FGF21 gene therapy as treatment for obesity and insulin resistance

Abstract

Prevalence of type 2 diabetes (T2D) and obesity is increasing worldwide. Currently available therapies are not suited for all patients in the heterogeneous obese/T2D population, hence the need for novel treatments. Fibroblast growth factor 21 (FGF21) is considered a promising therapeutic agent for T2D/obesity. Native FGF21 has, however, poor pharmacokinetic properties, making gene therapy an attractive strategy to achieve sustained circulating levels of this protein. Here, adeno-associated viral vectors (AAV) were used to genetically engineer liver, adipose tissue, or skeletal muscle to secrete FGF21. Treatment of animals under long-term high-fat diet feeding or of ob/ob mice resulted in marked reductions in body weight, adipose tissue hypertrophy and inflammation, hepatic steatosis, inflammation and fibrosis, and insulin resistance for > 1 year. This therapeutic effect was achieved in the absence of side effects despite continuously elevated serum FGF21. Furthermore, FGF21 overproduction in healthy animals fed a standard diet prevented the increase in weight and insulin resistance associated with aging. Our study underscores the potential of FGF21 gene therapy to treat obesity, insulin resistance, and T2D.

Keywords: AAV gene therapy; FGF21; insulin resistance; obesity; type 2 diabetes.

Figure 1. AAV8‐mediated liver gene transfer of FGF21 counteracts HFD‐induced obesity

1. A, B

   Evolution of body weight in animals treated with AAV8‐hAAT‐FGF21 as young adults (A) or as adults (B). C57Bl6 mice were fed a HFD for ˜10 weeks and then administered with 1 × 10¹⁰, 2 × 10¹⁰, or 5 × 10¹⁰ vg/mouse of AAV8‐hAAT‐FGF21 vectors. Control obese mice and control chow‐fed mice received 5 × 10¹⁰ vg of AAV8‐hAAT‐null.

2. C

   Weight of the epididymal (eWAT), inguinal (iWAT), and retroperitoneal (rWAT) white adipose tissue depots, the liver, and the quadriceps obtained from mice treated with AAV8‐hAAT‐FGF21 vectors as young adults (top panel) or as adults (bottom panel).

3. D

   Circulating levels of FGF21 at different time points after vector administration.

Data information: All values are expressed as mean ± SEM. In (A–D), young adults: AAV8‐hAAT‐null chow (n = 10 animals), AAV8‐hAAT‐null HFD (n = 8), AAV8‐hAAT‐FGF21 HFD 1 × 10¹⁰ vg (n = 9), and 5 × 10¹⁰ vg (n = 8). Adults: AAV8‐hAAT‐null chow (n = 7), AAV8‐hAAT‐null HFD (n = 7), AAV8‐hAAT‐FGF21 HFD 1 × 10¹⁰ vg (n = 7), 2 × 10¹⁰ vg (n = 8) and 5 × 10¹⁰ vg (n = 7). In (A–D), data were analyzed by one‐way ANOVA with Tukey's post hoc correction. *P < 0.05, **P < 0.01, and ***P < 0.001 versus the chow‐fed null‐injected group. ^# P < 0.05, ^## P < 0.01, and ^### P < 0.001 versus the HFD‐fed null‐injected group. HFD, high‐fat diet.

Figure 2. Reversal of WAT hypertrophy and inflammation by AAV8‐hAAT‐FGF21 treatment

1. A

   Representative images of the hematoxylin–eosin staining of the eWAT from animals fed a chow or a HFD and administered with either AAV8‐hAAT‐null or 5 × 10¹⁰ vg/mouse AAV8‐hAAT‐FGF21 vectors as young adults (left panels) or adults (right panels). While HFD‐fed, null‐injected mice had larger adipocytes, HFD‐fed, FGF21‐treated animals had adipocytes of reduced size. Scale bars: 100 μm.

2. B

   Morphometric analysis of the area of WAT adipocytes in animals treated as young adults or as adults.

3. C, D

   Circulating levels of adiponectin (C) and leptin (D).

4. E

   Immunohistochemistry for the macrophage‐specific marker Mac2 in eWAT sections from animals that received 5 × 10¹⁰ vg/mouse AAV8‐hAAT‐FGF21 as adults. The micrographs illustrate the presence of crown‐like structures (red arrows and inset) in the eWAT of HFD‐fed, null‐injected animals but not in the eWAT of HFD‐fed, FGF21‐treated mice. Scale bars: 200 μm and 50 μm (inset).

