CRISPR-Cas9-mediated generation of obese and diabetic mouse models.

Mouse models of obesity (ob/ob) and diabetes (db/db) in which the leptin (Lep) and leptin receptor (Lepr) genes have been mutated, respectively, have contributed to a better understanding of human obesity and type 2 diabetes and to the prevention, diagnosis, and treatment of these metabolic diseases. In this study, we report the first CRISPR-Cas9-induced Lep and Lepr knockout (KO) mouse models by co-microinjection of Cas9 mRNA and sgRNAs that specifically targeted Lep or Lepr in C57BL/6J embryos. Our newly established Lep and Lepr KO mouse models showed phenotypic disorders nearly identical to those found in ob/ob and db/db mice, such as an increase in body weight, hyperglycemia, and hepatic steatosis. Thus, Cas9-generated Lep and Lepr KO mouse lines will be easier for genotyping, to maintain the lines, and to use for future obesity and diabetes research.

Introduction

Obesity and type 2 diabetes are two of the risk factors of death worldwide [19] and represent significant economic burden. Because the incidence of these diseases continually increasing, the predicted cost for treating them is expected to rise to $48 billion per year by the year 2030 [32]. Additionally, comorbidities including hypertension, hyperlipidemia, and cancer represent significant economic burden and loss of life quality [13], making metabolic disorders a significant public health concern. Therefore, the identification of new methods for disease prevention, diagnosis, and cure is critical. New animal models of metabolic disease are being continually developed in an effort to unravel the mechanism of action behind several genes and functional pathways implicated in the development of metabolic disorders and will prove powerful tools for understanding obesity and diabetes pathology, as well as for developing new treatments and preventive approaches.

There are five obesity-related loci in mouse models of metabolic disease; of these five, the obese (ob) and diabetes (db) loci are the most commonly investigated [9]. The genes at these two loci were identified as leptin (Lep) and leptin receptor (Lepr) in 1994 and 1996, respectively [3, 33]. LEP is a 167-amino acid peptide that is primarily produced in white adipose tissues (WAT). Upon entering the bloodstream, signal peptide removal reduces the length of LEP to 147 amino acids [31]. This protein is a key player in regulating energy balance and in body weight control [26]. The brain recognizes the circulating 147 amino acid-long form of leptin through the leptin receptor, which is encoded by the Lepr gene [11]. The long Lepr isoform, OB-Rb, is mainly expressed in the hypothalamus, while the short isoform, OB-Ra, is expressed throughout the body [8]. Mutations in Lep or Lepr lead to early-onset obesity and accompanying hyperglycemia with insulin resistance in mice [18], and mutations in LEP and LEPR, though rare, has been reported in humans [21]. Although the existing ob/ob and db/db mouse models have strong, relevant phenotypes and made important contributions to our understanding of metabolic disorders, it is difficult to maintain these lines due to their specific genotyping strategy, which requires additional steps other than PCR [14]. Therefore, the development of easier to maintain obese diabetic mouse lines would represent a significant contribution to metabolic disorder research.

CRISPR-Cas9 is a third-generation engineered nuclease that causes double-strand DNA breaks, eliciting the non-homologous end joining and homologous recombination repair systems [25]. Here, using the CRISPR/Cas9 system, we generate and phenotype Lep and Lepr knockout (KO) mouse models. These mice are obese and exhibit a diabetic phenotype, even in the first generation, and are phenotypically very similar to the existing ob/ob and db/db lines such as gross body weight increasement, hepatic steatosis, and hyperglycemia.

Materials and Methods

In vitro synthesis of Cas9 mRNA and sgRNAs

CRISPR-Cas9 constructs were synthesized as described previously [28]. Briefly, Cas9 mRNA was synthesized in vitro from linear DNA templates using the mMESSAGE mMACHINE T7 Ultra kit (Ambion) according to the manufacturer’s instructions. DNA templates for sgRNAs were also synthesized in vitro using PCR; RNA was synthesized from these templates using a MEGAshortscript T7 kit (Ambion) according to the manufacturer’s instructions. The constructs were diluted RNase-free injection buffer (0.25 mM EDTA, 10 mM Tris at pH 7.4) prior to microinjection. Following sequence information was used for sgRNA synthesis; sgRNA#1 of Lep, AAGCCACAGGAACCGACACA; sgRNA#2 of Lep, TGAGGGTTTTGGTGTCATCC; sgRNA#1 of Lepr, GAGTCATCGGTTGTGTTCGG; sgRNA#2 of Lepr, AGAAGCCCCCTTCAAAGCCG.

