Diabetes Research Institute, University of Miami, 1450 NW 10 Avenue (R-134), Miami, FL 33136. E-mail: Dberman2@med.miami.edu
Objective: To test the hypothesis that the greater β-adrenoceptor (β-AR)-stimulated lipolysis and sensitivity (half-maximal lipolytic response) in abdominal (ABD) adipocytes, greater gluteal (GLT) adipose tissue-lipoprotein lipase (AT-LPL) activity, and dyslipidemia associated with obesity in older women are modifiable by weight loss (WL) and are not due to menopause or aging.
Research Methods and Procedures: The metabolic effects of 6 months of hypocaloric diet and low-intensity walking WL program on the regional regulation of in vitro lipolysis and AT-LPL activity in subcutaneous ABD and GLT adipocytes were measured in 34 obese (48.7 ± 0.7% body fat, mean ± SE) postmenopausal (59 ± 1 years) white women.
Results: The lipolytic responsiveness to the β-AR agonist isoproterenol and basal lipolysis in the presence of 1 U/mL adenosine deaminase-uninhibited (lipolysis) were greater (p < 0.01) in ABD than GLT adipocytes before and after WL, but there were no regional differences in postreceptor (dibutyryl 3′, 5′-cyclic adenosine monophosphate)-stimulated lipolysis. β-AR sensitivity was greater in ABD than GLT adipocytes before (p < 0.01) but not after WL. Regional AT-LPL did not change after WL, but the change in the activity of ABD (but not GLT) AT-LPL correlated with the baseline adenosine deaminase-uninhibited lipolysis (r = 0.38, p = 0.03). There were no relationships between the declines in plasma triglyceride or increases in high-density lipoprotein cholesterol associated with WL and the changes in regional fat cell metabolism.
Discussion: Thus, despite improving lipoprotein lipid profiles in obese, postmenopausal women, WL does not affect the regulation of regional fat metabolism, and a greater tonic inhibition of basal lipolysis by endogenous adenosine may increase the activity of AT-LPL after WL and predispose older women to develop ABD adiposity.
The storage of triglycerides (TGs)1 in fat cells results mainly from a balance between the incorporation of free fatty acids (FFAs) released by the hydrolysis of circulating TG by the enzyme adipose tissue-lipoprotein lipase (AT-LPL) (1) and lipolysis of stored TG to FFAs and glycerol by the rate-limiting enzyme hormone-sensitive lipase (2). These processes regulate fat cell size and are reciprocally regulated, suggesting an inverse relationship between the activity of AT-LPL and hormone-sensitive lipase (3).
The only hormones that acutely affect lipolysis in human adipocytes are catecholamines (epinephrine and norepinephrine) and insulin (4). The regulation of lipolysis by catecholamines involves adrenoceptor (AR) stimulation of adenylate cyclase with β-AR (β1-, β2-, and β3-AR) and inhibition by α2-AR (5, 6). In addition, paracrine mediators such as adenosine have powerful antilipolytic actions (7). Endogenously released adenosine binds to A1 receptors in the plasma membrane of adipocytes to inhibit the activity of adenylate cyclase, providing a tonic inhibition of lipolysis (8, 9). Deamination of adenosine to inosine by the enzyme adenosine deaminase (ADA) usually increases adipocyte lipolysis (8) and provides an indirect assessment of the tonic inhibition of lipolysis by adenosine.
In postmenopausal women, there is a tendency for weight gain to occur in the abdominal (ABD) region (10, 11), and the magnitude of the increase in ABD adiposity correlates better with obesity-related risk factors (hypertension, hyperlipidemia, and type 2 diabetes) for cardiovascular disease than the absolute degree of obesity (12). ABD adipocytes are more sensitive and responsive to β-AR stimulation of lipolysis than gluteal (GLT) adipocytes, whereas GLT adipocytes have higher AT-LPL activity than ABD adipocytes in older women, suggesting there are regional differences in fat cell metabolism with aging (13, 14, 15, 16). In addition, we recently showed that postmenopausal women have lower regional basal lipolysis and higher regional activities of AT-LPL compared with obesity-matched perimenopausal women (15), suggesting that the lower basal lipolysis and the higher regional AT-LPL may predispose older women to gain body fat after menopause. However, it is unclear how regional differences in lipolysis and AT-LPL contribute to the development of ABD obesity in older women.
