Hepatokine ERAP1 impairs skeletal muscle insulin sensitivity via ADRB2/PKA pathway

The current study aimed to investigate the role of endoplasmic reticulum aminopeptidase 1 (ERAP1), a novel hepatokine, in whole-body glucose metabolism. Here, we found that hepatic ERAP1 levels were increased in insulin-resistant leptin-receptor-mutated (db/db) and high-fat diet (HFD)-fed mice. Consistently, hepatic ERAP1 overexpression attenuated skeletal muscle (SM) insulin sensitivity, whereas knockdown ameliorated SM insulin resistance. Furthermore, serum and hepatic ERAP1 levels were positively correlated, and recombinant mouse ERAP1 or conditioned medium with high ERAP1 content (CM-ERAP1) attenuated insulin signaling in C2C12 myotubes, and CM-ERAP1 or HFD-induced insulin resistance was blocked by ERAP1 neutralizing antibodies. Mechanistically, ERAP1 reduced ADRB2 expression and interrupted ADRB2-dependent signaling in C2C12 myotubes. Finally, ERAP1 inhibition via global knockout or the inhibitor thimerosal improved insulin sensitivity. Together, ERAP1 is a hepatokine that impairs SM and whole-body insulin sensitivity, and its inhibition might provide a therapeutic strategy for diabetes, particularly for those with SM insulin resistance.


Background
The number of type 2 diabetes (T2D) patients is increasing rapidly worldwide, and this is often associated with many metabolic complications, such as hypertension and hyperlipidemia, among others1,2. T2D begins with the development of peripheral insulin resistance, and a previous study has shown that it can commonly originate within the skeletal muscle (SM)3. Because the SM is responsible for more than 80% of insulin-induced glucose uptake and disposal under normal conditions, it is considered the most important organ for whole-body glucose homeostasis, including insulin sensitivity4.
Many factors contribute to SM insulin resistance, such as increased serum triglyceride (TG) and free fatty acid (FFA) levels, as well as in ammation, among others5-7. However, insulin sensitivity in SM can also be regulated by secreted proteins from other tissues, such as the liver7-11.
The liver can secrete protein factors called hepatokines to regulate metabolism in other tissues, including the SM7-11. For example, it secretes broblast growth factor 2112, fetuin A11, ectodysplasin A8, leukocyte cell-derived chemotaxin 29, and selenoprotein P10 into circulation and regulates insulin sensitivity in the SM. Recent proteomic studies show that the liver might release hundreds to thousands of proteins into circulation13, suggesting that hepatokines are worthy of further investigation.
Endoplasmic aminopeptidase 1 (ERAP1) is a multifunctional enzyme belonging to the M1 family of zinc metallopeptidases14. Its subcellular localization is considered to be inside the endoplasmic reticulum (ER), soluble in the matrix, or anchored to the plasma membrane15. ERAP1 has the potential to trim peptide antigens to optimal lengths for binding to MHC class I molecules in the ER16. Besides, it can bind directly to some in ammatory cytokine receptors, such as type I tumor necrosis factor receptor, and promote their ectodomain shedding to generate soluble receptors and suppress in ammatory response17. ERAP1 also plays a vital role in many autoimmune diseases, such as type I diabetes18, ankylosing spondylitis19, psoriasis20, and Behcet's disease21. Further, it is abundantly expressed in the liver22 and can also be secreted from hepatocytes as a hepatokine13. The lack of hepatic ERAP1 promotes hepatocellular carcinoma growth in immune-de cient recipients and reduces the e cacy of adoptive T-cell therapy in mice23. However, a function for hepatic and serum ERAP1 in the regulation of whole-body glucose metabolism has not been indicated. Given the fact that another member of the M1 aminopeptidase family, namely placental leucine aminopeptidase, is associated with glucose uptake in adipocytes and SM cells24 and the liver is an important organ that regulates whole-body glucose homeostasis, we hypothesized that hepatic ERAP1 might play an important role in glucose metabolism as a hepatokine.
The aim of our current study was thus to investigate the role of hepatic ERAP1 in the regulation of wholebody glucose metabolism. Our work demonstrates a novel function of hepatic ERAP1 as a hepatokine that regulates SM insulin sensitivity. Moreover, the inhibition of ERAP1 might provide a new therapeutic strategy for diabetes, particularly with respect to SM insulin resistance.

