Creation and Preliminary Characterization of a Leptin Knockout Rat

Abstract

Leptin, a cytokine-like hormone secreted mainly by adipocytes, regulates various pathways centered on food intake and energy expenditure, including insulin sensitivity, fertility, immune system, and bone metabolism. Here, using zinc finger nuclease technology, we created the first leptin knockout rat. Homozygous leptin null rats are obese with significantly higher serum cholesterol, triglyceride, and insulin levels than wild-type controls. Neither gender produced offspring despite of repeated attempts. The leptin knockout rats also have depressed immune system. In addition, examination by microcomputed tomography of the femurs of the leptin null rats shows a significant increase in both trabecular bone mineral density and bone volume of the femur compared with wild-type littermates. Our model should be useful for many different fields of studies, such as obesity, diabetes, and bone metabolism-related illnesses.

Leptin, encoded by the obese (ob or Lep) gene, is a hormone synthesized primarily by adipocytes and regulates energy homeostasis, the balance between food intake and energy expenditure, through binding to the leptin receptor, encoded by the diabetes, or db gene (1). Accordingly, leptin-deficient (ob/ob) or leptin receptor-deficient (db/db) mice are obese and diabetic and have been used in studies on diabetes, obesity, and cardiovascular and metabolic syndromes. In addition, leptin plays a proinflammatory role in the regulation of innate and adaptive immune responses (2) and regulates fertility (3) and bone mass (1, 4).

The rat is the preferred model for many disease areas, such as cardiovascular and neurodegenerative diseases, due to its closer resemblance to human physiology as well as with the availability of the wealth of physiological data accumulated in the past (5). However, the rat has been sidelined as an animal model system because of the lack of convenient genome engineering methods until the recent application of zinc finger nucleases (ZFNs) (6, 7) and later transcription activator-like effector nucleases (8). Engineered nucleases are designed to sequence-specifically bind and cleave DNA. The resulting double strand breaks stimulate homologous recombination or nonhomologous end joining repair mechanisms that lead to desired gene modifications. Instead of targeting embryonic stem cells, ZFNs allow direct and efficient editing of the genome in single-cell embryos, leading to high germline transmission rates that translate to shortened timelines. The rat equivalent of the db/db mouse is the Zucker rat, also known as fa/fa rat (9), which develops severe obesity associated with hyperphagia and hyperinsulinemia and has been similarly used in diabetic, cardiovascular, and metabolic syndrome research. However, a leptin knockout rat would allow studies on leptin functions as well as leptin therapy in ways that are not available with the Zucker rat, such as phenotypic rescue through exogenous leptin administration.

Here, we report the creation and characterization of the first Lep knockout rat via pronuclear microinjection of ZFNs. The Lep knockout rat, designated as Lep^Δ151/Δ151, is obese, diabetic, sterile, and with compromised immune system and increased bone mineral density (BMD), as observed in ob/ob and db/db mice. We believe that this knockout rat will expand the suite of animal models for biomedical and pharmacological research, such as obesity, type 2 diabetes, and osteoporosis.

Materials and Methods

Leptin ZFNs were obtained from Sigma's CompoZr product line. In vitro transcription of ZFN mRNA was performed as described previously (6).

Embryo microinjection and animal care

Founder animals were obtained via pronuclear microinjection of 5 ng/μl ZFN mRNA into rat embryos and embryo transfer, at Xenogen Biosciences (Cranbury, NJ, now part of Taconic Farms, Inc., Hudson, NY). The knockout rat colonies were maintained at Sigma Advanced Genetic Engineering Labs, which operates under approved animal protocols overseen by Sigma Advanced Genetic Engineering's Institutional Animal Care and Use Committee. Sprague Dawley (SD) rats were housed in static cages maintained on a 12-h light, 12-h dark cycle with ad libitum access to food and water.