Most of the metabolic abnormalities associated with ABD obesity in postmenopausal women improve after weight loss (WL). These include reductions in total and low-density lipoprotein cholesterol (LDL-C), TGs, and blood pressure, a rise in high-density lipoprotein cholesterol (HDL-C), and an improvement in glucose tolerance and insulin sensitivity (14, 17, 18, 19). Therefore, it was our hypothesis that regional differences in lipolysis and AT-LPL activity are secondary effects of obesity in older women and, thus, modifiable by WL.
To test this hypothesis, we examined the effect of a 6-month hypocaloric and low-intensity walking program on the regional regulation of in vitro lipolysis and AT-LPL activity in subcutaneous ABD and GLT adipocytes in obese postmenopausal women.
Research Methods and Procedures
As previously described (14), overweight and obese (BMI > 25 kg/m2) white women of similar age were recruited. The women provided informed written consent to participate in this study according to guidelines of the University of Maryland Institutional Review Board for Human Research. The women were postmenopausal, i.e., had not menstruated for at least 1 year, and had plasma follicle-stimulating hormone levels > 30 mIU/mL. None of the women were on estrogen replacement therapy or medications affecting lipid or glucose metabolism for at least 1 year. All women were healthy, with no evidence of cardiovascular disease; diabetes (fasting blood glucose level < 126 mg/dL; 2-hour glucose < 200 mg/dL after an oral glucose tolerance test); hyperlipidemia, liver, renal, or hematologic disease on medical exam and blood chemistries. A graded exercise test according to a modified Bruce protocol was performed to exclude subjects with an abnormal cardiovascular response to exercise (20). Maximal oxygen consumption (Vo2max) was measured during a second test, as previously described (14). All women were sedentary (<20 min of exercise two times per week), weight stable (<2 kg of weight change in past year), and had not smoked cigarettes for at least 5 years. We report findings on 34 of the 55 white women who completed the WL intervention and had complete measures of regional AT-LPL activity and adipocyte basal and stimulated lipolysis.
Before beginning metabolic testing, women completed a 7-day food record to provide information about their dietary habits. To establish dietary control before metabolic testing and to eliminate changes in dietary composition as a confounder of metabolic changes with WL, all women met weekly with a registered dietitian for 6 to 8 weeks for instructions and weight stabilization on diets according to the principles of the American Heart Association (AHA) Step 1 diet (21). Subjects were weight stable and compliant to this diet for ≥2 weeks before and throughout the period of data collection. The dietitian monitored compliance by weekly review of 7-day food records and 24-hour dietary recalls. After 2 weeks of weight stabilization on this diet, nutrient intake was controlled by providing subjects with 2 days of a eucaloric diet composed of 50% to 55% carbohydrate, 15% to 20% protein, and 30% fat ≤ 300 mg cholesterol and a polyunsaturated to saturated fat ratio of 0.6 to 0.8 before each fat biopsy (pre- and post-WL).
During the 6-month WL intervention, all subjects met weekly with a registered dietitian for instruction on the principles of a hypocaloric diet (250 to 350 kcal/d deficit) that followed the AHA guidelines. The WL program focused on eating behavior, stress management, control of portion sizes, modification of binge eating, and other adverse eating habits and also encouraged low-intensity walking 3 d/wk for 30 to 45 min to increase their physical activity. The women walked 1 d/wk on a treadmill at our exercise facility, at a leisurely pace of 50% to 60% heart rate reserve, under the supervision of an exercise physiologist, and were instructed to walk 2 d/wk at this intensity on their own. The women were encouraged to maintain this amount of physical activity throughout the 6-month intervention. After the intervention, women were weight stabilized (<0.5 kg change) on a eucaloric diet that followed the AHA guidelines for a period of 2 weeks before retesting. They had to be weight stable on this diet for ≥2 weeks before they were allowed to be tested, or the period was extended. They maintained their 3 d/wk of walking during this period, and the fat biopsy was performed ≥36 h after their last reported walking session.
Body Composition and Fat Distribution
Waist-to-hip ratio was measured as the ratio of the minimal waist circumference to the hip circumference at the maximal GLT protuberance, and the mean of three values within 2 mm of each other was used as the measurement. Percent body fat, fat mass, and lean mass were measured using DXA (model DPX-L; Lunar Radiation Corporation, Madison, WI). Intra-ABD and subcutaneous fat areas were measured using a single-slice computed tomography scan taken midway between L4 and L5 performed on a GE Hi-Light computed tomography scanner (Milwaukee, WI), as described previously (20).
