The reversal of obesity is a difficult feat at best and is a growing problem as the obesity epidemic increases worldwide. Considerable focus has been made on the arcuate nucleus (Arc) in the control of body and lipid mass and food intake. To test the role of the Arc in body fat mobilization, we compared the effects of food deprivation on white adipose tissue (WAT) mass in adult Siberian hamsters by making exocytotic lesions of the Arc via neonatal subcutaneous injections of monosodium glutamate (MSG). MSG-treated hamsters had significantly increased body mass, total and individual WAT pad masses, and serum leptin concentrations compared with their vehicle-injected counterparts. MSG produced marked reductions in Arc Nissl staining, tyrosine hydroxylase-immunoreactive (ir) neurons, and neuropeptide Y (NPY)- and agouti-related protein (AgRP)-ir fibers compared with controls. MSG significantly decreased hypothalamic paraventricular nucleus (PVN) NPY- and AgRP fiber-ir compared with controls, likely because of Arc projections to this nucleus. MSG treatment also reduced area postrema (AP) tyrosine hydroxylase (TH)-ir fibers compared with controls. MSG treatment did not, however, block food deprivation-induced decreases in WAT pad mass compared with controls. Thus, despite considerable damage to the Arc and some of its projections to the PVN, as well as the AP, body fat was mobilized apparently normally, bringing into question the necessity of these structures for food deprivation-induced lipid mobilization. These data support recent evidence that chronically decerebrate rats, in which the forebrain is surgically isolated from the caudal brainstem, show normal food deprivation responses, including lipid mobilization.
- white adipose tissue
- sympathetic nervous system
- neuropeptide Y
- agouti-related protein
- tyrosine hydroxylase
obesity is a major health concern, and factors associated with the excessive accumulation of white adipose tissue (WAT) are linked to serious secondary adverse health consequences such as non-insulin-dependent diabetes mellitus, heart disease, stroke, and certain types of cancer (29, 57). Obesity is not relegated to adults; there are growing concerns associated with the increases in childhood obesity (37, 49). Imbalance in energy intake and expenditure underlies the development and reversal of obesity, but the brain mechanisms underlying the control of these processes and the resulting alterations in WAT lipid deposition or mobilization are not precisely known. The preponderance of research on energy intake/energy expenditure has focused on hypothalamic forebrain sites/circuits (for review, see Refs. 20, 28, 69). An even narrower focus has been placed on the hypothalamic arcuate nucleus (Arc) for numerous reasons, not the least of which is the identification of the largely adipocyte-derived cytokine leptin (74) and the localization of substantial numbers of leptin receptors on Arc neurons (e.g., Refs. 32, 41, 58). Within the ARC, there are two major cell types that are targets for leptin (27, 38). One cell type produces neuropeptide Y (NPY) and agouti-related protein (AgRP), two orexigenic peptides, whereas the other cell type produces proopiomelanocortin (POMC) and cocaine-associated related transcripts (CART), two anorexigenic peptides (for review, see Refs. 12, 20).
The role of these Arc cell types in altering energy intake is indisputable, and increasing emphasis also has been placed on their involvement in energy expenditure as well (e.g., Refs. 40, 60, 70). Of its projections, the Arc sends prominent NPY and AgRP efferents to the hypothalamic paraventricular nucleus (PVN) (1, 21). The activation of these Arc neurons by energy-related stimuli is clear, given that Arc NPY (13, 42) and AgRP gene expression (25, 42) is increased with food deprivation, whereas Arc POMC and CART gene expression is decreased (16, 38, 42). Thus these Arc peptide systems appear to be likely candidates not only for the modulation of energy intake and expenditure but also perhaps participating in other energy-related responses such WAT lipid mobilization. In terms of the latter, the Arc is a component of the sympathetic nervous system (SNS) outflow circuits to WAT, as revealed by transneuronal viral tract tracing using the pseudorabies virus (2, 15, 59, 61). The Arc has projections to the PVN, including a significant α-melanocyte-stimulating hormone (α-MSH) projection (e.g., Ref. 14, 21), and, as with the Arc, the PVN also is part of the SNS outflow to WAT (2, 15, 59, 61). Of the PVN neurons comprising the multisynaptic SNS circuit innervating WAT, they possess high numbers and percentages of melanocortin 4 receptor (MC4-R) mRNA (62). The role of the sympathetic innervation of WAT as the principal initiator of WAT lipolysis (for review, see Refs. 7, 9, 11) suggests the possible involvement of the Arc and its projections to the PVN in lipid mobilization triggered by a variety of stimuli, including food deprivation.