Founder identification

Tail or toe biopsies were used for genomic DNA extraction and analysis as described previously (6). Primers flanking the target site (forward 5′-ATCCACAGCCTACAGCAGGT, reverse 5′-AAGGCGCTCACTGTGATTCT) were used to amplify a wild-type (WT) amplicon of 337 bp, using 60 C for annealing temperature.

Western blottings

White adipose tissue was homogenized and centrifuged. The supernatant was mixed with 2× Laemmli buffer (Sigma-Aldrich Corp., Saint Louis, MO), denatured, resolved on a denaturing PAGE, and transferred to a nitrocellulose membrane. The antileptin primary antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) was used at 1:200 dilution. The antirabbit secondary antibody (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) was used at 1:50,000 dilution. SuperSignal West-Pico kit (Thermo Fisher Scientific, Waltham, MA) was used to develop the signal.

Body weight and flood consumption

WT and Lep^{Δ151/ Δ151} rats of both genders as well as Lep^Δ151/+ males were weighted every other week from birth to 19 wk of age. WT and Lep^Δ151/Δ151 rats were provided with free access to a standard laboratory chow diet and water. Food was weighed, and average daily intake was calculated between 7 and 10 wk of age.

Serum chemistry, glucose tolerance test, and serum insulin levels

Rats of 7 and 10 wk of age were fasted for 16 h, and blood was collected by tail vein puncture and analyzed at IDEXX RADIL (Columbia, MO) for a standard serum chemistry panel.

For glucose tolerance test, the animals were fasted for 6 h before administered ip injection with d-glucose at 2 g/kg, and blood was collected by tail vein puncture and sent to IDEXX RADIL for glucose levels.

For serum insulin levels, 4-wk animals were fasted for 16 h, euthanized, and blood was drawn from the vena cava. Blood was left to clot for 1 h at room temperature, and serum was obtained after centrifugation. The samples were then sent to IDEXX RADIL for insulin level measurements.

Flow cytometry analysis and intracellular cytokine staining

Heparinized whole blood from 12-wk-old WT and Lep^{Δ151/ Δ151} rats was incubated for 5 h in medium alone or in medium containing 25 ng/ml phorbol 12-myristate 13-acetate, 0.5 μg/ml ionomycin, and 5 μg/ml protein transport inhibitor brefeldin A. For cell surface staining, peripheral blood samples were incubated with allophycocyanin-conjugated CD3, fluorescein isothiocyanate-conjugated CD4, and BD Horizon V450-conjugated CD8a antibodies. After red blood cell lysis and fixation, the cells were permeabilized for intracellular staining and incubated with phycoerythrin-conjugated interferon (IFN)-γ antibody. Stained and fixed samples were analyzed with a MACS Quant Analyzer (Miltenyi Biotec, Auburn, CA), and data analysis was carried out using FlowJo cytometry analysis software (TreeStar, Ashland, OR). The differential expression of the cell surface markers phenotypically distinguish the CD3+/CD4+ (T-helper) and CD3+/CD8+ (T-suppressor/cytotoxic) lymphocyte subpopulations. The percentage of IFN-γ-producing cells was determined within these subpopulations.

Microcomputed tomography (μCT) analysis and histomorphometry

Eight-week-old rats were euthanized with CO₂, and femurs were dissected, transferred to 70% ethanol, and scanned in a Scanco Medical Micro-CT40 system (Scanco Medical, Wangen-Brüttisellen, Switzerland) at 55 kVp, 145 mA, and 16 mm resolution. Gauss sigma of 1.2, Gauss support of 2, lower threshold of 237, and upper threshold of 1000 were used for all the analyses. Regions of interest were selected as 150 slices above the growth plate of the distal femur to evaluate the trabecular compartment. A three-dimensional analysis was done to determine bone volume, trabecular number, trabecular thickness, trabecular separation, cortical thickness, cortical bone fraction, and cortical porosity.

