Abstract

Objective

   B7-H4 (B7S1, B7x, VTCN1) is an important immune checkpoint molecule
   that maintains immune homeostasis and is also expressed in pancreatic β
   cells. The polymorphism of B7-H4 influences the prevalence of Type 2
   diabetes (T2D), suggesting a potential role of B7-H4 in the
   physiological function of pancreatic β cells and the pathogenesis of
   T2D.

Methods

   β-cell-specific B7-H4 knockout mice (B7-H4 cKO mice) and their
   wild-type littermates were used to investigate the in vivo effects of
   B7-H4 on pancreatic β-cell morphology and function. AAV2/8-ins2-B7H4
   and a control virus were infused via the pancreatic intraduct into
   high-fat diet (HFD)-treated mice to elucidate the therapeutic effect of
   B7-H4. RNA sequencing was conducted on primary islets. A Luminex assay
   was used to quantify cytokine changes in B7-H4 cKO mice. Electron
   microscopy imaging was used to observe insulin secretory vesicles in
   pancreatic β cells.

Results

   Lesion of B7-H4 in β cells results in glucose intolerance due to
   reduced β-cell mass and deficient insulin secretion, whereas
   overexpression of B7-H4 in β cells ameliorates glucose intolerance in
   HFD-fed mice. Mechanistically, B7-H4 deficiency activates signal
   transducer and activator of transcription 5 (Stat5) signalling, which
   inhibits the expression of apolipoprotein F (Apof), leading to reduced
   cholesterol efflux and accumulated cholesterol in β cells, thereby
   impairing insulin processing and secretion. Overexpression of Apof in β
   cells or intraperitoneal injection of a Stat5 inhibitor reverses the
   metabolic phenotype and insulin secretion deficiency in B7-H4 cKO mice.

Conclusion

   Our study demonstrated that B7-H4 plays an important role in regulating
   β-cell mass and insulin secretion, which may shed new light on the
   development of novel strategies for T2D treatment.

   Keywords: B7-H4, Insulin secretion, Cholesterol metabolism, Apof, Type
   2 diabetes

Graphical abstract

   [45]Image 1
   [46]Open in a new tab

Highlights

     * •
       The expression of B7-H4 was evaluated in β cells from HFD-induced
       obese and genetically modified T2D mice.
     * •
       Lesions of B7-H4 in β cells result in glucose intolerance due to a
       reduced β-cell mass and deficient insulin secretion.
     * •
       B7-H4 regulates insulin secretion by modulating cholesterol
       metabolism through Stat5 signaling-regulated Apof expression.
     * •
       B7-H4 is an important linker for β cells to maintain islet immune
       homeostasis and their normal functions.

1. Introduction

   Type 2 diabetes (T2D) accounts for approximately 90% of the 537 million
   diabetes cases worldwide and has become one of the most serious public
   health problems, with increasing morbidity and mortality [[47]1]. Islet
   inflammation is believed to be a hallmark of diabetes and is involved
   in β-cell dysfunction in patients with T2D [[48]2,[49]3]. Infiltrated
   cells, molecules and cytokines in the islet microenvironment are
   critical for islet inflammation. Therapies targeting cytokines have
   yielded encouraging results [[50]4,[51]5], but molecular targets are
   still rare. To develop therapies that target the fundamental
   pathological processes of T2D, new molecular targets and therapeutic
   strategies are needed.

   Immune checkpoint molecules play important roles in regulating immune
   responses to maintain immune homeostasis. Some immune checkpoint
   molecules are also expressed in human islet β cells and play metabolic
   roles in regulating insulin secretion [[52][6], [53][7], [54][8],
   [55][9]]. Moreover, anticancer therapies that block certain immune
   checkpoint molecules can lead to immune checkpoint inhibitor-associated
   diabetes (ICI-DM), which goes beyond autoimmune insulin-dependent
   diabetes [[56]10,[57]11]. Moreover, glucolipotoxicity initiates β-cell
   death through certain immune checkpoint pathways [[58]12]. These
   studies suggest that immune checkpoint molecules may play important
   roles in the physiological function of pancreatic β cells and even in
   the pathogenesis of T2D. We previously explored the differentially
   expressed immune checkpoint genes in the pancreas of T2D patients and
   healthy controls and identified six immune checkpoint genes, including
   CD44, CD47, SIRPA, HAVCR2, TNFSF9, and VTCN1 [[59]13], that might link
   islet inflammation and β-cell dysfunction in the pathological processes
   of T2D.

   B7-H4 (B7x, B7S1 and VTCN1), the protein encoded by the VTCN1 gene, is
   a member of the B7 family of proteins expressed on the surfaces of
   antigen-presenting cells (APCs) and negatively regulates the activities
   of T cells [[60]14,[61]15]. B7-H4 is also highly expressed in numerous
   tumour tissues and is positively correlated with tumour growth and
   metastasis due to evasion of immune surveillance [[62]16]. Studies from
   our laboratory and many other groups have shown that B7-H4 has a direct
   effect on tumorigenesis by regulating the cellar behaviour of tumour
   cells [[63]17,[64]18]. B7-H4 has been reported to contribute to
   maternal–foetal immune tolerance [[65]19]. It is also important for
   immune homeostasis in pancreatic islets. The reduction in B7-H4
   expression on APCs leads to the loss of peripheral protection in
   pancreatic islets and is associated with the development of autoimmune
   diabetes, whereas early treatment of NOD mice with B7-H4 reduces the
   incidence of autoimmune diabetes [[66][20], [67][21], [68][22],
   [69][23], [70][24]]. Because B7-H4 is also expressed in pancreatic
   islet cells and colocalizes with insulin [[71]25], B7-H4 may play a
   physiological role in the function of β cells. Moreover, the
   polymorphism of B7-H4 influences the prevalence of T2D [[72]26],
   suggesting that B7-H4 may play important roles in the pathogenesis of
   T2D.

   The production and secretion of insulin is an important function of
   pancreatic β cells and involves complex processes that include the
   synthesis of preproinsulin, trans-Golgi network sorting of proinsulin,
   the formation of immature secretory vesicles, the maturation of insulin
   secretory vesicles, the transport of insulin secretory vesicles and the
   exocytosis of insulin vesicles at the plasma membrane [[73]27].
   Cholesterol, an important component of the phospholipid bilayer of the
   cell membrane, is critical for the maintenance of the β-cell membrane
   structure and is thus essential for glucose-stimulated insulin
   secretion. Alterations in the cholesterol content of β cells lead to
   enlarged insulin secretory vesicles and reduced glucose-stimulated
   insulin secretion [[74][28], [75][29], [76][30], [77][31]]. Cholesterol
   homeostasis is also important for immune cell functions. Apolipoprotein
   L, a specific and minor component of HDL, is a conserved gene family
   involved in innate immunity. APOL genes have recently been reported as
   novel regulators of islet inflammation and may contribute to β-cell
   damage during the development of diabetes [[78]32]. These studies
   suggest that cholesterol metabolism may be linked to islet inflammation
   and β-cell dysfunction.