5. F–H

   Quantification by qRT–PCR of the expression of the markers of inflammation F4/80 (F), IL‐1β (G), and TNF‐α (H) in the group of animals that initiated the HFD feeding and received FGF21 vectors as adults.

Data information: All values are expressed as mean ± SEM. In (B), n = 4 animals/group. In (C, D, F–H), young adults: AAV8‐hAAT‐null chow (n = 10 animals), AAV8‐hAAT‐null HFD (n = 8), AAV8‐hAAT‐FGF21 HFD 1 × 10¹⁰ vg (n = 9), and 5 × 10¹⁰ vg (n = 8). Adults: AAV8‐hAAT‐null chow (n = 7), AAV8‐hAAT‐null HFD (n = 7), AAV8‐hAAT‐FGF21 HFD 1 × 10¹⁰ vg (n = 7), 2 × 10¹⁰ vg (n = 8), and 5 × 10¹⁰ vg (n = 7). In (B–D, F–H), data were analyzed by one‐way ANOVA with Tukey's post hoc correction. *P < 0.05, **P < 0.01, and ***P < 0.001 versus the chow‐fed null‐injected group. ^# P < 0.05, ^## P < 0.01, and ^### P < 0.001 versus the HFD‐fed null‐injected group. HFD, high‐fat diet.

Figure 3. AAV8‐hAAT‐FGF21‐mediated increased energy expenditure and decreased fat accumulation in iBAT and iWAT

1. A

   Histogram depicting the food intake of animals fed a chow or a HFD and administered with either AAV8‐hAAT‐null or AAV8‐hAAT‐FGF21 vectors as young adults or adults.

2. B

   Energy expenditure was measured with an indirect open circuit calorimeter in all experimental groups 2 months after AAV8‐hAAT‐null or AAV8‐hAAT‐FGF21 vector delivery. Data were taken during the light and dark cycles.

3. C

   Assessment of the locomotor activity through the open‐field test in animals that had been subjected to HFD feeding since ˜2 months of age and were treated with either null or FGF21‐encoding vectors 2 months later (young adults).

4. D

   Hematoxylin–eosin staining of BAT tissue sections obtained from the same cohort of animals as in (A). Scale bars: 50 μm.

5. E

   Western blot analysis of UCP1 content in BAT from the same cohort of animals as in (C). A representative immunoblot is shown (top). The histogram depicts the densitometric analysis of two different immunoblots (bottom).

6. F

   Hematoxylin–eosin staining of iWAT tissue sections obtained from the same cohort of animals as in (A). Scale bars: 100 μm.

7. G

   Quantification by qRT–PCR of the expression of Phospho1 in iWAT in the groups of animals that initiated the HFD feeding and received FGF21 vectors as young adults or adults.

Data information: All values are expressed as mean ± SEM. In (A–C, G), young adults: AAV8‐hAAT‐null chow (n = 10 animals), AAV8‐hAAT‐null HFD (n = 8), AAV8‐hAAT‐FGF21 HFD 1 × 10¹⁰ vg (n = 9), and 5 × 10¹⁰ vg (n = 8). Adults: AAV8‐hAAT‐null chow (n = 7), AAV8‐hAAT‐null HFD (n = 7), AAV8‐hAAT‐FGF21 HFD 1 × 10¹⁰ vg (n = 7), 2 × 10¹⁰ vg (n = 8), and 5 × 10¹⁰ vg (n = 7). In (E), n = 4 animals/group. In (A–C, E, G), data were analyzed by one‐way ANOVA with Tukey's post hoc correction. *P < 0.05, **P < 0.01, and ***P < 0.001 versus the chow‐fed null‐injected group. ^# P < 0.05 and ^### P < 0.001 versus the HFD‐fed null‐injected group. HFD, high‐fat diet.Source data are available online for this figure.

Figure 4. Treatment with FGF21‐encoding vectors reverses hepatic steatosis and fibrosis

1. A

   Representative images of the hematoxylin–eosin staining of liver sections obtained from animals fed a chow or a HFD and administered with either AAV8‐hAAT‐null or 5 × 10¹⁰ vg/mouse AAV8‐hAAT‐FGF21 vectors. HFD clearly induced the deposition of lipid droplets in the liver, and this was reverted by AAV8‐hAAT‐FGF21 treatment both in young adults and in adults. Scale bars: 100 μm.

2. B, C

   Fed hepatic triglyceride and cholesterol content in the same cohorts of animals.

