In this manuscript we provide data from human, mouse and rat studies that suggest that RGS9-2, a brain specific RGS family member, can regulate body weight and adiposity. First we link an intronic deletion in the RGS9 gene to increased body mass index in humans (BMI). We provide mechanistic insights into this association by showing i) that the intronic deletion is coincident with a binding motif for the polypyrimidine tract binding protein (PTB), a protein involved in regulating the processing of RNA and ii) that the intronic variation can alter the splicing of the RGS9 gene product. Further confirmation for a role of RGS9 in regulating body weight is provided by observations showing that RGS9 knockout mice are heavier than their wild-type littermates, have increased adiposity and adipocytes with approximately doubled cross-sectional area. Finally we identify a possible site for the action of RGS9-2 in regulating body-weight by showing that overexpression of RGS9-2 in the nucleus accumbens of rats can decrease body-weight.
A major finding of our study is that an intronic deletion polymorphism in the RGS9 gene (ΔTTTCT
) is associated with significantly increased body mass index (BMI) in humans. The increase in mean BMI (24.9 versus 24.2) represents an increase in weight from 68.2 kg (150 lb) to 70 kg (154 lb), or a 2.7% increase in weight, for an individual 168 cm (5 ft 6 inch) tall. Thus, the influence of the ΔTTTCT
deletion polymorphism on body weight is quite mild: a World Health Organization panel has suggested that the obesity cut-off definition for Asians should be lowered to a BMI of 25 
, compared to 35 for Caucasians, but the mean BMI of the ΔTTTCT
deletion positive East Asian individuals in our study is less than even this lowered BMI value.
An examination of shows that the increase in mean BMI of the ΔTTTCT
deletion positive subjects relative to the deletion negative individuals is greatest among the Chinese, is very small in Pacific Islanders and is reversed in Southeast Asians. However, the above differences and the differences in BMI between deletion positive and negative individuals within each ethnic sub-group were not statistically significant. Nevertheless, it is interesting to note that since Pacific Islanders (Polynesians) are descendants of voyagers from Southeast Asia—either Taiwanese aborigines or former inhabitants of New Guinea and the surrounding islands 
, they are more closely related genetically to Southeast Asians than to Chinese, Japanese or Koreans. Hence, it is possible that within these two former genetically similar ethnic groups, the ΔTTTCT
deletion is not correlated to BMI, or the effect of the deletion is masked by other factors. Future studies with larger samples sizes will be required to determine whether the effect of ΔTTTCT
deletion on BMI is restricted to a subset of ethnic groups.
The association between the ΔTTTCT intronic deletion polymorphism in the RGS9 gene and increased BMI was studied in the East Asian population because the deletion was rare in other groups. However, these results do not imply that RGS9-2 contributes to body weight set-points only in this ethnic group. Other RGS9-related variations, including yet unidentified variations in the RGS9 gene and regulatory regions, may contribute to variations in cellular levels of RGS9-2 protein and thus modify body-weight in individuals regardless of ethnicity.
It is also interesting to note that the mean weights of the female wild-type and knockout mice diverged significantly from each other at an earlier time point than for the male mice (day 30 after weaning for the females versus day 40 after weaning for the males). In addition, close to the end of the experiment, on day 50, the mean percent increase in weight of the female knockout mice compared to the female wild-type mice was 16.5% while the mean percent increase for the male knockout mice relative to the wild-type mice was 6.8%, but this difference was not statistically significant. Similarly, the difference in mean BMI of the ΔTTTCT
deletion positive relative to the deletion negative East Asian individuals was greater in females than in males () and the difference between the genotypes was marginally significant in females but not in males (p
0.07 for females versus p
0.45 for males, t test). Future studies with larger samples sizes will be required to determine whether the effect of RGS9
gene variations on weight is more prominent in females.
The location of the ΔTTTCT
deletion polymorphism and the identity to a binding motif for the polypyrimidine tract binding protein (PTB) suggested that this naturally occurring allele could affect the splicing of RGS9 mRNA. Indeed, our data shows that the deletion can alter splicing and produces a substantial reduction in correctly-spliced RGS9 gene transcript (). The ΔTTTCT
polymorphism lies within intron 13. The exons that flank intron 13 code Exons 13 and 14 which encode the RGS domain 
. The RGS domain is conserved among all the RGS family members and mediates the canonical Gα GTPase accelerating (GAP) function 
. Thus, altered splicing due to the ΔTTTCT
deletion polymorphism, such as exon skipping or intron retention, will result in a protein product that is truncated at this region and therefore lacks the important class-defining GAP function. These results, in turn, suggest that individuals carrying the deletion polymorphism have a higher BMI due to a reduction in the functional levels of brain RGS9-2.
