Seipin knockout (SKO) mice with germline disruption of the Bscl2 gene were generated as previously described on a C57BL/6J background (19). Unless otherwise stated, mice had ad libitum access to water and food consisting of a standard rodent chow diet (CRM (P) 801722, Special Diets Services). Mice were group-housed by gender at 20-24°C, 45-65% humidity and exposed to a 12-hour light/12-hour dark period. Fed blood glucose levels were determined at 10:00 AM from tail punctures using a glucometer (AlphaTrak^® II, Zoetisus). Body composition (fat mass and lean mass) was measured using a EchoMRI-500 analyser (Zinsser Analytic). Tissues were rapidly dissected post-mortem and either snap-frozen in liquid nitrogen and stored at -70°C or fixed in 10% neutral buffered formalin for molecular analysis and histology, respectively. All procedures were carried out in accordance with the U.K. Animals (Scientific Procedures) Act 1986 and following local ethical approval at University of Aberdeen.

Male and female mice (22- to 25-week-old) were placed in clean cages and food withheld for five hours. Mice were randomly assigned to treatment and control groups. 0.2 mg/kg liraglutide (Tocris) was intraperitoneally injected to the treatment group 45 minutes before the tolerance tests (-45 min). Phosphate-buffered saline (PBS) was used as a vehicle and injected in control mice at the same time. Basal glucose levels (0 min) were determined by glucometer readings (AlphaTrak^® II, Zoetisus) from tail punctures. Mice then received an intraperitoneal bolus of 2 mg/g D-Glucose (Sigma-Aldrich) for glucose tolerance tests (GTT), 0.75 IU/kg of human insulin (Actrapid 100 IU/mL, Novo Nordisk) for insulin tolerance tests (ITT) or 2 g/kg pyruvate (Sigma-Aldrich) for pyruvate tolerance tests (PTT) as in (20). Blood glucose levels were then monitored at 15, 30, 60, 90 and 120 minutes. Mice had ad libitum access to water throughout the experiment. For the sub-chronic treatment, mice received a daily intraperitoneal injection of either 0.2 mg/kg liraglutide (Tocris) or vehicle (PBS) 30 minutes before lights out for 14 days.

Mice were singly housed in cages automatically collecting food and water intake data (PhenoMaster, TSE Systems). Mice acclimatized to the cages for 3 days prior to measurements. Food intake and water consumption were measured over 8 consecutive days as in (20). Body weights were recorded daily.

Blood was collected by cardiac puncture in SST™ Amber tubes (BD Microtainer^®) from five-hour fasted mice. After 30 minutes at room temperature, samples were centrifuged at 12,000 x g for 10 minutes at 4°C and serum collected. Serum triglyceride contents were determined using the Triglyceride Liquid Assay (Sentinel Diagnostics) following manufacturer’s instructions. Serum insulin levels were measured using a mouse insulin ELISA kit (Mercodia).

Frozen liver tissue samples were weighed (~100 mg) and homogenised by rapid agitation in 1 mL of phosphate-buffered saline with 1 mm diameter zirconia beads (BioSpec products, #11079110zx) using Precellys-Bertin homogeniser (Bertin Technologies). Liver lysates were then centrifuged at 12,000 x g for 10 minutes at 4°C. Triglycerides levels were determined from the supernatants using the Triglyceride Liquid Assay (Sentinel Diagnostics) following manufacturer’s instructions as in (21).

RNA was extracted from frozen liver tissues using the RNeasy mini kit (Qiagen), treated with DNase I (Sigma-Aldrich), then reverse-transcribed with M-MLV reverse transcriptase (Promega). Real-time quantitative PCR was carried out on the CFX384 Touch Real-Time PCR Detection System (BIO-RAD). Gene expression was normalised to the geometric mean of three stable reference genes (Nono, Ywhaz and Hprt). Sequences and details of primers and assay probes are provided in the Supplementary Table S2 .