3. D

   Immunostaining for the macrophage‐specific marker Mac2 of liver sections from animals fed a HFD that received either AAV8‐hAAT‐null or 5 × 10¹⁰ vg/mouse AAV8‐hAAT‐FGF21 vectors. Red arrows indicate the presence of crown‐like structures. Scale bars: 200 μm and 50 μm (inset).

4. E, F

   Quantification by qRT–PCR of the expression of collagen 1 in the liver in the group of animals that initiated the HFD feeding and received FGF21 vectors as young adults (E) or adults (F).

Data information: All values are expressed as mean ± SEM. In (B, C, E, F), young adults: AAV8‐hAAT‐null chow (n = 10 animals), AAV8‐hAAT‐null HFD (n = 8), AAV8‐hAAT‐FGF21 HFD 5 × 10¹⁰ vg (n = 8). Adults: AAV8‐hAAT‐null chow (n = 7), AAV8‐hAAT‐null HFD (n = 7), AAV8‐hAAT‐FGF21 HFD 5 × 10¹⁰ vg (n = 7). In (B, C, E, F), data were analyzed by one‐way ANOVA with Tukey's post hoc correction. *P < 0.05, **P < 0.01, and ***P < 0.001 versus the chow‐fed null‐injected group. ^# P < 0.05, ^## P < 0.01, and ^### P < 0.001 versus the HFD‐fed null‐injected group. HFD, high‐fat diet.

Figure EV1. AAV8‐hAAT‐FGF21 treatment improves liver fibrosis

Analysis of hepatic fibrosis through PicroSirius staining in animals fed a HFD that received 5 × 10¹⁰ vg/mouse of either AAV8‐hAAT‐null or AAV8‐hAAT‐FGF21 vectors. AAV8‐hAAT‐FGF21 treatment (right panels) markedly decreased the detection of collagen fibers that were readily detectable (in red) in animals treated with the null vector (left panels). Scale bars: 50 μm.

Figure 5. Treatment with AAV8‐hAAT‐FGF21 improves insulin sensitivity and glucose tolerance

1. A

   Fasted and fed blood glucose levels in mice fed a chow or HFD and injected with either AAV8‐hAAT‐null or different doses of AAV8‐hAAT‐FGF21 vectors as young adults or as adults.

2. B

   Fasted and fed serum insulin levels in the same groups of animals as in (A).

3. C

   β‐Cell mass in the group of animals that initiated the HFD feeding and received FGF21 vectors as adults.

4. D, E

   Insulin sensitivity was determined in all experimental groups after an intraperitoneal injection of insulin (0.75 units/kg body weight). Results were calculated as the percentage of initial blood glucose levels. During the ITT, basal (t = 0 min), peak (t = 60 min), and final (t = 90 min) glycemia were 166.4 ± 5.7, 37.9 ± 5.2, and 62.4 ± 8.2 mg/dl, respectively, in HFD‐fed mice treated as young adults with 5 × 10¹⁰ vg of AAV8‐hAAT‐FGF21 vectors.

5. F

   Glucose tolerance was studied in the group of mice that initiated the HFD feeding and received FGF21 vectors as young adults after an intraperitoneal injection of glucose (2 g/kg body weight).

6. G

   Serum insulin levels during the glucose tolerance test shown in (F).

Data information: All data represent the mean ± SEM. In (A, B, D, E), young adults: AAV8‐hAAT‐null chow (n = 10 animals), AAV8‐hAAT‐null HFD (n = 8), AAV8‐hAAT‐FGF21 HFD 1 × 10¹⁰ vg (n = 9), and 5 × 10¹⁰ vg (n = 8). Adults: AAV8‐hAAT‐null chow (n = 7), AAV8‐hAAT‐null HFD (n = 7), AAV8‐hAAT‐FGF21 HFD 1 × 10¹⁰ vg (n = 7), 2 × 10¹⁰ vg (n = 8), and 5 × 10¹⁰ vg (n = 7). In (C), adults: AAV8‐hAAT‐null chow (n = 5), AAV8‐hAAT‐null HFD (n = 5), AAV8‐hAAT‐FGF21 HFD 1 × 10¹⁰ vg (n = 4), 2 × 10¹⁰ vg (n = 5) and 5 × 10¹⁰ vg (n = 5). In (F, G), n = 6 animals/group. In (A–G), data were analyzed by one‐way ANOVA with Tukey's post hoc correction. *P < 0.05, **P < 0.01, and ***P < 0.001 versus the chow‐fed null‐injected group. ^# P < 0.05, ^## P < 0.01, and ^### P < 0.001 versus the HFD‐fed null‐injected group. HFD, high‐fat diet.

