Introduction

Melanocortins are bioactive peptides derived from a common prehormone, pro-opiomelanocortin (POMC); the central melanocortin system plays a critical role in the regulation of energy balance and glucose metabolism [1–6]. Reduced expression of hypothalamic Pomc is associated with obesity syndromes caused by: (1) mutations in any of several genes, including leptin receptor [7, 8], tubby [9] and Nhlh2 [10]; (2) hypothalamic damage [11]; and (3), perhaps most commonly, by ageing [12]. Reduced hypothalamic Pomc mRNA may be one contributor to the obese phenotypes in these models because mutations in the Pomc gene cause obesity in mice [13] and humans [14]. However, there is still little evidence that normalisation of central POMC tone can reverse obese phenotypes. The present study aimed to address this question using an age-related obese rat model.

A common form of obesity, age-related obesity, is characterised by a progressive increase in body weight and visceral adiposity. This increase in body weight with age is a major risk factor for insulin resistance, diabetes and atherosclerotic cardiovascular disease [15]. Aged mice and rats are associated with reduced hypothalamic Pomc expression [12], and the induction of Pomc by exogenous leptin is impaired with ageing [16]. The male F1 hybrid of Fischer 344-Brown Norway rats (F344/BN) is a useful model for age-related obesity because these rats have a relatively long lifespan and gain a substantial amount of body weight and fat mass with ageing [17]. F344/BN rats also display age-associated impairments in glucose metabolism and insulin responsiveness [18]. Although chronic pharmacological treatment of melanocortin agonists in rodents, including aged rats, reduces food intake and increases energy expenditure, its effectiveness is limited by the rapid tachyphylaxis of the melanocortin responses [19–21]. On the other hand, we have demonstrated previously that Pomc gene delivery mediated by recombinant adeno-associated virus (rAAV) elicited a sustained anorexic response (up to 38 days) in obese Zucker rats with defective leptin receptors [22]. Using the F344/BN rat as a model of age-related obesity, we sought to examine whether overproduction of POMC in the hypothalamus could reverse the metabolic impairments associated with age, by reducing body fat and weight while improving glucose metabolism. In addition, we hypothesised that long-term targeted overexpression of Pomc in the hypothalamus would result in a prolonged anorexic response compared with the pharmacological administration of melanocortin agonists.

The advantages of using rAAV to obtain long-term transgene expression include site-specific integration within a defined region of human chromosome 19, the ability to transduce postmitotic tissues efficiently, and a lack of pathogenicity and immunogenicity [23]. We have demonstrated previously that a serotype 2 rAAV vector encoding Pomc reduces visceral adiposity and improves insulin sensitivity in genetically obese Zucker rats [22]. A new serotype rAAV, rAAV type 5, has been developed recently and shown to be more efficient in transducing select tissues in vivo [24, 25]. In the present study, we used a serotype type 5 rAAV vector encoding murine Pomc (rAAV-Pomc) to assess the long-term consequences of Pomc gene delivery on energy balance, glucose metabolism, brown adipose tissue thermogenesis and mRNA levels of hypothalamic neuropeptides and melanocortin receptors in aged obese F344/BN rats.

Materials and methods

Construction of rAAV vector plasmids

pTR-Pomc encodes the full-length 935-bp murine Pomc cDNA [26] under the control of the hybrid cytomegalovirus immediate early enhancer/chicken β-actin promoter [27]. The woodchuck hepatitis virus post-transcriptional regulatory element is placed downstream of the Pomc transgene to enhance its expression [28] (Fig. 1). The control plasmid, termed pTR-control, is similar to pTR-Pomc except for the incorporation of the cDNA encoding an enhanced form of green fluorescent protein instead of Pomc cDNA. The control vector has been described previously [29].

