CD59 drives diet-induced obesity and glucose intolerance, insulin resistance, and metabolic dysfunction-associated steatotic liver disease Authors: Marian Zeibak, Netanel Karbian, Yael Riahi, Saja Baraghithy, Labiba Ahmad, Adi Tabib, Ifat Abromovitz, Bella Agranovitz, Hadar Benyamini, Eyal Gottlieb, Joseph Tam, Yuval Dor, Gil Leibowitz, Dror Mevorach Published: 2026 ABSTRACT CD59 is known as a membrane-bound regulator of the complement system that prevents the formation of the membrane attack complex on host cells. Here we report the metabolic consequences of CD59a knockout (KO) in mice fed a high-fat diet (HFD). Mice lacking CD59a were protected from the development of insulin resistance, glucose intolerance, hyperinsulinemia, obesity, and fatty liver. Mutants fed an HFD had elevated adiponectin levels and reduced leptin levels in plasma. Data from metabolic cages suggested decreased appetite and an increase in voluntary wheel activity in mutants. Liver transcriptome analysis showed a marked decrease of inflammatory and fibrotic pathways in CD59a KO mice on an HFD, and plasma and liver metabolomics were remarkably similar, indicating close correspondence between systemic and hepatic metabolic profiles. In conclusion, we uncover a noncanonical role of CD59a in the development of diet-induced insulin resistance, hyperinsulinemia, glucose intolerance, and obesity. Introduction Metabolic syndrome is defined as an unfavorable metabolic profile including abdominal obesity, dyslipidemia, and insulin resistance leading to the development of type II diabetes mellitus (T2D) and metabolic dysfunction-associated steatosis liver disease (MASLD). Obesity often induces low-grade inflammation due to activation of the innate immune system1,2. This, in turn, may contribute to the development of insulin resistance and MASLD, also known as nonalcoholic fatty liver disease (NAFLD), thereby increasing cardiovascular risk3,4. Metabolic syndrome leads to diabetes mellitus (DM), a major global human health burden, with more than 570 million people affected and an increasing incidence worldwide. T2D is the most common form of disease, accounting for >90% of diabetic patients3. T2D is a complex metabolic disease associated with insulin resistance and obesity, all of which increase the risk of developing cardiovascular complications and mortality3,5. Complement is an important defense mechanism in innate immunity, contributing to tissue homeostasis and modulation of the adaptive immune response6,7. Canonical activation of complement is known to occur through the classical, alternative, and lectin pathways, which are based on sequential cascades of enzymatic reactions. Once activated, the complement cascade generates effector compounds that are delivered to any cell surface in an indiscriminate manner; however, progression of the cascade and activation of the effector molecules are strictly regulated by multiple activators and inhibitors that modulate the complement response. These regulators are localized on cell membranes and discriminate between self- and non-self surfaces8. CD46, CD55, CD35, and CD59 are all cell membrane regulatory proteins; however, CD59 is the most important regulator of the membrane attack complex (MAC). CD59 is a membrane-bound regulator that interacts with C8 in the C5b-8 complex and inhibits the incorporation of C9, thereby preventing the formation of the lytic MAC9,10. Secondary CD59 and CD55 deficiencies are common findings in adult patients with paroxysmal nocturnal hemoglobinuria(PNH)11. A distinct primary homozygous Cys64Tyr CD59 missense mutation in seven patients was first reported by our group12–17. These patients suffered from chronic hemolysis and recurrent episodes of Guillain-Barré syndrome (GBS)-like disease beginning in early infancy. Surprisingly, refs. 18,19 showed that CD59 plays a noncanonical role in the regulation of insulin secretion. Knockdown of CD59 in β-cells impaired insulin secretion. They further showed that CD59 mediated its effects on insulin secretion via non-GPI-anchored CD59 isoforms19. Coimmunoprecipitation and proximity ligation assay showed that CD59 interacts with several proteins involved in insulin exocytosis, such as SNAP-25 and VAMP2. These interactions are essential for the proper docking and fusion of insulin-containing granules with the plasma membrane, leading to insulin secretion18,19, The Cys64Tyr CD59 mutant, which was originally described by us, rescued insulin secretion in CD59 knockdown β-cells18 despite failure of the CD59 mutant to reach the cell membrane, suggesting distinct intracellular and cell-surface functions of CD5920. In mice, CD59 exists as two isoforms—CD59a and CD59b. CD59a is broadly expressed across various tissues and effectively inhibits the formation of MAC, thereby protecting host cells from complement-mediated damage. In contrast, CD59b exhibits a more restricted expression pattern, predominantly in reproductive tissues like the testis. In humans, however, only a single form of CD59 is present21,22. Of note, CD59a knockout (KO) mice do not develop diabetes, suggesting that deficient insulin secretion might be offset by increased insulin sensitivity. In the current report, we studied the role of CD59a in diet-induced obesity. We show that while whole-body CD59a KO does not cause metabolic alterations, CD59a KO mice fed a high-fat diet (HFD) are protected from the development of obesity, insulin resistance, hyperinsulinemia, fatty liver, and steatohepatitis, likely via a central mechanism. This indicates that CD59a has a novel role in regulating systemic metabolism. Results The noncanonical role of CD59a in mice fed an HFD Observations by Krus18 and Golec, King, and Blom19,23 suggested that CD59 may have noncanonical functions, regulating insulin secretion from β cells. To further investigate this possibility, we tested the systemic metabolic effects of a 16-week HFD in CD59a KO and WT mice. HFD typically contained 45% of calories from fat, compared with 18% of calories from fat in a standard diet (SD). The HFD is a well-established approach for investigating metabolic health, including glucose metabolism, weight gain, and NAFLD. The high-fat content of the diet can induce lipid accumulation in the liver, resembling the pathogenesis of human NAFLD24. Mice placed on an HFD typically experience rapid weight gain, followed by insulin resistance and impaired glucose clearance. As expected, the body weight of WT mice fed the HFD increased compared to mice that were fed with an SD (Fig. 1). The body weight of CD59a KO mice on an SD was similar to that of the WT animals; however, on the HFD, the weight gain of CD59a KO mice was >25% less than the gain for WT mice (Fig. 1B).Fig. 1CD59a KO mice on HFD were protected from gaining weight.A Weekly measurements of murine body weight change for 16 weeks in WT and CD59a KO mice. Mice fasted for 2 h before weighing. B Percentage of initial weights of mice. Data presented as mean ± SEM for eight randomly chosen mice per group. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. CD59a KO mice on HFD were protected from gaining weight. A Weekly measurements of murine body weight change for 16 weeks in WT and CD59a KO mice. Mice fasted for 2 h before weighing. B Percentage of initial weights of mice. Data presented as mean ± SEM for eight randomly chosen mice per group. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Metabolic cages indicate that CD59a KO protects against HFD-induced obesity by reduced food intake, increased fat oxidation, and increased activity To clarify the mechanisms involved, we assessed food intake and energy expenditure of WT mice and CD59a KO mice fed a SD or a HFD using metabolic cages. In CD59a KO mice fed an SD, food consumption and meal size (Fig. 2A and Supplementary Fig. 1A) as well as carbohydrate oxidation (Fig. 2C) were increased compared to WT mice. In contrast, in CD59a KO mice fed an HFD, food consumption and meal size were markedly reduced compared to WT mice (Fig. 2A and Supplementary Fig. 1A). This was accompanied by increased fat oxidation in the KO mice, along with decreased carbohydrate oxidation in both dark and light hours (Fig. 2B, C), suggesting a metabolic shift from glucose to fat oxidation. This may explain the attenuation of diet-induced obesity in CD59a KO mice. The Respiratory Exchange Ratio (RER) is the ratio of carbon dioxide (CO2) produced to oxygen (O2) consumed during respiration. It is a measure of the type of fuel being metabolized by an organism. A respiratory exchange rate close to 1 indicates utilization of carbohydrates as the primary energy source, while values closer to 0.7 indicate that fat oxidation is predominant, and 0.8 reflects protein metabolism. RER values in CD59a KO mice on a HFD remained around 0.7 throughout the light and dark periods, lower than the values measured in WT mice on the HFD, indicating preferential fat oxidation throughout the day (Supplementary Fig. 1B).Fig. 2Metabolic effect on food consumption and utilization of CD59a KO mice.A Cumulative food intake. B Fat oxidation. C Carbohydrate oxidation. Representative coronal hypothalamic section of POMC Cre:td tomato. Mice were subjected to fluorescent hybridization for CD59a (green) and POMC (red). Magnification bars: 200 μm (n = 4). Data represent the mean ± SEM from four mice per group. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Metabolic effect on food consumption and utilization of CD59a KO mice. A Cumulative food intake. B Fat oxidation. C Carbohydrate oxidation. Representative coronal hypothalamic section of POMC Cre:td tomato. Mice were subjected to fluorescent hybridization for CD59a (green) and POMC (red). Magnification bars: 200 μm (n = 4). Data represent the mean ± SEM from four mice per group. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Cumulative measurements of total energy expenditure (TEE) showed increased energy expenditure in mice fed on an HFD compared to those on an SD, with a similar HFD effect on TEE in CD59a KO and WT mice (Fig. 3A). However, running distance and speed on the wheel were notably higher in CD59a KO mice compared to WT mice, on both the SD and the HFD (Fig. 3B–D). Ambulatory and voluntary activities, including moving in a defined space, eating, drinking, and grooming, as well as pedestrian locomotion and speed, were similar in CD59a KO mice and WT mice fed a HFD (not shown).Fig. 3Metabolic effect on energy expenditure and activity of CD59 KO mice.A Cumulative total energy expenditure (TEE). B total wheel-running distance accumulated over 24 h. C Distance measurements representing the average distance traveled by mice on a running wheel. D Wheel speed measurement (n = 4). Data represent the mean ± SEM from four mice per group. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Metabolic effect on energy expenditure and activity of CD59 KO mice. A Cumulative total energy expenditure (TEE). B total wheel-running distance accumulated over 24 h. C Distance measurements representing the average distance traveled by mice on a running wheel. D Wheel speed measurement (n = 4). Data represent the mean ± SEM from four mice per group. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. To more accurately interpret the effects of genotype and diet on TEE, we performed an ANCOVA analysis. ANCOVA analysis controlling body weight revealed that TEE was not significantly affected by HFD in KO mice, whereas WT mice exhibited a significant increase in TEE under an HFD. This finding refines the earlier observation that HFD elevated cumulative TEE in both genotypes, suggesting that the apparent rise in KO mice is largely attributable to their inherently higher locomotor activity rather than a genuine increase in metabolic rate. Indeed, KO mice displayed markedly greater voluntary running distance and speed under both diets, consistent with a hyperactive behavioral phenotype and circadian dysregulation, while their non-exercise activities and pedestrian locomotion remained comparable to WT. Thus, the ANCOVA-adjusted results indicate that when body weight and activity differences are accounted for, the TEE response to an HFD is primarily evident in WT animals, whereas KO mice maintain elevated activity-driven energy expenditure independent of diet (Supplementary Fig. 1C). KO mice exhibited higher locomotor and wheel-running activity than WT animals during the light phase under both SD and HFD conditions, accompanied by reduced sleep time, suggesting disrupted circadian regulation and diminished rest–activity segregation. In WT mice, HFD increased daytime activity and reduced sleep compared to SD, indicating diet-induced attenuation of normal nocturnal rhythmicity. In contrast, KO mice showed minimal changes in activity and sleep patterns in response to an HFD, implying that their circadian system is already dysregulated and less responsive to dietary cues (Supplementary Fig. 2). Collectively, these findings indicate that both genotype and diet influence circadian rhythmicity, with KO mice displaying blunted diurnal cycles and further HFD-associated dampening of rhythmicity primarily in WT animals. Due to the surprising finding of reduced appetite in the KO mice, we stained the hypothalamus and arcuate nucleus for CD59a. As shown in supplementary Fig. 3, CD59a is strongly expressed in neurons, including neurons that express POMC in the hypothalamus. Finally, to verify that decreased intake was a significant contributor to the effect of CD59a on reduced food intake, we performed a paired feeding test. When WT and KO animals consumed the same amount of food for 8 weeks, there was a clear reduction in the percent of weight gained by the WT mice (144.71 ± 5.53 versus 157.6 ± 5.32, p < 0.0024), indicating that reduced food intake was a significant contributor to preventing weight gain in CD59a KO (Supplementary Fig. 4). Taken together, our findings show that CD59a KO protects against HFD-induced obesity by reduced food intake and increased fat oxidation as well as behavioral changes, including increased activity. CD59a KO protects mice from diet-induced obesity, glucose intolerance, and insulin resistance Next, we studied the effects of CD59a deficiency on systemic glucose metabolism. WT mice fed on an HFD gradually developed glucose intolerance. The increase in glucose area under the curve (AUC) on the intraperitoneal glucose tolerance test (IPGTT) was seen already after 1 month and became statistically significant after 2 months on an HFD. CD59a KO attenuated the development of glucose intolerance (Fig. 4A). WT mice fed a HFD also developed insulin resistance, as shown by insulin tolerance tests (ITT) (Fig. 