   Signal transducer and activator of transcription 5 (Stat5, Stat5A and
   Stat5B in mice) is a member of the STAT transcription factor family.
   Stat5 is an essential regulator of cytokines or growth factors that
   induce cell survival and proliferation, as well as the crosstalk
   between STAT signalling and metabolism. Stat5 directly and indirectly
   regulates various metabolism-related processes in cancer cells and the
   metabolism of immune cells in the tumour microenvironment, playing
   crucial roles in the metabolic reprogramming of tumours [[79]33]. Stat5
   is also involved in metabolic disorders. Knockout of Stat5 in T cells
   ameliorates high-cholesterol and high-fat diet-induced
   hypercholesterolemia by influencing cholesterol metabolism in the liver
   [[80]34]. Deletion of the liver-specific Stat5 gene alters the
   expression of bile acid metabolism genes and reduces liver damage in
   lithogenic diet-fed mice [[81]35]. Studies on the role of Stat5 in
   β-cell functions are rare. Previous studies have indicated that Stat5
   activity in β cells influences susceptibility to experimentally induced
   type 1 and type 2 diabetes [[82]36], but the mechanism is still
   unclear.

   In the present study, we demonstrated that the immune checkpoint
   molecule B7-H4 is an important molecular linker for β cells to maintain
   islet immune homeostasis and their normal functions, regulating β-cell
   mass and insulin processing and secretion. Specifically, knocking out
   B7-H4 in β cells results in glucose intolerance due to decreased β-cell
   mass and deficient insulin secretion, whereas overexpressing B7-H4 in β
   cells ameliorates glucose intolerance in HFD-fed mice. The role of
   B7-H4 in regulating insulin processing and secretion is attributed to
   cholesterol metabolism regulated by the Stat5 signalling pathway. B7-H4
   decreases Stat5 signalling to upregulate the expression of Apof,
   followed by increased cholesterol outflow and glucose-induced insulin
   secretion. These findings provide new information for the development
   of strategies to treat T2D.

2. Materials and methods

2.1. Mice

   C57BL/6J genetic background B7-H4 fl/fl mice, in which exon 2 of the
   Vtcn1 gene is flanked by two loxP sites, were generated by Cyagen
   Biosciences (Suzhou, China) with CRISPER/Cas9 system.
   B6.Cg-Tg(Ins2-cre)25Mgn/J, commonly called Rip-Cre mice in which Cre
   recombinase is under the control of the rat insulin II promoter and
   B6.FVB-Tg(Pdx1-cre)6Tuv/J, commonly called Pdx-Cre mice in which Cre
   recombinase is under the transcriptional control of the mouse Pdx1
   promoter, were purchased from the Jackson Laboratory. 8-weeks old
   genetic type 2 diabetic models db/db mice were purchased from Shanghai
   laboratory animal center. Mice homozygous for the floxed B7-H4 allele
   (Flox) were crossed with Rip-Cre or Pdx-Cre transgenic mice to create
   β-cell–specific B7-H4 knockout mice (Rip-Cre; B7-H4 fl/fl; B7-H4 cKO or
   Pdx-Cre; B7-H4 fl/fl; PB7-H4 cKO) and their control littermates.
   Genotyping was identified by PCR on genomic DNA obtained from toes.
   Assessment of deletion efficiency were performed by quantitative PCR on
   cDNA obtained from islets. To avoid interference of the Cre transgene
   in the mouse metabolic phenotype [[83]37], we used Cre and Flox mice as
   littermate controls. All animal experiments were approved by the
   Laboratory Animal Ethics Committee in Ninth People's Hospital
   Affiliated to Shanghai Jiao Tong University School of Medicine
   (Approval No.: SH9H-2023-A824-1), according to the Guide for the Care
   and Use of Laboratory Animals prepared by the National Academy of
   Sciences and published by the National Institutes of Health.

   The mice were housed in groups of 4–6 per cage under standard housing
   conditions with a temperature-controlled (22 °C ± 2 °C) animal room on
   a 12-hour light/dark cycle with food and water available ad libitum.
   Mice were fed with a normal chow diet (NCD) or high-fat diet (HFD) (60%
   fat, 20% protein, and 20% carbohydrate) (catalog no. D12492; Research
   Diets, New Brunswick, NJ).

2.2. Cell line culture and high glucose, palmitate and IL-1β treatment

   The murine insulinoma derived MIN6 cells were cultured in Dulbecco's
   Modified Eagle Medium (DMEM) (Gibco, 11965092) with 15% fetal bovine
   serum (Gibco, A3160801), 100 mg/mL streptomycin, 100 units/mL
   penicillin (Gibco, 15140-122), 10 mmol/L HEPES (Gibco, 15630-080),
   1 mmol/L Sodium pyruvate (Gibco, 11360-070), and 50 μM
   β-Mercaptoethanol (Sigma-Aldrich, 444203) at 37 °C in 5% CO2
   incubators. For high-glucose treatment, MIN6 cells were incubated in
   modified DMEM medium (Gibco, 11885084, contains 5.5 mM glucose) for 48h
   with various concentrations of glucose (low glucose: 5.5 mmol/L
   glucose + 19.5 mmol/L mannitol; high glucose: 25 mmol/L glucose). For
   palmitate and IL-1β treatment, MIN6 cells were incubated in completed
   DMEM medium for 48h with various concentrations of palmitate
   (Sigma-Aldrich, 0, 0.1, 0.2, 0.3, 0.4, 0.5 mmol/L) and recombinant
   mouse IL-1β (Peprotech, 0, 2, 4, 6, 8, 10 ng/mL).