Fig. 1
figure 1

Schematic presentation of the rAAV-Pomc construct. TR AAV terminal repeat sequence; CMV ie enhancer cytomegalovirus immediate-early enhancer; WPRE woodchuck hepatitis virus post-transcriptional regulatory element; bGH poly(A) bovine growth hormone polyadenylation signal

Full size image

Packaging of rAAV vectors

The plasmids pTR-Pomc and pTR-control were packaged in serotype 5 rAAV capsids. Serotype 5 was produced by the process known as ‘pseudotyping’, by using the helper plasmid pXYZ5 [30], containing the rAAV 5 capsid genes. Vectors were packaged, purified, concentrated and titred as described [30]. The titres for the rAAV-Pomc and rAAV-control vectors used in this study were 2.51×1013 physical particles/ml.

Animals

Male F344/BN rats aged 22 months were obtained from Harlan Sprague–Dawley (Indianapolis, IN, USA) under contract with the National Institute on Aging. Animals were cared for in accordance with the principles of the NIH Guide to the Care and Use of Experimental Animals. Rats were housed individually with a 12:12 h light:dark cycle (lights on 0700 to 1900 hours). Free access was provided to standard Purina 5001 rodent diet and water.

Administration of rAAV vector

Under anaesthesia induced by 6 mg/kg xylazine (Phoenix Pharmaceutical, St Joseph, MO, USA) and 60 mg/kg ketamine (Monarch Pharmaceuticals, Bristol, TN, USA), rats were bilaterally administered rAAV-Pomc (n=12) or rAAV-control (n=12), 2.51×1010 particles/injection in 1 μl, or artificial cerebrospinal fluid (surgical sham control, n=5, 1 μl per injection) into the basomedial hypothalamus with coordinates targeting the arcuate nucleus, as described previously [22]. Using an UltraMicropump II system (World Precision Instruments, Sarasota, FL, USA), a 1-μl volume of virus stocks was delivered over 5 min to each site. The needles remained in place at the injection site for a further 5 min. At the time of surgery, rats were injected with the analgesic Buprenex (buprenorphine hydrochloride, 0.05 mg/kg; Reckitt and Colman, Richmond, VA, USA).

Intraperitoneal glucose tolerance test

An intraperitoneal glucose tolerance test (IPGTT) was performed on days 20 and 36 after vector administration. Rats were fasted overnight and injected intraperitoneally with glucose (2 g/kg body weight) at 10.00 hours. blood was taken from the tail vein immediately before glucose injection and 15, 30, 60 and 120 min after injection. Blood glucose was measured by One Touch SureStep glucose meter (LifeScan, Milpitas, CA, USA). Plasma insulin concentrations during IPGTT were also measured by a rat insulin ELISA kit (Linco Research, St Charles, MO, USA).

Tissue collection and preparation

Rats were killed by cervical dislocation under 100 mg/kg pentobarbital anaesthesia at days 22 and 42. Cardiac blood, hypothalamus, brown adipose tissue and perirenal, retroperitoneal and epididymal white adipose tissues were obtained as described previously [22]. For dissection of hypothalamic nuclei, two groups of rats (n=6 for rAAV-Pomc and rAAV-control, respectively) were killed and brains were rapidly excised, chilled on ice-cold saline and sliced using a Stoelting tissue slicer. Brains were sectioned at 0, −2 and −5.0 mm relative to the anterior commissure, corresponding to the brain atlas of Paxinos and Watson [31]. The paraventricular nucleus was removed from the slices taken at 0 and −2 mm, and the arcuate nucleus, dorsomedial hypothalamic nucleus and lateral hypothalamic area were removed from the −2 and −5 mm slices. Tissues were stored at −80°C until analysis. For western blot analysis, brown adipose tissue was homogenised as previously described [22]. Protein was determined using the DC protein assay kit (Bio-Rad, Hercules, CA, USA).

Serum leptin, NEFA, triglyceride and cholesterol

Serum leptin was measured using a rat radioimmunoassay kit (Linco Research). Serum NEFA, triglyceride and total cholesterol levels were determined by enzymatic colorimetric kits from Wako Chemicals (Richmond, VA, USA).