Tibias were dissected from the same animals, fixed overnight in formalin, and decalcified in 14% EDTA. Paraffin-embedded tibias sections were stained with hematoxylin and eosin or for tartrate-resistant acid phosphatase (TRAP) activity, and trabecular bone volume, osteoclast surface, and osteoclast number were measured using Bioquant Osteo (Bioquant Image Analysis Corp., Nashville, TN). Bone samples or sections were coded before analysis, which was performed by an observer blinded to genotype.

Ovaries from 4-wk-old females were dissected, and hematoxylin and eosin histological sections were prepared.

Data analysis

Values reported graphically are expressed as mean ± se, with numbers of samples indicated in figure legends. A P value was obtained through the use of unpaired one-tailed Student's t test. P values are indicated in each figure, and values less than or equal to 0.05 were considered significant.

Results

Generation and validation of Lep^−/− rats

A pair of ZFNs targeting exon 1 (Supplemental Fig. 1A, published on The Endocrine Society's Journals Online web site at http://endo.endojournals.org) of the Lep gene was used to create Lep^−/− rats in the SD background via pronuclear microinjection, as described previously (6). A founder carrying a 151-bp deletion flanking the target site (Supplemental Fig. 1B) was chosen to establish a colony. The deletion spans the exon 1/intron 1 junction and likely leads to aberrant splicing of the leptin transcript. When the founder was bred to a WT rat, six out of nine F1 pups inherited the mutant allele. Homozygous offspring of the founder, designated as Lep^Δ151/Δ151, were visually different from their WT or heterozygous littermates as early as 3 wk of age. Figure 1A shows a 20-wk knockout rat and a WT littermate. Western blot analysis of total protein from white adipose tissue of a Lep^Δ151/Δ151 rat confirmed the absence of leptin protein (Fig. 1C). Repeated effort to mate the Lep^Δ151/Δ151 rats to WT rats failed to result in pregnancy. We then analyzed histological sections of ovaries from Lep^Δ151/Δ151 and WT rats at 4 wk of age and observed no difference (Supplemental Fig. 2C). Nor did we visually observe differences on sperm counts and motility between the WT and Lep^Δ151/Δ151 rats at 10 wk of age.

Fig. 1.

Lep^Δ151/Δ151 rats are obese. A, Picture of a WT (Lep^+/+) and a leptin-deficient rat (Lep^Δ151/Δ151) rat at 20 wk of age. B, Western blot analysis of total proteins extracted from white adipose tissues of WT and Lep^Δ151/Δ151 rats. Actin was used as loading control. The expected size for leptin is 16 kDa.

Body weight

To quantify the level of obesity, body weight measurements were taken between birth and 18 wk of age. The Lep^Δ151/Δ151 males were about 30% heavier than WT and heterozygotes at 7 wk and about 50% heavier by 18 wk (Fig. 2A). The female Lep^Δ151/Δ151 rats were 40% heavier than WT at 4 wk and 52% by wk 16. The increased body weight of Lep^Δ151/Δ151 rats was associated with elevated daily food consumption in both genders (Fig. 2B).

Fig. 2.

Leptin deficiency-induced obesity, glucose intolerance, and hyperinsulinemia. A, Body weight was measured over 18 wk for WT (Lep^+/+, n = 4), heterozygous (Lep^+/Δ151, n = 8), and homozygous (Lep^Δ151/Δ151, n = 45) males and WT (n = 2) and homozygous (n = 4) females. *, P < 0.01 vs. controls. B, Means ± se of daily food intake for Lep^+/+ (n = 6) and Lep^Δ151/Δ151 (n = 6) measured over 3 wk of time. *, P < 0.05 vs. controls. C, Serum insulin levels in 4-wk-old WT and Lep^Δ151/Δ151 rats (n = 6 for each group). *, P < 0.05 vs. controls. D, Six- and 10-wk old WT and Lep^Δ151/Δ151 rats (n = 10 for each) were administrated with d-glucose, and serum glucose levels were determined at 0, 15, 30, 60, and 120 min after administration. *, P < 0.05 vs. controls.