2.3. Metabolic assays

   Body weight and food intake were estimated every week from 5 weeks old
   to 10 weeks. For glucose tolerance tests (GTTs), mice were injected
   with glucose (2 g per kg of body weight for NCD-fed mice and 1 g per kg
   of body weight for HFD-fed mice) intraperitoneally (i.p.) after 16 h of
   fasting. Blood glucose levels were measured at 0,15, 30, 60, 90, and
   120 min after injection from the tail vein with a glucometer monitor
   (Jonson), and plasma insulin levels were measured at 0, 5, 15, 30, and
   60 min after injection using Highly Sensitive Mouse Insulin ELISA kit
   (EZassy, MS100). For insulin tolerance tests (ITTs), mice were fasted
   for 6 h in the morning and i.p. injected with human insulin (Novorapid,
   Novo Nordisk) (1unite per kg of body weight) diluted in saline
   solution, and measured blood glucose levels at 0, 15, 30, 60, and
   90 min after injection. Blood was obtained after an overnight fast to
   assess the fasting glucose, insulin, glucagon, triglyceride (TG), and
   total cholesterol (TC). Serum TG were evaluated using Triglyceride
   Assay Kit (Nanjing Jiancheng, A110-1-1) and TC were measured using
   LabAssay Cholesterol Kit (Wako, KBR0082). Serum glucagon was evaluated
   using Mouse Glucogon EIL Kit (RayBio, EIAM-GLU-1).

2.4. Pancreatic intraductal viral infusion

   To elucidate the therapeutic effect of B7-H4 on obesity associated
   metabolic dysfunction, 6- weeks-old Flox mice fed with 8 weeks of NCD
   or HFD were then injected with adeno-associated virus serotype 2/8
   vector (AAV2/8)-rat insulin2 promoter (Rip)-specific B7-H4-ZSGreen
   (AAV2/8-ins2-B7H4) or with control virus (AAV2/8-Ctrl) (Hanbio
   Biotechnology Co.Ltd., Shanghai, China). To clarify the role of Apof in
   β cell function affected by β-cell–specific B7-H4 knockout, 6-weeks-old
   Flox and B7-H4 cKO mice fed with 8 weeks of HFD were injected with
   AAV2/8-Apof-3xFlag-eGFP or control virus (AAV2/8-eGFP) (Genomeditech,
   Shanghai, China).The viruses (10^12 genome copy particles/mL) were
   infused at a rate of 6 μL/min by pancreatic intraductal viral infusion
   as described previously [[84]38].

2.5. Stat5 inhibitor treatment

   To ascertain whether enhanced Stat5 is involved in β cell dysfunction
   in B7-H4 cKO mice, 8 weeks-old B7H4 cKO mice were subsequently
   randomized to intraperitoneally injected with either vehicle control
   (VC) or pimozide, Stat5 inhibitor (Selleck, S4358, CAS: 2062-78-4,
   10 mg/kg body weight daily) for 2 weeks.

2.6. Primary pancreatic islet isolation and culture

   To isolate mouse islets, pancreata were perfused through the common
   bile duct with a HBSS collagenase solution (1  g/L collagenase P
   33768624 and 1.5  g/L DNAse I 1014159001, Roche) and digested in the
   same solution in a 37 °C water bath for 25 min. After shaking for 15 s,
   pancreata were washed three times with HBSS supplemented with 0.5%
   bovine serum albumin (BSA) with centrifuge 1500g for 1 min. Islets were
   handpicked under a dissection microscope using a pipette with a 10 μL
   tip and pooled into a dish with DMEM (Gibco, 11885084) containing
   5.5 mM glucose, 100 units/mL penicillin, 100 mg/mL streptomycin and 10%
   FCS. Islets were used directly for Western blot, RNA isolation or cell
   culture in humid environment containing 5% CO2.

2.7. In vitro glucose stimulated insulin secretion assays

   For in vitro insulin secretion stimulation assays, 10 islets were
   placed in each well of the 48-well plate and cultured overnight in
   300 μl complete medium and then pre-incubated for 30 min in modified
   Krebs-Ringer bicarbonate buffer (KRB; 115 mM NaCl, 4.7 mM KCl, 2.6 mM
   CaCl[2]·2(H2O), 1.2 mM KH[2]PO[4], 1.2 mM MgSO[4]·(2H[2]O), 10 mM
   HEPES, 0.5% bovine serum albumin, pH 7.4) containing 2.8 mM glucose.
   KRB was then replaced by KRB with 2.8 mM glucose and collected after
   1 h to determine the basal insulin release. This was followed by 1 h
   incubation in KRB with 16.7 mM glucose to determine the glucose
   stimulated insulin release (GSIS). After supernatant collection, islet
   protein content was extracted with 0.18 N hydrochloride acid in 70%
   ethanol to measure insulin and protein concentration (10 islets in
   100 μL) with Wide Filed Mouse Insulin ELISA kit (EZassy, MS300) and BCA
   Kit (Solarbio, PC0020) respectively.

2.8. Histological analysis

   For islet β cell and α cell area analysis, the pancreata (divided into
   three parts including the head, body and tail, three sections per
   animal, [85]Figure 3A) of Flox and B7H4 cKO mice were fixed overnight
   in 4% paraformaldehyde at 4 °C, followed by paraffin embedding.
   Sections were deparaffinized, re-hydrated and incubated overnight at
   4 °C with primary antibody against insulin (1:200) (Abcam,
   [86]Ab181547) and glucagon (1:200) (Cell Signaling Technology, 2760S),
   followed by detection with FITC- (1:1,000) (Abcam, [87]Ab150077) and
   Cy3-conjugated secondary antibody (1:500) (Abcam, Ab6939) respectively,
   counterstained with DAPI and mounted with Mounting medium (Dako).
   Pancreas sections were analyzed using a Hamamatsu flash 4.0 camera with
   a 4× Nikon Objective NA 0.2 and automatically scanned using a Nikon NiE
   with Prior PL-200 slide loading robot. Insulin+ and glucagon + area and
   distribution were calculated with NIS-Elements software (Nikon).
   Anti-B7-H4(1:200) (Abcam, [88]Ab108336), anti-CD31(1:400) (Abcam,
   [89]Ab182981), anti-insulin, anti-glucagon and anti-F4/80 (1:400)
   (Abcam, Ab6640) were used for detection of islet B7-H4 expression and
   location, Anti-Cleaved Caspase-3(1:400) (Cell Signaling Technology,
   9661) and anti-insulin were used for detection of β cell apoptosis. All
   images in TIFF format were analyzed in ImageJ and Imaris (Oxford
   instruments) after background noise reduction and contrast adjustment.
   Measurements such as area, integrated density, and Feret's diameter
   were recorded for regions exceeding the threshold. Parameters for β
   cells were set based on their size and shape for cell counting.
   Colocalization analysis was performed to study the relationship between
   two antigens. For quantitative analysis, regions of interest (ROIs)
   were selected, and the integrated density was measured. Data obtained
   from these measurements were exported and compiled in a spreadsheet or
   statistical software for further analysis.

Figure 3.

   [90]Figure 3
   [91]Open in a new tab

   β cell mass and insulin processing and secretion were impaired in B7-H4
   cKO mice.

   (A) Sketch map of pancreas scanning which was divided into head, body,
   and tail.