Elevated ERAP1 expression in the liver is related to insulin resistance
Leptin-receptor-mutated (db/db) mice are a genetic mouse model with severe insulin resistance25. The liver, as an important organ in the regulation of whole-body glucose homeostasis, exhibits distinct gene expression patterns between db/db and C57 BL/6J wild-type (WT) mice26. Interestingly, ERAP1 expression was elevated in the livers of db/db mice, but not in white adipose tissue (WAT) or SM, which are also crucial to maintain glucose homeostasis ( Figure 1A and Figure S1A). As diet-induced glucose homeostasis dysregulation is more common in the clinic, we further examined ERAP1 expression in mice fed a high-fat diet (HFD) or a control diet for 12 weeks. Similar results were obtained as observed in db/db mice ( Figure 1B and Figure S1B).

Overexpression of ERAP1 in the liver impairs SM insulin sensitivity in WT mice
To verify the role of liver ERAP1 in regulating physical glucose homeostasis, WT mice were injected with adenovirus expressing ERAP1 (Ad-ERAP1) or control green uorescent protein (Ad-GFP) via the tail vein. First, we observed that ERAP1 was overexpressed in the livers of mice injected with Ad-ERAP1 (Figure 2A and 2B). Consistent with a role in glucose metabolism regulation, Ad-ERAP1 increased blood glucose and serum insulin levels under both at fed and fasting status, as well as the homeostatic model assessment of insulin resistance (HOMA-IR) index ( Figure 2C-2E). By performing glucose tolerance tests (GTTs) and insulin tolerance tests (ITTs), we found that the blood glucose levels decreased much more slowly the following challenge with glucose or insulin in mice injected with Ad-ERAP1 compared to that in control mice ( Figure 2F and 2G).
For further investigation, we conducted an in vivo insulin signaling assay by examining the insulinstimulated phosphorylation of insulin receptor (IR) on Tyr 1150/1151 (p-IR), protein kinase B on Ser473 (p-AKT) and glycogen synthase kinase 3β on Ser 9 (p-GSK3β)27. Consistent with the change in systemic insulin sensitivity, insulin signaling in the SM was signi cantly impaired, as demonstrated by the decreased levels of p-AKT and p-GSK3β compared to those in control mice ( Figure 2H). To our surprise, insulin signaling was not signi cantly changed in the liver or WAT, though p-IR was slightly increased in the liver of Ad-ERAP1 mice ( Figure S2).

Liver-speci c knockdown of ERAP1 ameliorates SM insulin resistance
Because the increased level of liver ERAP1 disrupted SM insulin sensitivity, we wondered if decreasing the ERAP1 level in the livers of diabetic mice could improve SM insulin sensitivity. To test this possibility, WT mice injected with negative control adenovirus (Ad-NC) or adenovirus expressing small-hairpin RNA speci c for mouse ERAP1 (Ad-shERAP1) were fed a control diet or HFD. ERAP1 protein expression was knocked down in the liver of mice injected with Ad-shERAP1 compared to that in control mice under HFD conditions ( Figure 3A). The downregulation of ERAP1 blocked HFD-increased blood glucose and serum insulin levels under both fed and fasting conditions, as well as the HOMA-IR index ( Figure 3B-3D). Blood glucose levels decreased much more quickly the following challenge with glucose or insulin in HFD mice injected with Ad-shERAP1 compared to that in control mice as measured by GTTs and ITTs and comparing the differences at each time point, respectively ( Figure 3E and 3F).
We also conducted in vivo insulin signaling assays and found that SM insulin signaling was signi cantly enhanced, as demonstrated by the increased protein levels of p-AKT and p-GSK3β ( Figure 3G). The changes in insulin signaling in the liver or WAT were not affected or not consistent with whole-body insulin sensitivity, respectively ( Figure S3). Similar results were observed in WT mice injected with Ad-shERAP1 under control diet conditions ( Figure S4). Moreover, Ad-shERAP1 also improved the parameters above in db/db mice ( Figure S5).

