Most published studies on GK rats involve animals of a single age. By using a time series from 4 weeks to 20 weeks, we observe the progression of differences between WKY control and GK animals as diabetes develops. With respect to body weight, a difference is apparent by 8 weeks of age. While not different at 4 weeks, body weights began to diverge at 8 weeks, with WKYs significantly heavier than GKs from 12 weeks onwards. This difference occurs despite the fact that GKs consume more calories per gram body weight than WKYs. Our results show that adipose tissue is a major contributor to differential weight gain in GK rats, as visceral fat accumulation halts by 12 weeks of age, and even declines somewhat at later ages (). It is also significant that epididymal fat showed a similar pattern () suggesting a generalized defect in adipose development. In contrast, our previous work demonstrated that as the animals grow, the liver is larger in the GKs relative to total body weight 
. Although the GK rat is often described as “lean” (i.e., non-obese), our results clearly demonstrate that the GK strain of rats have an age-specific failure to accumulate body fat.
Previously we conducted a time series gene array analysis of the livers from these same animals 
. A functional analysis of the gene expression in the livers from these animals led to two major conclusions. The first was that in the liver, there is chronic natural immune activation. This chronic activation of natural immunity is apparent in both the LPS cascade response to gram negative bacteria and the interferon response cascade to viruses. The second was that in the GK population there was a disruption in lipid metabolism. In the present report we describe a similar functional gene expression analysis of WAT from the animals. The array data for genes involved in the use of energy substrates show that there is divergence between the two populations of animals that starts at 12 weeks of age, the same time when adipose tissue development diverges. For example, after 12 weeks, the expression of fatty acid synthase (Fasn
) and ATP citrate lyase (Acl
) are markedly higher in the WKY population (). Additional relevant genes with almost the exact same temporal pattern of expression include ELOVL family member 6, elongation of long chain fatty acids (Elovl6
), transketolase (Tkt
), butyryl Coenzyme A synthetase 1 (Bucs1
), malic enzyme 1 (Me1
), and uncoupling protein 1 (Ucp1
) (Table S2
). Taken together, the results suggest that adipose tissue in the adult GK population is incapable of carrying out the mature adipocyte function of storing triglycerides and releasing free fatty acids.
The likelihood of impaired adipocyte differentiation in the GK population is reinforced by the expression of several genes involved in transcription and translation. For example, the expression of both MAX gene associated (Mga
) and kruppel-like factor 9 (Klf9
) is higher at all ages in the adipose tissue of the GK population. Max is a heterodimeric partner of MYC, which has been shown to inhibit preadipocyte to adipocyte differentiation 
. Similarly, KLF9 transiently increases during preadipocyte differentiation but decreases with triglyceride accumulation 
. Consistent with the conclusion that in the adipose tissue of the GK population the preadipocyte to mature adipocyte transition is impaired is the observation that the expression of Stat3
is considerably higher at all ages in the adipose tissue of the WKY population. Stat3
is activated by leptin and stimulates adipogenesis 
. Perhaps one of the more intriguing observations relevant to transcription and translation differences between the two populations is the consistently higher expression of trinucleotide repeat containing 6 (Tnrc6
) in WKY animals. This same difference was also observed in the livers from these animals. TNRC6 is a component of the miRNA-containing ribonucleoprotein silencing complexes 
. Recently it has been demonstrated that miRNA mediated gene silencing plays a significant role in adipocyte differentiation 
Our conclusion from both the fat development data and the gene expression analysis is that in the GK population there is impairment in preadipocyte to mature adipocyte differentiation. Consistent with this conclusion is the documented decreased numbers of large mature lipid engorged adipocytes in 3–4 month old GK animals reported previously 