Consistent with this notion that the Arc/PVN may be involved in lipid mobilization is the obesity that results from destruction of either nucleus by electrolytic lesions (e.g., Refs. 4, 19, 39, 73). Although conventional lesions of the PVN can be complete, are relatively confined to the structure, and are not technically demanding, making similar lesions of the Arc that are complete and confined is difficult because of its elongated shape, but it is possible (19). Arc lesions also carry the high probability of destruction of the median eminence/pituitary, which greatly complicates the interpretation of the effects of Arc lesions. An alternative to conventional Arc lesions is their destruction by neonatal administration of monosodium glutamate (MSG). Although not completely selective to the Arc, destruction of Arc by neonatal MSG produces a profound decrease in the number of Arc neurons (up to 80–90%; Ref. 50), an effect due to the underdeveloped blood-brain barrier (BBB) in this area allowing MSG to penetrate the brain (50, 52). Not surprising, therefore, is the destruction of other areas with a weak BBB by neonatal MSG, including the area postrema (AP) in the caudal brain stem (36, 50, 63). One of the most notable effects of neonatal MSG treatment is obesity, usually without overeating, in adult laboratory mice (e.g., Refs. 23, 50) and rats (e.g., Refs. 18, 46). Such lesions, therefore, are consistent with the hypothesized role of the Arc in energy intake and expenditure.
In addition to the MSG-induced obesity without overeating (e.g., Ref. 18), this neonatal treatment produces an adult with short stature, likely associated with decreases in growth hormone (64), and sympathetic nervous system dysfunction, most notably decreases in brown adipose tissue (BAT) thermogenesis (e.g., Ref. 44, 72). More specifically in terms of the latter, MSG-induced obesity is associated with decreases in BAT norepinephrine turnover, a measure of sympathetic drive (47, 54, 71) Collectively, these data and those described above suggest that both lipid mobilization and thermogenesis are compromised in MGS-treated animals.
Therefore, the purpose of the present experiment was to test whether destruction of the Arc, and likely other areas developmentally weak in the BBB (e.g., AP), interferes with the ability of food deprivation to mobilize body fat stores in Siberian hamsters. Siberian hamsters were chosen because of the extensive information existing on the sympathetic innervation of WAT in this species (for review, see Refs. 3, 6, 10). This was accomplished by neonatal administration of MSG followed by a food deprivation test when the hamsters reached adulthood, measurement of changes in the body mass, food intake, mass of several of the major WAT depots, and circulating serum leptin concentrations. In addition, we documented the MSG-induced lesions of the Arc, AP, and Arc projections to the PVN by staining for cell bodies (Nissl staining), as well as immunohistochemistry (IHC) for several neuropeptides (NPY, AgRP, α-MSH) or enzymes of synthesis [tyrosine hydroxylase (TH) for catecholamines] found in these brain sites and known to be involved in energy balance.
MATERIALS AND METHODS
Animals, housing, and MSG treatment.
All experiments were approved by the Georgia State University Institutional Animal Care and Use Committee (IACUC) in accordance with National Institutes of Health and United States Department of Agriculture guidelines. Thirty-three neonatal female Siberian hamsters from our breeding colony, the lineage of which was described recently (15), were injected subcutaneously into the dorsal dermis area just below the interscapular region from postnatal day 1 to 5 (postnatal day 0 = day of birth) with 25 μl of MSG (Sigma-Aldrich, St. Louis, MO) to deliver 4 mg/g body mass or with vehicle [0.1 M phosphate buffer (PBS), pH 7.4] according to the method of Ebling et al. (26) for successful MSG treatment in this species. An additional 17 animals (8 PBS, 9 MSG) were included for supplementary histological measures (α-MSH staining). All animals within a litter were injected with MSG or PBS to minimize pup cannibalization that occurs in this species with differential treatment of members of the same litter (unpublished observations). A 13% death rate occurred in the MSG-injected pups, compared with no deaths in the 20 pups treated with PBS. Eight animals (3 MSG, 5 PBS), however, had to be excluded from the analysis because their body mass was low, being greater than two standard deviations from the mean, resulting in 18 MSG and 15 PBS animals. Hamsters were weaned at 21 days postpartum and housed in single polypropylene cages (27.8 × 17.5 × 13.0 cm) containing corn cob bedding (The Andersons, Maumee, OH) and cotton nestlets (Ancare, Belmore, NY) until used in the present experiment. The temperature, relative humidity, and photocycle of the vivarium were 20 ± 1.5°C, 50 ± 5%, and 16:8-h light-dark (lights on at 0300), respectively, throughout the experiment. Animals were given Purina lab chow 5001 (TestDiet, Brentwood, MO) and tap water ad libitum unless otherwise noted.
