Comparative transcriptome and mutation analyses of the pancreatic islets of a rat model of obese type 2 diabetes identifies a frequently distributed nonsense mutation in the lipocalin 2 gene

Abstract

Type 2 diabetes (T2D) is a multifactorial disease caused by insulin resistance and impaired insulin secretion from pancreatic β-cells, but the precise mechanisms remain to be elucidated. To identify primary genetic factors of T2D in a rat model, we performed comparative transcriptome and mutation analyses of the pancreatic islets between the obese Zucker fatty rat and the Zucker fatty rat-derived T2D model Zucker fatty diabetes mellitus (ZFDM) rat. Among differentially expressed genes irrespective of obesity and glucose intolerance states, we identified a nonsense mutation, c.409C > T (p.Gln137X), in the lipocalin 2 (Lcn2) gene which encodes a secreted protein called neutrophil gelatinase-associated lipocalin, a well-known biomarker for inflammation. We examined the relevance of the Lcn2 mutation with T2D in the ZFDM rat by using genome editing and genetic linkage analysis and confirmed that the Lcn2 mutation exhibits no significant association with the onset of T2D. Interestingly, we found that the Lcn2 mutation is distributed widely in rat species, such as commonly used DA and F344 strains. Our data indicate that several rat strains would serve as Lcn2 deficient models, contributing to elucidate the pathophysiological roles of Lcn2 in a wide variety of phenotypes.

Introduction

Type 2 diabetes (T2D) is a multifactorial disease caused by insulin resistance and impaired insulin secretion from pancreatic β-cells; β-cell failure is a central element in the development and progression of T2D. The β-cell failure in T2D is a complex process involving multiple factors, such as insulin resistance and β-cell overload, lipotoxicity, chronic inflammation, oxidative stress, endoplasmic reticulum stress, and genetic and epigenetic factors. However, the precise mechanisms remain to be elucidated.

Transcriptome analysis of the pancreatic islets offers a powerful technique not only to reveal the pathophysiological state of β-cells but also to identify candidate genes in T2D. There have been several studies to identify candidate genes for T2D by transcriptome analyses in pancreatic islets from individuals with T2D and nondiabetic controls. PAX5 has been reported as a potential transcriptional regulator of many T2D-associated differentially expressed genes (DEGs) in human islets.¹ Transcriptome profiling in db/db mice revealed that a reduction in the expression of Glut2, Ins1, Ins2, MafA, Mt1, and Pdx1 was indicative of dedifferentiation in db/db islets.² The Zucker diabetic fatty (ZDF) rat, a model of obese T2D, carries a genetic defect in β-cell gene transcription, in which insulin promoter activity and insulin gene expression were reduced even in lean animals.³ They concluded that the genetic reduction in β-cell gene transcription in ZDF rats likely contributes to the development of diabetes in the setting of insulin resistance. Transcriptome profiling of ZDF rats showed an increase in the genes encoding proteases and extracellular matrix components that are associated with tissue remodelling and fibrosis.⁴ In addition, ZDF rats showed an increase in vascular endothelial growth factor-A and Thrombospondin-1 genes, suggesting that an inability of the islet to maintain vascular integrity may contribute to β-cell failure.⁵ However, the primary genetic factors have not been clarified yet.

Among animal models of T2D, the Zucker fatty diabetes mellitus (ZFDM) rat has been derived from the obese Zucker fatty (ZF) rat harbouring a missense mutation (fatty, fa) in the leptin receptor (Lepr) gene.⁶ Animals homozygous for the fatty mutation in both ZF and ZFDM strains exhibit obesity, whereas only male ZFDM rats develop T2D accompanied with histopathological changes in the pancreatic islets such as loss of islet structure and β-cell destruction.^6,7 In spite of the same origin, there is a significant difference in genetic profiles between the 2 strains⁸; genetic factors involved in the development of T2D remain unknown.