RT-PCR

Expression levels of Pomc, neuropeptide Y (Npy), agouti-related protein (Agrp), melanocortin 3 receptor (Mc3r), melanocortin 4 receptor (Mc4r) in the hypothalamus were identified by relative quantitative RT-PCR using the QuantumRNA 18S Internal Standards kit (Ambion, Austin, TX, USA) as described previously [22, 29]. Relative quantitative PCR was performed by multiplexing corresponding primers (Pomc sense 5′-GCTTGCAAACTCGACCTCTC-3′, antisense 5-CTTGATGATGGCGTTCTTGA-3′; Npy sense 5′-ATGGGGCTGTGTGGACTGACC-3′, antisense 5′-GTCAGGAGAGCAAGTTTCATTT-3′; Agrp sense 5′-AGGGCATCAGAAGGCCTGACCA-3′, antisense 5′-CTTGAAGAAGCGGCAGTAGCAC-3′; Mc3r sense 5′-AGCAACCGGAGTGGCAGT-3′; antisense 5′-GGCCACGATCAAGGAGAG-3′; Mc4r sense 5′-AGTCTCTGGGGAAGGGGCA-3′; antisense 5′-CAACTGATGATGATCCCGAC-3′), 18S primers and competimers, and coamplifying. The linearity for all amplicons was determined to be between 20 and 29 cycles. The optimum ratio of 18S primer to competimer was 1:5 for Pomc, 1:7 for Npy, 1:4 for Agrp, 1:6 for Mc3r and 1:9 for Mc4r. PCR was performed at 94°C denaturation for 60 s, 59°C annealing temperature for 50 s, and 72°C elongation temperature for 50 s for 26 (Pomc), 22 (Npy), 25 (Agrp), 23 (Mc3r) or 27 (Mc4r) cycles. The PCR product was electrophoresed on acrylamide gel and stained with SYBR green (Molecular Probes, Eugene, OR, USA). Gels were scanned using a Storm fluorescent scanner (Molecular Dynamics, Sunnyvale, CA, USA) and data analysed using ImageQuant (Molecular Dynamics). The relative value of each mRNA was derived by dividing the signal obtained for the corresponding amplicon by that for the 18S amplicon.

UCP1 protein

Immunoreactive uncoupling protein 1 (UCP1) was determined with an antibody specific to rat UCP1 (Linco Research). Brown adipose tissue homogenates (20 μg) prepared as described previously [22] were separated on an SDS-PAGE gel and electrotransferred to nitrocellulose membrane. Immunoreactivity was assessed with an antibody specific to UCP1. Immunoreactivity was visualised by enhanced chemiluminescent detection (Amersham Pharmacia Biotech, Piscataway, NJ, USA) and quantified by densitometry.

Statistical analysis

Results are presented as means±SE. Repeated measures ANOVA was used for analyses of body weight and food intake. Two-way ANOVA was employed for analyses of Pomc gene expression in different areas of the hypothalamus. When the main effect was significant, a post hoc test (t-test or the Bonferroni method) was applied to determine individual differences between means. For all other data, unpaired two-tailed Student’s t-test was employed. A value of p<0.05 was considered significant.

Results

Pomc expression in the hypothalamus of obese aged rats

To verify overexpression of the Pomc transgene following central bilateral viral delivery, Pomc mRNA was measured in the whole hypothalamus by RT-PCR (Fig. 2a). Forty-two days after vector delivery, hypothalamic Pomc mRNA levels were elevated 12-fold in obese aged F344/BN rats given rAAV-Pomc compared with those given rAAV-control (p<0.001). Using the brain punch technique, in two additional groups of rats we identified Pomc mRNA only in the arcuate nucleus of the control rats, while in rAAV-Pomc rats the highest levels of Pomc expression were located in the arcuate nucleus (an 11-fold increase over the counterpart in controls). In addition, a significant amount of Pomc mRNA was also found in the dorsomedial hypothalamic nucleus and lateral hypothalamic area, and the lowest levels were observed in the paraventricular nucleus (Fig. 2b).