Food deprivation, serum, and tissue harvesting.
At ∼12 wk of age, animals had food and bedding removed from their pouches (depouched), were placed in new bedding, and were food deprived for 56 h (IACUC approved). In our previous studies, we used food deprivation ranging from 12 to 56 h (IACUC approved). The 56-h food deprivation length may appear extreme or “nonphysiological” at first; however, Siberian hamsters are ∼50% body fat (e.g., Refs. 8, 67), and behavioral responses triggered by food deprivation, for example, foraging and food hoarding, are not engaged until body fat levels have appreciably decreased (e.g., Refs. 5, 68); the threshold for these responses is ∼24 h of food deprivation and reaches peaks at 48–56 h. Food intake was measured weekly after weaning. Body mass was measured before and after food deprivation. At 56 h post-food deprivation, animals were lightly anesthetized with isoflurane and orbital blood was taken for serum measurement of leptin. Serum was collected into tubes, centrifuged at 4°C for 30 min at 2,000 rpm, and the serum aliquots were stored at −20°C until analysis. A commercially available mouse leptin ELISA kit (LINCO Research, St. Charles, MI) previously validated for Siberian hamsters (24) was used. Animals were then anesthetized (pentobarbital sodium, 50 mg/kg ip), and the right inguinal WAT (IWAT), bilateral retroperitoneal WAT (RWAT), bilateral parametrial WAT (PWAT), and interscapular brown adipose tissue (IBAT) depots were removed and weighed.
Animals were then perfused transcardially with ∼125 ml of isotonic saline, followed by ∼125 ml of 4% paraformaldehyde in 0.1 M PBS solution (pH 7.4). Seventeen additional animals were perfused transcardially with ∼125 ml of isotonic saline, followed by ∼125 ml of 4% paraformaldehyde (pH 7.4) and 2.5% acrolein in phosphate buffer (pH ∼6.8) for improved histological quantification. All animals were perfused between 1200 and 1600. Brains were stored in 4% paraformaldehyde and transferred to sucrose (30%, and 0.1% sodium azide) after 24 h until they were processed for histology. Brains were sectioned in the coronal plane on a freezing microtome at a thickness of 40 μm. Brain sections were stored in 0.1 M PBS with 0.1% sodium azide at 4°C if histology was to be performed within 30 days; otherwise, sections were stored in cryoprotectant at −20°C.
To reduce background staining, sections were incubated in 0.6% H2O2 in 0.1 M PBS with 0.4% Triton X-100 (PBTx) for 30 min and rinsed with 0.1 M PBS (3 times). A blocking step was introduced using 10% normal goat serum (Biomeda, Foster City, CA) in PBTx. Sections were then incubated overnight at room temperature in primary antibodies for TH, NPY, and AgRP in 2% normal goat serum in PBTx. Antibodies (mouse anti-TH, 1:10,000; rabbit anti-NPY, 1:10,000; rabbit anti-AgRP, 1:10,000) were obtained commercially (Immunostar, Hudson, WI). This procedure was followed by incubation in the secondary antibodies of biotinylated goat anti-mouse (TH) and biotinylated goat-anti-rabbit (NPY, AgRP; Vector Laboratories, Burlingame, CA) for 2 h at room temperature in 2% normal goat serum in PBTx. In addition, sections were bathed in avidin biotin solution (Vector Laboratories, Burlingame, CA) for 1 h at room temperature. Slides were then rinsed in 0.1 M PBS (3 times), and a chromagen was generated using 3.3′-diaminobenzidine tetrahydrochloride (DAB) and 0.00024% H2O2 with 0.1 M PBS. Finally, the sections were rinsed with 0.1 M PBS (3 times), mounted onto microscope slides (Superfrost; Fisher Scientific, Suwanee, GA), and dried overnight. The slides were then dehydrated, delipidated, and coverslipped using Permount (Fisher Scientific). An additional set of sections was stained with cresyl echt violet for Nissl body staining. Acrolein-perfused brains were used for α-MSH