We have previously performed transcriptome analysis of the pancreatic islets in ZFDM rats to examine the mechanism underlying functional differences between non-large and enlarged islets. Together with the insulin secretion experiment and metabolome analysis, we found that enlarged pancreatic islets show tumour cell-like metabolic features of glucose metabolism accompanied with reduced β-cell specific gene expressions and glutamate production, which could contribute to the development of incretin unresponsiveness in obese T2D.⁹ However, the pathogenesis of dysfunction of the pancreatic islets and genetic factors involved in the development of T2D still need to be clarified.

In the present study, to elucidate the gene expression profile of the pancreatic islets and primary genetic factors of T2D in the ZFDM rat, we performed comparative transcriptome and mutation analyses on the pancreatic islets between ZF and ZFDM rats. Among DEGs, we identified a nonsense mutation in a strong candidate gene for T2D, which is found to be existed frequently in rat species. We also evaluated the relevance of the mutation with T2D by both genome editing and genetic linkage analysis.

Materials and methods

Ethics approval and consent to participate

All animal experiments were approved by the Committee on Animal Experimentation of Kobe University and Kyoto University and carried out in accordance with the Guidelines for Animal Experimentation at Kobe University and Kyoto University.

Animals

For isolation of the pancreatic islets and genetic linkage analysis, male ZF rats (Slc:Zucker-Lepr^fa/fa and -Lepr^+/+) were purchased from Japan SLC, Inc. and male ZFDM rats (Hos:ZFDM-Lepr^fa/fa and -Lepr^fa/+) were provided by Hoshino Laboratory Animals, Inc. All animals were maintained under specific pathogen-free conditions with a 12 h light-dark cycle and were provided with a commercial diet CE-2 (CLEA Japan, Inc.) at the Animal Facility of Kobe Biotechnology Research and Human Resource Development Center of Kobe University. At the end of the experiments, animals were sacrificed by overdose of anaesthesia with pentobarbital sodium (2018 or before).

For genome editing experiment, female and male SD rats (Slc:SD) were purchased from Japan SLC, Inc. and female ZFDM rats (Hos:ZFDM-Lepr^fa/+) and male ZFDM rats (Hos:ZFDM-Lepr^fa/fa and -Lepr^fa/+) were provided by Hoshino Laboratory Animals, Inc. All animals were maintained under specific pathogen-free conditions with a 14 h-light and 10 h-dark cycle and were provided with a commercial diet F-2 (Oriental Yeast Co., Ltd.) at the Institute of Laboratory Animals, Graduate School of Medicine, Kyoto University. At the end of the experiments, animals were sacrificed by carbon dioxide inhalation (2019 or later).

Isolation of the pancreatic islets

Pancreatic islets were isolated by the collagenase digestion and Ficoll gradient method.^10,11 Isolated pancreatic islets were cultured for 3 days in RPMI1640 (Sigma-Aldrich) before experiments.

RNA sequencing and data analysis

Total RNA was extracted from the pancreatic islets using RNeasy Micro kit (Qiagen). RNA sequencing (125 bp paired-end) was performed on 1 μg each of total RNA, using an Illumina HiSeq 2500 system by Eurofins Genomics. Sequence reads were cleaned using trimmomatic (ver. 0.39),¹² the quality were checked using FastQC (version 0.11.9),¹³ and then aligned to the rat genome (mRatBN7.2) using STAR (version 2.7.10).¹⁴ Data were transformed into BAM format using Samtools (version 1.15.1),¹⁵ and raw read counts were calculated using featureCounts (version 2.0.3).¹⁶ The following analyses were performed using R (version 4.1.3) (https://www.r-project.org/). Filtering low expression genes, TMM (trimmed mean of M values) normalization, and extraction of DEGs were done using edgeR (version 3.40.1).^17,18 DEGs were extracted using a quasi-likelihood F-test of edgeR with a threshold of FDR (false discovery rate) < 0.01 and fold change > 2. Information of DEGs was obtained using biomaRt (version 2.54.0).^19,20 Gene Ontology (GO) and pathway analyses were performed using Database for Annotation, Visualization, and Integrated Discovery (DAVID).^21,22 GO analysis was also performed using Metascape.²³ The RNA sequencing data have been deposited in DDBJ Sequence Read Archive (DRA) with the accession numbers DRA007109,⁹ DRA007371,⁹ and DRA012479.