4B), and had higher fasting insulin levels (Fig. 5A, B), indicating insulin resistance in both fed and fast states. In contrast, insulin sensitivity remained normal, and serum insulin levels did not increase in the CD59a KO mice, even when they were fed an HFD (Figs. 4B and 5A).Fig. 4The development of glucose intolerance, insulin resistance, and increased gluconeogenesis in CD59a KO mice on a HFD.A IPGTT for WT and KO CD59a mice. Mice fasted for 16 hours (overnight fast) before the experiment. Following overnight fasting, glucose solution (0.225 g/ml) was injected intraperitoneally (1.5 g glucose per 1 kg). Glucose blood levels were measured at time 0 and 15-, 30-, 60-, and 120-min post injection. (Upper panel) Blood glucose levels (IPGTT) at different ages of WT and KO CD59a mice. (Lower panel): Calculated area under the curve (AUC). Values represent mean ± SEM for 6−9 randomly chosen male mice per group. Statistics were analyzed using one-way ANOVA with Brown–Forsythe and Welch ANOVA post hoc tests. Representative experiment of three experiments. B Insulin tolerance tests in WT and CD59a KO mice (ITT). Upper panel: ITT at various time points during progression of the diet. Lower panel: corresponding AUC. Values are shown as mean ± SEM. Mice fasted for 16 h (overnight fast) before the experiment. Following overnight fasting, insulin solution (1U/kg) was injected intraperitoneally. Values represent mean ± SEM for 6−9 randomly chosen male mice per group. Glucose blood levels were measured at time 0 and 15-, 30-, 60-, and 120-min post injection of WT and CD59a KO mice. Upper panel: ITT at various time points during progression of the diet. Lower panel: corresponding area under the curve (AUC). Values represent mean ± SEM for 6−9 randomly chosen male mice per group. Statistics were analyzed using one-way ANOVA with Brown–Forsythe and Welch ANOVA post hoc tests. Representative experiment out of three experiments. C Pyruvate tolerance test (PTT) in WT and CD59a KO mice. Upper panel: PTT at various time points during progression of the diet. Lower panel: corresponding AUC. Values represent mean ± SEM for 6−9 randomly chosen male mice per group. Mice fasted for 6 h prior to IP injection of pyruvate. 100 ul/10 g mice of 10% pyruvate solution was injected and glucose blood levels were measured at time 0 and 15-, 30-, 60-, and 120-min post injection. Statistics were analyzed by using one-way ANOVA with Brown–Forsythe and Welch ANOVA post hoc test. Representative of two experiments. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.Fig. 5Insulin resistance in mice on a HFD.A Insulin plasma levels. Blood was collected at week 18 on an HFD. We collected two blood samples from each WT and CD59a KO mice, either during ad libitum feeding or following a 4-h fast. ELISA was used for measurements. a. Insulin plasma levels of fed mice b. Insulin plasma levels of mice after 4 h of fasting. Representative of two experiments. c. HOMA-IR index values. d. QUICKI index values. Values are shown in truncated violin plots. Data represent mean ± SEM for 6−9 randomly chosen mice per group. Statistics were analyzed using one-way ANOVA with Brown–Forsythe and Welch ANOVA post hoc tests. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. B Insulin signaling pathway activation in the liver following injection of insulin in WT and CD59a KO mice. To perform Western blot studies, tissues were taken 5 min after insulin injection. Anti-pAKT did not appear in WT mice on an HFD. Anti-p110 and anti-pS6 are shown. There was a weak-to-no appearance in WT mice on an HFD but not in CD59a KO mice. Values represent mean ± SEM for 6−9 randomly chosen male mice per group. Representative of three samples. The development of glucose intolerance, insulin resistance, and increased gluconeogenesis in CD59a KO mice on a HFD. A IPGTT for WT and KO CD59a mice. Mice fasted for 16 hours (overnight fast) before the experiment. Following overnight fasting, glucose solution (0.225 g/ml) was injected intraperitoneally (1.5 g glucose per 1 kg). Glucose blood levels were measured at time 0 and 15-, 30-, 60-, and 120-min post injection. (Upper panel) Blood glucose levels (IPGTT) at different ages of WT and KO CD59a mice. (Lower panel): Calculated area under the curve (AUC). Values represent mean ± SEM for 6−9 randomly chosen male mice per group. Statistics were analyzed using one-way ANOVA with Brown–Forsythe and Welch ANOVA post hoc tests. Representative experiment of three experiments. B Insulin tolerance tests in WT and CD59a KO mice (ITT). Upper panel: ITT at various time points during progression of the diet. Lower panel: corresponding AUC. Values are shown as mean ± SEM. Mice fasted for 16 h (overnight fast) before the experiment. Following overnight fasting, insulin solution (1U/kg) was injected intraperitoneally. Values represent mean ± SEM for 6−9 randomly chosen male mice per group. Glucose blood levels were measured at time 0 and 15-, 30-, 60-, and 120-min post injection of WT and CD59a KO mice. Upper panel: ITT at various time points during progression of the diet. Lower panel: corresponding area under the curve (AUC). Values represent mean ± SEM for 6−9 randomly chosen male mice per group. Statistics were analyzed using one-way ANOVA with Brown–Forsythe and Welch ANOVA post hoc tests. Representative experiment out of three experiments. C Pyruvate tolerance test (PTT) in WT and CD59a KO mice. Upper panel: PTT at various time points during progression of the diet. Lower panel: corresponding AUC. Values represent mean ± SEM for 6−9 randomly chosen male mice per group. Mice fasted for 6 h prior to IP injection of pyruvate. 100 ul/10 g mice of 10% pyruvate solution was injected and glucose blood levels were measured at time 0 and 15-, 30-, 60-, and 120-min post injection. Statistics were analyzed by using one-way ANOVA with Brown–Forsythe and Welch ANOVA post hoc test. Representative of two experiments. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Insulin resistance in mice on a HFD. A Insulin plasma levels. Blood was collected at week 18 on an HFD. We collected two blood samples from each WT and CD59a KO mice, either during ad libitum feeding or following a 4-h fast. ELISA was used for measurements. a. Insulin plasma levels of fed mice b. Insulin plasma levels of mice after 4 h of fasting. Representative of two experiments. c. HOMA-IR index values. d. QUICKI index values. Values are shown in truncated violin plots. Data represent mean ± SEM for 6−9 randomly chosen mice per group. Statistics were analyzed using one-way ANOVA with Brown–Forsythe and Welch ANOVA post hoc tests. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. B Insulin signaling pathway activation in the liver following injection of insulin in WT and CD59a KO mice. To perform Western blot studies, tissues were taken 5 min after insulin injection. Anti-pAKT did not appear in WT mice on an HFD. Anti-p110 and anti-pS6 are shown. There was a weak-to-no appearance in WT mice on an HFD but not in CD59a KO mice. Values represent mean ± SEM for 6−9 randomly chosen male mice per group. Representative of three samples. The homeostatic model assessment of insulin resistance (HOMA-IR) quantifies insulin resistance by assessing fasting insulin and glucose levels25. We found significantly elevated levels of insulin levels in WT HFD mice compared to their KO HFD counterparts (Fig. 5A, C). The quantitative insulin sensitivity check index (QUICKI), a marker of insulin sensitivity that estimates insulin sensitivity based on fasting insulin and glucose levels26,27, was significantly higher in CD59a KO mice fed on HFD when compared to WT mice (Fig. 5A, D). These findings provide further evidence of preserved insulin sensitivity in CD59a KO mice fed a HFD, indicating that CD59a KO protects against the development of HFD-induced insulin resistance. The metabolic syndrome is associated with increased hepatic glucose production28. We further studied the effects of CD59a KO on gluconeogenesis in animals fed an HFD compared to control animals using the pyruvate tolerance test (PTT). In WT animals, the HFD increased glucose production, but this was attenuated in the CD59a KO mice (Fig. 4C). Given that pyruvate can be processed by multiple tissues and may not exclusively represent hepatic output, we further examined hepatic gluconeogenic gene expression. Transcriptomic analysis of liver tissue revealed marked downregulation of key gluconeogenic enzymes, including Pck1 and G6pc, in CD59a KO mice fed a HFD (Supplementary Fig. 7), supporting a reduction in hepatic gluconeogenesis in the absence of CD59a. Normal insulin signaling in CD59a KO mice that are fed an HFD To further investigate insulin resistance mechanisms, we examined components of the insulin signaling pathway. Insulin controls processes such as glycogen storage, glucose production inhibition, lipid synthesis, and glucose uptake. The phosphatidylinositol 3 kinase (PI3K) pathway, activated by insulin binding to its receptors, activates AKT/PKB kinase, facilitating the translocation of glucose transporter 4 (GLUT4) to the cell membrane for glucose uptake in muscles and adipose tissues. In the liver, insulin signaling suppresses gluconeogenesis and promotes glycogen synthesis. Disruptions in this pathway contribute to insulin resistance, which impairs the liver’s ability to regulate glucose homeostasis. We evaluated insulin resistance by extracting liver and muscle tissues from mice post-insulin injection and performing Western blot analysis using AKT, pAKT, PI3 kinase (p110), and pS6 antibodies (Fig. 5B). We found abrogation of insulin-stimulated Akt, PI3 kinase, and S6 phosphorylation in vivo. In contrast, in CD59a KO mice fed an HFD, insulin signal transduction was preserved. Collectively, we found that CD59a KO protected mice from diet-induced obesity, glucose intolerance, and insulin resistance. CD59a KO completely prevented the development of fatty liver and macrophage infiltration We assessed the impact of CD59a KO on whole-body fat accumulation and distribution and the development of MASLD using MRI. The HFD markedly increased visceral abdominal and hepatic fat accumulation (Fig. 6A). Total body fat volume on an HFD was ~60% greater in WT mice than in CD59a KO mice (Fig. 6B). In CD59a KO mice, total body fat was 27.5% less than in controls (Fig. 6B). Strikingly, CD59a KO completely prevented the development of hepatic steatosis (Fig. 6C–E). Immunofluorescence for the macrophage marker F4/80 showed that WT mice fed a HFD developed macrophage infiltration in addition to steatohepatitis, indicating an inflammatory response29,30, whereas CD59a KO mice were protected from the development of fatty liver and steatohepatitis (not shown).Fig. 6Liver and fat changes in WT and CD59a KO mice on an SD or an HFD.A Fat content and appearance on MRI imaging. B Total fat area from MRI scanning, n = 4. C Macroscopic appearance of liver. D Liver weight as a percentage of total body weight, n = 5–7. E H&E histology staining of the liver showing extensive fat distribution in WT and CD59a KO mice on an HFD, and quantification of liver lipid content (% lipid area) confirms reduced hepatic steatosis in CD59a KO mice under HFD conditions, n = 3. Statistics were analyzed by using two-way ANOVA with Šídák multiple comparison post hoc test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Representative of three experiments. Liver and fat changes in WT and CD59a KO mice on an SD or an HFD. A Fat content and appearance on MRI imaging. B Total fat area from MRI scanning, n = 4. C Macroscopic appearance of liver. D Liver weight as a percentage of total body weight, n = 5–7. E H&E histology staining of the liver showing extensive fat distribution in WT and CD59a KO mice on an HFD, and quantification of liver lipid content (% lipid area) confirms reduced hepatic steatosis in CD59a KO mice under HFD conditions, n = 3. Statistics were analyzed by using two-way ANOVA with Šídák multiple comparison post hoc test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Representative of three experiments. Elevated levels of adiponectin and reduced serum leptin levels in CD59a KO mice We measured two key hormones regulating body weight and insulin sensitivity: adiponectin and leptin. Adiponectin levels were higher in the plasma of CD59a KO mice fed a HFD compared to WT HFD mice (Fig. 7A). While no notable differences were detected between WT and KO mice under SD conditions, CD59a KO mice fed a HFD exhibited markedly reduced serum leptin levels compared to WT controls (Fig. 7B). Taken together, the preservation of adiponectin levels in CD59a KO mice fed a HFD indicates prevention of adipose tissue dysfunction and leptin levels are likely decreased because fat mass is reduced.Fig. 7KO of CD59a impacts the secretion of adipocytokines.A Adiponectin plasma levels (n = 3−5). B Leptin serum levels (n = 2−4). Blood was collected in week 18 and ELISA was used for measurements. Values are shown in truncated violin plots. Values represent mean ± SEM for 6−9 randomly chosen male mice. Statistics were analyzed using one-way ANOVA with Šídák multiple comparison post hoc test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Representative of two experiments. KO of CD59a impacts the secretion of adipocytokines. A Adiponectin plasma levels (n = 3−5). B Leptin serum levels (n = 2−4). Blood was collected in week 18 and ELISA was used for measurements. Values are shown in truncated violin plots. Values represent mean ± SEM for 6−9 randomly chosen male mice. Statistics were analyzed using one-way ANOVA with Šídák multiple comparison post hoc test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Representative of two experiments. Bulk RNA liver extracts in CD59a KO show a reduction in the cytokine and fibrotic signaling pathways We performed transcriptome analysis in liver and muscle extracts from WT and CD59a KO mice fed an SD or an HFD. As shown in Fig. 8A, gene expression was similar in the KO SD and WT SD groups, indicating that the absence of CD59a did not significantly alter hepatic gene expression under standard diet conditions. However, the KO mice on HFD showed a gene expression pattern that was similar to both the KO SD and WT SD groups, suggesting that the high-fat diet did not induce significant changes in gene expression in the absence of CD59a. In contrast, WT mice on an HFD exhibited a completely different gene expression pattern, highlighting the impact of the HFD on gene expression in the presence of CD59a.Fig. 8RNA bulk sequencing and metabolomic analysis of murine liver.A A heatmap showing alterations in gene expression between WT HFD and CD59a KO HFD