staining only. Sections were rinsed in PBS (4 times) and incubated in 1% Na borohydride for 20 min. Sections were then rinsed (10 times) in PBS and placed in primary antibody (rabbit anti- α-MSH; Immunostar) diluted in PBTx for 48 h at 4°C. After removal from the antiserum and rinsing in PBS (10 times), sections were placed in biotinylated secondary antibody (biotinylated goat-anti-rabbit; Vector Laboratories) used at 1:600 for 1 h. Sections were cleared from the secondary antibody via rinsing with PBS (4 times) before being placed in avidin biotin solution (Vector Laboratories) at 4.5 μl/ml for 1 h at room temperature. Sections were then rinsed in sodium acetate (0.5 M; 3 times) and placed in nickel sulfate-DAB chromogen solution for 10–15 min. The diluted nickel sulfate-DAB chromagen solution was prepared using 0.002 g of DAB, 0.25 g of nickel sulfate, and 8.3 μl of 3% H2O2 in 40 ml of sodium acetate. After final rinsing in sodium acetate (3 times) and PBS (3 times) to stop the reaction, sections were processed as mentioned previously.
Density and cell counts.
TH and α-MSH immunoreactivity (ir) for cell bodies was quantified visually in the Arc and AP using a microscope with magnification of ×20 (Nikon eclipse E800, 34357; camera: Retiga EXi Qimaging, 32-0062A-143). Optical density of NPY- and AgRP-ir fiber staining, as well as Nissl staining, was determined using a commercial computer-assisted imaging system (IPLab Scientific Image Processing, version 3.9.1, Mac OS 10.4.7; Scanalytic, Fairfax, VA). Three levels of the Arc were analyzed using a mouse brain atlas to guide in the identification of brain regions (51) approximately corresponding to the stereotaxic coordinates of the Arc anterior region (bregma, −1.46 mm), medial region (bregma, −1.70 mm), and posterior region (bregma, −2.18 mm). In addition, three levels also were chosen for PVN analysis (anterior region: bregma, −0.82 mm; medial region: bregma, −1.06 mm; and posterior region: bregma, −1.22 mm). Only one level of the AP was analyzed (bregma, −7.48). Each brain site was bilaterally scored three times, averaged, and subtracted from an averaged (3 measures) background. All quantification was done with the observer blind to the experimental condition.
Data were analyzed using a two-way ANOVA [2 × 2: treatment (MSG/PBS) × food deprivation/non-food deprivation group] with Bonferroni post hoc tests (GraphPad Prism version 4.00; GraphPad Software, San Diego, CA) when appropriate. Food intake was analyzed using a repeated measures ANOVA (MSG/PBS) with Bonferroni post hoc tests. Differences between means were considered significant if P < 0.05. Exact probabilities and test values were omitted for simplicity and clarity of presentation.
Effects of MSG and food deprivation on Arc, PVN, and AP neuroanatomy.
MSG treatment significantly decreased Nissl staining (Fig. 1A), NPY- and AgRP-ir fibers (Fig. 1, B and C), and TH-ir cells (Fig. 1D) at all three levels of the Arc (P < 0.05). MSG treatment significantly reduced α-MSH cell bodies in the medial and posterior levels of the ARC (P <0.05; Fig. 1E), but not in the anterior level of the ARC (P < 0.1). Food deprivation alone did not alter NPY or AgRP fiber-ir; however, the decrease in anterior AgRP fiber-ir of the ARC by MSG was exaggerated with food deprivation compared with that in PBS-treated animals (P < 0.05; Fig. 1C). Finally, thinning of the medial eminence and retraction of necrotic tissue around the third ventricle were observed in MSG-treated animals (Fig. 2, A and B).
MSG treatment significantly reduced TH- ir (P < 0.05; Figs. 5 and 6), but not α-MSH-ir in cells of the AP (data not shown). No other significant interactions between MSG treatment and food deprivation were observed.
Body and WAT pad masses, food intake, and serum leptin concentrations.