Mutation analysis using RNA sequencing data

Single nucleotide polymorphisms (SNPs) between ZF and ZFDM rats were detected from BAM files of RNA sequencing data of the pancreatic islets at 8 weeks of age using the Genome Analysis Toolkit (GATK) (version 4.2.6.1)²⁴ with the GATK Best Practice.^25,26 Resulting Variant Calling Format (VCF) files were summarized using VCFtools (version 0.1.16).²⁷ Variants were annotated using Variant Effect Predictor (VEP) (version 108).²⁸ SNPs were defined as those fixed for distinct homozygous states between ZF and ZFDM rats.

Mutation screening of the lipocalin 2 gene

Genome DNA was extracted from tail tip samples using the Wizard SV Genomic DNA Purification System (Promega). All amino acids coding regions in the lipocalin 2 gene of ZF and ZFDM rats were sequenced by the Sanger method. A polymerase chain reaction (PCR)-restriction fragment length polymorphism (RFLP) system was developed for genotyping of the Q137X nonsense mutation found in ZFDM rats: PCR product (669 bp) amplified by using a primer set (FW, 5′-aaccctgggtatgacctgaa-3′; RV, 5′-ctggggcctggattattgta-3′) is digested by restriction enzyme XspI (CTAG), resulting in 464 bp, 114 bp, and 91 bp fragments in ZF rats (wildtype allele) while the 464 bp fragment is further divided into 347 bp and 117 bp fragments in ZFDM rats (mutant allele). The PCR-RFLP system was also applied for genotyping of the mutation among rat species. The genome DNA of 157 inbred rat strains (Supplementary Table S13) were provided by the National Bioresource Project-Rat (NBRP-Rat), Kyoto University (Kyoto, Japan). The genome DNA of the SDT/Jcl rat has been obtained previously.²⁹ The ZDF-Lepr^fa/CrlCrlj rat was purchased from the Jackson Laboratory Japan, Inc. and genome DNA was extracted.

CRISPR/Cas9-mediated genome editing in ZFDM rats

Lcn2 knock-in ZFDM rats, in which the Q137X nonsense mutation was replaced with a wild-type nucleotide, were generated by CRISPR /Cas9-mediated genome editing as described previously with some modification.³⁰ Briefly, female ZFDM fa/+ rats were superovulated by intraperitoneally injection with 15 IU (International Unit) of pregnant mare serum gonadotropin (NIPPON ZENYAKU KOGYO Co., Ltd.) and 15 IU of human chorionic gonadotropin (ASKA Pharmaceuticals Co., Ltd.), and then the female rats were mated with male ZFDM (fa/+ or fa/fa) rats. The next day, pronuclear-stage embryos were collected from superovulated rats and cultured in a modified Krebs-Ringer bicarbonate (m-KRB) culture medium before microinjection. The Lcn2 target sequence (5′-TGACTACGACTAGTTTGCCA-3′; Fig. 3A) was designed using CRISPOR (http://crispor.gi.ucsc.edu). The recombinant Cas9 protein and crRNA and tracrRNA were purchased from Integrated DNA Technologies (Coralville, Iowa, USA). The chemically synthesized single-strand oligo-DNA (ssODN; 5′-AAGTGGCCGACACTGACTACGACCAGTTTGCCATGGTATTTTTCCAGAAGACCTCTGAAA-3′, Exigen) was used for replacing the nonsense mutation with the wild-type allele on the Lcn2 locus. The recombinant Cas9 protein (50 ng/mL), chemically synthesized crRNA (25 ng/mL) and tracrRNA (25 ng/mL), and the ssODN (100 ng/μL) were co-injected into the cytoplasm of pronuclear stage embryos. The injected embryos were cultured in m-KRB culture medium overnight. The 2-cell embryos were transferred into the oviduct of pseudopregnant SD rats anaesthetized using a mixture of 0.15 mg/kg medetomidine, 2.0 mg/kg midazolam, and 2.5 mg/kg butorphanol during operation. After the surgery, 0.15 mg/kg atipamezole (NIPPON ZENYAKU KOGYO Co., Ltd.) and 5 mg/kg enrofloxacin (Kyoritsu Seiyaku) were administered. Analgesics (0.01 mg/kg buprenorphine; Otsuka Pharmaceutical Co. Ltd.) were administered subcutaneously twice on the day of surgery and the following day.