. However, this study only examined the adipocyte population in adult animals, and currently a similar analysis in younger animals, before the growth transition at 12 weeks, would be of value. Likewise, no information is currently available on adipose tissue stem cell populations in these animals. However, it has previously been shown that GK rats have a reduced regenerative ability after partial pacreatectomy, suggesting a reduced stem cell population in that tissue 
However, the data do provide an indication of the cause of the impaired adipocyte differentiation. Although excess adipose tissue (obesity) is thought to be a cause of chronic inflammation because of the attraction of macrophages into the environment of lipid engorged adipocytes, the presence of chronic inflammation actually retards adipocyte differentiation 
. In addition to our previously reported data showing early chronic inflammation in the liver, in the present report we provide data demonstrating the white cell count is elevated in the GK population as early as 4 weeks and remains elevated throughout the experiment (). The gene expression data on WAT similarly indicates early inflammation. About seventy percent of genes that might be considered markers for immune cells were more highly expressed at later ages in the adipose tissue of the WKY population. The higher percentage of immune cell markers at later ages in WKY is consistent with the greater mass of adipose tissue in these control animals. However, a detailed analysis focusing on those genes that reflect inflammation suggests chronic inflammation in the adipose tissue of the GK population. For example, shows that the expression of Cxcl14
is much more highly expressed in adipose tissue of the GK population as early as 4 weeks as compared to the WKY population. CXCl14 is a secreted protein involved in immunoregulatory and inflammatory processes 
and is involved in attracting macrophages into WAT. This gene was also more highly expressed in the livers of these animals at the same age. Two other genes with a relationship to inflammation with early higher expression in the GK population are interferon-induced protein with tetratricopeptide repeats 1 (Ifit1
) and interferon-inducible GTPase (Iigp1
). Interferons are involved in natural immunity. Additional genes associated with inflammation such as the macrophage secreted galectins, Lgals5
similarly suggest heightened inflammation in the GK population (Table S2
). A plausible hypothesis to be formulated from the data as a whole is that early chronic inflammation impairs adipocyte maturation in the GK population. This hypothesis is supported by the earlier observation that in the rat, the number of mature adipocytes expands for about the first 12 weeks of life 
together with our current data demonstrating that 12 weeks is the age when the GKs cease to accumulate WAT.
One potential clue to the immunological differences between the WKY and GK strains is the considerably different expression in both liver and adipose tissue of MHC class II genes Db1, Bb, Ba. The coding sequences of these genes are located very close to each other on rat chromosome 20 
. Polymorphisms amongst rat strains in these genes are associated with a susceptibility to developing a variety of autoimmune diseases in the rat which include type 1 diabetes, thyroditis, arthritis and autoimmune encephalomyelitis 
Previously we provided data that the GK population was transiently hyperinsulinemic and that plasma glucose was elevated at 4 weeks and climbed to a high hyperglycemic plateau by 12 weeks. We chose to maintain our rats in a normal fed state in order to avoid possible variations in physiological measurements induced by fasting. Studies in the literature, particularly on glucose in diabetic rats, differ in whether measurements are taken in the fed state or following varying lengths of fasting (6 hours to 18 hours, or more often a less defined “overnight fast”). Since rats are nocturnal, and approximately 90% of their food consumption occurs in the dark (i.e. overnight) period 
, an overnight fast is actually more equivalent to 24 hours or more, and such animals may be in early phases of starvation 
. For this reason, we maintained our animals in a fed state and collected samples between 1.5 and 3.5 h after lights on. Because blood glucose has a circadian rhythm 
, a measurement at any particular time under a particular condition is only relative. In contrast, HbA1c provides a very good indication of glycemic control over time. shows a continuous increase in glycosylated hemoglobin throughout the 20 week experimental period, demonstrating a chronic disruption of glucose homeostasis in the GK population.