Serum chemistry

Diabetes-associated parameters were measured through serum chemistry (Table 1). Triglyceride levels in Lep^−/− rats were twice as high as those of WT at 6 wk and over 10 times as high to 972.9 ± 72.82 mg/dl at 10 wk. The increases in cholesterol, high-density lipoprotein, and low-density lipoprotein levels were less dramatic, yet statistically significant. Serum glucose levels in fasting Lep^Δ151/Δ151 rats were statistically higher than those of WT at 10 wk of age (93.5 ± 4.60 mg/dl over 79.3 ± 5.19 in WT) but not at 6 wk. Taken together, the Lep^Δ151/Δ151 rats are hyperlipidemic and slightly hyperglycemic. We also measured serum creatinine levels and observed no significant difference between 6- and 10-wk-old WT and leptin-deficient rats, indicating normal kidney functions in the Lep^Δ151/Δ151 rats up to 10 wk of age.

Table 1.

Fasting serum chemistry at 6 and 10 wk of age

	Six weeks		Ten weeks
	WT	LepΔ151/Δ151	WT	LepΔ151/Δ151
Triglycerides (mg/dl)	98.1 ± 9.69	234.9 ± 32.5b	87 ± 11.66	972.9 ± 72.82c
Glucose (mg/dl)	95.6 ± 3.84	104.6 ± 5.96	79.3 ± 5.19	93.5 ± 4.60c
Cholesterol (mg/dl)	110.2 ± 4.57	145.8 ± 4.03b	114 ± 5.47	197.4 ± 10.49c
HDL (mg/dl)	40.1 ± 1.55	65 ± 3.14c	35.5 ± 1.22	68.5 ± 4.19c
LDL (mg/dl)	37.8 ± 1.73	43.6 ± 2.63a	28.1 ± 2.09	46.5 ± 2.40c
Creatinine (mg/dl)	0.32 ± 51	0.27 ± 32	0.29 ± 32	0.26 ± 33.7

HDL, High-density lipoprotein; LDL, low-density lipoprotein. Values are means ± se. (n = 10).

^aP < 0.05 vs. controls.

^bP < 0.005.

^cP < 0.001.

Hyperinsulinemia

As early as 4 wk of age, the Lep^Δ151/Δ151 rats showed a dramatic increase in serum insulin levels, at 0.855 ± 0.436 ng/ml compared with 0.035 ± 0.012 ng/ml in the WT rats (Fig. 2C).

Glucose intolerance

We tested glucose tolerance in Lep^Δ151/Δ151 and WT rats at 6 and 10 wk of age. After ip injection of d-glucose, blood was collected at 15, 30, 60, and 120 min to determine serum glucose concentration. Serum glucose levels were higher by over 40% (Fig. 2D) in the 6-wk-old Lep^Δ151/Δ151 rats at 60 min (253.9 ± 35.8 over 174.3 ± 15.7 mg/dl in WT) and at 120 min (136.8 ± 16.7 vs. 95.6 ± 6.1 mg/dl in WT). By 10 wk, a 30% increase was observed at 30 min after glucose administration (411.6 ± 26.98 vs. 299.4 ± 6.69 mg/dl in WT), and there was an over 50% increase at 60 min (361.2 ± 35.67 vs. 213.9 ± 6.44 mg/dl in WT) and a 30% increase at 120 min (181.6 ± 18.45 mg/dl vs. the WT value of 119.7 ± 2.62 mg/dl).

Reduced T-cell differentiation and elevated IFN-γ levels

It has been shown that leptin regulates thymic homeostasis by promoting Th1 cell differentiation and cytokine production in the mouse (10). We looked at T-cell differentiation as well as IFN-γ production in the absence of leptin. Lep^Δ151/Δ151 rats had about 30% reduction in percentage of CD3-positive (24.93 ± 3.5 vs. 35.87 ± 2.4 in the WT), CD4-positive (17.16 ± 2.20 vs. 24.8 ± 1.28 in the WT), and CD8-positive T cells (5.7 ± 1.2 vs. 8.5 ± 1.2 in the WT) (Fig. 3A). Also observed was a 2-fold increase in the percentage of IFN-γ-expressing CD4-positive (0.84 ± 0.11 vs. 0.42 ± 0.05 in the WT) and CD8-positive T cells (7.01 ± 0.78 vs. 5.18 ± 0.38 in the WT) in the Lep^Δ151/Δ151 rats (Fig. 3B).