Fig. 2
figure 2

Hypothalamic Pomc expression 42 days after rAAV-Pomc or control vector delivery in aged-obese rats. a Representative image of relative quantitative RT-PCR analysis of whole hypothalamic Pomc mRNA with 18S rRNA as an internal standard. b Quantification of Pomc mRNA normalised to 18S RNA. c Comparisons of Pomc mRNA levels in the paraventricular nucleus (PVN), arcuate nucleus (ARC), dorsomedial hypothalamic nucleus (DMH) and lateral hypothalamic area (LH) of control (open bars) and rAAV-Pomc-treated (closed bars) rats. Data are means±SE from six rats per group. *p<0.001 vs control by unpaired t-test; † p<0.001 vs all others by two-way ANOVA with post-hoc Bonferroni test

Full size image

Food intake and body weight

Bilateral delivery of rAAV-Pomc into the basomedial hypothalamus resulted in a sustained reduction in body weight and a transient (19 days) suppression of food intake in aged obese F344/BN rats (Fig. 3). Both rAAV-Pomc- and rAAV-control-treated rats had transient anorexia after surgical administration of vectors but completely recovered their food intake from day 5 after vector injection. Pomc gene delivery significantly reduced food consumption from day 10 after vector injection, and these rats ate 2–4 g less per day compared with rAAV-control rats. However, this anorexia attenuated by day 30 (Fig. 3a). The patterns of feeding and weight change of the surgical sham control were very similar to those of rAAV-control rats during the whole experimental period (data not shown). There was a steady decrease in body weight following Pomc gene delivery during most of the period when food consumption was diminished. Near the end of and after attenuation of the anorexic response, body weight reached a plateau. In contrast to the return of food consumption to control level, the weight reduction was sustained for the duration of the experiment (Fig. 3b). Before and on the day of vector delivery, average body weight of rAAV-Pomc-treated rats was comparable to that of rAAV-control rats (591±23 vs 569±18 g at day 0, p=0.5). Immediately after vector delivery, both rAAV-Pomc- and rAAV-control-treated rats lost 20–25 g of body weight. This is probably a result of surgical insult, although we cannot rule out the possibility that a viral infection in the hypothalamus may play a minor role. While body weight of rAAV-control rats remained steady throughout the whole experimental period, rats given rAAV-Pomc lost more weight starting from day 10, and the difference in body mass between the two groups gradually increased over the 42 days (Fig. 3b). At the end of the experiment (day 42), rAAV-Pomc rats had lost an average of 9.5% of their initial body weight compared with only 2.7% in control rats (weight change −56.8±9.9 vs −14.8±5.0 g, p<0.01).

Fig. 3
figure 3

Food consumption (a) and body weight change (b) after rAAV-Pomc (closed circles) or rAAV-control (open circles) delivery in aged obese rats. The vectors were injected at day 0. Data are means±SE of six rats per group. p<0.001 for difference in food intake (a) and weight change (b) with treatment by repeated measures ANOVA. *p<0.05 for difference in food intake (a) or weight change (b) by unpaired t-test at the individual time point

Full size image

Visceral adiposity and serum leptin levels

Because Pomc gene therapy reduced the body weight of the aged obese rats, body adiposity levels were assessed. Forty-two days after central Pomc gene delivery, there were significant reductions in the visceral adiposity, as reflected by a 19% reduction in the sum of the perirenal, retroperitoneal and epididymal white adipose tissues (p<0.05) in rAAV-Pomc-treated compared with control rats (Fig. 4a). Given the difference in overall body weight, we also normalised the sum of the three regions of adipose tissues to the total body weight. By this calculation, there was also a significant reduction in visceral adiposity with Pomc gene delivery compared with controls (4.56±0.25% of total body weight vs 5.21±0.22%, p<0.05). Serum leptin levels, one indicator of body fat mass [32], were 33% lower in the rAAV-Pomc group compared with the control group (Fig. 4b). In rats killed on day 22 after central Pomc gene delivery, there was only a non-significant 11% reduction in visceral adiposity (23.4±1.4 g vs rAAV-control 26.4±1.5 g, p=0.17) and a 27% decrease in serum leptin (15.2±1.7 ng/ml vs rAAV-control 20.8±1.0 ng/ml, p<0.05).