groups, along with the WT SD and CD59a KO SD groups (n = 4). Red and blue colors indicate high and low gene expression, respectively. B Bar plot of IPA analysis revealing significant downregulation in hepatic gene levels of CD59 KO HFD mice in comparison to WT HFD mice. The X-axis corresponds to the statistical significance of the overrepresentation of the pathway in the differentially expressed genes. All the selected pathways are predicted to be downregulated (predicted activation z-score <−2). The darker the color, the higher the absolute value of the z-score. RNA bulk sequencing and metabolomic analysis of murine liver. A A heatmap showing alterations in gene expression between WT HFD and CD59a KO HFD groups, along with the WT SD and CD59a KO SD groups (n = 4). Red and blue colors indicate high and low gene expression, respectively. B Bar plot of IPA analysis revealing significant downregulation in hepatic gene levels of CD59 KO HFD mice in comparison to WT HFD mice. The X-axis corresponds to the statistical significance of the overrepresentation of the pathway in the differentially expressed genes. All the selected pathways are predicted to be downregulated (predicted activation z-score <−2). The darker the color, the higher the absolute value of the z-score. IPA analysis (Fig. 8B) showed a marked reduction in inflammatory cytokine signaling pathways in CD59a KO mice fed a HFD when compared to WT HFD mice, consistent with the reduction in the inflammatory response in the mutants. IFN-g signaling and the TH1 pathway, as well as signaling by TH2, IL-4, IL-13, TNF-a, NF-kB, and IL-17, were markedly inhibited. There was also a notable decrease in IL-12 signaling and production in macrophages, and reduced MHC-II antigen presentation, as well as neutrophil degranulation, inhibition of the IL-8 pathway, and NK cytotoxicity. Furthermore, a marked decrease in PDGF, an initiator of fibrosis, was observed together with a clear reduction in the idiopathic pro-fibrotic signaling pathway, including TGFβ signaling pathways. In relation, the GP6 signaling pathway, a major signaling receptor for collagen, as well as assembly of collagen fibrils and collagen biosynthesis, were all markedly reduced in CD59a KO on an HFD when compared to WT HFD mice. MASLD signaling pathways were significantly reduced, including the HIF1-α signaling pathway. Additional changes are shown in Supplementary Fig. 5. In muscle, there were only seven genes that significantly changed in mutants in comparison to WT on HFD: CD59a, CD59b, OSBPL6, GM43063, ADAMTS1, and SRP54a were reduced, while BNIP3 was upregulated. Thus, the muscle response of CD59a KO mice to HFD seems to be relatively mild, while major changes occurred in the liver. Liver metabolomic assessments of CD59a- deficient mice fed a HFD show increased glycine and decreased serine, oleic, and palmitoleic acid, consistent with protection from MASLD Lastly, a metabolomic analysis using LC-MS metabolomics was performed in liver and plasma (Fig. 9). Metabolomic analysis using a liver heatmap (Fig. 9A) illustrates the normalized concentrations of 100 metabolites that were found to be significantly different between the various experimental groups based on ANOVA analysis (FDR of <0.05). In enrichment analysis (Supplementary Fig. 6A), the main metabolic pathway changes are shown. There was a significant increase in glycine levels in HFD-fed mice compared to both WT HFD and WT SD mice. Interestingly, it was recently shown that early metabolic changes in MASLD that limit glycine availability also sensitize mice to xenobiotics, even at the reversible stage of this disease31, supporting the protective effect seen in CD59a KO mice against MASFLD. Of note, the levels of serine, which is a precursor of glycine, were also relatively reduced in HFD-fed WT mice.Fig. 9Mice liver and plasma metabolomic analysis.A A heatmap illustrating the normalized concentrations of 100 metabolites in the liver for which there were significantly different levels in the experimental groups based on ANOVA analysis (FDR <0.05). B Glycine, palmitoleic acid, oleic acid, and NAD+ expression in the liver in KO HFD mice compared to WT on HFD mice. C A heat map displaying the normalized levels of 100 metabolites in the plasma for which there were significantly different levels in the experimental groups according to ANOVA analysis (FDR <0.05). D Glycine and acetyl-L-carnitine levels in plasma. In B, D Values are shown as mean ± s.e.m. Statistics were analyzed by using ordinary one-way ANOVA with Šídák multiple comparison post hoc test. levels measured using LC-MS (n = 4) Mice liver and plasma metabolomic analysis. A A heatmap illustrating the normalized concentrations of 100 metabolites in the liver for which there were significantly different levels in the experimental groups based on ANOVA analysis (FDR <0.05). B Glycine, palmitoleic acid, oleic acid, and NAD+ expression in the liver in KO HFD mice compared to WT on HFD mice. C A heat map displaying the normalized levels of 100 metabolites in the plasma for which there were significantly different levels in the experimental groups according to ANOVA analysis (FDR <0.05). D Glycine and acetyl-L-carnitine levels in plasma. In B, D Values are shown as mean ± s.e.m. Statistics were analyzed by using ordinary one-way ANOVA with Šídák multiple comparison post hoc test. levels measured using LC-MS (n = 4) Oleic and palmitoleic acids were significantly reduced in CD59a KO mice when compared to WT HFD mice (Fig. 9B). Both are increased as a part of the metabolic syndrome signature and are typical in metabolically unhealthy obese individuals32. In addition, the CD59a KO HFD group demonstrated upregulation of NAD+ and its precursor, tryptophan (Fig. 9B). Elevated NAD+ levels may support oxidative phosphorylation and mitochondrial function33, along with increased fatty acid oxidation. Both are protective against weight gain and insulin resistance34. In CD59a KO HFD mice, coenzyme A, cholate, and various amino acids such as lysine, phenylalanine, threonine, leucine, and tyrosine, were upregulated (Supplementary Fig. 6B). The higher levels of acetyl coenzyme A further support the observation of increased fatty acid oxidation34. This is consistent with our earlier findings in the metabolic cages, which demonstrated elevated fatty acid oxidation (Fig. 2B). Increased levels of cholate, a bile acid, suggest a protective effect against diet-induced body mass gain35 and improved lipid digestion, potentially preventing lipid accumulation in the liver. The upregulation of bile acid metabolism in CD59a KO mice was also evident in our transcriptomic analysis. Figure 9C shows the differences in plasma metabolomics. Among the metabolites, acetyl-L-carnitine (ALCAR) was significantly elevated in KO HFD mice (Fig. 9D). ALCAR facilitates the transport of fatty acids into mitochondria for beta-oxidation, thus promoting their utilization as an energy source. In the context of the HFD, where there is an abundance of dietary fat, ALCAR can enhance fatty acid oxidation, potentially helping to mitigate lipid accumulation in tissues such as the liver and adipose tissue36. Among the enriched pathways in HFD KO mice was fatty acid degradation. Other pathways that were upregulated included linoleic acid metabolism, glycine, serine, threonine metabolism, and