Support for the suggestion that RGS9-2 is important in regulating body-weight is provided by the finding that mice with the RGS9
gene deletion have elevated body weight () and that conversely overexpressing RGS9-2 in the rat nucleus accumbens (NAc), via herpes simplex virus (HSV)-mediated gene transduction, lowers body weight relative to control animals (). RGS9-2 is expressed in the vast majority of striatal medium spiny neurons 
, which comprise 90–95% of the neurons in the striatum 
. Hence, the percentage of infected non-target cells that do not normally express RGS9-2 is small and consequently, this technique has been successfully utilized to define physiological and pathophysiological functions for RGS9-2 using both rodent and primate models 
Indeed, the results presented above mirror those from previous studies, where HSV-mediated RGS9-2 over-expression in the rat striatum has been shown to produce functional responses that are opposite to those exhibited by the RGS9
knockout mice. For example, RGS9
knockout mice exhibit increased cocaine-induced locomotion, while HSV-mediated overexpression of RGS9-2 in rats dampens cocaine-induced locomotion 
. In addition, RGS9
knockout mice show accelerated development of drug-induced dyskinesia 
, while HSV-mediated overexpression of RGS9-2 in the striatum of rats and monkeys diminished intensity of drug-induced dyskinesia 
The specificity of RGS9-2 overexpression effects are highlighted by the parallel experiments with RGS7 and RGS11 
, closely related members of the R7 RGS protein family. Though RGS11 is thought to be specifically expressed in retinal bipolar neuron, the two proteins have been shown in vitro
to act as GTPase accelerating proteins for the same G proteins 
. The opposite effects of RGS9-2 and RGS7 and RGS11 overexpression on body weight () suggest that the effects cannot be attributed solely to the GAP function which is common to the three proteins. While we do not yet understand the mechanism for the opposing action of these proteins on body weight, it is interesting to note that R7 RGS proteins, such as RGS7, RGS9 and RGS11, can compete for their obligate binding partners (Gβ5 and R9AP or R7BP). In fact, knock-out of these binding partners leads to marked reduction in all R7 RGS proteins 
. Therefore, one hypothesis explaining the relative increase in weight produced by RGS7 or RGS11 overexpression, is that RGS7 and RGS11 competes with RGS9-2 for the available Gβ5 and R7BP, destabilizing native RGS9-2 protein and producing an effect on weight similar to that seen in RGS9
knock-out mice. Results similar to those we report have been observed at the cellular level: previous studies have reported that, despite their structural similarity, RGS7 and RGS11 produce opposite effects to those produced by RGS9 on cellular signaling pathways 
While there is one report of RGS9 transcript expression in blood lymphocytes 
, all other studies examining the tissue distribution and function of RGS9 gene products have reported only on the expression of RGS9 in the brain and the retina 
. Thus, the absence of reports of RGS9 expression in peripheral tissue, interpreted in conjunction with, i) our observations with rats demonstrating that RGS9-2 overexpression in the NAc lowers body weight compared to control animals () and ii) undetectable levels of RGS9 protein expression in fat tissue (), suggest that the RGS9 acts to regulate body-weight and adiposity via expression in the brain.
RGS9-2 exhibits extremely dense expression in rodent striatum, and pleasure, desire and reward circuits operating in the striatum are thought to be important in obesity and eating disorders 
. However, the actions of RGS9-2 in regulating body-weight and adiposity are likely to involve alternative striatal connections, since we report that the food intake of the heavier RGS9
knockout mice was similar to that of wild-type mice.
The dopamine-sensitive NAc reward centers are densely interconnected with the hypothalamus, a brain region that is critical for regulating energy expenditure. Hypothalamic neurons, including orexin 
and melanin concentrating hormone neurons 
, make connections with the NAc, and the NAc can influence hypothalamic functions 
. Thus, RGS9-2 via expression in the NAc could modulate the activity of hypothalamic centers that control energy expenditure and future experiments will examine the total and resting oxygen consumption in the RGS9
knockout mice. Interestingly, in addition to extremely dense expression in striatum, RGS9-2 is also localized to medial hypothalamus 
. However, it is clear from our experiments that modulating RGS9-2 levels specifically in the NAc can alter body weight ().
Human genetic studies have linked variations in two other R7 RGS proteins, RGS6 
and RGS7 
to obesity. The above human genetic studies provide a precedent for R7 RGS family proteins regulating body-weight through their expression in the brain since these proteins have been reliably detected only in the brain or in excitable tissue such as retina and heart 
More recently, it has been reported that mice with a targeted deletion of one copy of the G protein Gβ5 subunit gene are heavier and have increased adiposity when compared to their wild-type counterparts, even though their food intake was not different and their locomotor activity levels were enhanced 
. Similarly, we found that the food intake of the heavier RGS9 knockout mice was not different from their wild-type littermates, and also observed a small but significant increase in initial activity of the male RGS9
knockout mice when placed in the movement measurement cages (). Gβ5 protects R7 RGS family proteins from proteolysis and expression of R7 RGS family proteins, including RGS9, is eliminated or hugely reduced in the absence of Gβ5 
. Therefore our results raise the possibility that the increased adiposity and weight phenotype of heterozygous Gβ5 knockout mice is produced as a result of a Gβ5 knockdown-mediated reduction in brain RGS9-2 protein levels.
Human and mouse studies have also implicated RGS2, 4 and 5, in the regulation of body weight and obesity 
. However these RGS proteins and RGS9 likely regulate body weight through different mechanisms: RGS2, 4 and 5 are members of a different R4 subfamily of RGS proteins, have a different molecular architecture and strikingly different cellular and tissue expression patterns compared to RGS9 
. It has been reported that knock-in mice homozygous for a mutant Gαi2
G protein subunit that does not bind RGS protein are resistance to diet-induced obesity 
. However, the molecular mechanism underlying the body-weight phenotype of these mice is also likely to be different from that in the RGS9
knockout mice: Gαi2, is expressed ubiquitously in peripheral tissues and in many brain regions and does not have the restricted expression pattern of the RGS9
gene products 
Previously, altered RGS9-2 levels have been shown to be involved in drug-addiction and the reward response to psychostimulants 
and our results are significant because we implicate the same molecule in the control of body weight and adiposity.