Fig. 4
figure 4

Visceral adiposity (a) and fasting serum leptin (b) 42 days after rAAV-Pomc or rAAV-control delivery in aged obese rats. Visceral adiposity levels are represented by the sum of perirenal, retroperitoneal and epididymal white adipose tissues. Data are means±SE of six rats per group. *p<0.05 vs. control by unpaired t-test

Full size image

Intraperitoneal glucose tolerance test

There is mounting evidence suggesting that the central melanocortin system plays a role in the regulation of glucose metabolism. Therefore, we performed an IPGTT on days 20 and 36 after vector injection. On day 20, when the differences in body weight and adiposity were much smaller than those on day 42, Pomc gene delivery had not altered the glucose metabolism (Fig. 5a). As shown in Fig. 5b, although the fasting blood glucose levels of rAAV-control and rAAV-Pomc-treated rats were similar, Pomc gene delivery significantly accelerated glucose clearance after glucose administration on day 36 (p<0.05 by repeated ANOVA). The capacity to metabolise glucose in these aged obese rats, as indicated by the area under the curve in the IPGTT, was improved by 19% following rAAV-Pomc treatment. Consistent with these findings, plasma insulin levels at each time point were reduced in rAAV-Pomc-treated rats compared with rAAV-control rats following the glucose challenge (Fig. 5c).

Fig. 5
figure 5

Levels of blood glucose on day 20 (a) and day 36 (b) and plasma insulin on day 36 (c) after rAAV-Pomc (closed circles) or rAAV-control (open circles) delivery in aged obese rats following intraperitoneal administration of glucose (2 g/kg body weight; time 0). Data are means±SE of six rats per group. p<0.05 for difference in blood glucose (b) and plasma insulin (c) with treatment by repeated measures ANOVA. *p<0.05; **p<0.01 vs control by unpaired t-test

Full size image

Serum NEFA, triglyceride and cholesterol

Forty-two days of Pomc gene delivery reduced serum NEFA and triglyceride by 30% and 15%, respectively (Fig. 6a,b). Meanwhile, serum cholesterol levels were comparable between rAAV-Pomc and control groups (Fig. 6c).

Fig. 6
figure 6

Serum NEFA (a), triglyceride (b) and cholesterol (c) levels 42 days after rAAV-Pomc or rAAV-control delivery in aged obese rats. Data are means±SE of six rats per group. *p<0.05 vs control by unpaired t-test

Full size image

Brown adipose tissue

Induction of UCP1 in brown adipose tissue is an important marker for enhanced thermogenesis and thus energy expenditure in rodents [33, 34]. In the present study, we examined the UCP1 protein levels 42 days after Pomc gene delivery. Total brown adipose tissue weight markedly declined with rAAV-Pomc treatment, whereas the protein concentration (per unit of that tissue) increased dramatically, suggesting that the reduction in brown adipose tissue mass was due to the lipolysis associated with the activation of this tissue. This was supported by a 55% increase in brown adipose tissue UCP1 protein level in the rAAV-Pomc-treated compared with the control rats (Table 1).

Table 1 Uncoupling protein 1 and brown adipose tissue parameters 42 days after rAAV-Pomc or rAAV-control delivery
Full size table

Hypothalamic Npy, Agrp, Mc3r and Mc4r mRNA levels

To assess the effects of Pomc gene delivery on the expression of hypothalamic neuropeptides, Npy and Agrp mRNA levels were measured 42 days after Pomc vector delivery (Table 2). RT-PCR revealed that the expression of Agrp, the endogenous antagonist of melanocortin receptors, was unchanged in rats given rAAV-Pomc compared with control rats, whereas the mRNA levels of another potent orexigenic neuropeptide, Npy, tended towards a decrease (p=0.14).

Table 2 Hypothalamic neuropeptides, melanocortin 3 receptor (Mc3r) and melanocortin 4 receptor (Mc4r) 42 days after rAAV-Pomc or rAAV-control delivery
Full size table

MC3R and MC4R are the predominant melanocortin receptors in the hypothalamus and mediate the effects of POMC-derived α-melanocyte-stimulating hormone (α-MSH) on the homeostatic regulation of body weight. The delivery of rAAV-Pomc reduced the expression levels of hypothalamic Mc3r and Mc4r by 17% (p=0.09) and 25% (p<0.05), respectively, compared with rAAV-control (Table 2).