Generation and maintenance of Lep and Lepr KO mouse lines

Lep and Lepr KO mouse models were generated using the CRISPR-Cas9 system as described previously [29]. Briefly, CRISPR-Cas9 constructs and sgRNAs were microinjected into fertilized C57BL/6J embryos. Mutations on Lep and Lepr were confirmed by Sanger sequencing (Cosmo Genetech). Mice were maintained on a normal diet (PicoLab® Rodent Diet 20, Orientbio) under a 12 h light/dark cycle. All animal experiments were performed in accordance with Korean Food and Drug Administration guidelines and protocols were reviewed and approved by the Institutional Animal Care and Use Committee of the Yonsei University (201507-390-01) and Gachon University (2015-0042).

Genotyping of Lep and Lepr KO mouse lines

The genotypes were confirmed by PCR using the following primers; 5′- TCC CAG GGA GGA AAA TGT GCT -3′ (forward for Lep), 5′- TGA CAT GTT TCT CAG ACT CTG GTT -3′ (reverse for Lep), 5′- CTG CTG GAG CCC CAA ACA ATG C -3′ (forward for Lepr), 5′- TTC AAC AAT TGC TTC AGA AGC C -3′ (reverse for Lepr). PCR was performed as follows: denaturation at 94°C, 30 cycles of 30 s at 94°C, 30 s at 60°C, and 30 s at 72°C; 7 min at 72°C. Electroporation was conducted on 2–3% of agarose gels and visualized with a Bio-Rad ChemiDoc system.

Mouse management and experiments

At the experimental endpoint, mice were euthanized at the beginning of the light cycle and weighed. Blood samples were collected via the hepatic portal vein, and plasma was obtained by centrifugation at 300 × g for 15 min at 4°C. Livers and gonadal fat were removed, weighed, and either snap-frozen in liquid nitrogen or formalin-fixed. Liver and plasma samples were stored at –80°C until RNA isolation or biochemical analysis. Body composition was assessed by 1H-magnetic resonance spectroscopy (BioSpin, Bruker, Billerica, MA). Mouse lines of Lep and Lepr KO are available in the Korea Ministry of Food & Drug Safety (Lep KO, KNL-HYD-TG0623; Lepr KO, KNL-HYD-TG0624).

Hematoxylin and eosin staining

Liver and WAT tissues were fixed in neutral-buffered formalin and embedded in paraffin blocks according to standard procedures, and sections were stained with hematoxylin and eosin (H&E) as described previously [24].

Real-time quantitative PCR

Total RNA was isolated from the liver, brain, and WAT using TRIzol reagent (Ambion), and then cDNA was synthesized using a RevertAid First strand cDNA Synthesis Kit (Thermo). Quantitative gene expression analyses were performed using a CFX384 Real-Time PCR system (Bio-Rad, Berkeley, CA) using SYBR1 Premix Ex TaqTM II, ROX Plus (Takara). Primers used for RT-qPCR were as follows: 5′-AGC TGC AAG GTG CAA GAA GAA-3′ (Lep forward), 5′-GGA ATG AAG TCC AAG CCA GTG AC-3′ (Lep reverse), 5′-TGA TGT GTC AGA AAT TCT ATG TGG T-3′ (Lepr forward), 5′-AGC AAC AGT GGA AGA CTG TTT TG-3′ (Lepr reverse), 5′-CAA GAA TAC CAA AGT GCG ATC AA-3′ (Pparg forward), 5′-GAG CTG GGT CTT TTC AGA ATA ATA AG-3′ (Pparg reverse)’; GGA CAA CAC GCA TTT CAT GAT C-3′ (Cidea forward), 5′-GGC TAT TCC CGA TTT CTT TGG-3′ (Cidea reverse), 5′-GGA GAT GGC ACA GGA GGA A-3′ (Cyclophilin forward), and 5′-GCC CGT AGT GCT TCA GCT T-3′ (Cyclophilin reverse).

