Genetics of the Spontaneously Diabetic Torii Rat

The Spontaneously Diabetic Torii (SDT) rat has recently been established as a novel model of nonobese type 2 diabetes. SDT rats exhibit inflammation and fibrosis in and around the islets during development of the disease. To clarify the genetic basis of the disease, we previously performed quantitative trait locus (QTL) analysis of glucose tolerance at 20 weeks of age using backcrossed progeny produced from the (BNSDT)F1SDT cross. The analysis identified three major QTLs (Gisdt1, Gisdt2, and Gisdt3) on rat chromosomes 1, 2, and X, respectively. To examine genetic factors for diabetes, glucose tolerance, islet inflammation, and fibrosis in the SDT rat, we also performed genetic analysis of these traits at 60 weeks of age using intercrossed progeny produced from the (F344SDT)F2 cross. Genetic analysis of diabetes identified a major locus, Dmsdt1, on chromosome 3. QTL analysis of blood glucose levels revealed, in addition to Dmsdt1, three other loci (Dmsdt2, Dmsdt3, and Dmsdt4) on chromosome 8, 13, and 14, respectively. Analysis of a congenic strain for Dmsdt1 (F344.SDT-Dmsdt1) indicates that the dominantly acting SDT allele induces islet inflammation and fibrosis. These data clearly demonstrate that development of diabetes in the SDT rat is controlled by the combination of several QTLs with considerable effects. Identification of the genes responsible would provide greater understanding of the pathogenesis and pathophysiology of diabetes.


INTRODUCTION
Type 2 diabetes mellitus is a multifactorial disease involving interaction of multiple genetic and environmental factors.Since the genetic basis of type 2 diabetes has not been fully clarified yet, analysis using spontaneous animal models would provide useful information on genetic factors involved in the development of the disease.There have been many obese models of type 2 diabetes such as the Otsuka Long-Evans Tokushima fatty (OLETF) rat [1], Wistar fatty rat [2], Zucker diabetic fatty (ZDF) rat [3], db/db mouse [4], KK-A y mouse [5], Nagoya Shibata Yasuda (NSY) mouse [6], New Zealand obese (NZO) mouse [7], TallyHo (TH) mouse [8], and the Tsumura Suzuki obese diabetic (TSOD) mouse [9], but there is no known model of type 2 diabetes without obesity except for the Goto-Kakizaki (GK) rat [10].The GK rat has, therefore, been exclusively utilized as a nonobese model of the disease.
The Spontaneously Diabetic Torii (SDT) rat was recently established as a novel nonobese model of type 2 diabetes [11].Male SDT rats show 100% incidence of diabetes by 40 weeks of age.Before the onset of diabetes, pathological changes such as inflammation and fibrosis in and around the islets continue for several months, and are accompanied by a decrease in the number of pancreatic -cells [12].
To clarify the genetic basis of diabetes, glucose intolerance, islet inflammation, and fibrosis in SDT rats, we performed genetic analysis of these traits using two different crosses: the (BN SDT) F1 SDT backcross [13] and the (F344 SDT) F2 intercross [14].Analysis of the former cross identified three QTLs (Gisdt1, Gisdt2, and Gisdt3) for glucose intolerance on rat chromosomes 1, 2, and X, respectively.Analysis of the latter cross revealed four QTLs (Dmsdt1, Dmsdt2, Dmsdt3, and Dmsdt4) for blood glucose levels on rat chromosome 3, 8, 13, and 14, respectively.Further congenic analysis clarified Dmsdt1 as a major locus for islet inflammation and fibrosis.

GENETIC ANALYSIS OF GLUCOSE INTOLERANCE USING (BN SDT) F1 SDT BACKCROSSED PROGENY
To investigate genetic control of diabetes and glucose intolerance in SDT rats, we produced (BN SDT)F1 SDT backcrossed (N2) progeny, performed oral glucose tolerance test (OGTT) at 20 weeks of age, and phenotyped for the onset of diabetes up to 25 weeks of age [13].The cumulative incidence of diabetes in N2 rats was only 1.9% (6/319) at 25 weeks of age.
By selective genotyping on N2 rats with marked glucose intolerance (n = 26) and normal glucose tolerance (n = 30), followed by chi-square test, we found significant differences (P < 0.01) in the genotype frequencies at several markers on chromosomes 1, 2, 6, 7, 8, 11, 14, 18, and X.These chromosomes were then subjected to QTL analysis using all of the N2 rats.We found three major QTLs affecting postprandial blood glucose levels on chromosomes 1, 2, and X, designated Gisdt1 (glucose intolerance in SDT rat 1), Gisdt2, and Gisdt3, respectively.These loci have been registered as Gluco13, Gluco14, and Gluco15, respectively (Rat Genome Database, RGD).We also found a major QTL for body weight at the Gisdt1 region.The recessively acting SDT allele at Gisdt1 (Gluco13) on chromosome 1 was involved in the higher blood glucose levels and higher body weight in N2 rats.The recessively acting SDT allele at Gisdt2 (Gluco14) on chromosome 2 and the single SDT allele at Gisdt3 (Gluco15) on X chromosome were involved in the higher blood glucose levels in N2 rats.The genetic analysis using N2 rats revealed that there are at least three major QTLs affecting postprandial blood glucose levels in SDT rats.