more (Supplementary Fig. 6A). Interestingly, low levels of glycine are known to lead to impairment of glutathione synthesis, ROS production, and ER stress, which leads to MASLD31. In addition, in alcoholic liver disease, the linoleic acid pathway, shown to ameliorate MASLD31, was downregulated. In contrast, linoleic acid increased in CD59a KO mice. There was a significant increase in linoleic acid, glycine, and threonine, with a possible impact on the development of MASLD31. Plasma results were remarkably similar to the findings for liver metabolites, indicating a close correspondence between systemic and hepatic metabolic profiles. Notably, certain metabolites were more prominently emphasized in plasma, underscoring their potential significance in influencing overall metabolic outcomes. Discussion Here, we showed a novel noncanonical role of the complement membrane regulator CD59a. Surprisingly, the lack of CD59a prevented the development of insulin resistance, glucose intolerance, obesity, and fatty liver, and mice lacking CD59a were protected. Metabolomic and transcriptome analyses of the liver and plasma supported the physiologic CD59a appearance in HFD with reversal of the alterations observed in WT mice fed a HFD, suggesting that CD59a is an important mediator of diet-induced obesity and its associated inflammation and fibrosis. Reduced feeding of the mice on an HFD, together with aversion from fat in CD59a-deficient mice and increased voluntary wheel activity, may suggest involvement of CNS mechanisms. CD59a KO mice show increased food intake on a standard diet but reduced intake on an HFD, suggesting a diet-specific regulatory role. CD59a is expressed in the hypothalamus (https://www.google.com/url?q=https://www.proteinatlas.org/ENSG00000085063-CD59/brain&sa=D&source=editors&ust=1765214865562177&usg=AOvVaw15VSLP6ZVlTq-a8Asaaorx), including POMCpos neurons (Supplementary Fig. 3). This raises the possibility that dietary cues modulate its expression and influence feeding behavior. Although we did not directly assess CD59a dynamics in response to diet, potential mechanisms may involve altered hypothalamic insulin or leptin sensitivity, central lipid sensing, or developmental compensation, highlighting avenues for future investigation. Systemic effects of CD59a deficiency appear to preserve insulin sensitivity under an HFD, reduce insulin secretion, and enhance glucose clearance, leading to lower hepatic fat accumulation and inflammation. Adiponectin levels are maintained, suggesting protection against adipose dysfunction, while reduced leptin likely reflects decreased fat mass and may indicate attenuated leptin resistance. Systemically, obesity promotes low-grade inflammation and the accumulation of immune cells in adipose tissue, liver, skeletal muscle, and pancreas, ultimately impairing glucose homeostasis. It also alters the bone marrow niche and hematopoietic stem-cell differentiation, favoring myeloid progenitor expansion and increasing marrow adiposity. Collectively, these processes remodel the bone marrow microenvironment and influence systemic metabolism37,38. The specific contribution of CD59 within hematopoietic cells, however, remains to be elucidated. Overall, CD59a may integrate central and peripheral mechanisms to regulate energy balance and metabolic homeostasis, and inflammation, but direct mechanistic studies in the CNS and bone marrow are needed to confirm its role. What is the relevance to human biology? In our cohort of 15 patients17,20 who have a CD59 Cys64Tyr mutation and completely lack functional cell surface CD59 in all cells, neither glucose intolerance, diabetes, nor fatty liver developed during a follow-up period of almost 12 years. However, all our patients are young. Further clinical follow-up, and a larger cohort are needed. Gut hormone-based medications, including GLP-1 receptor agonists (GLP-1 RA), GIP receptor agonists (GIP-RA), and glucagon receptor agonists (glucagon-RA), have demonstrated robust effects on obesity, diabetes, MASLD and cardiovascular risk in patients with and without diabetes39–41. However, these medications may have significant gastrointestinal side effects, and the clinical response varies between individuals. Given the ongoing obesity epidemic and its public health impact, there is an urgent need to develop novel therapeutic strategies. We propose that inhibition of CD59a and/or its signaling pathway could represent an innovative approach to obesity treatment, potentially narrowing the efficacy gap compared to bariatric surgery and providing an alternative therapeutic option. Further research to elucidate the mechanisms regulating energy intake, expenditure, and inflammation, which may be at least partially mediated by CD59a, remains a priority. Methods Animal studies Eight-week-old C57BL/6 and B6.129-Cd59atm1Bpm (CD59−/−) male mice were studied. CD59−/− Mice were a generous gift from Professor Morgan, Cardiff University. All experiments were performed in the pathogen-free unit of the Hebrew University-Hadassah Faculty of Medicine (Jerusalem, Israel). All animal procedures were performed in accordance with guidelines established by the Hebrew University Ethics Committee for Animal Experiments (IACUC research approval no. MD-20-16058-4). High-fat diet Mice were given unrestricted access to a pelleted HFD (45%kcal fat diet, 21% MF, 2% SBO, TD.08811, Envigo Teklad Diets, Indianapolis, IN, USA) or a standard chow diet (SD) for 17 weeks42. Mice were fasted for 2 h weekly before measuring body weight, glucose levels, and food intake. Tail vein blood was collected to monitor glucose levels using a glucometer (ACCU CHEK PERFORMA, DYN Diagnostics, Israel). Pair feeding experiment Mice were single-housed at 8 weeks of age, 2 weeks before being subjected to pair-feeding with an HFD (45%kcal fat diet, 21% MF, 2% SBO, TD.08811, Envigo Teklad Diets). Pair-fed animals, CD59 WT mice, received an equivalent amount of food based on the previous day’s intake of CD59a KO mice. Throughout the 9-week experimental period, food weight was recorded daily, and body weight was measured weekly. Food intake was calculated by subtracting the remaining food content in the hoppers from the amount added the previous day. This pair-feeding protocol ensured identical caloric intake between groups. Multi-parameter metabolic assessment Following 16 weeks of follow-up, mice were put in metabolic cages for assessment of food and water intake, energy expenditure and locomotor activity, as well as fat and carbohydrate oxidation, using the Promethion high-definition behavioral phenotyping system (Sable Instruments, Las Vegas, NV, USA). MetaScreen software version 2.2.18.0 was used for data acquisition and instrument control. Raw data were processed using ExpeData version 1.8.4 with an analysis script detailing all aspects of data transformation. Mice with free access to food and water were subjected to a standard 12 h light/12 h dark cycle, which consisted of a 48-h acclimation period followed by 24 h of sampling. Respiratory gases were measured using the GA-3 gas analyzer (Sable Systems, Inc., Las Vegas, NV, USA) employing a pull-mode negative-pressure system. Airflow was measured and controlled by FR-8 (Sable Systems) at a flow rate of 2000 mL/min. Water vapor was continuously measured, and its dilution effect on O2 