MSG treatment during the neonatal period significantly increased food intake over the course of 3 mo (Fig. 7A) and increased body mass in ∼3-mo-old Siberian hamsters (P < 0.05) compared with the PBS controls (Fig. 7B). In addition, MSG significantly increased WAT pad mass for all depots assayed (IWAT, RWAT, and gonadal WAT; P < 0.05; Fig. 8) as well as IBAT (Fig. 8). Total dissected WAT also was significantly increased by MSG treatment compared with PBS-treated controls (P < 0.05; Fig. 9).
Neonatal MSG injections did not block or diminish the food deprivation-induced lipid mobilization from any of the WAT pads assayed as evidenced by their decreased masses compared with the PBS-treated controls. Thus both MSG- and PBS-treated animals had significant food deprivation-induced decreases in IWAT and RWAT mass (P < 0.05) compared with their respective ad libitum-fed counterparts, but GWAT mass did not decrease with food deprivation for either group compared to their respective ad libitum-fed controls (Fig. 8).
Leptin concentrations were significantly increased in MSG-treated hamsters compared with their PBS-treated controls (P < 0.05; Fig. 10), and both MSG- and PSB-treated hamsters showed similar significant food deprivation-induced decreases in leptin concentrations (P <0.05; Fig. 10).
In this study, we tested whether MSG-induced destruction of the Arc (and other brain sites showing increased sensitivity to this neonatal treatment such as the AP) inhibited WAT lipid mobilization in food-deprived Siberian hamsters, as indicated by decreases in WAT pad mass, an integrative measure of adiposity. Despite significant damage to the Arc, as well as to some of its projections to the PVN (AgRP and NPY) and to the AP in the brain stem (see directly below), and despite these areas being involved in the SNS outflow to WAT as revealed by retrograde transneuronal viral tract tracing after virus inoculation in WAT (2, 15, 30, 59, 61, 62), MSG-treated animals mobilized body fat stores comparably to their PBS-treated controls when food deprived. This uncompromised ability of food-deprived adult hamsters to mobilize WAT lipid was not due to ineffective neonatal treatment with MSG, because MSG-treated hamsters had significantly decreased Arc Nissl (mean: approximately −43.2% across the 3 levels), indicating overall cell loss, as well as significantly decreased Arc TH-ir cells (mean: approximately −42% across the 3 levels) and Arc NPY and AgRP fiber-ir (means: approximately −34 and −46%, respectively, across the 3 levels). In addition, likely Arc NPY and AgRP projections to the PVN also were compromised by the lesion, as evidenced by significantly decreased fiber-ir for both NPY and AgRP (means: approximately −45 and −53%, respectively, across the 3 levels). The AP was significantly compromised as well, as indicated by significant decreases in TH-ir cells (mean: approximately −33.4%). Therefore, the MSG-produced Arc lesions and likely Arc projections to the PVN, as well as AP lesions, were substantial, and the normal function of these areas likely was compromised, yet food deprivation-induced WAT lipid mobilization appeared normal. Thus it appears that an intact Arc/PVN and/or AP are not necessary for food deprivation-induced lipid mobilization from WAT but could play a role in lipid mobilization triggered by other energy-demanding conditions such as exercise or cold exposure. Food deprivation-induced lipid mobilization in neonatally treated MSG laboratory rats (53, 56) and mice (48) also occurs unimpeded, although the degree of MSG-induced neural destruction was not verified. In addition, MSG-induced Arc destruction does not block the decrease in body and lipid mass of Siberian hamsters exposed to short “winterlike” days as occurs in intact animals (26).