GENETIC ANALYSIS OF DIABETES AND GLUCOSE INTOLERANCE USING (F344 SDT)F2 INTERCROSS-ED PROGENY
To perform genetic analysis of diabetes and glucose intolerance in SDT rats, we produced (F344 SDT) F2 rats, phenotyped for the onset of diabetes up to 60 weeks of age, and performed OGTT at the same age [14].F2 rats developed diabetes at 25 weeks of age or later, and the cumulative incidence of diabetes reached 19% (31/167) at 60 weeks of age.Interestingly, postprandial blood glucose levels of (F344 SDT) F1 rats were significantly higher than those of control F344 rats, indicating that dominantly acting SDT alleles are involved.Although none of the F1 rats developed diabetes till 75 weeks of age, F1 rats did show pathological inflammatory changes in the pancreas, such as inflammation and fibrosis in and around the pancreatic islets.
By selective genotyping on diabetic F2 rats (n = 31) and non-diabetic F2 rats with the area under the blood glucose response curve value of the lowest 20% (n = 33), followed by chi-square test, we found highly significant differences (P < 0.001) in the genotype frequencies at several markers on chromosomes 3, suggesting the presence of a major locus responsible for diabetes.The locus was designated Dmsdt1 (diabetes mellitus locus in the SDT rat No.1), and has been registered as Gluco35 (RGD).A dominantly acting SDT allele at Dmsdt1 is involved in the development of diabetes in F2 rats.
By QTL analysis using all of the F2 rats, in addition to Dmsdt1 we identified three major QTLs affecting blood glucose levels on chromosome 8, 13, and 14, designated Dmsdt2, Dmsdt3, and Dmsdt4, respectively, and one major QTL affecting body weight on chromosome 3, designated Bwsdt1.These loci have been registered as Gluco36, Gluco37, Gluco38, and Bw82, respectively (RGD).The dominantly acting SDT allele at Dmsdt1 (Gluco35) on chromosome 3 was involved in the higher postprandial blood glucose levels in F2 rats.The recessively acting SDT allele at Dmsdt2 (Gluco36) on chromosome 8 was involved in the higher fasting and postprandial blood glucose levels in F2 rats.The recessively acting SDT allele at Dmsdt3 (Gluco37) on chromosome 13 was involved in the higher fasting blood glucose levels in F2 rats.The additively acting SDT allele at Dmsdt4 (Gluco38) on chromosome 14 was involved in the higher postprandial blood glucose levels in F2 rats.The dominantly acting SDT allele at Bwsdt1 (Bw82) on chromosome 3 was involved in the higher body weight in F2 rats.
By QTL analysis on insulin levels at OGTT, we identified three significant QTLs on chromosome 3, 13, and 14 (Fig. 1).The LOD score plot suggested that the loci on chromosome 3 and 14 correspond to Dmsdt1 (Gluco35) and Dmsdt4 (Gluco38), respectively, while the locus on chromosome 13 is different from Dmsdt3 (Gluco37).
To clarify the role of the dominantly acting SDT allele at Dmsdt1 (Gluco35) on the development of diabetes, we produced a congenic strain carrying an SDT allele on the control F344 genetic background (F344.SDT-Dmsdt1) [14].At 35 weeks of age, inflammation in and around the pancreatic islets and hemosiderin deposition and fibrosis in the islets were frequently found in congenic rats, indicating that a dominantly acting SDT allele at Dmsdt1 (Gluco35) induces the pathological inflammatory changes in the pancreas.
Genetic analysis using F2 rats clarified that at least four major QTLs are involved in impaired regulation of blood glucose levels in SDT rats.These loci are completely different from those identified by the analysis using N2 rats, except for Dmsdt2 (Gluco36), the recessively acting locus on chromosome 8.These data indicates Dmsdt2 (Gluco36) to be a common genetic factor to the two different crosses, but does not indicate the locus is the most definite one.The inconsistency between the two studies is likely due to differences in crossing, control strains, and age of phenotyping.It is much more straightforward and efficient to analyze complex genetic traits through a backcross (N2) rather than an F2 intercross, since each backcrossed (N2) animal have one of two genotypes at each locus, while each F2 animal can have one of three genotypes at each locus.Although the backcross (N2) analysis is concentrated on detecting recessively-acting genetic factors, the analysis using F2 cross makes it possible to detect both recessivelyand dominantly-acting genetic factors.