Glucose and Lipid Metabolic Risk Factors
Venous blood samples for measurement of lipoprotein lipid, glucose, and insulin levels were drawn after a 12-hour fast on 2 separate days, and the mean of the two measurements of lipoprotein lipids is reported. Plasma FFA concentrations were measured in 27 women using a colorimetric assay (Wako Chemicals USA, Inc., Richmond, VA). Total cholesterol, HDL-C, LDL-C, and TG levels were measured as previously described (22). Plasma glucose concentration was measured in duplicate using the glucose oxidase method (Beckman Glucose Analyzer; Beckman, Fullerton, CA). Plasma insulin (Linco, St. Louis, MO) was measured in duplicate by radioimmunoassays with intra-assay and interassay coefficients of variation of 5.0% and 9.0%, respectively.
Isolation of Adipocytes
After an overnight fast, 2 to 4 grams of subcutaneous AT were obtained under local anesthesia (1% xylocaine) from both the ABD and GLT regions by aspiration with a 16-gauge needle (23). Adipocytes were isolated by collagenase digestion of AT fragments in Krebs-Ringer Hepes buffer containing 4% FFA-free bovine serum albumin, 5 mM glucose, 0.1 mM ascorbic acid, and 200 nM adenosine using a modification of the Rodbell method (24). The isolated cells were filtered through a 250-μm nylon mesh and washed three times with the same Krebs-Ringer Hepes buffer to remove collagenase from the medium. The adipocytes were resuspended in the same buffer so that the final cell concentration was ∼4% to 5%. The mean fat cell lipid weight was calculated using the average mean adipocyte diameter and SD (25), and total lipid content in the suspension was determined gravimetrically after extraction (26). Cell number was obtained by dividing total lipid weight of the suspension by the mean fat cell lipid weight.
Heparin-releasable AT-LPL activity was measured as previously described (27). AT-LPL was eluted from 30- to 50-mg fragments of AT into 2.5 mL of Krebs-Ringer-Phosphate buffer containing 5 U heparin during a 45-min incubation at 37 °C. Triplicate 0.5-mL aliquots of the eluate were incubated with 0.1 mL of substrate prepared by sonication of 4 μCi of 1-14C-glycerol triolein, 5 mg of unlabeled triolein, and 240 μg of lecithin in 4 mL of 0.5 M Tris buffer (pH 8.2) containing 2% FFA-free bovine serum albumin and 0.25 mL of fasting serum. The enzyme reaction was stopped after 45 min at 37 °C by addition of Belfrage's extraction mixture (28) to separate the product, labeled FFA, from unreacted substrate. The labeled FFA were quantitated by liquid scintillation counting and, after correction for recovery during the extraction, AT-LPL activity was expressed as nanomoles of FFA produced per minute by 106 cells and per cell surface area in picomoles FFA per micrometer squared per minute by 108 cells. The 34 subjects reported in this manuscript are a subset of 36 patients previously reported by us (14) who had complete measures of lipolysis and AT-LPL activity performed at baseline and after WL.
Glycerol released from adipocytes was used as the index of lipolysis (29). Triplicate 0.75-mL aliquots of the diluted cell suspension were incubated in plastic vials with gentle shaking in a water bath at 37 °C. Pharmacological agents were added just before the beginning of the incubation in 10-μL aliquots in vehicle to obtain the indicated final concentration. After 2 hours, the lipolysis reaction was stopped with 76 μL of 2.5 M perchloric acid, and 80 μL of the infranatant was removed for the measurement of glycerol concentration by an enzymatic fluorometric technique (30). The lipolytic response of isolated adipocytes was measured after incubation with 10−10 to 10−5 M isoproterenol (nonselective β-AR agonist) and with a maximal concentration (2 mM) of dibutyryl 3′, 5′-cyclic adenosine monophosphate (db-cAMP) (phosphodiesterase-resistant cAMP analogue). Variations in adenosine metabolism have a major influence on the measurement of lipolysis (8, 31); therefore, in experiments measuring the stimulatory effect of isoproterenol, ADA was added to remove adenosine present in the medium. However, to control for the activation of lipolysis after adenosine removal, which might prevent or attenuate the response to isoproterenol, isoproterenol stimulation of lipolysis was evaluated in the presence of 1 U/mL ADA and 100 nM N6-(1–2-phenylisopropyl)-adenosine, a potent adenosine receptor agonist that is neither a substrate nor an inhibitor of ADA, as shown previously (9). Lipolysis was expressed per cell number in micromoles glycerol per 106 cells per 2 hours and per cell surface area in nanomoles glycerol per micrometer squared per 108 per 2 hours. The maximal lipolytic effect, or responsiveness, was calculated as the difference between basal glycerol released and the glycerol released at the maximally effective concentration of the lipolytic agent. The concentration of isoproterenol giving half-maximal lipolytic response (EC50), an index of sensitivity, was obtained by computer fitting of individual dose-response curves to isoproterenol using SigmaPlot software (Jandel Scientific, Chicago, IL).