Fig. 3.

Lep^Δ151/Δ151 rats have lower CD3 and CD4-positive T-cell population (A) and increased production of IFN-γ by CD4 and CD8-positive T cells (B). Whole-blood samples were stimulated with phorbol 12-myristate 13-acetate (PMA)/ionomycin and analyzed by flow cytometry for CD3, CD4, and CD8 cells subpopulations and for IFN-γ intracellular staining in CD4 and CD8 in Lep^Δ151/Δ151 (n = 11) and WT rats (n = 12). *, P < 0.05 vs. controls; **, P < 0.005.

Changes in bone architecture

To assess the effects of leptin disruption on rat bone metabolism, we analyzed the femur bone architecture of 8-wk-old rats by μCT scan. The three-dimensional image revealed visually noticeable increased trabecular content in the Lep^Δ151/Δ151 rats (Fig. 4A) that was confirmed by statistically significant change in each individual parameter: trabecular bone volume from 0.39 ± 0.01 (WT) to 0.67 ± 0.03 mm; trabecular thickness from 0.1 ± 0.002 (WT) to 0.16 ± 0.01 mm; BMD from 302.1 ± 5.9 (WT) to 441.1 ± 19.1 mg/cm³; and femur trabecular spaces from 0.2 ± 0.006 (WT) to 0.12 ± 0.006 mm (Fig. 4B). Cortical thickness, cortical bone fraction, and cortical porosity were not changed (Supplemental Fig. 2). In addition, we did TRAP staining of the tibias (Fig. 4C), which demonstrated a significant increase in osteoclast number from 44.75 ± 4.64 (WT) to 64.25 ± 3.38/mm and osteoclast surface from 13.55 ± 1.12 (WT) to 18.46 ± 1.6% (Fig. 4D) associated with a 2-fold increase of bone volume per tissue volume (BV/TV) from 20.93 ± 1.12 (WT) to 45.02 ± 1.88% (Fig. 4E).

Fig. 4.

Leptin disruption affected bone metabolism. A, Snapshots of μCT scan of trabecular bone of the femurs of 8-wk-old WT and Lep^Δ151/Δ151 rats. B, BV/TV (bone volume per tissue volume), Tb.N (trabecular number), Tb.Th (trabecular thickness), Tb.Sp. (trabecular spacing), BMD, and cortical area obtained from the μCT scan. C, Tibias from the same animals were fixed sectioned and TRAP (tartrate resistant acid phosphatase) stained for histomorphometric analysis. D, Osteoclast-related parameters, Oc.N (number of osteoclasts per bone surface), and Oc.S/BS (osteoclast surface per bone surface) were counted. E, BV/TV was also measured by histomorphometry on histological sections of the tibias. n = 8 for both WT and Lep^Δ151/Δ151 rats. *, P < 0.05 vs. controls.

Discussion

Obesity and obesity-related diseases impose an ever increasing risk to public health (11). The discovery of leptin in 1994 (12) and its anorexic potential were highly promising, and much has since been learned on the physiology of leptin using animal models like ob/ob and db/db mice (2). In addition to regulating body weight, leptin is involved in insulin regulation (2) and bone metabolism (4) and enhances CD4 T-cell proliferation by binding to its receptor expressed on CD4 cells (13). However, to date, leptin's clinical applications are limited (14). Leptin models in a species other than mice should be helpful.

The mouse has become the most popular animal model system owing to the convenience in its genome manipulation. However, the rat was the premium model of choice in medicine for many decades because its physiology reflects that of humans more closely in disease areas, such as cardiovascular disease, diabetes, arthritis, and many autoimmune, neurological, behavioral, and addiction disorders (15).