COMPARISON WITH REPORTED QTLS IN RATS, MICE, AND HUMANS
There have been reported hundreds of QTLs affecting blood glucose levels or body weights in rats, mice, and humans.QTLs mapped to the rat genomic regions harboring SDT QTLs, or those mapped to the orthologous regions in mice and humans are listed (Table 1).Several QTLs, such as Niddm40 and Niddm45 affecting postprandial glucose levels in OLETF rats [15,16] have been mapped to the Gisdt1 (Gluco13) region on rat chromosome 1.In the Gisdt2 (Gluco14) region on chromosome 2, no QTL has been reported in rats and mice.On chromosome X, Niddm16 affecting glucose levels in OLETF rats has been reported [15,17] on the genomic region overlapped with Gisdt3 (Gluco15).Dmsdt1 (Gluco35) induces inflammation and fibrosis in the pancreatic islets that are the main pathogenic features of SDT rats, while uncommon in human type 2 diabetes.Although Gluco39 affecting glucose levels in Wistar Ottawa Karlsburg W (WOKW) rats has been reported [18] in the Dmsdt1 (Gluco35) region on chromosome 3, no such loci has been reported in the homologous region in humans.Dmsdt2 (Gluco33) on chromosome 3 is involved in both fasting and postprandial glucose levels in SDT rats.Interestingly, Gluco43 affecting fasting glucose levels in Wistar Kyoto (WKY) rats [19] and Niddm21 affecting postprandial glucose levels in OLETF rats [20] have been mapped to this region.No QTL has been reported in rats and mice mapped to the Dmsdt3 (Gluco37) region on chromosome 13.In the Dmsdt4 (Gluco38) region on chromosome 14, Niddm20 affecting postprandial glucose levels in OLETF rats has been reported [21].
In the Bwsdt1 (Bw82) region on chromosome 3, several QTLs affecting body weights have been reported, such as Bw24 in Dahl Salt Sensitive (SS) rats [22], Bw31 and Bw36 in Lyon Hypertensive (LH) rats [23], and Bw81 in Genetically Hypertensive (GH) rats [24].There are several QTLs thought to be common for SDT and OLETF rats, while there seems to be no such QTL for SDT and GK rats [25,26].To our surprise, the loss of function polymorphism in Prlhr (prolactin releasing hormone receptor, also known as Gpr10) affecting hyperphagia, dyslipidaemia, and obesity in OLETF rats [27] has been found in SDT and GK rats (Yokoi N, unpublished observation).In fact, there was a significant association with body weight around the Gpr10 region on chromosome 1 [13].Although SDT rat is a nonobese model of type 2 diabetes, SDT alleles at several other loci have also been involved in the higher body weight in the N2 or F2 rats.Since SDT rats showed larger body size as compared to BN or F344 rats which were used as control strains in the genetic studies [13,14], the SDT alleles affecting body weight may well be involved in large body size rather than obesity.These data indicate that large body size may have some role on the phenotype of SDT rats.

CONCLUSION
So far, we have identified seven QTLs (Gisdt1, Gisdt2, Gisdt3, Dmsdt1, Dmsdt2, Dmsdt3, and Dmsdt4) affecting blood glucose levels on chromosome 1, 2, X, 3, 8, 13, and 14, respectively, and one QTL Bwsdt1 affecting body weight on chromosome 3 in SDT rats.Among them, Dmsdt1 has been found to be involved in islet inflammation and fibrosis, the major pathogenic features of this model.Since retinal inflammation is among the early pathological changes linked causally to the development of retinopathy, Dmsdt1 might also be involved in the pathogenesis of retinopathy, the major diabetic complication of SDT rats.At present, there is no data which suggests the relation of other QTLs to retinopathy.Further characterization of Dmsdt1 and other QTLs are needed to identify the genes responsible and to clarify the mechanism of pathogenesis of diabetes in SDT rats.