Data are means ± SE. Because ADA-uninhibited lipolysis is not normally distributed (Shapiro-Wilk test), it was logarithmically transformed to yield normal distributions before parametric analyses. Student's paired t test was used to compare ABD vs. GLT values in the same individual and differences between variables before and after WL. Pearson's product-moment correlation coefficients were used to determine statistically significant relationships between variables. All analyses were performed using Jump Software (SAS Institute, Inc., Cary, NC). Differences were considered significant when p < 0.05.
Metabolic Effects of WL
The 34 sedentary, obese women had a mean age of 59 ± 1 years (range: 51 to 68 years). The 6-month hypocaloric diet and walking program resulted in a WL of 6.3 ± 0.6% of body weight (Table 1), with decreases in fat mass (14.0 ± 3.0%) but no significant change in lean body mass (−2.1 ± 3.5%). Visceral and subcutaneous ABD fat decreased by 17.0 ± 3.7% and 15.1 ± 3.6%, respectively. There was no change in Vo2max after the intervention. Fasting glucose values decreased significantly after WL (p = 0.01) (Table 2), and there was an average 6% nonsignificant reduction in fasting insulin values. WL was associated with a 9% decrease in TG levels (p = 0.001), 4% increase (p < 0.05) in HDL cholesterol levels, but no changes in total or LDL-C. Fasting FFA levels decreased significantly after WL.
Table 1. . Physical characteristics of the subjects before and after weight loss
GLT adipocytes were larger than ABD (GLT 0.89 ± 0.02 vs. ABD 0.84 ± 0.02 μg TG/cell, p < 0.01) at baseline, and WL resulted in comparable decreases (p < 0.0001) in cell size (ABD: 10.8 ± 1.8% vs. GLT: 11.4 ± 2.3%). To control for differences in fat cell size after WL, measurements of lipolysis and AT-LPL activity are expressed both per cell and per cell surface area.
WL had no effect on basal lipolysis, ADA-uninhibited lipolysis, β-adrenergic agonist-stimulated lipolysis, postreceptor-stimulated lipolysis, or mean levels of AT-LPL activity in ABD or GLT adipocytes (Table 3). Similar results were found when data were expressed per cell surface area (not shown).
Table 3. . Effect of WL on Regional AT-LPL activity and lipolysis
Values are mean ± SE of the difference between values after and before WL (Δ). Adipose tissue LPL activity is expressed in nanomoles of free fatty acid per 106 cells per minute; lipolysis is expressed in micromoles glycerol per 106 cells per 2 hours. Maximal response (Δ) and sensitivity (EC50) to isoproterenol were calculated from dose-response curves (10−10 to 10−5 M) in the presence of 1 U/ml ADA and 100 nM N6-(1-2-phenylisopropyl)-adenosine.
−0.36 ± 0.52
0.05 ± 0.74
0.20 ± 0.18
−0.11 ± 0.13
Basal + ADA (1 U/ml)
0.44 ± 0.39
0.03 ± 0.23
0.40 ± 0.42
0.06 ± 0.31
−43.7 ± 35.6
−62.6 ± 36.8
Δ db-cAMP (2mM)
0.06 ± 0.43
−0.15 ± 0.42
Although there were no regional differences in basal lipolysis either before or after WL, when adenosine was removed from the medium by adding 1 U/mL ADA (ADA-uninhibited lipolysis), the basal lipolytic rate increased 1.5 fold more in the ABD than in the GLT cells both before and after WL (p < 0.05, Figure 1); this suggests greater adenosine-induced inhibition of lipolysis in ABD adipocytes. The maximal lipolytic response to isoproterenol also was higher in ABD vs. GLT cells both before and after WL (p < 0.001) (Figure 2A); however, the sensitivity (EC50) to isoproterenol was 38% higher in ABD cells before (p < 0.01) but not after WL (Figure 2B). There were no regional differences in postreceptor (db-cAMP)-stimulated lipolysis before or after WL (Figure 2A). Similar results were found when data were expressed per cell surface area (not shown).