Liver tissue and pancreata were fixed in 10% neutral buffered formalin solution for 24 hours then subsequently embedded in paraffin. 5 µm-thick sections were stained with hematoxylin and eosin (H&E) or picrosirius red solution (Abcam #ab150681) following manufacturer’s protocol as previously described (22). For insulin staining, pancreatic sections were incubated overnight at 4°C with anti-insulin antibody (1:500, Clone L6B10, Cell Signalling). Chromogenic detection was performed with an HRP-conjugated mouse secondary antibody using 3,3’-diaminobenzidine (DAB, Vector Laboratories, #SK-4100) substrate. For β-cell area quantification, between 6 and 10 pancreata per gender and genotype and for each treatment were analysed. For each pancreas, 3 sections 150 μm apart from each other were stained for insulin. The total area occupied by insulin-positive cells was measured using QuPath software (23) and expressed as a percentage of the total pancreatic area. Researchers were blind to the group allocation and outcome assessment.

All the data are presented as mean ± SEM and analysed by unpaired two-tailed Student’s t-test or two-way ANOVA with Bonferroni or Tukey’s post-hoc test as appropriate using GraphPad Prism. P < 0.05 was considered as statistically significant.

We first investigated the acute effects of GLP-1 agonism in male and female seipin knockout (SKO) mice. Liraglutide (0.2 mg/kg) was intraperitoneally injected 45 min before performing insulin, glucose, and pyruvate tolerance tests (ITT, GTT and PTT respectively). As expected, SKO mice failed to respond to insulin during an ITT, consistent with severe insulin resistance ( Figures 1A–D ). Liraglutide treatment significantly improved insulin sensitivity in SKO mice compared to PBS-treated SKO controls. Area under the curve (AUC) analysis showed significant decrease in SKO-treated mice, returning to values similar to those observed in wild-type (WT) PBS-treated mice ( Figures 1B, D ). During a GTT, SKO mice showed impaired glucose clearance compared to WT littermate controls ( Figures 1E–H ), indicating glucose intolerance. Glucose clearance rates were significantly improved in liraglutide-treated female SKO mice and were similar to those observed in liraglutide-treated WT mice ( Figures 1E, F ). Glucose clearance in liraglutide-treated male SKO returned to levels observed in WT PBS-treated controls ( Figures 1G, H ). As reported previously, liraglutide significantly improved glucose tolerance in WT mice (24–26). A PTT is used as an indirect measure of gluconeogenesis. In response to pyruvate injection, significantly greater increases in blood glucose levels were observed in female SKO mice injected with PBS compared with similarly treated WT controls. This was completely abolished by liraglutide treatment, leading to blood glucose levels lower than those observed in WT PBS-treated mice ( Figures 1I, J ). Liraglutide treatment also partially prevented this glucose excursion in male SKO mice, but this failed to reach statistical significance ( Figures 1K, L ).

Collectively, these data show that acute liraglutide treatment significantly improves glucose tolerance, insulin sensitivity and impaired hepatic gluconeogenesis in SKO mice, with the greatest effects being observed in female SKO mice.