Figure EV2. AAV8‐hAAT‐FGF21‐mediated reversal of islet hyperplasia

1. A

   Fasted glucagon levels in the group of animals that initiated the HFD feeding and received FGF21 vectors as young adults.

2. B

   Representative images of the immunostaining against insulin in pancreas sections from animals that received 5 × 10¹⁰ vg/mouse of AAV8‐hAAT‐FGF21 as adults. Scale bars: 400 μm. Inset scale bars: 100 μm.

3. C

   Representative images of the double immunostaining against insulin (in green) and glucagon (in red) in pancreas sections from animals that received 5 × 10¹⁰ vg/mouse of AAV8‐hAAT‐FGF21 as young adults (upper panel) or adults (lower panel). Scale bars: 100 μm.

Data information: All values are expressed as mean ± SEM. In (A), young adults: AAV8‐hAAT‐null chow (n = 10 animals), AAV8‐hAAT‐null HFD (n = 10), AAV8‐hAAT‐FGF21 HFD 1 × 10¹⁰ vg (n = 9), and 5 × 10¹⁰ vg (n = 10). In (A), data were analyzed by one‐way ANOVA with Tukey's post hoc correction. ^# P < 0.05 and ^## P < 0.01 versus the HFD‐fed null‐injected group. HFD, High‐fat diet.

Figure EV3. No bone abnormalities were observed in AAV8‐hAAT‐FGF21‐treated animals

The long‐term effects of FGF21 gene transfer on bones were studied by comparison of HFD‐fed mice treated with the highest dose (5 × 10¹⁰ vg/mouse) of FGF21 vectors as young adults or adults with null‐injected, chow or HFD‐fed animals.

1. A

   Total naso‐anal length.

2. B

   Tibial length.

3. C–O

   Micro‐computed tomography (μCT) analysis of the epiphysis (C–J) and the diaphysis (K–O) of tibiae obtained at the time of sacrifice, that is, when animals were 18 months of age, from HFD‐fed mice administered with either null or FGF21‐encoding AAV vectors.

4. P, Q

   Circulating IGFBP1 (P) and IGF1 (Q) levels measured by ELISA.

Data information: All data represent the mean ± SEM. In (A, P, Q), Young adults: AAV8‐hAAT‐null chow (n = 10 animals), AAV8‐hAAT‐null HFD (n = 8), AAV8‐hAAT‐FGF21 HFD 1 × 10¹⁰ vg (n = 9), and 5 × 10¹⁰ vg (n = 8). Adults: AAV8‐hAAT‐null chow (n = 7), AAV8‐hAAT‐null HFD (n = 7), AAV8‐hAAT‐FGF21 HFD 1 × 10¹⁰ vg (n = 7), 2 × 10¹⁰ vg (n = 8), and 5 × 10¹⁰ vg (n = 7). In (B–O), n = 4 animals/group. In (A, B, P, Q), data were analyzed by one‐way ANOVA with Tukey's post hoc correction. In (C–O), data were analyzed by unpaired Student's t‐test. **P < 0.01 and ***P < 0.001 versus the chow‐fed null‐injected group. HFD, high‐fat diet; BMD, bone mineral density; BMC, bone mineral content; BV, bone volume; BV/TV, bone volume/tissue volume ratio; BS/BV, bone surface/bone volume ratio; Tb.N, trabecular number; Tb.Th, trabecular thickness; Tb.Sp, trabecular separation.

Figure 6. Reduced obesity and improved insulin sensitivity in ob/ob mice treated with AAV8‐hAAT‐FGF21 vectors

1. A, B

   Follow‐up over the course of 5 months of the body weight (A) and body weight gain (B) of ob/ob animals injected at 2 months of age with either 5 × 10¹¹ vg/mouse of AAV8‐hAAT‐null vectors or 1 × 10¹¹ or 5 × 10¹¹ vg/mouse of AAV8‐hAAT‐FGF21 vectors.

2. C

   Circulating levels of FGF21 were measured 2 and 5 months after vector administration.

3. D

   Representative images of the hematoxylin–eosin staining of eWAT tissue sections obtained from ob/ob animals injected with either null or FGF21‐encoding AAV vectors at 2 × 10¹⁰ or 5 × 10¹⁰ vg/mouse. Scale bars: 100 μm.

4. E

   Serum adiponectin levels in all groups.

5. F

   Representative images of the hematoxylin–eosin staining of liver tissue sections obtained from ob/ob animals injected with either null or FGF21‐encoding AAV vectors at 2 × 10¹⁰ or 5 × 10¹⁰ vg/mouse. Scale bars: 200 μm.