ERAP1 acts as a hepatokine
We next wondered how altered ERAP1 levels in the liver would affect SM insulin sensitivity. Because ERAP1 can be secreted from the liver as a hepatokine13, we speculated that circulating levels might regulate SM insulin sensitivity. As predicted, serum ERAP1 levels were elevated in db/db, HFD-fed, and Ad-ERAP1 mice, and decreased in mice with the liver-speci c knockdown of ERAP1 ( Figure 4A-4D). ERAP1 expression in the WAT and SM were not changed in Ad-ERAP1 mice ( Figure S6A), suggesting that secreted ERAP1 could not enter the WAT or SM.
To con rm that circulating ERAP1 could regulate SM insulin sensitivity, C2C12 myotubes were incubated with recombinant mouse ERAP1 (rmERAP1), and insulin signaling was impaired, as shown by the decreased levels of p-IR, p-AKT, and p-GSK3β compared to those with control treatment ( Figure 4E). rmERAP1 had no in uence on insulin signaling in primary hepatocytes or C3H10T1/2 adipocytes ( Figure   S6B and S6C). The role of secreted ERAP1 was also tested using conditioned medium with a high content of ERAP1 (CM-ERAP1) from HepG2 cells infected with Ad-ERAP1 to treat C2C12 myotubes. We rst determined that Ad-ERAP1 increased ERAP1 levels in the CM-ERAP1 ( Figure 4F). As expected, CM-ERAP1 impaired insulin signaling in C2C12 myotubes and this effect was reversed by ERAP1 neutralizing antibodies ( Figure 4G). Furthermore, we used an ERAP1 neutralizing antibody to block the effect of circulating ERAP1 in vivo. Consistently, a single injection of the ERAP1 neutralizing antibodies could improve SM insulin sensitivity in HFD-fed mice, as demonstrated by ITTs and SM insulin signaling, with no effect on the liver and WAT insulin signaling ( Figure 4H and 4I, Figure S6D and S6E).

ERAP1 regulates SM insulin sensitivity by decreasing the β2-adrenergic receptor (ADRB2) expression
To investigate the cause of SM insulin resistance mediated by serum ERAP1, we performed RNA-seq on the SM of WT mice injected with Ad-GFP or Ad-ERAP1. By conducting KEGG enrichment analysis, we found that the cyclic adenosine monophosphate (cAMP) pathway was enriched remarkably in Ad-ERAP1injected mice ( Figure 5A). As shown previously28, we also observed higher levels of Adrb2 mRNA in the SM compared to Adrb1 and Adrb3 ( Figure S7A). In addition, the expression of Adrb1 or Adrb3 was not changed in mice with hepatic overexpression of ERAP1 ( Figure S7B), suggesting the altered cAMP pathway might be regulated by Adrb2. We then examined some of the genes in this pathway, including Adrb2, carnitine palmitoyltransferase 1B, lipoprotein lipase, and catalase29-32. Interestingly, we found that most of them were downregulated, except for Adrb2, which was increased ( Figure 5B). However, considering that ERAP1 is a proteolytic enzymes33, we speculated that the protein levels of ADRB2 might be downregulated. The protein levels of ADRB2 in the SM of Ad-ERAP1 mice were downregulated as expected, as well as its downstream protein kinase A (PKA) activity34, as demonstrated by phospho-PKA (p-PKA) substrate levels ( Figure 5C). Moreover, rmERAP1 also decreased ADRB2 expression and p-PKA substrate levels in C2C12 myotubes ( Figure 5D). To further con rm the role of ADRB2 in ERAP1-induced SM insulin resistance, we examined insulin signaling in C2C12 myotubes overexpressing ADRB2 or those stimulated with the adenylyl cyclase activator forskolin35 in the presence of rmERAP1. We found that the overexpression of ADRB2 or treatment with forskolin reversed the suppressive effect of rmERAP1 on insulin signaling and p-PKA substrate levels in C2C12 myotubes ( Figure 5E and 5F). In contrast, treatment with an agonist of the β-adrenergic receptor (β-ARs) isoprenaline36 could not reverse the attenuated insulin signaling and p-PKA substrate levels caused by rmERAP1 in C2C12 myotubes ( Figure 5G).