   (B) Representative images of pancreas immunofluorescence staining in
   the indicated groups. Green, Insulin; Red, Glucagon; Blue, DAPI.

   (C) Pancreatic islet number in head, body and tail of pancreas in the
   indicated groups. n = 3 per group.

   (D) The density of islet in head, body and tail of pancreas estimated
   by islet number/pancreas area in the indicated groups. n = 3–5 per
   group.

   (E) The size of islet in head, body and tail of pancreas in the
   indicated groups. n = 84–220 islets per group.

   (F) β cell area in head, body and tail of pancreas in the indicated
   groups. n = 3 per group.

   (G and H) mRNA expression of Ins1 and Ins2 in the indicated groups.
   n = 18 per group.

   (I) The mass of α cell estimated by α cell area/pancreas area in the
   indicated groups. n = 3–5 per group.

   (J) Representative images of pancreatic islet immunofluorescence
   staining in the indicated groups. Green, Insulin; red, CCaspase 3;
   blue, DAPI.

   (K) Cleaved Caspase3 intensity of islet in the indicated groups.

   (L) Ratio of Cleaved Caspase3 mRNA expression to pancreatic islet
   number. n = 3 per group.

   (M) Caspase 3, Bax, and Bcl mRNA expression in islets of the indicated
   groups. n = 9–18 per group.

   (N) Representative images of pancreatic islet transmission electron
   micrograph in the indicated groups in different concentrations of
   glucose.

   (O) Number of vesicles in β cells in the indicated groups incubated in
   the medium with various concentration of glucose. n = 3 per group.

   (P) Distance of vesicle in β cell to plasma membrane in the indicated
   group stimulated by various concentration of glucose. The vertical
   coordinate showed relative frequency of vesicle with the same distance
   to plasma membrane. n = 3 per group.

   (Q) mRNA expression of genes involved in insulin secretory granule
   transport and priming in islets of the indicated group. n = 12 per
   group.

   Data are shown as mean ± SEM. ∗: P < 0.05; ∗∗: P < 0.01; ∗∗∗:
   P < 0.001, ∗∗∗∗: P < 0.0001 n.s.: not significant. Unpaired two-tailed
   Student's t-test was used in (C)–(I), (K)–(M), (O), and (Q). Three
   independent experiments were performed.

2.9. Luminex assay

   Serum cytokines and chemokines of Flox and B7H4 cKO mice were
   quantified by the high-performance Luminex 31-plex method (Bio-Rad
   Laboratories, lot number 64469922). Test items included cytokines
   (IFN-γ, IL-1β, IL-2, IL-4, IL-6, IL-10, IL-16, and TNF-α), chemokines
   (CCL1, CCL2, CCL3, CCL4, CCL5, CCL7, CCL11, CCL12, CCL17, CCL19, CCL20,
   CCL22, CCL24, CCL27, CXCL1, CXCL1, CXCL5, CXCL10, CXCL11, CXCL12,
   CXCL13, CXCL16 and CX3CL1).

2.10. RNA extraction and qRT-PCR

   Total RNA was extracted using RNeasy Mini Kit (QIAGEN, 74104) and
   RNAiso Plus (Takara, 9108) for islets and MIN6 cells. cDNA was prepared
   with PrimeScript™ RT Master Mix (Takara, RR036A) according to
   manufacturer's instructions. Gene expression was determined with
   PowerUp™ SYBR™ Green and the realtime PCR system QuantStudio 6 Pro
   (Applied Biosystems). Data were quantified using the comparative 2-▵▵CT
   method using β-actin as a normalization control. The primer sequences
   are shown in [92]Supplementary Table 1.

2.11. RNA sequencing

   Total RNA was extracted from islets of Flox mice, B7H4 cKO mice and
   HFD-fed Flox mice, HFD-fed Flox with B7H4 overexpression mice for
   RNA-sequencing library construction. The RNA libraries were sequenced
   on the illumina NovaseqTM 6000 platform by OE Biotech, Inc., Shanghai,
   China. Gene expression levels are represented as fragments per kilobase
   per million mapped reads (FPKM). The raw sequence data have been
   submitted to the NCBI Gene Expression Omnibus (GEO) datasets with
   accession number [93]GSE268688
   .

2.12. Bioinformatics analysis of the genotype tissue expression (GTEx)
dataset

   The data used for the analyses described in this study were obtained
   from GTEx Portal on October13, 2023
   ([94]https://www.gtexportal.org/home/downloads/adultgtex/bulk_tissue_ex
   pression). To study how the effect of B7-H4 at the molecular level, we
   analyzed RNA-seq data from 328 human pancreas samples in the GTEx
   portal. We compared the pancreas gene expression by establishing two
   groups based on the B7-H4 expression level in the pancreas (highest
   quartile vs lowest quartile) using the R 4.2.1.

2.13. Cytoplasmic and nuclear protein extraction and western blot

   Pancreatic islets total proteins were extracted as previously described
   [[95]39]. 800 islets from each group were used to extract cytoplasmic
   and nuclear protein using Cytoplasmic and Nuclear Protein Extraction
   Kit (Solarbio,R0050). Primary antibodies were as follows: anti-B7-H4
   (1:500, R&D Systems, [96]AF2154), anti-Apof (1:250; Abcam, ab231585),
   anti-Stat5 (1:1,000; Cell Signaling Technology, 94205),
   anti-phospho-Stat5 (Tyr694) (1:1,000; Cell Signaling Technology, 4322),
   anti-Lamin B1 (1:5000; Proteintech, 12987-1-AP), anti-α-Tubulin
   (1:5000; Proteintech, HRP-66031), anti-GAPDH (1:2,500; Proteintech,
   60004-1-Ig), and anti-β-actin (1:5,000; Proteintech, 66009-1-Ig).

2.14. Transmission electron microscopy

   Islets were fixed overnight at 4 °C in 2.5% glutaraldehyde in 0.1 M
   phosphate buffer (pH 7.4) and then postfixed with 2 % osmium tetroxide
   for 1 h. Following serial alcohol dehydration (30, 50, 70, 80, 90, and
   100%), the samples were embedded in epon resin, cut into 70-nm
   sections, and stained with 2% uranyl acetate for examination with a
   Zeiss GeminiSEM 460 electron microscope. Insulin vesicles were
   segmented by Cellpose (Version: 2.2.3) [[97]40]. The diameters, density
   and distances of insuin vesicle from the cellular membrane were
   quantified using Image J software (Version: 2.9.0).