In addition to its insulinotropic action, GLP-1 has also been shown to reduce food intake and slow gastric emptying (27). To establish metabolic changes induced by liraglutide in mice with CGL, WT and SKO mice were given daily intraperitoneal injections of PBS (vehicle) or 0.2 mg/kg liraglutide for 14 days. We examined the effect of liraglutide administration on food intake, water consumption, body composition and non-fasted blood glucose levels, assessed over 8 consecutive days. In accordance with clinical features observed in CGL2 patients, SKO mice were hyperphagic ( Figure 2A ), attributable to the low circulating levels of leptin (28–32). Water intake was also significantly elevated in SKO mice when compared to WT mice ( Figure 2B ), which is consistent with polydipsia, a recognised feature of diabetes (33). The 2-week liraglutide administration did not significantly alter food intake in female SKO mice although there was a slight decrease around day 12 of treatment ( Figure 2C ). Liraglutide had a more pronounced effect in male SKO mice where it significantly decreased food intake by day 12 when compared to PBS-injected SKO mice ( Figure 2C ). SKO mice injected with liraglutide also showed a progressive normalisation of water consumption with a 45% and 62% reduction by day 12 in female and male SKO mice, respectively ( Figure 2D ). This likely reflects an anti-diabetic treatment effect of liraglutide on SKO mice. As shown previously (19, 28, 29, 34), lipodystrophic SKO mice display significantly reduced fat mass ( Figures 2E ; Supplementary Figure S1A ) and increased lean mass ( Figure 2F ; Supplementary Figure S1B ) relative to WT mice. When comparing metabolic parameters pre and post the 2-week treatment, we observed that liraglutide significantly lowered fat mass in male WT mice, however similar changes failed to reach statistical significance in female WT mice ( Figure 2E ). This was associated with a significant increase in lean mass in both male and female mice WT mice ( Figure 2F ). In contrast, body composition of SKO mice was not affected by liraglutide or vehicle treatment ( Figures 2E, F ; Supplementary Figures S1A, B ). We also observed increased tissue weights for heart and kidneys in female but not male SKO mice, consistent with previously reported organomegaly in this model (35). This was not altered by treatment with liraglutide for 14 days ( Supplementary Figures S1C, D ) Finally, blood glucose levels were examined in the non-fasted state before and after 10 days of treatment. Liraglutide did not alter blood glucose levels in WT mice, which remained around 10 mmol/L. Prior to the study, SKO mice were hyperglycaemic with blood glucose levels of approximately 16.4-21.4 mmol/L and 26.7-27.5 mmol/L for female and male SKO mice, respectively ( Figure 2G ). PBS-treated SKO mice had comparable blood glucose levels pre- and post-treatment. However, SKO hyperglycaemia was significantly lowered by liraglutide treatment with both male and female SKO mice displaying a ~35% reduction in blood glucose concentrations ( Figure 2G ).

These results demonstrate that liraglutide efficiently reduced polydipsia and hyperglycaemia without altering adiposity in mice with generalised lipodystrophy.

Due to an inability to appropriately store lipids, individuals with lipodystrophies typically develop metabolic dysfunction-associated steatotic liver disease (MASLD) which often progresses to hepatic steatosis and fibrosis. Recapitulating the human condition, SKO mice displayed significantly enlarged livers when compared to WT controls ( Figure 3A ) (30, 31, 36). This was associated with the substantial accumulation of lipids, principally triglycerides, within this tissue ( Figure 3B ). Daily liraglutide treatment for 14 days significantly improved liver health in SKO mice. Specifically, liver weights were decreased by 24% in male and 27% in female SKO mice ( Figure 3A ). Hepatic triglyceride content was significantly reduced by 32% in female SKO mice ( Figure 3B ) but was not different between PBS and liraglutide-treated male SKO mice ( Figure 3B ). Consistent with these findings, H&E staining of liver sections revealed significant lipid accumulation in SKO mice that was considerably reduced by liraglutide treatment in female SKO mice ( Figure 3C ) but this was not evident in liraglutide-treated male SKO mice ( Figure 3D ). Picrosirius red staining showed substantial fibrosis in both male and female SKO mice when compared to WT mice ( Figures 3C, D ). Administration of liraglutide qualitatively reduced the presence of collagen fibres in livers of SKO mice ( Figures 3C, D ). To quantitatively examine hepatic fibrosis in SKO mice, the expression of pro-fibrogenic markers were analysed. In agreement with histological staining, mRNA levels of several markers, including Mmp13, Timp1, Tgfβ and Col3a1 were significantly upregulated in both male and female PBS-treated SKO mice when compared to the PBS-treated WT group ( Figures 3E, F ). The levels of Mmp13, Timp1, PAI-1, Tgfβ and Col3a1, were significantly reduced by liraglutide treatment in female SKO mice, consistent with the histological observations ( Figure 3E ). Mmp13 expression was also significantly lower in liraglutide-treated male SKO mice ( Figure 3F ). Moreover, there was a trend toward a liraglutide-induced decrease in the expression of other markers of fibrosis examined in both male and female SKO mice, although these failed to reach statistical significance ( Figures 3E, F ). Liraglutide treatment did not alter expression of these genes in WT mice (data not shown).