Inhibition of ERAP1 improves insulin sensitivity
Finally, ERAP1-knockout (KO) mice and ERAP1 inhibitors were used to verify the possibility that ERAP1 could be exploited as a drug target for insulin resistance. We generated ERAP1-KO mice and found that ERAP1 levels were almost completely absent in the liver and serum ( Figure 6A and 6B). As expected, ERAP1-KO mice exhibited improved glucose metabolism with a decrease in fasting blood glucose levels and the HOMA-IR index, as well as improved GTTs and ITTs, though serum insulin was not changed, as compared to those in WT mice ( Figure 6C-6G). We then tested the effect of the ERAP1 inhibitor thimerosal on glucose metabolism in insulin-resistant db/db mice. We found that thimerosal signi cantly ameliorated insulin resistance in these mice, as shown by the corresponding changes in the aforementioned parameters examined ( Figure 7A-7E).

Discussion
Different tissues communicate with each other via secreted proteins, metabolites, microRNAs, or other molecules37. Prior studies have shown that proteins secreted from adipocytes (adipokines), such as leptin and adiponectin38,39, as well as those from SM (myokines), including irisin and myostatin40,41, regulate other tissues and whole-body glucose and lipid homeostasis. In addition to adipokines and myokines, hepatokines have gained much attention because some of them were shown to play important roles in many processes8-12, and the liver releases many hepatokines into circulation, with unknown functions that need to be identi ed13.
Our current study revealed an important function for the hepatokine ERAP1 in regulating insulin sensitivity in the SM and whole-body glucose metabolism. An important effect for hepatic ERAP1 on insulin sensitivity in the SM was shown by the observation that its expression was increased in the livers of insulin-resistant mice; further, the liver-speci c overexpression of ERAP1 impaired SM insulin sensitivity and the knockdown of ERAP1 had the opposite effects under basal or insulin-resistant conditions. While investigating the possible mechanisms underlying the hepatic ERAP1 control of insulin sensitivity in the SM, we speculated that some hepatokines might be involved in this regulation. We preferentially considered the involvement of ERAP1 in this regulation because it was previously shown that ERAP1 is a hepatokine13, and serum ERAP1 levels were positively associated with liver ERAP1 levels and increased under conditions of insulin resistance. This possibility was con rmed based on the inhibitory effect of rmERAP1 on insulin signaling in the SM in vitro, as well as the ability of neutralizing ERAP1 antibodies to reverse the effect on attenuated insulin signaling in the SM mediated by CM-ERAP1 incubation in vitro or HFD-fed mice in vivo. Individuals with insulin resistance and prediabetes present with SM insulin resistance as the earliest abnormality42, and thus, it is crucial to understand the underlying mechanism to prevent the further progression of glucose disorders. Our results provide important insights into the molecular mechanisms underlying SM insulin resistance. Moreover, our results help to understand the mechanisms of crosstalk between the liver and SM that synergistically control whole-body metabolism. Our results also suggest that extensive studies regarding hepatokines need to be carried out.
Though serum ERAP1 levels were increased under conditions of insulin resistance or in mice overexpressing ERAP1, ERAP1 expression was not increased in the SM, suggesting that it is unlikely to function by entering SM cells. As ERAP1 acts as a secreted factor, it is conceivable that it regulates SM insulin sensitivity through a membrane protein.
A previous study showed that ERAP1 could promote the shedding of some in ammatory cytokine receptors17. However, it is unlikely that ERAP1 regulates SM insulin sensitivity by targeting in ammatory cytokine receptors, as in ammation typically induces insulin resistance4. We, therefore, tried to identify other membrane proteins by conducting RNA-seq analysis of SM in Ad-ERAP1-injected WT mice and performing KEGG pathway analysis. In mice injected with Ad-ERAP1, we noticed that many of the differentially expressed genes were enriched in the cAMP signaling pathway, which is under control of β-ARs43. β-ARs are G protein-coupled receptors expressed in most tissues44. Norepinephrine binds to, and actives β-ARs and subsequently activates protein