Relationships among Changes in Regional Adipocyte Metabolism with WL
There was a positive relationship between the change in the activity of AT-LPL with WL and the pre-WL ADA-uninhibited lipolysis in ABD adipocytes, regardless of the mode of expression of AT-LPL activity (per cell, r = 0.38, p = 0.03; per cell surface area, r = 0.45, p < 0.01) but not in GLT adipocytes (per cell, r = 0.19, not significant; per cell surface area, r = 0.07, not significant). Thus, women with higher tonic inhibition of lipolysis by adenosine before WL in the ABD cells tended to increase their AT-LPL activity more after WL. There was no relationship between the change in AT-LPL activity with WL- and β-AR maximal-stimulated lipolysis or sensitivity, postreceptor-stimulated lipolysis before WL, or between regional changes in the activity of AT-LPL and changes in lipolysis after WL. There also were no relationships between the absolute or relative changes in any of the indices of lipolysis measured in ABD or GLT adipocytes and changes in subcutaneous (SC) fat or intra-abdominal (IA) fat area, fat mass, body weight, or in the measures of glucose or lipid metabolism after WL.
The results of this study show that the regional differences in the subcutaneous adipocyte metabolism of obese postmenopausal women persist after a 6-month hypocaloric diet and walking-induced moderate WL and may predispose older women to have an ABD fat distribution and develop the metabolic complications associated with central obesity. In addition to the previously reported higher β-AR maximal responsiveness and sensitivity in ABD vs. GLT adipocytes (13), our data show regional differences in ADA-uninhibited lipolysis in these women. Thus, ABD adipose cells had a higher tonic inhibition of lipolysis by adenosine compared with GLT cells as reflected by a higher ADA-stimulated lipolysis both before and after WL. Furthermore, the correlation between the change in the activity of AT-LPL after WL and ADA-uninhibited lipolysis at baseline suggests that women with greater adenosine inhibition of lipolysis in ADB cells may be prone to raise their AT-LPL activity after WL. Based on our previous findings in this population (14), women with these metabolic findings have less favorable metabolic responses to WL and are more prone to weight regain.
The results of studies examining the effects of WL treatment of obese subjects on β-AR stimulated lipolysis are controversial because some reports show no change (32, 33, 34, 35, 36, 37), an enhancement (31, 34, 38), or a reduction in lipolysis (39, 40) after WL. Most of these studies were performed in premenopausal women, and the disparity in the results may be due in part to differences in subject's age; the variance in study duration, from 3 to 4 weeks (31, 33, 34, 37) to as long as 12 to 15 weeks (32, 40); the concomitant use of a serotonin uptake inhibitor (32); and differences in the degree of caloric restriction, from 300 to 400 kcal/d (31, 33, 34, 37, 40) up to 800 kcal/d (32, 36). Furthermore, in some studies, the post-WL measurements were performed while the subjects were still on the hypocaloric diet (33, 34, 36, 37), whereas in others, subjects were weight stabilized on an isocaloric diet for 4 to 6 weeks before retesting (32, 40). In the present study, these obese postmenopausal women showed no change in regional basal, ADA-uninhibited basal, β-AR, or postreceptor-stimulated lipolysis when retested after WL after 2 weeks of weight stability on isocaloric diets. The similarity in β-AR responsiveness that we observed after WL is comparable with that reported by other investigators in premenopausal women after short- or long-term hypocaloric treatment in some (32, 33, 34) but not in other (31, 38, 39, 40) studies. Although Mauriege et al. (32) found similar β-AR responsiveness before and after 15 weeks of WL treatment in obese premenopausal women, they reported an increase in β-AR sensitivity (lower EC50) in adipocytes from both depots. However, we find a nonsignificant increase in β-AR sensitivity after WL in either ABD or GLT regions that eliminated the regional difference in β-AR sensitivity (ABD > GLT) observed before WL. The changes in regional fat cell metabolism might have been more dramatic had these women lost more weight and approached their ideal body weight.