2.15. Statistical analysis

   GraphPad Prism software (version 9.0.0; GraphPad, San Diego, CA, USA)
   was used to obtain graph presentations, curve fittings, statistics, and
   P values. For comparison of two groups, unpaired two-tailed Student's
   t-test was used for parametric, or Mann-Whitney test was used for
   non-parametric distribution. And for more than two groups, one-way
   ANOVA with Bonferroni's multiple comparison was used for parametrically
   distributed data. A P value < 0.05 was considered significant. Error
   bars in histograms indicated the standard error of the mean (SEM).

3. Results

3.1. B7-H4 expression is evaluated in diabetic pancreatic β cells

   Identification of functional immune molecules in diabetic islet cells
   may be critical for developing novel strategies for T2D treatment.
   Immune checkpoint molecules are important molecules that maintain
   immune homeostasis and have also been reported to play metabolic roles
   in human islet β cells, suggesting potential roles in islet metabolic
   inflammation in T2D. Previously, we screened islet transcriptomic data
   from patients with T2D and healthy controls and reported that six
   immune checkpoint genes, including CD44, CD47, SIRPA, HAVCR2, TNFSF9,
   and VTCN1, were highly expressed on human diabetic islet cells,
   although the functions and underlying mechanisms are not clear
   [[98]13]. Interestingly, one of the identified candidates, B7-H4 (B7x,
   B7S1 and VTCN1), the protein encoded by the VTCN1 gene, which is
   expressed primarily on splenic lymphocytes, was also expressed on
   islets, as determined by Western blot analysis ([99]sFigure 1A). Gene
   expression data from the GEO database, a public functional genomics
   data repository, revealed that patients with T2D from the dataset of
   [100]GSE76894 had increased mRNA levels of B7-H4 in pancreatic islets
   ([101]Figure 1A). In obese mice induced by HFD or in genetically
   modified mice with T2D (db/db mice), B7-H4 was abnormally highly
   expressed in the islets ([102]Figure 1B, C). To identify the cell
   source of elevated B7-H4 in diabetic islets, we detected the
   coexpression of B7-H4 with β cells (insulin), α cells (glucagon),
   vascular endothelial cells (CD31) and macrophages (F4/80) via
   two-colour immunofluorescence staining and found that B7-H4 was
   coexpressed mainly with insulin but not with glucagon, CD31 or F4/80
   ([103]Figure 1B and [104]sFigure 1B), indicating that B7-H4 is
   expressed mainly in diabetic β cells. Diabetic pancreatic β cells are
   in chronic inflammatory states with hyperglycaemia and hyperlipemia. To
   determine the factors that regulate B7-H4 expression in diabetic β
   cells, glucose, palmitate and IL-1β were added to MIN6 cell cultures.
   We observed that high glucose and IL-1β concentrations could stimulate
   B7-H4 expression ([105]Figure 1D, E), whereas palmitate had no direct
   effect on B7-H4 expression ([106]sFigure 1C).

Figure 1.

   [107]Figure 1
   [108]Open in a new tab

   B7-H4 expression is evaluated in diabetic pancreatic β cells.

   (A) RNA expression of B7-H4 in pancreatic islets samples of GSE 76894
   (n = 84 for non-diabetes, n = 19 for T2D).

   (B) Representative images of pancreatic islet immunofluorescence
   staining in the indicated groups. Red, B7-H4; Green, Insulin; Blue,
   DAPI.

   (C) The quantitation of B7-H4 estimated by integrated optical
   density/islet area in the indicated groups. non-diabetes, n = 20; HFD,
   n = 13; db/db, n = 15.

   (D) B7-H4 mRNA expression of MIN6 cell incubated in the medium with
   various concentrations of IL-1β for 48h. n = 12 per group.

   (E) B7-H4 mRNA expression of MIN6 cell incubated in the medium with
   various concentration of glucose for 48h. 5.5 mM, n = 9; 25 mM, n = 9.

   Data are shown as mean ± SEM. ∗: P < 0.05; ∗∗: P < 0.01; ∗∗∗:
   P < 0.001; n.s.: not significant. Mann-Whitney test was used in (A);
   unpaired two-tailed Student's t-test was used in (E); one-way ANOVA
   with Bonferroni's multiple comparison was used in (C) and (D). Three
   independent experiments were performed.

3.2. B7-H4 knockout in pancreatic β cells results in glucose intolerance and
insulin secretion deficiency

   To investigate the physiological function of B7-H4 in pancreatic β
   cells and whether it is involved in the pathological process of T2D, we
   generated β-cell-specific B7-H4 knockout (B7-H4 cKO) mice by crossing
   loxP-flanked B7-H4 fl/fl mice (the targeting strategy is shown in
   [109]sFigure 2A) with Rip-Cre transgenic mice. The genotypes of all
   offspring were identified via PCR ([110]sFigure 2B). mRNA
   quantification revealed that the B7-H4 level was 86% lower in islets
   from B7-H4 cKO mice than in those from Flox mice ([111]sFigure 2C).
   Western blotting further confirmed a significant reduction in B7-H4
   expression in islets from B7-H4 cKO mice ([112]sFigure 2D).

   We fed B7-H4 cKO and littermate homozygous Flox controls a NCD or HFD,
   and no differences were detected between B7-H4 cKO mice and controls in
   terms of body weight, food intake, serum TG, total cholesterol (TC) or
   fasted or random blood glucose when fed a NCD ([113]Figure 2A;
   [114]sFigure 2E–I). However, glucose tolerance tests (GTTs) revealed
   that, in both the NCD-fed and HFD-fed B7-H4 cKO mice, glucose disposal
   was impaired, and B7-H4 cKO mice presented higher blood glucose levels
   than the controls did at 15, 30, 60, 90, and 120 min after glucose
   injection ([115]Figure 2B, C). To rule out interference of the Cre
   transgene in glucose disposal, we conducted GTTs in Rip-Cre, Flox and
   B7-H4 cKO mice at 4, 8, 12 and 16 weeks of age, respectively. Glucose
   tolerance in Rip-Cre mice was impaired compared with that in Flox mice
   but was greater than that in B7-H4 cKO mice at 4 and 8 weeks of age. As
   time progressed, the clearance of glucose in Rip-Cre mice improved, and
   the blood glucose levels were slightly higher than those in Flox mice
   at 15 or 30 min after glucose injection at 12 and 16 weeks of age;
   however, in B7-H4 cKO mice, the clearance of glucose continued to be
   impaired, and the blood glucose levels were markedly greater than those
   in Flox and Rip-Cre mice at 15, 30, 60 and 90 min after glucose
   injection ([116]sFigure 3A–D), suggesting that although the Cre
   transgene had a slight impact on glucose tolerance in the early stage,
   the impaired glucose tolerance in B7-H4 cKO mice was caused mainly by
   the specific knockout of B7-H4 in β cells. We further verified this in
   Pdx-1-guided β-cell-specific knockout B7-H4 mice by crossing homozygous
   loxP-flanked B7-H4 fl/fl mice with Pdx-Cre mice. GTT results revealed
   that Pdx-1-guided β-cell-specific knockout B7-H4 mice (B7-H4 fl/fl;
   Pdx-Cre) also presented impaired glucose clearance ([117]sFigure 3E).