To confirm the genome-edited alleles, the Sanger sequencing analysis was performed on the target locus. The PCR products were obtained from tail tip genome DNAs by using the primer set (FW, 5′-aaccctgggtatgacctgaa-3′; RV, 5′-ctggggcctggattattgta-3′), and sequenced with each of the primers. Two female founder rats (G0-28 and -29) heterozygous for the genome-edited allele were obtained and were mated with male ZFDM rats to produce offsprings. The inheritance of the genome-edited allele was confirmed in the offsprings. Both lines were maintained for several generations to produce fa/fa animals homozygous or heterozygous for the genome-edited allele and those homozygous for the original mutant allele. Finally, the genome-edited allele was fixed in the homozygous state. The G0-28- and -29-derived genome-edited ZFDM rats, ZFDM-Lcn2^em1Nyo, and -Lcn2^em2Nyo, were deposited to the NBRP-Rat under deposition No.0972 and No.0973, respectively.

Phenotyping

The rats were checked for body weights and nonfasting blood glucose levels by a portable glucose metre (ANTSENSE Duo, HORIBA, Ltd.). Diabetes was defined as blood glucose levels equal to or higher than 300 mg/dL for 3 consecutive weeks under ad libitum dietary conditions. The week of the diabetes onset was defined as the first week in which blood glucose levels were equal to or higher than 300 mg/dL.

Genotyping of the fatty mutation in the leptin receptor gene

A PCR-RFLP system was used for genotyping the fatty mutation. PCR product (596 bp) amplified by using a primer set (FW, 5′-aagccatctcatttgctggt-3′; RV, 5′-ggcaggcagatctctcaatc-3′) is digested by restriction enzyme MspI (CCGG). The 596 bp fragment (wildtype allele) is further divided into 328 bp and 268 bp fragments in the fatty allele.

Plasma Lcn2 levels

Plasma Lcn2 levels were measured using ELISA kits (BioPorto Diagnostics A/S).

Plasma lipid parameters

Plasma lipid parameters were measured using the medium-sized biochemistry automatic analyser 7180 (Hitachi, Ltd.).

Genotyping of the missense variant in the growth hormone receptor gene

A PCR-RFLP system was developed for genotyping of the A546V missense variant found in ZFDM rats: PCR product (282 bp) amplified by using a primer set (FW, 5′-cagatgccaaaaagtgcatcgccg-3′; RV, 5′-ggtctgtgctcacatagccacat-3′) is digested by restriction enzyme AccII (CGCG), resulting in 192 bp, 66 bp, and 24 bp fragments in ZF rats (wildtype allele) while 216 bp and 66 bp fragments in ZFDM rats (mutant allele).

Genetic analysis

Female ZF fa/+ rats were crossed with male ZFDM fa/fa rats to produce F1 animals. Then, female ZFDM fa/+ rats were crossed with the male fa/fa F1 rats to produce backcross progenies. Male fa/fa backcross progenies were checked for nonfasting blood glucose levels until 60 weeks of age. Genotyping of the Lcn2 Q137X nonsense mutation and the Ghr A546V missense variant were performed by the PCR-RFLP system as described above.