Discussion

The present study examined the long-term consequences of central rAAV-mediated Pomc gene therapy in aged obese F344/BN rats. Our findings are in agreement with our earlier short-term pharmacological study indicating that direct activation of the central melanocortin system by the α-MSH analogue melanotan II (MTII) is effective in partially reversing the obese phenotype in aged rats [19]. Our present approach in particular resulted in long-term overexpression of Pomc in the basal hypothalamus using serotype 5 rAAV, and these aged rats responded with significant reductions in both food intake and body weight. To our knowledge, this is the first study reporting the effectiveness of gene delivery into the hypothalamus using serotype 5 rAAV. To date, eight serotypes of primate adeno-associated virus (AAV) have been identified [35]. Of these, type 2 AAV is best characterised and most employed in gene therapy studies. Type 5 AAV has a distinct advantage over type 2 because of a higher level of transduction efficiency in certain tissues [24, 25]. Our observation of the robust expression of transgene Pomc in the hypothalamus following type 5 rAAV-Pomc delivery suggests that type 5 rAAV is a viable option for long-term gene delivery in the hypothalamus.

The present study provides several distinct sets of salient findings. First, in aged obese rats, rAAV-mediated Pomc gene delivery results in a sustained reduction in body weight despite short-term suppression of food consumption. The decrease in food intake in rAAV-Pomc-treated rats commenced at day 10 after vector injection and abated on day 30 and afterwards. Meanwhile, starting from day 10 after vector administration, rats given rAAV-Pomc consistently lost more body weight when compared with rats given rAAV-control. For the most part, the greatest decrease in body weight occurred during the period when food intake was diminished. Soon after the anorexia attenuated, the average body weight of Pomc-vector-treated rats reached a nadir and then stabilised. Interestingly, food intake returned to precisely the same levels as before treatment. A recent report [36] suggested that melanocortins primarily regulate body adiposity rather than food intake. Based on this concept, the melanocortin-induced anorexic response is adjusted as necessary to achieve a specific level of adiposity. Thus, once a certain level of adiposity concordant with the degree of melanocortin signalling is achieved, as in the present study, the anorexic response wanes. The absence of a rebound in body weight in rAAV-Pomc-treated animals after the attenuation of anorexia also indicates that another component of body weight homeostasis, energy expenditure, must be persistently elevated in response to central Pomc gene therapy. We previously demonstrated that an increase in energy expenditure is sufficient to prevent the regain in body weight following anorexia [37]. In the present study, UCP1 protein levels were significantly elevated 42 days after Pomc vector delivery and 2 weeks after the attenuation of the POMC-mediated inhibition in food intake. UCP1-mediated non-shivering thermogenesis in brown adipose tissue represents an essential element in adaptive energy expenditure in rodents. An increase in UCP1 protein is indicative of increased brown adipose tissue-facilitated energy expenditure [33, 34]. It is known that pharmacological activation of the melanocortin system augments energy expenditure in rodents. For example, normal animals treated with the α-MSH analogue MTII have elevated levels of brown adipose tissue UCP1 expression compared with pair-fed controls [38]. In our previous study, a serotype 2 rAAV-Pomc vector also markedly stimulated brown adipose tissue thermogenesis in obese Zucker rats 38 days after vector delivery [22]. Moreover, transgenic MSH overexpression in lean and obese db/db mice reduced weight gain and adiposity without affecting food intake [39]. Although lacking direct evidence of whole-body energy expenditure, we suggest that, in addition to the hypophagia, an increase in energy expenditure, such as fat oxidation within brown adipose tissue, white adipose tissue or muscle, contributed to the amelioration of body weight and fat in aged obese rats following central Pomc gene therapy and, in particular, was instrumental in maintaining the lost weight after the anorexia attenuated.