The regional differences in adipose cell metabolism persisted after the moderate WL. Furthermore, the changes in SC or IA fat areas, percentage body fat, fat mass, or body weight did not correlate with the changes in fat cell metabolism. This suggests that the observed regional differences in adipocyte metabolism in obese, white, postmenopausal women are not secondary to obesity and may be related to either the menopause or aging in these women. However, because we only studied postmenopausal women within an age range of 51 to 68 years, future studies should consider studying women over time as they transition the menopause and again later in life to determine the relative contributions of menopause and aging to changes in body weight, fat distribution, and regional differences in fat cell metabolism.
Endogenous adenosine inhibits in vitro lipolysis with A1 adenosine receptors, and ADA catalyzes the irreversible deamination of adenosine to inosine (8, 31, 41). Bottini and Gloria-Bottini (42) reported that subjects with type 2 diabetes and BMIs higher than 34 kg/m2 have a high proportion of a genotype carrying the ADA*2 allele, which is associated with lower activities of ADA. AT of obese subjects has a higher content of adenosine than normal-weight subjects, which reduces the sensitivity and number of adenosine receptors through a desensitization mechanism (43). The higher lipolytic response to ADA observed in ABD vs. GLT fat cells before and after WL suggests a higher adenosine content and tonic inhibition of lipolysis in ABD than GLT cells, a finding that contrasts reports by Mauriege et al. (32, 44) of comparable ADA-uninhibited lipolysis in ABD and femoral adipocytes of obese premenopausal women. Our recent report of higher AT-LPL activity in ABD compared with GLT AT in post- compared with perimenopausal women (15), coupled with our finding of a higher tonic inhibition of lipolysis in ABD than GLT adipocytes, suggests that these abnormalities in fat cell metabolism may be associated with the menopause and predispose older women to accumulate fat in the ABD or central depot. This would predispose them to develop the metabolic syndrome, thus increasing their risk for cardiovascular morbidity and mortality.
The direction (increase or decrease) of the change in AT-LPL activity with WL has significant metabolic implications because those women who decrease their regional AT-LPL after WL seem to have better improvements in lipid risk factors and regain less weight after 6-month follow-up than women who increase their regional AT-LPL after WL (14). Our findings show that women with increased ADA-uninhibited lipolysis (an index of increased adipocyte adenosine content) before WL raise their ABD AT-LPL activity after WL and, thus, may be more susceptible to regain weight. The positive relationship between the change in ABD AT-LPL with WL and the pre-WL levels of lipolysis after removal of the inhibitory effects of adenosine suggests that factors that inhibit lipolysis may affect the responses of AT-LPL in ABD but not in GLT fat to WL. Although we did not directly measure adenosine content or the effect of A1 receptor agonist/antagonists on basal lipolysis, our results suggest that there is a relationship between the adenosinergic signaling system and responses of ABD (but not GLT) AT-LPL activity to WL.
Thus, contrary to our original hypothesis, our results suggest that regional differences in β-AR-stimulated lipolysis and sensitivity, ADA-uninhibited lipolysis, and AT-LPL activity, metabolic conditions that may predispose older women to accumulate fat centrally, are not modifiable by a moderate WL treatment. This suggests they are not secondary to obesity and may be related to the menopause and/or aging in these postmenopausal women. In addition, our findings also suggest that the regulation of ABD subcutaneous fat cell metabolism by the adenosinergic system, its relationship to the changes in AT-LPL activity with WL, and the long-term metabolic adaptations to WL may predispose some postmenopausal women to ABD adiposity and its metabolic and cardiovascular health consequences.
The authors thank all the women who participated in this study; Linda Bunyard and Kelly Barton for their assistance with dietary counseling; and the nurses, exercise physiologists, and laboratory technicians of the Geriatric Research, Education, and Clinical Center (GRECC) for assistance in the conduct of the research. This work was supported by grants from the National Institutes of Health (K01-AG00685 to D.M.B., RO1-NR03514 to K.E.D. and A.P.G., K07-AG00608 to A.P.G., R29-AG-14066 to B.J.N., and KO1-AG00747 to A.S.R.), the American Federation for Aging Research (A96178 to D.M.B.), and the Baltimore Veterans Affairs GRECC and Medical Research Service (VA MERIT Review) of the Department of Veterans Affairs.