Statistical analysis

Data are expressed as mean ± SEM (standard error of the mean). Differences among the groups were analysed with the Tukey–Kramer method as indicated in the figure legends. Association of the Lcn2 genotype with diabetes was analysed with a chi-square test. P < 0.05 was regarded as statistically significant. Statistical analysis was performed using R (version 4.4.1).

Results

Comparative transcriptome analysis of the pancreatic islets between ZF and ZFDM rats

To elucidate the gene expression profile of the pancreatic islets and primary genetic factors of T2D in ZFDM rats, we performed comparative transcriptome analysis between ZF and ZFDM rats at 8 and 12 weeks of age (Fig. 1A). ZFDM fa/fa rats at 8 weeks of age showed normoglycemia with very slight glucose intolerance upon oral glucose load, while those at 12 weeks of age exhibited apparent glucose intolerance.^6,7 Since there was a large variation in the size of the islets in fa/fa rats at 12 weeks of age,⁹ we compared gene expression in non-large and enlarged (exceeding 300 μm in diameter) islets separately. Among ~13,000 genes detected, we found DEGs for each comparison (Fig. 1B): 1,046 DEGs for ZF +/+ versus ZFDM fa/+ at 8 weeks of age (Supplementary Table S1), 359 DEGs for ZF fa/fa versus ZFDM fa/fa at 8 weeks of age (Supplementary Table S2), 1,355 DEGs for ZF +/+ versus ZFDM fa/+ at 12 weeks of age (Supplementary Table S3), 778 DEGs for non-large islets of ZF fa/fa versus ZFDM fa/fa at 12 weeks of age (Supplementary Table S4), 766 DEGs for enlarged islets of ZF fa/fa versus ZFDM fa/fa at 12 weeks of age (Supplementary Table S5). Common DEGs at each age may represent primary expression differences between strains irrespective of obesity and subsequent glucose intolerance conditions, which may be reflected by primary genetic differences. We performed gene enrichment analysis on these common DEGs. Common 127 DEGs (up: 58, down: 69) at 8 weeks of age (Fig. 1C; Supplementary Table S6) were enriched in the GO term associated with ‘response to stimulus’ and ‘extracellular matrix’ (Supplementary Table S7). Common 166 DEGs (up: 114, down: 52) at 12 weeks of age (Fig. 1C; Supplementary Table S8) were enriched in the GO term associated with ‘positive regulation of multicellular organismal process’, ‘response to stimulus’, and ‘extracellular matrix’ (Supplementary Table S9). Gene enrichment analyses using the DAVID and Metascape tools produced similar results. Due to the relatively small number of genes, there was no significant GO term for common 46 DEGs (up: 18, down: 28) for all the comparisons (Fig. 1C; Supplementary Tables S10 and S11). However, these 46 DEGs may serve as strong candidate genes for T2D in ZFDM rats.

Identification of a nonsense mutation in the Lcn2 gene in the ZFDM rat

To further elucidate primary genetic factors of T2D in ZFDM rats, we performed mutation analysis on RNA-seq data of the islets of ZF and ZFDM rats at 8 weeks of age and found that there were ~5,900 variants including 5 stop-gained variants (3 genes) and 921 missense variants as compared with the rat reference genome sequence (Supplementary Table S12). The ZF rat has stop-gained variants in Commd7 and RT1-Db1 while the ZFDM rat has those in Lcn2 and RT1-Db1. Among genes with stop-gained variants, lipocalin 2 (Lcn2) was included in the common 46 DEGs. In addition, the ZFDM rat has 2 missense variants (K541N and A546V) in the growth hormone receptor (Ghr) gene which was also included in the common 46 DEGs. Since stop-gained variants would have a more significant effect on the gene function as compared with missense variants, we focussed on Lcn2. In ZF rats, Lcn2 gene expression was much higher in the islets of fa/fa rats as compared with that of +/+ rats, and the expression levels increased with age (Fig. 2A). In contrast, the expression was significantly lower in the islets of both fa/+ and fa/fa in the ZFDM rat as compared with those of ZF rats. As for ZF rats, the expression levels in ZFDM fa/fa rats were higher than those in ZFDM fa/+ rats, but the difference was much lower than that of ZF rats.