Second, although the 19-day anorexic response to Pomc gene delivery attenuated, the onset of this tachyphylaxis to central Pomc gene delivery was markedly delayed compared with attenuation of the anorexic response following pharmacological administration of α-MSH or MTII in either normal or dietary obese mice and rats [20, 21]. The suppression of food consumption lasted no longer than 4 days in any of these latter studies. Additionally, in our recent study, tachyphylaxis to MTII occurred within 6 days after central infusion of MTII in aged obese F344/BN rats [19]. The mechanism of the rapid attenuation to melanocortin treatment in pharmacological studies is not clear, but may involve agonist-mediated receptor internalisation [40]. In contrast to these pharmacological studies, adeno-associated virus-mediated Pomc gene therapy suppressed food intake for up to 38 days in obese Zucker rats [22] and, in the present study, for 19 days in aged obese rats. Therefore, the prolonged anorexic response we observed in aged and obese Zucker rats may be unique to central Pomc gene delivery. Considering that Pomc is expressed at only two locations in the brain–the arcuate nucleus of the hypothalamus and the nucleus of the tractus solitarius of the brainstem–we delivered Pomc vector into the basal medial hypothalamus, aiming at the arcuate nucleus, where Pomc-expressing neurones are located. With this procedure, the overproduction of α-MSH derived from transgene Pomc expression is presumably assisted by a variety of endogenous enzymes, such as prehormone convertases, carboxypeptidase E and peptidyl α-amidating mono-oxygenase [41]. This endogenously regulated production of α-MSH may help prevent the rapid desensitisation witnessed in previous pharmacological studies. The prolonged anorexic response could also be attributed to potential overproduction of other peptides in addition to α-MSH, such as β-MSH and β-endorphin, both of which are normally derived from the POMC precursor. It has been suggested that these peptides also participate in the regulation of energy balance [42, 43]. However, with our approach, the rAAV-mediated overexpression of Pomc in the hypothalamus did result in the transfection of neural cells outside the arcuate nucleus, most notably in the dorsomedial hypothalamic nucleus and lateral hypothalamic area. The potential ectopic expression of Pomc in the brain might account for some of the responses observed.

Central Pomc gene therapy also appears to affect the expression of the two important central melanocortin receptors. When assessed at the time the animals were killed, hypothalamic Mc4r expression was significantly reduced, whereas Mc3r mRNA levels tended to decrease in the rAAV-Pomc rats. Interestingly, the expression of the two orexigenic neuropeptides, Npy and Agrp, did not change after Pomc gene delivery. These data suggest that prolonged exposure to Pomc gene overexpression in aged animals may downregulate hypothalamic melanocortin receptors. This may be one desensitisation mechanism for the attenuation of anorexia evoked by Pomc gene therapy.

Lastly, rAAV-Pomc gene delivery improves glucose metabolism and insulin sensitivity in aged obese rats. Aged F344/BN rats are associated with insulin resistance and glucose intolerance, as shown in several previous reports [15, 18, 44]. In the present study, we also observed impaired glucose tolerance, as indicated by the elevated blood glucose levels after glucose loading in the aged obese F344/BN rats. Central Pomc gene therapy partially normalised glucose levels during IPGTT, and also reduced serum insulin levels markedly at all time points after glucose loading. These data are indicative of improved glucose metabolism and insulin sensitivity by Pomc gene delivery and in agreement with previous findings that central melanocortin receptor activation suppresses insulin release from the pancreas and enhances glucose metabolism [2, 6, 22, 38, 45]. Since glucose metabolism was not significantly improved on day 20 after Pomc gene delivery, when the differences in body weight reduction and visceral adiposity levels were not as large as those on day 36, it suggests that the improvement in glucose metabolism is mainly the consequence of the decreased food consumption and body weight rather than the direct result of central Pomc overexpression. In addition to its impact on insulin and glucose, Pomc gene delivery also reduced serum triglyceride and NEFA levels in the obese aged rats. Such effects could be due to the reduced lipogenesis and/or increased lipolysis in white fat tissues, along with the enhanced non-shivering thermogenesis in brown adipose tissue following Pomc gene therapy.

In conclusion, targeted Pomc gene delivery to the hypothalamus suppressed food intake, diminished body weight and reduced visceral adiposity in aged obese rats. This treatment also improved glucose and fat metabolism and insulin sensitivity. The induced hypophagia and stimulated brown adipose tissue thermogenesis are the likely mechanisms underlying these improvements, and these data suggest that long-term activation of the central melanocortin system may be a viable strategy to combat age-related obesity.