In addition to the histological evidence for the effectiveness of neonatally MSG-treated hamsters, these animals also displayed the typical obesity syndrome seen in similarly treated laboratory rats and mice (e.g., Refs. 17, 18, 35, 45, 55, 56). Body mass was significantly increased at ∼3 mo of age following MSG treatment compared with PBS-treated controls, as were all assayed WAT pads (IWAT, GWAT, RWAT) and serum leptin concentrations. MSG-treated laboratory rats and mice usually are not absolutely heavier than their controls; instead, they have greater adiposity, indicating a redistribution of carcass components (18). Unlike the present study, the first study to test the effects of neonatal MSG treatment in Siberian hamsters (26) found no increase in body mass; instead, laboratory rat- and mouse-like MSG-induced decreases in body mass were observed with no significant increase in EWAT mass. The reason for the discrepancies in the effects of MSG on these measures between the two studies may reside with the differences in the sex of the hamsters, because we used females and Ebling et al. (26) used males. As noted above, females tend to be more sensitive to the body and lipid mass-promoting effects of neonatal MSG treatment (50). The genesis of the MSG-induced obesity in the present study is precisely unknown, because we did not measure energy expenditure, only food intake. In identically MSG- or PBS-treated male Siberian hamsters (Dailey M and Bartness T, unpublished observations), we found significant decreases in food intake similar to that seen in MSG studies of laboratory rats and mice (22, 43, 46, 64) in contrast to the significant increases in food intake reported in the present study. In MSG-treated laboratory rats and mice, there are decreases in thermogenic and/or sympathetic activity in their BAT (43, 44, 65, 66, 71, 72) that thereby promote obesity via decreases in energy expenditure in the presence of hypophagia. Therefore, in the present study, it appears that because food intake by MSG-treated hamsters is increased, this likely contributes to the obesity along with a probable decrease in thermogenesis, but again the latter was not measured.
The overall purpose of this study was to test whether food deprivation-induced lipid mobilization from WAT, as evidenced by the integrative measure of decreases in WAT mass, was compromised by destruction of the Arc and some of its projections to the PVN, as well as other circumventricular areas susceptible to MSG destruction (e.g., AP). Despite MSG-induced destruction of the Arc and its likely projections to the PVN, as well as AP damage, food deprivation-induced lipid mobilization was not different from that in PBS-treated controls. While this study was in progress, we also found that food deprivation-induced lipid mobilization was uninfluenced by a much more significant lesion than produced by the MSG. Specifically, we made complete transections of the neuroaxis in laboratory rats at the mesodiencephalic juncture (i.e., chronic decerebration; for review of this model, see Ref. 31). When food deprived, chronic decerebrate rats show similar responses to their neurally intact sham-operated controls, with both groups reducing energy expenditure, decreasing respiratory quotient (indicating lipid utilization), defending body temperature, decreasing serum leptin and insulin concentrations, and, most importantly for the present study, mobilizing lipid from WAT (33). Thus, in these chronically decerebrate rats in which all forebrain structures, not just the Arc, are disconnected from the brain stem, thereby blocking forebrain descending sympathetic outflow to WAT as well as ascending inputs to forebrain sympathetic control areas (e.g., PVN, dorsomedial nucleus, suprachiasmatic nucleus), food deprivation-induced increases in lipid mobilization proceed unabated. Similarly, in another set of studies conducted while the present work was in progress, we found that bilateral electrolytic lesions of the PVN failed to affect food deprivation-induced lipid mobilization in Siberian hamsters (Foster M and Bartness T, unpublished observations). Furthermore, microknife cuts that sever the descending projections of the PVN to the brain stem also fail to block food deprivation-induced lipid mobilization (34). To some extent, these collective results are surprising, because the Arc, PVN, and AP all are part of the SNS outflow from brain to WAT (2, 15, 30, 59, 61, 62). It may be, however, that not all of the labeled neurons in these circuits participate in all conditions of WAT lipid mobilization; rather, different collections of neurons may participate under more restrictive conditions (i.e., one set for food deprivation, a different set for cold exposure and the like). Collectively, the present data, our contemporary data (Ref. 33; Foster M and Bartness T, unpublished observations), and the data of others (e.g., Ref. 34) suggest that the Arc, PVN, or, indeed, the forebrain more generally (33) are not necessary for the normal food deprivation-induced increases in lipid mobilization but that the caudal brainstem contains the sufficient neurocircuitry for such responses. This is not to dismiss the role of these forebrain structures in normal food deprivation-induced increases in lipid mobilization, because it would be surprising if they did not refine and coordinate responses between peripheral tissues involved in energy balance (e.g., muscle, liver, WAT, and BAT) to stimuli that challenge energy balance.
This research was supported, in part, by National Institute of Diabetes and Digestive and Kidney Diseases Grant R01 DK-35254 (to T. J. Bartness).
We thank the animal care staff at Georgia State, Karma Mehta for help during the MSG injection period, Dr. Anne Murphy (Georgia State University) for the use of the imaging system, and Dr. C. Kay Song for assistance with immunohistochemical procedures.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society