Sanger sequencing confirmed the stop-gained variant in Lcn2 and verified that the ZFDM rat has a nonsense mutation at the 137th glutamine codon (CAG to TAG), c.409C > T; p.Q137X (Fig. 2B). Lcn2 consists of 6 exons spanning ~3.3 kb genomic region on rat chromosome 3 and the Q137X mutation is located on exon 4 (Fig. 2C). Lcn2 protein belongs to the lipocalin family, and the protein structure is characterized by a single polypeptide chain that forms a barrel-like structure composed of 8 β-sheets, which create a central cavity (Fig. 2D). The mutation disrupts the protein after the sixth β-sheets, deleting the seventh and eighth β-sheets and C-terminal α-helix domain. The mutant and wildtype alleles were clearly distinguished by PCR-RFLP analysis (Fig. 2E). Since Lcn2 is known to be secreted into the blood, plasma Lcn2 levels were examined in ZF and ZFDM rats. Plasma Lcn2 proteins were detected in both +/+ and fa/fa of the ZF rat with no significant changes with age. In contrast, Lcn2 proteins were hardly detected in the plasma of both fa/+ and fa/fa of the ZFDM rat (Fig. 2F).

To clarify the frequency of the mutation in rat species, we searched for the mutation among 159 inbred and 2 outbred rat strains. Interestingly, the mutant allele was detected in 32 inbred and 1 outbred strains including well-known inbred rat strains such as BB, DA, and F344 (Fig. 2G; Supplementary Table S13).

Evaluation of the relevance of the Lcn2 mutation with T2D by genome editing technique

To evaluate directly that the Lcn2 mutation is responsible for the development of T2D in ZFDM rats, we generated Lcn2 knock-in rats to correct the stop-gained variant in the Lcn2 gene in ZFDM rats using CRISPR/Cas9-mediated genome editing. Among ~30 founder (G0) animals, there were 2 animals (G0-28 and -29) harbouring an edited allele in which the nonsense mutation was replaced with a wild-type nucleotide (Fig. 3A). Successful correction of the Q137X mutation was confirmed by the fact that animals in both G0-28- and -29-derived lines heterozygous for the edited allele showed the expression of the Lcn2 protein in the plasma (Fig. 3B). We then produced male fa/fa animals homozygous or heterozygous for the edited allele and compared phenotypes with animals homozygous for the mutant allele (Fig. 3C). There was no significant difference in body weights among animals with the 3 genotypes. Blood glucose levels also showed no significant difference among the 3 genotypes. Accordingly, there was no significant difference in the cumulative incidence of diabetes among them. Although Lcn2 has been reported to be related with plasma lipid levels in humans,³¹ there was no clear difference in plasma lipid parameters among these animals (Supplementary Fig. S1).

Genetic linkage analysis of the Lcn2 mutation with T2D

To further confirm the relevance of the Lcn2 mutation with T2D, we also performed a genetic linkage analysis by using a genetic cross between ZF and ZFDM rats. Since none of the male F1 animals homozygous for fa/fa developed diabetes, we produced backcross progenies by crossing between female ZFDM and male F1 (Fig. 4A). Among 100 male backcross progenies (fa/fa), 57 animals developed diabetes by 60 weeks of age, suggesting that a recessively acting autosomal allele in ZFDM rats is involved in the development of T2D in the cross (Fig. 4B). We determined the Lcn2 genotypes in each backcross progeny and performed a chi-square test between the genotype and diabetic phenotype (Fig. 4C). It has been revealed that there was no significant relationship of the Lcn2 genotype with the onset of diabetes. In addition, we also examined the relevance of the Ghr missense variants with T2D and clarified that there was also no significant relationship with T2D (Fig. 4D).

