinger, H. Zwick, and K. Irsigler. Pulmonary gas exchange and oxygen uptake during exercise in patients with type 1 diabetes mellitus. Diabetes Med.
25. Wanke, T., H. Lahrmann, M. Auinger, et al. Endogenous opioid system during inspiratory loading in patients with type 1 diabetes. Am. Rev. Respir. Dis.
26. Wanke, T., T. Paternostro-Sluga, W. Grisold, et al. Phrenic nerve function in type 1 diabetic patients with diaphragm weakness and peripheral neuropathy. Respiration
27. Wei, M., L. W. Gibbons, T. L. Mitchell, J. B. Kampert, C. D. Lee, and S. N. Blair. The association between cardiorespiratory fitness and impaired fasting glucose and type 2 diabetes mellitus in men. Ann. Int. Med.
28. Woodiwiss, A. J., W. J. Kalk, and G. R. Norton. Habitual exercise attenuates myocardial stiffness in diabetes mellitus in rats. Am. J. Physiol.
RUNNING; DIABETES; SKELETAL MUSCLE; ENDURANCE; HYPERGLYCEMIA
©2004The American College of Sports Medicine
What does "Remember me" mean?
By checking this box, you'll stay logged in until you logout. You'll get easier access to your articles, collections,
media, and all your other content, even if you close your browser or shut down your
To protect your most sensitive data and activities (like changing your password),
we'll ask you to re-enter your password when you access these services.
What if I'm on a computer that I share with others?
If you're using a public computer or you share this computer with others, we recommend
that you uncheck the "Remember me" box.
Highlight selected keywords in the article text.
Data is temporarily unavailable. Please try again soon.