In addition to insulin, glucocorticoids play a major role in regulating glucose homeostasis. Corticosterone which is the primary glucocorticoid in the rat varies with a circadian rhythm driven by light input to suprachiasmic nucleus of the hypothalamus. Plasma corticosterone peaks in the rat shortly after the beginning of the dark period and reaches a low point shortly after the beginning of the light period 
. We measured plasma corticosterone at a time when the plasma concentration should be near its lowest point. In the control WKY population, the plasma corticosterone was relatively stable ranging between 100 and 150 ng/ml. over the 20 week experimental period. Given the constant time of sampling, the relative constancy is expected. In contrast, plasma corticosterone in the GK population was extremely high at 4 weeks (about 600 ng/ml) and declined asymptotically to a low point at 20 weeks of around 200 ng/ml. To put this into perspective, in our previous work on circadian rhythms using Wistar rats obtained from Harlan, the peak plasma concentration of corticosterone measured shortly after the beginning of the dark period was about 300 ng/ml. The results indicate that either the GK population has extremely elevated concentrations of plasma corticosterone or that their circadian rhythms are disturbed. Glucocorticoids enhance gluconeogenesis in the liver and inhibit glycerolgenesis in adipose tissue 
. The rate limiting enzyme for both processes is phosphoenol pyruvate carboxykinase (PEPCK). In some individuals with T2DM hyperglycemia is driven by excessive gluconeogenesis in the liver. In our previous report on the livers from these animals we found elevated expression of Pepck
in the liver only at 4 weeks. In the present study on adipose tissue we found no significant difference between GK and WKY in the expression of Pepck
(data not shown). Given the very large differences in plasma corticosterone between the two populations, the essential lack of difference in expression of Pepck
in both tissues is abnormal. It is relevant that disturbance of circadian rhythms can cause diabetes in a mammal 
and that glucocorticoids cause insulin resistance 
. The possibility of a disrupted circadian rhythm in GK rats is reinforced by the recent report of Bach et al. documenting a disturbance in the pineal melatonin circadian rhythm in these animals 
Adipose tissue is a source of a variety of hormone-like factors (adipokines) that play a role in systemic energy flow, including leptin and adiponectin. Leptin is an adipokine associated with decreased appetite and increased metabolic rate, and plasma concentrations have been shown to increase with increasing fat mass 
. In our study, plasma leptin concentrations in WKY animals increased through 20 weeks, as did adipose mass. In contrast, plasma leptin concentrations in the GK population increased only through 8 weeks, and even decline at 20 weeks. However, despite the fact that leptin output mirrors fat accumulation leptin levels, particularly at early ages, are higher in GK animals. Since GK animals consumed more food per gram body weight than WKYs, this suggests leptin resistance, at least in its effect on appetite suppression. Maekawa and colleagues have also previously reported increased food intake in GK rats, which they demonstrated was associated with central leptin resistance 
. However, because the plasma concentration of leptin, like corticosterone, has a circadian rhythm, interpretation of our results must be qualified. It is relevant that plasma leptin peaks in late dark near the beginning of the light period and reaches a low point in the late light near the beginning of the dark period. Thus, this result may also reflect some disruption of circadian rhythms in the GK population.
In contrast to leptin, plasma concentrations of the insulin sensitizing hormone adiponectin are inversely related to body fat mass 
. In both populations of animals, plasma adiponectin decreases asymptotically with age. What is unusual about these results is that at the younger ages when fat mass in both populations is similar, adiponectin concentrations are modestly higher in the GK population and this trend continues at the older ages when the WKY population has substantially greater body fat. Although adiponectin is normally insulin sensitizing, the GK population is hyperglycemic even at 4 weeks of age.
Gene arrays simply provide a snapshot of the concentration of mRNAs at one point in time. The results reflect the potential available at that time for the production of proteins. When used in a time series as we have done in this experiment, we are able to observe how that potential changes with disease development and progression. In addition, a time series design such as this greatly reduces the number of false positives that might be obtained in array studies if only two groups were compared. In fact, it is possible that our rather stringent filtering criteria may eliminate some genes of potential interest. In addition, because of the heterogeneity of the cells in WAT, we cannot directly determine if a particular message is from a particular cell type. Although we cannot directly ascertain whether steady-state concentrations of message are due to mRNA production or destruction, we can obtain insight from examining messages for proteins involved in functional processes that can either contribute to hyperglycemia or may reflect the impact of hyperglycemia.
Each animal in this study had a unique identifier and all samples are indexed to this identifier. Table S1
, which provides the raw data for each individual probe set identified as differentially regulated in adipose tissue from these animals, also provides an animal code for each individual chip. Chip identification in conjunction with the GEO make all our data completely available to others. It is our intent to provide this code information for all reported measurements in all publications from our ongoing studies, whether they be gene arrays, biochemical or physiological. From a systems biology perspective this will make these studies most useful to other investigators.