and CO2 was mathematically compensated. Respiratory exchange rate (RER) was calculated as the ratio of CO2 produced (VCO2) to O2 consumed (VO2) using Eq. (1):1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\rm{RER}}={{\rm{VCO}}}_{2}/{{\rm{VO}}}_{2}$$\end{document}RER=VCO2/VO2 Total energy expenditure (TEE) was calculated using VO2 and RER, according to Eq. (2):2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\rm{TEE}}={{\rm{VO}}}_{2}\times (3.815+1.232\times {\rm{RER}})$$\end{document}TEE=VO2×(3.815+1.232×RER) Fat oxidation (FO) and carbohydrate oxidation (CHO) were calculated using VO2 and VCO2 based on Eqs. (3) and (4), respectively:3\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\rm{FO}}=1.69\times {{\rm{VO}}}_{2}-1.69\times {{\rm{VCO}}}_{2}$$\end{document}FO=1.69×VO2−1.69×VCO24\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\rm{CHO}}=4.57\times {{\rm{VCO}}}_{2}-3.23\times {{\rm{VO}}}_{2}$$\end{document}CHO=4.57×VCO2−3.23×VO2 Sleep_hrs Sum of the animal’s sleep, in hours. The animal is considered sleeping when the animal has been “quiet” for more than 40 s. An animal’s “quiet” time is defined as time during which the animal was not engaged in eating, drinking, grooming, or locomotion. Sleep_pct Sum of animal’s sleep, as a percentage of cycle (%). The animal is considered sleeping when the animal has been “quiet” for more than 40 s. Animal’s “quiet” time is defined as time in which the animal was not engaged in eating, drinking, grooming, nor locomotion. Still_pct Animal’s “quiet” time, which is the time at which the animal was not engaged in eating, drinking, grooming, or locomotion, as a percentage of the cycle (%). Still_pct is always greater than Sleep_pct. Subtract Sleep_pct from Still_pct to get the time the animal is “quiet,” but not asleep. Walk_pct Sum of animal’s time spent in ambulatory locomotion, as a percentage of cycle (%). Run_pct Sum of animal’s time spent running on wheel, as a percentage of cycle (%). AllMeters Sum of all distances traveled within the beam break system, not including any distance run on the wheel (WheelMeters), in meters (m). This includes fine movement (such as grooming and scratching) as well as direct locomotion. Wheel running and locomotive activity. Assessment of wheel running and locomotor activity was performed using the Promethion high-definition behavioral phenotyping system (Sable Systems). Wheel revolutions were measured with a monitor that recorded voluntary wheel running activity, and locomotor activity was quantified using disruptions of infrared XYZ beam arrays with a beam spacing of 0.25 cm43. Metabolic tests Glucose tolerance44, insulin sensitivity44, and pyruvate tolerance test45, were examined after HFD feeding for 4, 8, 12, and 16 weeks. Prior to the performance of metabolic tests, mice were transferred to clean cages with access to drinking water but with no food in the hopper or cage bottom. The mice were fasted overnight for 16 h prior to performance of intraperitoneal glucose tolerance (IPGTT), 4 h prior to performance of insulin tolerance test (ITT), and 6 h prior to performance of pyruvate tolerance test (PTT). IPGTT, ITT, and PTT were performed by consecutive measurements of blood glucose following IP injection of 1.5 g of glucose/kg body weight, 1 U/kg insulin (NovoRapiD), and pyruvate (1 g/kg), respectively. Blood glucose was measured at 0, 15, 30, 60, and 120 min after the intraperitoneal injection. Insulin detection Mice were fasted for 4 h before collecting blood samples. Insulin levels were determined using a commercial Insulin ELISA Kit (Crystal Chem, Chicago, IL, USA, 90095). Equations for insulin sensitivity indices. HOMA-IR:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\frac{{\mathrm{fasting}}\,{\mathrm{glucose}}\left(\frac{\mathrm{mM}}{\mathrm{L}}\right)\times {\mathrm{fasting}}\,{\mathrm{insulin}}\left(\frac{\mathrm{pU}}{\mathrm{mL}}\right)}{22.5}$$\end{document}fastingglucosemML×fastinginsulinpUmL22.5 QUICKI:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\frac{1}{\log \left[{\mathrm{fasting}}\,{\mathrm{insulin}}\left(\frac{{\rm{\mu }}\mathrm{U}}{\mathrm{mL}}\right)\right]+\log \left[{\mathrm{fasting}}\,{\mathrm{glucose}}\left(\frac{\mathrm{mg}}{\mathrm{dL}}\right)\right]}$$\end{document}1logfastinginsulinμUmL+logfastingglucosemgdL Western blot analysis and sample preparations Prior to animal sacrifice at week 16 of HFD feeding, mice were intraperitoneally injected with insulin solution (2.5 U/kg). Five minutes post injection, murine liver and muscle were harvested, immediately snap-frozen in liquid nitrogen, and stored at −80 °C until further use. After adding beads and the RIPA buffer, the frozen tissues were homogenized for 10 min. To perform Western blot (WB), the samples were diluted 1:30 by X3 sample buffer (0.5 M Tris-HCl, 6.8 pH, SDS, glycerol, and bromophenol blue), heated at 95 °C for 5 min using 10% polyacrylamide gels, and then transferred to nitrocellulose/PVDF membranes. 5% BSA solution was used as a blocking solution for 45 min. Membranes were incubated overnight at 4 °C with the first antibody, followed by extensive washes and incubation with HRP-conjugated secondary antibody (1:10,000) for 45 min20. For insulin pathway signaling, we used antibodies from Cell Signaling (Danvers, MA, USA): AKT (catalog #4691), pAKT (catalog #4060), P110 (catalog #C73F8), and PS6 (catalog #D68F8). ECL reaction imaging was done using the ChemiDoc XRP+ system (Bio-Rad, Hercules, CA, USA). Quantification of band intensity was done by ImageLab Software (Bio-Rad). The statistical comparison of band intensity was processed through an analysis of variance by Prism 6.0 (GraphPad Software) with data expressed as mean ± SEM. In vivo microPET-MRI scanning Experiments were performed at the Wohl Institute for Translational Medicine of the Hadassah-Hebrew University Medical Center. PET-MRI images were acquired on a 7 T 24 cm bore cryogen-free MR scanner based on proprietary dry-magnet technology (MR Solutions, Guildford, UK) with a 3-ring PET insert that uses the latest silicon photomultiplier (SiPM) technology. The PET subsystem contains 24 detector heads arranged in three octagons of 116 mm diameter. For MRI acquisition, a mouse quadrature RF volume coil was used. Mice were anesthetized with isoflurane vaporized with O2. Isoflurane was used at 3.0% for induction and at 1.0–2.0% for maintenance. The mice were positioned on a heated bed that allows for continuous anesthesia and breathing rate monitoring. To determine the distribution of [18F]-FDG in mice, the tracer was injected into the tail vein (230 ± 30 mCi in 200 mL). Images were analyzed using VivoQuant preclinical image postprocessing software (Invicro, Needham, MA, USA). PET-MRI raw data were processed using the standard software provided by the manufacturers. Manual tissue segmentation of kidneys, liver, muscle, heart, brain and bladder was carried out on coregistered 3D MR images. Regions of interest (ROIs) were used to calculate tissue radiotracer uptake from the reconstructed PET images43. Immunofluorescence and immunohistochemistry Adult CD59 WT and CD59a KO mice were anesthetized, and different tissues were excised and placed in 10% formalin for 24 h. Tissues were then incubated in 70% ethanol for 24 h, followed by paraffin embedding. Paraffin blocks were cut into 5-μm sections, and