Figure 2.

   [118]Figure 2
   [119]Open in a new tab

   B7-H4 Knock out in pancreatic β-cells damages glucose disposal and
   insulin secretion.

   (A) Body weight of mice in the indicated groups. n = 6 per group.

   (B and C) GTT results of mice in the indicated groups. n = 5–8 per
   group.

   (D and E) ITT results of mice in the indicated groups. n = 6–9 per
   group.

   (F and G) In vivo GSIS results of mice in the indicated groups. n = 5–8
   per group.

   (H and J) In vitro GSIS results in the indicated groups. (H), NCD; (J),
   HFD. n = 4–6 per group.

   (I and K) Insulin contents in the indicated groups. (I), NCD; (K), HFD.
   n = 6 per group.

   Data are shown as mean ± SEM. ∗: P < 0.05; ∗∗∗: P < 0.001; n.s.: not
   significant. Unpaired two-tailed Student's t-test was used in (A), (C),
   (E), (G–K). Three independent experiments were performed.

   Insulin insensitivity, insulin insufficiency or elevated glucagon are
   the main causes of hyperglycaemia. To determine whether impaired
   glucose clearance in B7-H4 cKO mice is due to insulin resistance,
   insulin tolerance tests (ITTs) were performed in B7-H4 cKO and control
   mice. Both B7-H4 cKO mice and control mice exhibited normal insulin
   tolerance ([120]Figure 2D, E). The concentrations of serum glucagon
   were comparable between these mice ([121]sFigure 2J). Therefore,
   insulin insensitivity or elevated glucagon can be ruled out as a cause
   of hyperglycaemia upon B7-H4 knockout in β cells. To detect insulin
   secretion ability, glucose-induced insulin secretion (GSIS) tests were
   performed in vivo, and the results revealed that B7-H4 cKO mice had
   lower fasting and glucose-induced insulin release levels than did the
   controls in both the NCD and HFD groups and that the AUC of insulin
   release was significantly lower in the B7-H4 cKO mice ([122]Figure 2F,
   G). In line with the in vivo data, the in vitro GSIS tests using the
   same number of islets isolated from B7-H4 cKO and control mice revealed
   lower glucose-induced insulin secretion by B7-H4 cKO islets both on a
   NCD ([123]Figure 2H) and a HFD ([124]Figure 2J). However, the insulin
   content was greater in islets from B7-H4 cKO mice than in those from
   Flox mice ([125]Figure 2I, K). These results suggest that the inability
   of B7-H4 cKO mice to dispose of blood glucose is due to insulin
   insufficiency and secretion deficiency.

3.3. β-cell mass and insulin processing and secretion were impaired in B7-H4
cKO mice

   The number and function of β cells determine insulin concentrations in
   the blood. To determine whether the attenuated insulin secretion in
   B7-H4 cKO mice is due to a decreased β-cell number, we examined the
   islet morphology of adult mice (12 weeks of age) by immunostaining. The
   whole pancreas was divided into three parts: the head, body and tail
   ([126]Figure 3A). Notably, the number of islets decreased in both the
   body and tail of the pancreas in B7-H4 cKO mice, with a significant
   decrease, especially in the tail region ([127]Figure 3B, C). The
   density of islets calculated by islet number/pancreatic area was also
   lower in B7-H4 cKO mice ([128]Figure 3D), but the size of islets was
   slightly higher, although there was no significant difference between
   the two groups ([129]Figure 3E), suggesting compensatory hyperplasia in
   B7-H4 cKO islets. The reduced insulin+ β-cell areas in B7-H4 cKO mice,
   especially in the body and tail of the pancreas, suggest that the
   hyperplasia of islets did not compensate for the loss of β-cell mass
   ([130]Figure 3F). The lower mRNA expression of the Ins-1 and Ins-2
   genes in islets from B7-H4 cKO mice further verified the impaired
   β-cell mass in B7-H4 cKO mice ([131]Figure 3G, H). The mass of α cells
   was similar in B7-H4 cKO and control mice ([132]Figure 3I). Cleaved
   caspase 3 is a marker of apoptosis. We used cleaved caspase 3 and
   insulin to double-stain the pancreas and showed that although the ratio
   of caspase 3^+ islets was similar between the two groups, there was a
   striking increase in the expression levels of cleaved caspase 3 in the
   islets of B7-H4 cKO mice, suggesting that more apoptotic β cells were
   present in B7-H4 cKO islets ([133]Figure 3J–L). The results of qPCR
   also revealed increased caspase 3 and Bax and decreased Bcl2 mRNA
   expression in islets from B7-H4 cKO mice ([134]Figure 3M). These
   findings suggest that B7-H4 affects β-cell numbers by regulating their
   apoptosis.

   B7-H4 cKO mice not only presented significant defects in the baseline
   value of insulin but the peak of insulin release also almost completely
   disappeared after glucose stimulation ([135]Figure 2F), suggesting that
   the insufficient serum insulin in B7-H4 cKO mice was not only due to a
   decrease in β-cell mass. Indeed, the in vitro GSIS tests using the same
   number of islets from the two groups of mice revealed a noticeable
   reduction in insulin secretion but not insulin content in B7-H4 cKO
   islets ([136]Figure 2H–K). Transmission electron microscopy revealed
   that even in a low-glycaemic environment, there were more vesicles in β
   cells from B7-H4 cKO mice than in those from control mice
   ([137]Figure 3N–O). Moreover, the distance from vesicles to the plasma
   membrane was greater in B7-H4 cKO β cells than in control cells, and
   after stimulation with high glucose, the percentage of docked vesicles
   0–2 μm apart from the plasma membrane strikingly decreased
   ([138]Figure 3P), suggesting that B7-H4 knockout affected the
   exocytosis process of insulin secretory vesicles. qPCR results also
   revealed that the expression of genes involved in insulin secretory
   vesicle transport (Rab3a) and priming (Snap25, Vamp2) was significantly
   lower in islets from B7-H4 cKO mice ([139]Figure 3Q).