Discussion

In the present study, by using ZFDM rat as a model of obese T2D, we found that (i) 46 genes exhibit significant expression differences between ZF and ZFDM islets irrespective of obesity and glucose intolerance states; (ii) among these common DEGs, the ZFDM rat has a nonsense mutation in the Lcn2 gene, the mutation is distributed widely in rat species; and (iii) the Lcn2 mutation is, however, not involved in the development of T2D.

At first, we performed a comparative transcriptome analysis of the pancreatic islets between ZF and ZFDM rats at 8 and 12 weeks of age. The common DEGs of the comparison (ZF +/+ vs. ZFDM fa/+) and (ZF fa/fa vs. ZFDM fa/fa) at each age group were associated with ‘extracellular matrix’: Fgb, Fgg, Hapln4, Lgals3, Mmp13, Mmp19, and P3h2 showed lower expression, while Adamts16, Ccn4, Col7a1, Col9a1, Cspg4, Cthrc1, Fmod, Olfml2a, Ptx3, S100a4, S100a6, Serpinf1, Spon2, Srpx2, Tgfb1, Tnn, and Wnt5a showed higher expression in ZFDM rats as compared with ZF rats.

Matrix metalloproteinases (MMPs) degrade collagenous extracellular matrix, which are associated with tissue remodelling and fibrosis. In ZDF rats, as compared with lean control rats, gene expressions of Mmp2, -12, and -14 in diabetic fa/fa rats were increased with the onset of islet dysfunction and diabetes.⁴ In ZFDM rats, we found that most of Mmps exhibit higher expressions in fa/fa rats as compared with +/+ or fa/+ rats (Supplementary Fig. S2A). In addition, the expressions of most Mmps are higher in ZFDM rats as compared with those in ZF rats, while the expressions of Mmp13 and Mmp19 show opposite profiles (Supplementary Fig. S2B), indicating that these Mmps might have some roles in suppressing the development of islet dysfunction and diabetes.

Gene expression profiles of collagens (Supplementary Fig. S3A), the main component of extracellular matrix, show a similar pattern with those of Mmps, indicating that fa/fa animals have more extracellular matrix than +/+ or fa/+ animals, the degree is higher in ZFDM rats as compared with that in ZF rats. The expressions of Col7a1 and Col9a1 are significantly higher in ZFDM rats as compared with those in ZF rats (Supplementary Fig. S3B). Since these collagens consist of minor collagen components, the roles in islet dysfunction and diabetes remain to be elucidated. However, the common DEGs for all the comparisons may correspond to primary gene expression differences between strains irrespective of obesity and subsequent glucose intolerance conditions, which could represent primary genetic differences.

Secondly, we therefore searched for functional variants in genes expressed in ZF and ZFDM islets. Among a total of 5 stop-gained variants, only Lcn2 was found to be included in the common 46 DEGs and ZFDM has a Q137X nonsense mutation, serving Lcn2 as a strong candidate for T2D. The mutation deletes significant C-terminal domains of the protein, which may lead to a loss of function of the protein. In addition, the gene expression of the mutant Lcn2 was strongly reduced due to the nonsense-mediated mRNA decay, resulting in no detectable protein in the plasma. These findings indicate that the Q137X nonsense mutation causes a deficiency of the Lcn2 protein.

To our surprise, the Lcn2 mutation is revealed to be distributed widely in inbred rat strains including BB, a model of type 1 diabetes (T1D); DA, a model of collagen-induced arthritis, adjuvant-induced arthritis, and experimental autoimmune encephalomyelitis; F344, a well-used control strain; NAR, a model of analbuminemia; WBN, a model of nonobese T2D; and ZDF, a model of obese T2D. Especially, the fact that F344 strains have the mutation needs to be noticed since F344 strains frequently serve as background strains for genome editing experiments and as standard strains for preclinical drug safety assessment. Regarding models of diabetes, T1D model KDP rat and nonobese T2D model SDT rat have wildtype alleles, suggesting no apparent association of the mutation with both T1D and T2D. By database search, we found that the mutation has already been registered as rs3323635808 in public databases such as the Rat Genome Database (https://rgd.mcw.edu/rgdweb/homepage/), along with information on its widespread distribution in 50 laboratory rat strains including ACI, DA, F344, M520, and WN.