6. G

   Fed blood glucose levels.

7. H

   Fed serum Insulin levels.

8. I

   Insulin tolerance test after intraperitoneal injection of insulin at a dose of 0.75 units/kg body weight. Results were calculated as the percentage of initial blood glucose. Ob/ob mice treated with AAV8‐hAAT‐FGF21 vectors at any of the doses showed greater insulin sensitivity.

Data information: All data represent the mean ± SEM. In (A–C, E, G–I), AAV8‐hAAT‐null (n = 10 animals), AAV8‐hAAT‐FGF21 1 × 10¹¹ vg (n = 10), and 5 × 10¹¹ vg (n = 9). In (A–C, E, G–I), data were analyzed by one‐way ANOVA with Tukey's post hoc correction. *P < 0.05, **P < 0.01, and ***P < 0.001 versus null‐injected ob/ob group.

Figure 7. Intra‐eWAT administration of FGF21 vectors in ob/ob mice

1. A, B

   Progression of body weight (A) and body weight gain (B) in ob/ob animals treated at 11 weeks of age with an intra‐WAT injection of either AAV8‐CAG‐null vectors (1 × 10¹² vg/mouse) or AAV8‐CAG‐FGF21‐dmirT vectors at four different doses (1 × 10¹⁰, 5 × 10¹⁰, 2 × 10¹¹, 1 × 10¹² vg/mouse).

2. C

   Serum levels of FGF21 measured at the end of the 14‐week follow‐up period.

3. D

   Quantitative PCR analysis of FGF21 expression in the eWAT and iWAT fat pads of the same cohorts as in (A). The qPCR was performed with primers that specifically detected vector‐derived FGF21 mRNA.

4. E, F

   Representative images of the hematoxylin–eosin staining of (E) eWAT and (F) liver tissue sections obtained from ob/ob animals injected intra‐eWAT either null or FGF21‐encoding AAV8 vectors at all doses tested. Scale bars: 100 μm for eWAT (E) and 200 μm for liver (F).

5. G

   Glycemia in the fed state.

6. H

   Insulinemia in the fed state.

7. I

   Insulin tolerance test after intraperitoneal injection of insulin at a dose of 0.75 units/kg body weight. Results were calculated as the percentage of initial blood glucose.

Data information: All values are expressed as mean ± SEM. In (A, B, I) AAV8‐hAAT‐null (n = 7 animals), AAV8‐hAAT‐FGF21 1 × 10¹⁰ vg (n = 6), 5 × 10¹⁰ vg (n = 6), 2 × 10¹¹ vg (n = 7), and 1 × 10¹² vg (n = 8). In (C, D, G, H), AAV8‐hAAT‐null (n = 7), AAV8‐hAAT‐FGF21 1 × 10¹⁰ vg (n = 6), 5 × 10¹⁰ vg (n = 4), 2 × 10¹¹ vg (n = 7), and 1 × 10¹² vg (n = 8). In (A–D, G–I), data were analyzed by one‐way ANOVA with Tukey's post hoc correction. *P < 0.05, **P < 0.01 and ***P < 0.001 versus the null‐injected group.

Figure 8. Gene transfer of FGF21 to the skeletal muscle of healthy animals

1. A

   Circulating levels of FGF21 measured 40 weeks after injection of 3 × 10¹¹ vg/mouse of either AAV1‐CMV‐null or AAV1‐CMV‐FGF21 vectors to the skeletal muscle of healthy animals fed a chow diet.

2. B

   AAV‐derived FGF21 expression in the muscles and liver of healthy animals injected intramuscularly with AAV1‐CMV‐null or AAV1‐CMV‐FGF21 vectors.

3. C

   Evolution of the body weight in the 40‐week follow‐up period.

4. D

   Wet tissue weight of different muscles, adipose pads, and liver.

5. E, F

   Hepatic triglyceride and cholesterol content in the fed state.

6. G

   Fed serum insulin levels.

7. H

   Insulin sensitivity assessed through intraperitoneal injection of insulin (0.75 units/kg body weight) and represented as percentage of initial blood glucose.

Data information: All values are expressed as mean ± SEM. In (A–H), AAV1‐CMV‐null (n = 5 animals) and AAV1‐CMV‐FGF21 (n = 7). In (A–H), data were analyzed by unpaired Student's t‐test. *P < 0.05, **P < 0.01, and ***P < 0.001 versus the null‐injected group.