Biochemical measurements

Blood glucose levels during the glucose tolerance test were measured using Allmedicus, Gluco Dr.TM plus+ (AGM-3000). Plasma insulin and leptin levels were measured using the Mouse Ultrasensitive Insulin ELISA kit (ALPCO, 80-INSMSU-E01) and the Mouse/Rat Leptin ELISA kit (ALPCO, 22-LEPMS-E01), respectively. Homeostasis model assessment IR (HOMA-IR) index was calculated from glucose (mmol/L) X insulin (mU/L)/22.5 [20]. Plasma triglyceride (TG), total cholesterol (T-CHO), high-density lipoprotein cholesterol (HDL), low-density lipoprotein cholesterol (LDL), alanine transaminase (ALT), aspartate transaminase (AST), and lactate dehydrogenase (LDH) levels were determined by automated analysis (Model AU-480; Olympus, Tokyo, Japan) [22].

Statistical analyses

Data were expressed as mean ± standard error of the mean (SEM). Statistical analysis was performed using SPSS (Version 17.0; SPSS Inc., Chicago, IL). Data were analyzed using Kruskal-Wallis tests. Group comparisons were performed with Mann-Whitney U tests and corrected with Bonferroni.

Results

Creation of Lep KO mice by Cas9 microinjection

We designed a pair of independent sgRNAs targeting exon 2 of the Lep locus to establish a homozygous Lep KO mouse model. To minimize off-target sgRNA binding while maximizing target efficiency, we thoroughly analyzed exon 2 of the Lep locus using sgRNA designer and CRISPR Design [5, 27] while designing the sgRNAs. Co-injection of the two independent sgRNAs with Cas9 mRNA caused a 79-bp deletion mutation on exon 2 of the Lep locus (Fig. 1A), causing a frameshift mutation and exposing a premature stop codon. Compared with the wild type (WT) LEP protein, composed of 167 amino acid residues, the mutant peptide, if translated, encodes only three original amino acid residues and 41 abnormal residues (Fig. 1B). A BLAST search of the abnormal residues revealed no match to known annotated murine proteins, suggesting that the mutant leptin does not function normally. Because the deletion mutation removed a larger portion of the Lep locus than in existing ob/ob mouse models, we were able to confirm the genotype via PCR using a single primer pair (Fig. 1C). Additionally, although we also successfully mutated both Lep alleles simultaneously (2- and 77-bp deletion mutations, respectively) in first generation animals, which were severely obese (data not shown), the mice were infertile and therefore the line could not be maintained.

Fig. 1.

CRISPR-Cas9-mediated generation of Lep KO mice. (A) A representative view of the CRISPR-Cas9 targeting strategy for generating Lep KO mice. The Lep locus (top panel) and Lep nucleotide and amino acid sequences (bottom panel) are shown. Exons are indicated with black boxes and introns are indicated by black lines. The protospacer adjacent motif (PAM) of SpCas9 is italicized and underlined (bottom panel) and mutant amino acids are shown in italics. (B) The length and amino acid sequence of WT Lep and the predicted length and amino acid sequence of the mutated form. Italicized amino acids indicate predicted mutant peptides. The asterisk indicates a translation stop signal. (C) PCR genotyping of Lep WT (219 bp), heterozygous (219 bp for WT and 140 bp for KO), and KO (140 bp) mice.