3.4. Overexpression of B7-H4 in pancreatic β cells ameliorates glucose
tolerance and improves insulin secretion

   The metabolic disorders of B7-H4 cKO mice indicated that B7-H4 has
   important physiological functions in islets for the maintenance of
   β-cell mass and cell functions. We further investigated whether B7-H4
   could be used as a therapeutic treatment for obesity-associated
   metabolic dysfunction. As shown in [140]Figure 4A, B7-H4 fl/fl mice
   were fed a HFD for 8 weeks and then infused with AAV2/8-Ins2-B7-H4 or
   control virus from the pancreatic intraduct to specifically overexpress
   B7-H4 in β cells. As the virus carries the ZSGreen fluorescent protein,
   islets isolated from mice after 5 weeks of virus infusion presented
   strong green fluorescence ([141]sFigure 4A), and B7-H4 mRNA and protein
   levels were significantly elevated in islets from mice infused with
   AAV8-Ins2-B7-H4 compared with those from mice infused with control
   virus ([142]sFigure 4B, C). To assess the effects of excess B7-H4 on
   the functions of healthy β cells, we also treated healthy mice fed a
   NCD with the same amount of virus. Two weeks after virus infusion,
   HFD-fed mice exhibited marked improvement in glucose clearance. Blood
   glucose levels were significantly lower in HFD-fed mice with
   β-cell-specific B7-H4 overexpression than in HFD-fed control mice at
   30, 60, and 90 min after glucose injection ([143]Figure 4B). The AUC of
   the IPGTT also revealed ameliorated glucose tolerance in
   β-cell-specific B7-H4-overexpressing HFD-fed mice ([144]Figure 4C).
   However, there was no significant effect on healthy mice fed a NCD
   ([145]Figure 4B, C). After 2 more weeks, β-cell-specific B7-H4
   overexpression significantly improved glucose clearance in both HFD-fed
   and NCD-fed mice ([146]Figure 4D, E), and insulin secretion improved
   markedly. Serum insulin concentrations increased significantly in
   B7-H4-overexpressing mice at every time point after glucose
   stimulation, even during fasting ([147]Figure 4F, G). However, there
   were no significant changes in insulin sensitivity detected by ITTs
   ([148]sFigure 4D, E). Thus, β-cell-specific B7-H4 overexpression has
   robust effects on insulin secretion and glucose tolerance, indicating
   that B7-H4 may be an ideal molecular target for improving β-cell
   function.

Figure 4.

   [149]Figure 4
   [150]Open in a new tab

   B7-H4 overexpression in pancreatic β-cells improves glucose disposal
   and insulin secretion.

   (A) Detailed treatment to B7-H4 fl/fl mice.

   (B and C) GTT results after 2 weeks of virus infusion in the indicated
   groups. n = 5–7 per group.

   (D and E) GTT results after 4 weeks of virus infusion in the indicated
   groups. n = 5–7 per group.

   (F and G) GSIS results after 4 weeks of virus infusion in the indicated
   groups. n = 5–7 per group.

   Data are shown as mean ± SEM. ∗: P < 0.05; ∗∗: P < 0.01; ∗∗∗:
   P < 0.001; n.s.: not significant. Unpaired two-tailed Student's t-test
   was used in (C), (E), and (G). Three independent experiments were
   performed.

3.5. RNA-seq of islets indicates that B7-H4 changes the expression of β-cell
function and insulin secretion markers

   To further explore the molecular mechanisms underlying the protective
   effects of B7-H4, we isolated pancreatic islets from B7-H4 cKO and Flox
   mice fed a NCD and from HFD-fed obese mice after 5 weeks of
   AAV2/8-Ins2-B7-H4 or AAV2/8-Ctrl infusion. Total RNA was extracted from
   these islets, and unbiased transcriptomic profiling was performed.
   Differential expression analysis with an absolute log2-fold change (FC)
   over 1 and a q value below 0.05 revealed 202 upregulated and 133
   downregulated genes from B7-H4 cKO and Flox mice ([151]Figure 5A, B)
   and 182 upregulated and 520 downregulated genes from AAV2/8-Ins2-B7-H4
   and AAV2/8-Ctrl mice whose Vtcn1 (B7-H4) gene was dramatically
   upregulated ([152]Figure 5C, D). Kyoto Encyclopedia of Genes and
   Genomes (KEGG) enrichment analysis of the differentially expressed
   genes (DEGs) in B7-H4 cKO and Flox mice revealed that the downregulated
   genes were enriched in cytokine–cytokine receptor interactions, the
   adipocytokine signalling pathway and the PPAR signalling pathway, and
   the upregulated genes were enriched in the calcium signalling pathway
   and retinol metabolism ([153]Figure 5E, F). DEGs in the
   AAV2/8-Ins2-B7-H4 and AAV2/8-Ctrl mice were also enriched in
   cytokine–cytokine receptor interactions, the adipocytokine signalling
   pathway, the AGE‒RAGE signalling pathway in diabetes, the calcium
   signalling pathway and cholesterol metabolism ([154]Figure 5G, H). Gene
   Ontology (GO) analysis revealed significant enrichment in several
   biological processes, cellular components, and molecular functions.
   Most notable among these genes were cell adhesin, potassium ion
   transport, and cholesterol efflux, which were downregulated in B7-H4
   cKO mice and upregulated in B7-H4-overexpressing mice ([155]sFigure 5A,
   B). Gene set enrichment analysis (GSEA) revealed downregulation of the
   Golgi-to-plasma membrane transport and insulin secretion pathways in
   B7-H4 cKO mice and upregulation of pancreatic secretion and positive
   regulation of insulin secretion pathways in B7-H4-overexpressing mice
   ([156]Figure 5I).

Figure 5.

   [157]Figure 5
   [158]Open in a new tab

   RNA-seq on islets indicates B7-H4 changed the expression of β cell
   function and insulin secretion markers.

   (A and B) Heatmap and volcano plot displaying differentially expressed
   genes between Flox (n = 3) and cKO (n = 3).

   (C and D) Heatmap and volcano plot displaying differentially expressed
   genes in AAV2/8-Ins2-B7H4 (n = 3) and AAV2/8-Ctrl (n = 3).

   (E and F) KEGG pathway enrichment analysis of differentially expressed
   genes in Flox (n = 3) and cKO (n = 3), which showed that downregulated
   genes enriched in cytokine–cytokine receptor interaction, adipocytokine
   signaling pathway and PPAR signaling pathway, and upregulated genes
   enriched in calcium signaling pathway and retinol metabolism.

   (G and H) KEGG pathway enrichment analysis of differentially expressed
   genes in AAV2/8-Ins2-B7H4 (n = 3) and AAV2/8-Ctrl (n = 3), which showed
   that DEGs in AAV2/8-Ins2-B7-H4 and AAV2/8-Ctrl mice were enriched in
   cytokine–cytokine receptor interaction, adipocytokine signaling
   pathway, AGE-RAGE signaling pathway in diabetes, calcium signaling
   pathway and cholesterol metabolism.