Using ZFN technology, we created the first rat lacking leptin and observed substantial phenotypical differences from the Zucker rats. The Lep^Δ151/Δ151 rats are obese as expected and developed hyperinsulinemia as early as 4 wk, hyperlypidemia, and mild hyperglycemia at 10 wk, whereas the Zucker rats developed glucose intolerance and insulin resistance within 12 wk of age in a mixed genetic background (a cross between Merck 3M and Sherman rats) (16), and demonstrated a delayed onset to 21–23 wk when bred into the SD background (17). We did not observe abnormal serum creatinine levels, indicating lack of nephropathy development, in our leptin-deficient rats up to 10 wk of age. Immunostaining of the whole blood showed that leptin deficiency led to reduction of CD3+, CD4+, and CD8+ T cells and increased IFN-γ production by CD8+ and CD4+ T cells, confirming leptin's involvement in immune response in rats (2). Consistent with the bone phenotype of the db/db and ob/ob mice, our Lep^Δ151/Δ151 rat demonstrated a dramatic increase in BV/TV and BMD in the femur compared with WT littermates, at as early as 8 wk of age, supporting leptin's role as an inhibitor of bone formation. On the other hand, tibias from Lep^Δ151/Δ151 rats had also higher osteoclast number per bone surface, similar to that observed in ob/ob and db/db mice (4), where hypogonadal conditions and increased corticosterone levels led to reduction of estrogen and therefore increased osteoclast activity. Interestingly, in the Zucker rat, no difference in the BMD and very mild decrease in the BV/TV of the tibia at 9 and 15 wk of age were observed compared with lean controls (18). In summary, our leptin-deficient rat is phenotypically similar to the ob/ob and db/db mice but quite distinct from the Zucker rats.

Several reasons may likely contribute to the different phenotypes observed in our Lep^Δ151/Δ151 rats and the Zucker rats. First, because leptin does not always function through the leptin receptors (19), a leptin disruption could have a broader effect than that seen in a leptin receptor knockout. Next, the Zucker rats used in some studies were a cross between Merck 3M and Sherman rats (9), and our Lep^Δ151/Δ151 is in SD background. The differences in genetic background may lead to disparate phenotypes. Being in a more commonly used background, our model allows usage of the vast amount of physiological data accumulated in sd rats. Finally, the Zucker rat carries a missense mutation in the db gene, which does not affect expression level of leptin receptor or its affinity to leptin binding (20).

Obesity-induced insulin resistance has been shown to lead to human type 2 diabetes (21). The discovery of leptin more than a decade ago created hope that it might be used therapeutically in the treatment of obesity. However, except for rare leptin-deficient individuals, both diet-induced rodent models of obesity and obese humans are minimally responsive to leptin treatment due to the development of leptin resistance in the brain and defects in transportation of leptin across the blood brain barrier (22). A leptin knockout rat would be useful for pharmacological studies aimed to overcome leptin resistance, similar to the work by Ozcan et al. (23). They elegantly demonstrated that endoplasmic reticulum stress interfered leptin signaling, leading to leptin resistance. By pretreating high-fat diet-induced obese mice and ob/ob mice with chemicals inhibiting endoplasmic reticulum stress, the same level of weight loss was achieved by significantly lower leptin doses that were otherwise required. Meanwhile, the db/db mice did not respond to the same treatments (23). Our leptin-deficient rats can be used in similar studies that the Zucker rats would not be suitable.

In conclusion, our initial characterization shows that disruption of the leptin gene in SD rats leads to increased body weight, hyperinsulinemia, glucose intolerance, immunosuppression, and higher bone mass. Our model confirms the antiosteogenic function of leptin, suggesting that this model can be used to study central control of bone remodeling. Thus, it may support the design of drugs enhancing bone formation without affecting body weight, valuable for the development of osteoporosis treatment as well as providing a new platform for preclinical testing of potential drugs against obesity and diabetes.