Physiological comparisons of Lep KO and ob/ob mouse models

Due to the premature stop codon exposed by sgRNA-Cas9 mRNA co-injection, we would expect nonsense-mediated mRNA decay [2] and concurrent oblation of gene expression. Indeed, Lep mRNA was not detected in WAT in homozygous Lep KO mice, but was detected in both WT and heterozygous KO mice (Fig. 2A). Phenotypically, the CRISPR-Cas9-generated Lep KO mouse is highly comparable to the existing ob/ob mouse model. At 12 weeks of age, the Lep KO mice weigh about 1.8 to 2 times more than WT mice (Figs. 2B and C); similarly, ob/ob mice are about two times heavier than control mice [4]. Both males and females exhibited drastic changes in gross weight (Figs. 2B, C, and Supplementary Fig. S1A), with a 30% increase in weight at 6 weeks old and a 54% increase in weight at 12 weeks old compared to WT mice (Fig. 2C, Supplementary Fig. S1A). This weight gain compared to WT is likely due to the excessive food intake characterizing the KO animals (Fig. 2D), which is consistent with the ob/ob phenotype [10]. Although there was no change in lean mass, fat and liver mass increased in the Lep KO mice compared to WT (Fig. 2E, Supplementary Figs. S1B and C). Together, these data suggest that the CRISPR-Cas9-induced Lep mutation is sufficient to cause obesity in these mice.

Fig. 2.

CRISPR-Cas-9 mediated Lep KO causes obesity, hyperglycemia, and hyperinsulinemia. (A) RT-qPCR analysis of Lep mRNA levels in the WAT of 10-week-old mice. mRNA levels were normalized to β-actin. (B) A representative photograph of 9-month-old mice. Scale bar: 1 cm. (C) Monitoring of body weight over 13 weeks in male mice. The mice were experienced glucose tolerance test at 14 weeks of age. (D) Average daily food intake in 8-week-old mice. (E) Body composition measured by 1H-NMR in 15-week-old male mice. (F, G) Representative image of the macroscopic appearance of livers in 9-month-old mice (F) and H&E staining image of liver and WAT in 16-week-old mice (G). White bar: 1 cm; Black bar: 100 µm. (H) Plasma concentration of ALT, AST, and LDH in male and female mice (ALT: alanine transaminase, AST: aspartate transaminase, LDH: Lactate dehydrogenase). (I, J) Blood glucose (I) and plasma insulin (J) levels in 14-week-old mice after 16 h of fasting. (K) Results of a glucose tolerance test administered in 14-week-old Lep WT, heterozygous and homozygous KO male mice fasted for 16 h. Data are presented as the mean ± SEM. WT (n=4), Het (n=5), and Lep KO (n=4). Data were analyzed using Kruskal-Wallis test and group comparisons with Mann-Whitney U test. #P<0.01 vs. WT and Het group (A, C, D, and K), *P<0.05 vs. Het group (E, H–J).

Macroscopic and microscopic analyses of the liver and WAT revealed gross morphological changes due to fat accumulation in Lep KO mice (Figs. 2F, G, and Supplementary Fig. S1D). The liver was approximately 3-fold larger in Lep KO mice compared to WT, and was also a light salmon hue rather than the ruddy color that characterizes normal liver tissue. Consistent with these morphological observations, the two-adipogenic markers peroxisome proliferator-activated receptor gamma (Pparg) and cell death-inducing DNA fragmentation factor α-like effector A (Cidea) were upregulated in Lep KO mice (Supplementary Figs. S1E and F). Plasma levels of the liver damage markers alanine aminotransferase (ALT), aspartate aminotransferase (AST), and lactate dehydrogenase (LDH) were also significantly increased in Lep KO mice (Fig. 2H). Plasma TG, T-CHO, LDL, and HDL levels were also markedly higher in Lep KO mice (Supplementary Fig. S1G), which is consistent with previous reports on ob/ob mice [1]. These data therefore suggest that our Lep KO mouse model is an efficient model for obesity research as the mice exhibit a strong obesity phenotype with concurrent hepatic steatosis similar to that of ob/ob mice.