hematoxylin and eosin (H&E) staining was carried out. The sections of tissue used for immunofluorescence or immunohistochemistry staining were first antigen-retrieved, blocked, and then incubated overnight at 4 °C with a variety of primary antibodies: For immunofluorescence, rat anti-F4/80 (ab6640; Abcam) and rabbit anti-Glut4 (ab33780, Abcam) were used. Finally, slides were mounted with Elvanol and analyzed. Fluorescence images were obtained with a Zeiss confocal microscope fitted with a Hamamatsu ORCA-ER CCD camera (Hamamatsu City, Shizuoka Pref, Japan). Images were acquired and processed using the Zen2012 (Carl Zeiss, Oberkochen, Germany) and Photoshop software (Adobe, San Jose, CA, USA). RNA sequencing RNA was isolated from liver and muscle tissues using a TRI reagent-based protocol. A total of 1–2 µg RNA was used for the library construction. For quality control of RNA yield and library synthesis products, the RNA ScreenTape and High Sensitivity D1000 ScreenTape kits (both from Agilent Technologies), Qubit® RNA HS Assay kit, and Qubit® DNA HS Assay kit (both from Invitrogen) was used for each specific step. The KAPA Stranded mRNA-Seq Kit Illumina® Platforms kit (Roche) was used for library preparations according to the manufacturer’s protocol, and the final DNA library was eluted in 30 µL of elution buffer. Libraries were normalized and pooled together. A multiplex sample pool was loaded into a NovaSeq 6000 SP Reagent Kit v1.5 (100 cycles) (Illumina). Run conditions were in single-end and 122 bp length and loaded on the NovaSeq 6000 system machine (Illumina). Binary Base Call (BCL) output files from a Novaseq600 machine were converted to FASTQ format raw reads, using BCL to FASTQ (bcl2fastq v2.20.0.422 Copyright (c) 2007–2017 Illumina, Inc.). GSEA (cut-off independent) To determine whether a priori defined sets of genes show statistically significant, concordant differences between the two biological states, whole differential expression data were subjected to gene set enrichment analysis using GSEA46. The hallmark gene set collection from the molecular signatures database (MSigDB) was used. IPA (cut-off dependent) Significant differences in gene expression were assessed using enrichment analysis with the Ingenuity Pathway Analysis (IPA®) (QIAGEN Inc., (https://digitalinsights.qiagen.com/products-overview/discovery-insights-portfolio/analysis-and-visualization/qiagen-ipa/). Significance threshold for differentially expressed genes: padj < 0.1, basemean >5, absolute log2 foldchange >5/sqrt(basemean) + 0.3. Targeted metabolomics and analysis by LC-MS Samples preparation47 Liver and muscle: Precellys Lysing Kits comprising prefilled tubes with beads (CK14 1.4 mm ceramic beads for liver tissues and CK28 for 2.8 mm ceramic beads for muscle) were filled with 1000 μL of cold (−20 °C) metabolite extraction solvent (Methanol: Acetonitrile: Water; 5:3:2) and kept on ice. Frozen liver tissue was sliced, and ~30 mg samples were added to the homogenization tube. Samples were homogenized at ~4 °C using Precellys 24 homogenizer with the following parameters: 3 × 30 s, with a 30 s gap between each of the three cycles; 6500 RPM. The homogenates were centrifuged at 16,100×g for 10 min at 4 °C. The supernatant was collected into a microcentrifuge tube, centrifuged again and transferred to HPLC glass vials and kept at −80 °C until LC-MS analysis. About 10 ul of plasma was added to a prefilled tube with 90 ul cold (−20 °C) extraction solvent (Methanol:Acetonitrile 75:25). The extracts were vortexed for 15 min at 4 °C and centrifuged at 16,100×g for 10 min (4 °C). The supernatant was collected into a HPLC glass vial (equipped with 150 ul insert) and kept at −80 °C until LC-MS analysis. LC-MS data acquisition For all experiments described above, the same analytical method was used for data acquisition. A Thermo Ultimate 3000 HPLC system coupled to a Q- Exactive Orbitrap Mass Spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) was used with a resolution of 35,000 at 200 mass/charge ratio (m/z), electrospray ionization, and polarity switching mode to enable both positive and negative ions across a mass range of 67–1000 m/z. Spray voltage 4.5 kV (ESI+) and 3.5 kV (ESI−), sheath gas (nitrogen) flow rate 25 units, auxiliary gas (nitrogen) flow rate 15 units and capillary temperature at 320 °C. HPLC setup consisted of a ZIC-pHILIC column (SeQuant; 150 mm × 2.1 mm, 5 μm; Merck, Darmstadt, Germany). Five microliters of biological extracts were injected, and the compounds were separated using a gradient of 15 min, starting at 20% aqueous (20 mm ammonium carbonate adjusted to pH 2.0 with 0.1 of 25% ammonium hydroxide) and 80% organic (acetonitrile) and terminated with 20% acetonitrile. Flow rate and column temperature were maintained at 0.2 mL/min and 45 °C, respectively, for a total run time of 27 min. All metabolites were detected using mass accuracy below 5 ppm. Thermo Xcalibur 4.1 was used for data acquisition. Metabolomics data analysis The peak areas of metabolites were determined by using the exact mass of the singly charged ions. The mass accuracy for each identified metabolite was below 5 ppm. Its identity was further confirmed by comparing experimental retention time (RT) to RTs predetermined by analyzing an in-house mass spectrometry metabolite library of standards, which includes the IROA Sigma-Aldrich MSMLS. A TraceFinder 4.1™ (Thermo Fisher Scientific) was used for analysis. All metabolites’ signals were normalized to the tissue weight and the total sum of the signals. Metaboanalyst (PMID: 34019663) and Prism 6.0 (GraphPad Software, San Diego, CA, USA) were used for data visualization. Heatmaps of annotated metabolomics Data were created using MetaboAnalyst software 6.0 (https://www.metaboanalyst.ca/). These heatmaps were generated with log-transformed intensities and auto-scaling for normalization. Metabolite enrichment analysis, performed exclusively on annotated metabolites (HMDB ID) showing statistically significant changes (p < 0.05), utilized the Small Molecule Pathway Database (SMPDB) library and the enrichment pathway tool in MetaboAnalyst software 6.0. For this analysis, data were log-transformed and normalized using range-scaling. Statistical analysis Statistical analyses were performed using GraphPad Prism version 10 (GraphPad Software, San Diego, CA, USA). Data were presented as mean ± SEM. Specific tests used to assess statistical significance are detailed in each figure legend. A p value of <0.05 was considered statistically significant. The EE ANCOVA analysis done for this work was provided by the NIDDK Mouse Metabolic Phenotyping Centers (MMPC, www.mmpc.org) using their Energy Expenditure Analysis page (http://www.mmpc.org/shared/regression.aspx) and supported by grants DK076169 and DK115255. Limitations of the study We did not provide molecular evidence for the role of CD59a expressed in the hypothalamus and hematopoietic cells. To support the conclusion of this study, further research is needed. In addition, the autonomous role of CD59a in the liver, muscle, and immune cells should be further investigated. The relevance to human biology is a key issue that should be addressed in future clinical studies. Supplementary information ARRIVE Compliance Questionnaire Supplementary Figures ARRIVE Compliance Questionnaire Supplementary Figures