kinase A (PKA)34, which has been shown to play an important role in regulating lipid and glucose metabolism45,46. There are three subtypes of β-ARs (β1, β2, and β3) and the SM of mice preferentially expresses β2AR28. Consistent with the results of a previous report28, we also observed higher levels of Adrb2 mRNA in the SM. Even though different tissues act distinctly on the stimulation of β2AR, glucose metabolism in the SM was improved in almost all prior studies28,47. In our work, we showed that ERAP1 causes SM insulin resistance by reducing β2AR expression as demonstrated by the fact that β2AR levels were reduced by rmERAP1; further, the overexpression of β2AR or stimulation of a β2AR downstream effector reversed the attenuated insulin signaling and β2AR levels inhibited by rmERAP1 in C2C12 myotubes. However, an in vivo study will be required to con rm this mechanism. Furthermore, the molecular mechanisms underlying the ERAP1 regulation of β2AR expression is unclear. As ERAP1 is a proteolytic enzyme that promotes the ectodomain shedding of some in ammatory cytokine receptors17, we hypothesized that it might similarly reduce β2AR expression. This possibility needs to be studied in the future.
The liver-speci c knockdown of ERAP1 improves SM insulin sensitivity in mice under normal or insulinresistant conditions, suggesting that ERAP1 might be a potential drug target to treat SM insulin resistance. To further explore this possibility, we inhibited ERAP1 activity through the global knockout of ERAP1 in mice. Consistent with a previous report48, we observed that ERAP1-KO mice appear normal. However, they exhibited improved insulin sensitivity. Furthermore, our work shows that the inhibition of ERAP1 via its inhibitors or a neutralizing antibody reverses SM and whole-body insulin resistance in db/db mice or HFD-fed mice. These results suggest that hepatic and serum ERAP1 might be a potential drug target to treat insulin resistance in the SM.
However, there are still several questions. We noticed that the phosphorylation of IR was not affected by Ad-ERAP1 in vivo, but it was signi cantly inhibited by rmERAP1 or CM-ERAP1 in the SM in vitro. We speculate that there are some other factors in the serum from other tissues that can also regulate the phosphorylation of IR. This possibility requires further study. Another question is that the distinct role of serum ERAP1 in regulating insulin signaling in different tissues, as we found that ERAP1 only attenuated insulin signaling in the SM but had no signi cant effect on insulin signaling in the WAT or liver. This is possibly due to the preferential expression of different isoforms of βARs in different tissues49. ERAP1 has signi cant effects on β2AR expression, which is the dominant isoform expressed in the SM; however, the WAT mainly expresses β3AR, and the liver expresses mainly β1AR and β2AR49. Therefore, ERAP1 might have different effects on insulin signaling in different tissues. However, these possibilities need to be studied in the future.
The reasons for the upregulated expression of hepatic ERAP1 in insulin-resistant mice also remain unknown. Recent work reported that interferon γ (IFN-γ) induces interferon regulatory factor 1 expression, which is a transcription factor that increases the expression of ERAP150. Considering that db/db mice and HFD-fed mice have increased levels of IFN-γ in the liver 51, we speculate that this pathway might be involved in the regulation of ERAP1 in the livers of insulin-resistant mice. Future work will be required to explore this possibility.
Taken together, our work demonstrated that hepatic, along with serum ERAP1 levels were elevated in db/db or HFD-fed mice. Increased serum ERAP1 interrupted with ADRB2/PKA signaling and caused SM insulin resistance ( Figure 7F). The inhibition of ERAP1 by a neutralizing antibody or inhibitors could improve whole-body glucose hemostasis, and especially SM insulin sensitivity. These results provide valuable insights into the molecular mechanisms underlying SM insulin resistance, especially in a tissue crosstalk manner. Our results also suggest a potential drug target to target insulin resistance, and particularly SM insulin resistance. Because SM insulin resistance is an early sign of whole-body insulin resistance3, increased ERAP1 might also be a possible biomarker to detect the early stage of insulin resistance, which will be important to prevent the progression of diabetes.