Taken together, these data indicate that liraglutide treatment significantly improves hepatic steatosis and liver fibrosis in female SKO mice. These effects were less pronounced in male SKO mice, although beneficial effects were still observed on hepatomegaly and fibrosis.

Liraglutide, and GLP-1 agonists more broadly, have been shown to improve β cell function by promoting insulin secretion, β cell survival and β cell proliferation (12, 13, 37–39). Therefore, we examined whether these mechanisms contribute to the beneficial effects of liraglutide on glucose tolerance and hepatic steatosis observed in SKO mice. Fasting circulating insulin levels were ~30-fold higher in male and female SKO mice when compared to WT control mice. Two weeks of daily injections with 0.2 mg/kg liraglutide further increased insulin levels in both male and female SKO mice ( Figure 4A ). To evaluate whether this increase in insulin secretion could be attributed to an increase in β cell mass and/or an enhanced secretory capacity of the islets, immunohistochemical analysis of the pancreas and glucose-stimulated insulin secretion (GSIS) assay were performed. Insulin staining of pancreatic sections were consistent with changes in insulin secretion observed in Figure 4A with an increase in β cell mass in SKO mice relative to WT mice ( Figure 4B ). As observed in other insulin-resistant mouse models, this increase in β cell mass is likely to act as a compensatory mechanism to counteract the insulin resistance in SKO mice. Detailed morphometric analysis of the pancreas revealed a 3-fold increase in pancreatic β cell area in male and female SKO mice when compared to WT control mice ( Figure 4C ). β cell area was defined as the percentage of the total pancreatic area that was occupied by insulin-positive cells. While liraglutide- and PBS-treated SKO mice displayed comparable density of β cell area as a proportion of pancreatic area ( Figure 4C ), liraglutide treatment significantly increased pancreas weight in female SKO mice leading to an increase in total β cell mass per animal when normalised to body weight ( Figures 4D, E ). Male SKO mice displayed significantly enlarged pancreases when compared to WT mice, but this was not altered by liraglutide treatment ( Figure 4D ). As a result, total β cell mass was not significantly changed by liraglutide treatment male SKO mice, although there was a similar trend to that observed in female mice ( Figures 4D, E ). To interrogate whether liraglutide improved β cell function, we assessed GSIS in vivo ( Figures 4F, G ; Supplementary Table S1 ). Mice were fasted for five hours, and blood samples collected before and after administration of liraglutide and D-glucose to assess circulating blood glucose and insulin levels. SKO mice exhibited fasting blood glucose comparable to their WT counterparts ( Figure 4F ; Supplementary Table S1 ). Fasting insulin levels were elevated in SKO mice when compared to WT controls but to a similar extent between vehicle- and liraglutide-treated groups ( Figure 4G ; Supplementary Table S1 ). Administration of liraglutide, in the absence of glucose challenge, modestly decreased glucose levels in WT mice but had no effect on SKO mice glucose levels. Plasma insulin concentrations increased during euglycemia in liraglutide-treated SKO mice, independently of any glucose stimulus. In response to a glucose bolus, a 2-fold increase in blood glucose levels was rapidly observed in WT and SKO mice treated with vehicle, with minimal fluctuation in plasma insulin levels ( Figures 4F, G ; Supplementary Table S1 ). Liraglutide markedly reduced these glycaemic excursions in both WT and SKO mice, and this occurred concomitantly with an increase in insulin secretion ( Figure 4G ; Supplementary Table S1 ). Although variable between mice, SKO mice administered with liraglutide released very high levels of insulin in response to glucose, 8 times higher than those of PBS-treated SKO mice and 40 times higher than those of liraglutide-treated WT mice.

Together, these data illustrate that daily injections of liraglutide strongly stimulate insulin secretion in lipodystrophic mice. Not only does liraglutide appear to modulate β cell mass, but it also enhances the capacity of islets of SKO mice to secrete insulin.