Development of hyperglycemia with insulin resistance in Lep KO mice

In both the ob/ob and db/db models, hyperglycemia develops before obesity becomes apparent [23]. We measured blood glucose levels in WT, heterozygous Lep KO, and homozygous Lep KO mice to determine whether our Lep KO mice were hyperglycemic in addition to obese. Fasting blood glucose and plasma insulin levels were also significantly increased in Lep KO mice compared to heterozygous KO mice (Figs. 2I and J). To quantify the extent of systemic insulin resistance in Lep KO mice, Lep WT, heterozygous and homozygous KO mice were underwent IPGTT (Fig. 2K and Supplementary Fig. S1H). Fifteen mins after glucose injection, blood glucose levels increased by approximately 2-fold in WT and heterozygous KO mice and increased by 4-fold in homozygous KO mice. Maximal blood glucose levels approached 600 mg/dl in the Lep KO mice but remained at levels around 200 mg/dl in WT and heterozygous KO mice. Additionally, analysis of insulin resistance showed significant increment of HOMA-IR index in Lep KO mice (Supplementary Fig. S1I). These data indicate that Lep KO is associated not only with obesity and hepatic steatosis, but also with hyperglycemia and insulin resistance, like the existing ob/ob model [31].

Creation of Lepr KO mice by Cas9 microinjection

In a similar approach to that used to generate Lep KO mouse, a pair of independent sgRNAs targeting exon 3 of the Lepr locus was co-microinjected with Cas9 mRNA to induce a deletion and found the deletion of five nucleotides near binding site for the second sgRNA (Fig. 3A). This deletion caused a frame shift mutation, and exposed premature stop codon. Compared with the WT LEPR protein, which is composed of 1162 amino acid residues, the mutant protein, if translated, encodes only 27 N-terminal amino acid residues (Fig. 3B). Despite the small deletion of only 5 bp, the genotype was successfully confirmed by PCR using only a single primer pair to detect both WT and mutant forms (Fig. 3C).

Fig. 3.

CRISPR-Cas9-mediated generation of Lepr KO mice. (A) A representative view of the CRISPR-Cas9 targeting strategy for generating Lepr KO mice. The Lepr locus (top panel) and Lepr nucleotide and amino acid sequences (bottom panel) are shown. Exons are indicated with black boxes and introns are indicated by black line. The PAM of SpCas9 is italicized and underlined (bottom panel) and mutant amino acids are shown in italics. The asterisk indicates a stop codon. (B) The length and amino acid sequence of WT Lepr and the predicted length and amino acid sequence of the mutated form. Italicized amino acids indicate predicted mutant peptides. The asterisk indicates a translation stop signal. (C) PCR genotyping of Lepr WT (59 bp), heterozygous (59 bp for WT and 54 bp for KO), and KO (54 bp) mice.

Depletion of Lepr leads to obesity and liver steatosis

Similar to the Lep KO model, oblation of Lepr mRNA expression was achieved in the epididymal WAT (eWAT), the liver, and the brain in Lepr KO mice (Fig. 4A). Phenotypically, the CRISPR-Cas9-generated Lepr KO mouse is highly comparable to the existing db/db mouse model [10]. The Lepr KO mice weigh about 1.8 to 2 times more than WT mice (Figs. 4B and C), which is similar to db/db [4]. Both males and females exhibited gross changes in body shape and weight compared to control animals (Figs. 4B, C, and Supplementary Fig. S2A), which is likely due to increased food intake by KO mice (Fig. 4D). Liver and WAT are also weighty notably more in Lepr KO mice compared to control mice (Fig. 4E and Supplementary Fig. S2B), and liver morphology was severely altered (Fig. 4F). Microscopic and molecular analyses revealed hepatic steatosis with increased expression of adipogenic markers Pparg and Cidea in the Lepr KO livers (Fig. 4G, Supplementary Figs. S2C and D); however, no changes in WAT were observed. Although plasma TG and LDH levels were unchanged, plasma ALT, AST, LDL, HDL, and T-CHO levels were significantly increased in Lepr KO mice compared to controls (Fig. 4H and Supplementary Fig. S2E). To compensate for defective leptin receptor in Lepr KO mice, plasma leptin levels were significantly increased (Fig. 4I). Together, these data suggest that the CRISPR-Cas9-induced Lepr mutation induces obesity.

Fig. 4.