Finally, we evaluated the relevance of the Lcn2 mutation with T2D by both genome editing and genetic linkage analysis. Our analyses revealed that the mutation is not involved in the development of T2D in ZFDM rats. In addition to adipose tissue, liver, and immune cells that primarily produce Lcn2, various tissues and cells, including pancreatic β-cells, also produce Lcn2 under inflammation or stress conditions.³² Lcn2 is thought to have a beneficial role in the regulation of various aspects of energy metabolism: protection from diet-induced obesity and insulin resistance,³³ high-fat diet-induced adipose tissue remodelling,³⁴ and brown fat activation.³⁵ Serum LCN2 levels positively correlate with energy expenditure and fatty acid oxidation in normal weight but not obese women.³⁶ In contrast, other reports showed that serum Lcn2 levels are associated with obesity and insulin resistance in humans and mice.^37–39

Lcn2 has been also reported as a bone-derived hormone with metabolic regulatory effects⁴⁰: osteoblast-derived Lcn2 maintains glucose homeostasis by inducing insulin secretion, improving glucose tolerance and insulin sensitivity, and inhibiting food intake. In addition, Lcn2 counteracts metabolic dysregulation in obesity and diabetes, suggesting a distinct beneficial effect of Lcn2 on β-cell function and adaptive β-cell proliferation during toxicity or onset of obesity.⁴¹ A model of the compensatory homeostatic role of Lcn2 has been proposed, in which Lcn2 counteracts insulin resistance progression, prevents obesity, and suppresses diabetes, while once this mechanism is overwhelmed, obesity increases and diabetes develops.⁴¹ Taken together, our findings in the ZFDM rat, an extreme obese and severe T2D model, suggest that the compensatory homeostatic mechanism is overwhelmed and Lcn2, therefore, could not exert any beneficial effects on β-cell function and proliferation during the development of obesity and T2D. The role of Lcn2 needs to be examined in other rat models of mild obesity or diabetes.

In conclusion, we here find a nonsense mutation in the Lcn2 gene in a rat model of obese T2D, the mutation is distributed widely in rat species. Although we could not show the relevance of the Lcn2 mutation with the development of T2D, several rat strains would serve as Lcn2 deficient models, contributing to unravel normal and pathophysiological roles of Lcn2 in a wide variety of phenotypes.

Acknowledgments

We thank Hoshino Laboratory Animals, Inc. for providing ZFDM rats; Shihomi Hidaka, Ayako Kawabata, Takuro Yamaguchi, and Chihiro Seki for their excellent technical assistance. We also thank the National BioResource Project-Rat (http://www.anim.med.kyoto-u.ac.jp/NBR/) for providing the genome DNA of 157 inbred rat strains. Computations were partially performed on the National Institute of Genetics (NIG) supercomputer at the Research Organization of Information and Systems (ROIS) National Institute of Genetics.

Funding

This study was supported by JSPS KAKENHI (https://www.jsps.go.jp/j-grantsinaid/) Grant nos. JP18H02364 and JP21H02390 (to N.Y.), and partially supported by grants from the National Center for Global Health and Medicine (https://www.ncgm.go.jp/) Grant nos. 29-1001 and 23A1013 (to T.O.).

Conflict of interest

None declared.

Data availability

All raw RNA sequencing data have been deposited in the DDBJ Sequence Read Archive (DRA; https://www.ddbj.nig.ac.jp/dra/index-e.html) under the accession numbers PRJDB7245 and PRJDB12053.

References