Recombinant adenoviruses
The DNA fragments encoding ERAP1 and ADRB2 were ampli ed from mouse liver cDNA. The recombinant adenovirus expressing mouse ERAP1 (Ad-ERAP1) or ADRB2 (Ad-ADRB2) was generated using the AdEasy™ Adenoviral Vector System (Qbiogene, Irvine, CA, USA) and Ad-NC or Ad-shERAP1 was generated using the BLOCK-iT™ Adenoviral RNAi Expression System (Invitrogen, Carlsbad, CA, USA), according to the manufacturer's instructions. The shRNA sequence for mouse ERAP1 was 5′-CCAGCACCATTATTATGCATAGTCA-3′. Puri ed high-titer stocks of ampli ed recombinant adenoviruses were diluted in PBS and injected via the tail vein at a dose of 1 ´ 109 pfu /mice for a single injection54.

High-ERAP1 conditioned medium
HepG2 cells were infected with Ad-GFP or Ad-ERAP1 at a dose of 107 pfu/well in 12-well plates and changed with fresh medium 24 h later54; CM-ERAP1 was collected at 48 and 72 h, as described previously60.
Insulin resistance associated parameters In vivo insulin signaling assay Mice were fasted for 6 h prior to insulin injection, as previously described27. Small sections of the soleus muscle, WAT, and liver were excised from anesthetized live mice and kept as untreated controls. Insulin was injected at a dose of 2 U/kg into WT mice or at 5 U/kg in db/db mice via the portal vein; a small piece of the liver section was excised for western blot analysis after 3 min. Another side WAT and soleus muscle were excised after 4 and 5 min.

Relative RT-PCR and Illumina deep sequencing
Total RNA was extracted from mouse tissue samples using TRIzol reagent (Invitrogen, Waltham, MA, USA) as previously described54. mRNA levels were examined by RT-PCR with thebprimers described in Table S1. The samples were also sequenced using the Illumina HiSeq™ 4000 system at Shanghai Majorbio Bio-pharm Biotechnology Co., Ltd. (Shanghai, China). The data were analyzed with the free online platform Majorbio Cloud Platform (www.majorbio.com)

Quanti cation and statistical analysis
Statistical analysis was performed using GraphPad Prism, version 8.0 (GraphPad Software, San Diego, CA). All data are expressed as the mean ± SEM. Signi cant differences were assessed either by an unpaired two-tailed student t-test or one-way ANOVA followed by the Student-Newman-Keuls (SNK) test, as indicated. For GTTs and ITTs, a t-test or one-way ANOVA was used to compare the difference between groups at each time points examined. P < 0.05 was considered statistically signi cant.
24. Keller, S. R. The insulin-regulated aminopeptidase: a companion and regulator ofFrontiers in bioscience : a journal and virtual library 8, s410-420

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. SupplementalinformationmanuscriptNC.pdf