CRISPR-Cas9-mediated Lepr KO causes obesity, glucose intolerance, and hyperinsulinemia. (A) RT-qPCR analysis of Lepr mRNA levels in the WAT, liver, and brain of 10-week-old mice. Leptin receptor mRNA levels were normalized to β-actin. (B) A representative photograph of 9-month-old mice. (C) Monitoring of body weight over 14 weeks in male mice. The mice were experienced glucose tolerance test at 14 weeks of age. (D) Average daily food intake in 5-week-old male mice. (E) Liver/body weight ratio in 16-week-old male mice. (F) Representative image of the macroscopic appearance of livers in 9-month-old mice. White bar: 1 cm; Black bar: 100 µm. (G) Representative image of liver and WAT H&E staining of 16-week-old male mice. (H) Plasma concentration of ALT, AST, and LDH in 16-week-old WT and Lepr KO mice. (I–L) Plasma leptin (I), blood glucose (J), and plasma insulin (K) levels in mice in the fed and fasted (16 h) states. (L) Results of a glucose tolerance test administered in 14-week-old Lepr WT, heterozygous and homozygous KO male mice fasted for 16 h. Data are presented as the mean ± SEM. n≥4 individual mice per group. Data were analyzed using Kruskal-Wallis test and group comparisons with Mann-Whitney U test. *P<0.05 vs. WT group (A, C–E, H, I, and K), *P<0.05 and #P<0.01 vs. WT and Het group (L).

Development of hyperglycemia with insulin resistance in Lepr KO mice

As with Lep KO mice, we measured blood glucose to determine whether Lepr KO mice were hyperglycemic compared to WT and heterozygous KO mice. Fasting blood glucose and plasma insulin levels were higher in homozygous KO mice compared to WT or heterozygous KO animals (Figs. 4J and K). To quantify the extent of systemic insulin resistance in Lepr KO mice, Lepr WT, heterozygous and homozygous KO mice were underwent IPGTT (Fig. 4L). Blood glucose levels of Lepr KO mice were significantly higher than Lepr WT and heterogyzous mice throughout the course of the experiment. Furthermore, gross enhancement of HOMA-IR index was detected in Lepr KO mice (Supplementary Fig. S2F). These data indicate that Lepr KO is associated with hyperglycemia, significant glucose intolerance and insulin resistance, identically to the existent db/db model [17].

Genotyping of Lep and Lepr KO mice generated by Cas9

Even though the genotyping of the ob/ob and db/db mouse models can be determined by enzyme restriction, high resolution melting analysis, or a tetra-primer amplification refractory mutation system [7, 14,15,16, 30], the specific deletion mutations we introduced while generating the Lep (79-bp deletion) and Lepr (5-bp deletion) KO mouse models help genotyping of these animals, which can be determined via PCR using a single primer pair; the PCR products can be size differentiated using gel electrophoresis (Supplementary Table S1 and Supplementary Fig. S3).

Discussion

Here we describe the successful generation of two alternative ob/ob and db/db mouse models using the CRISPR-Cas9 system to induce deletion mutations in the Lep and Lepr genes. These models exhibit nearly identical obese and diabetic phenotypes to the existing ob/ob and db/db models, yet are easier to genotype and therefore, maintain, making them an attractive model for future metabolic disorder research. Lep and Lepr mutations have been associated with both obesity and diabetes in humans [6, 12], making our mouse models a relevant tool for understanding metabolic disorders. Notably, the similarity of the disease phenotypes of Lep and Lepr KO mouse models to those of the ob/ob and db/db models was achieved even though the mice are of different genetic backgrounds. Due to these similarities in body weight, hyperglycemia, and insulin resistance, we propose that our mouse models represent a good substitute for the existing ob/ob and db/db mouse models.

We hypothesize, due to the success we had using the CRISPR-Cas9 system to effectively target the Lep and Lepr loci, that the CRISPR-Cas9 system can also be used to target and alter SNPs linked to metabolic diseases, changing them to the WT sequence and thus correcting the problem. For example, a potential approach is CRISPR-Cas9 targeting of mutated loci in pluripotent stem cells isolated from obese patients. Through this, the economic burden associated with diagnosing and treating metabolic disorders may be significantly alleviated through the increased knowledge proffered by using our new obese and diabetic mouse models in investigations of metabolic disorders.