   (I) GSEA showed a downregulation in the Golgi to plasma membrane
   transport and insulin secretion pathways in B7-H4 cKO mice compared
   with Flox mice, and an upregulation in the pancreatic secretion and
   positive regulation of insulin secretion pathways in AAV2/8-Ins2-B7H4
   mice compared with AAV2/8-Ctrl mice. n = 3 per group.

   (J) Expression heatmap of β cell mature and differentiation related
   genes, disallowed genes, and insulin formation and secretion related
   genes in the indicated groups. n = 3 per group.

   To gain further insight into the effects of B7-H4 on β-cell function
   and insulin secretion, we analysed four groups of transcriptomic data
   together. Overall, the gene expression pattern of B7-H4 cKO mice was
   similar to that of HFD-fed mice injected with the AAV2/8-Ctrl virus,
   whereas HFD-fed mice that specifically overexpressed B7-H4 in β cells
   presented similar gene expression patterns to those of Flox mice.
   Markers of β-cell maturation, insulin formation- and secretion-related
   genes, and islet endocrine cell formation-related genes were
   significantly downregulated in B7-H4 cKO mice and HFD-fed obese mice.
   B7-H4 overexpression rescued the expression of these genes in HFD-fed
   obese mice. Although the expression of β-cell disallowed genes, which
   are typically repressed in mature adult β cells, was not changed in
   B7-H4 cKO mice, B7-H4 overexpression rescued the highly expressed
   disallowed genes caused by metabolic stress in HFD-induced obesity
   ([159]Figure 5J). Taken together, the results of RNA-seq of islets
   indicate that B7-H4 alters the expression of β-cell function and
   insulin secretion markers.

3.6. B7-H4 maintains β-cell function and insulin secretion by regulating
islet cholesterol metabolism

   To better investigate the molecular mechanisms by which B7-H4 regulates
   β-cell function and insulin secretion, we analysed the DEGs of B7-H4
   cKO and B7-H4-overexpressing islets jointly. There were 75 genes whose
   expression was significantly altered in B7-H4 cKO mice but whose
   expression was reversed by the overexpression of B7-H4
   ([160]Figure 6A). GO analysis of these genes revealed significant
   enrichment in cell adhesion and cholesterol metabolic processes and
   molecular functions enriched in cytokine receptor activity
   ([161]Figure 6B). KEGG enrichment analysis revealed enrichment of the
   JAK-STAT signalling pathway and calcium signalling pathway
   ([162]Figure 6C).

Figure 6.

   [163]Figure 6
   [164]Open in a new tab

   B7-H4 maintains β cell function and insulin secretion by regulating
   islet cholesterol metabolism.

   (A) Heatmap of the identified 75 genes which were significantly altered
   in B7-H4 cKO mice but were reversed by overexpression of B7-H4. n = 3
   per group.

   (B and C) GO and KEGG analysis of the identified genes, which showed a
   significant enrichment in cell adhesion and cholesterol metabolic
   process, and molecular functions enriched in cytokine receptor
   activity, and in JAK-STAT signaling pathway and calcium signaling
   pathway.

   (D and E) Serum cholesterol concentration and pancreatic islet
   cholesterol content in the indicated groups. n = 4–8 per group.

   (F) Venn diagram showed that Apof is the only common gene regulated by
   B7-H4 in both humans and mice.

   (G) Western blot analysis of Apof expression in pancreatic islets from
   the indicated group, which showed that Apof in islets was downregulated
   by B7-H4 KO and upregulated by B7-H4 overexpression.

   Flox: Mice homozygous for the floxed B7-H4 allele; cKO: β-cell–specific
   B7-H4 knockout mice; OE: Flox mice overexpressed B7-H4 through
   injecting with AAV2/8-Ins2-B7H4; Apof: Apolipoprotein F; B7-H4: B7
   family member, H4.

   (H and I) GTT results of mice in the indicated groups. n = 4–5 per
   group.

   (J and K) GSIS results of mice in the indicated groups. n = 4–5 per
   group.

   (L) Transmission electron micrograph showed that Apof overexpression
   made vesicles in of β cell closer to the plasma membrane. n = 3 per
   group. scale bar, 2um. n = 3 mouse per group.

   (M and N) Number and diameter of β cell vesicle in the indicated group.
   n = 3 mouse per group.

   (O) Distance of vesicle in β cell to plasma membrane in the indicated
   group. The vertical coordinate showed relative frequency of vesicle
   with the same distance to plasma membrane. n = 3 mouse per group.

   (P and Q) Serum cholesterol and islet cholesterol in the indicated
   group. n = 9–12 per group.

   Data are shown as mean ± SEM. ∗: P < 0.05; ∗∗: P < 0.01; ∗∗∗:
   P < 0.001; ∗∗∗∗: P < 0.0001; n.s.: not significant.

   Unpaired two-tailed Student's t-test was used in (D), (E), (I), (K),
   (M), (N), (P), and (Q). Three independent experiments were performed.

   Previous studies have reported that the cholesterol content in β cells
   affects the amount of insulin and the size of vesicles. We detected the
   serum cholesterol concentration and islet cholesterol content of B7-H4
   cKO mice and B7-H4-overexpressing mice. Interestingly, we found that
   the serum cholesterol concentration was low, but the islet cholesterol
   content was significantly higher in B7-H4 cKO mice, whereas B7-H4
   overexpression strongly reduced the serum and islet cholesterol
   concentrations in HFD-fed mice ([165]Figure 6D, E). To study how B7-H4
   regulates the cholesterol content in β cells at the molecular level, we
   analysed RNA-seq data from human pancreas samples in the GTEx portal.
   GO and KEGG analyses revealed that the DEGs associated with high (top
   25%) and low (bottom 25%) B7-H4 expression were enriched in
   cytokine–cytokine receptor interactions and in lipid, atherosclerosis
   and chemokine signalling pathways ([166]sFigure 6A, B). We compared the
   75 genes in mice whose expression was altered by B7-H4 knockout but
   reversed by B7-H4 overexpression with the differential genes in healthy
   humans with high and low expressions of B7-H4 in the GTEx database and
   found that Apof is a common gene regulated by B7-H4 in both humans and
   mice ([167]Figure 6F). We verified that the protein levels of Apof in
   islets were decreased by B7-H4 KO and increased by B7-H4 overexpression
   via Western blotting ([168]Figure 6G).

   Apof has been reported to reduce cholesterol levels [[169]41,[170]42].
   To study whether B7-H4 regulates insulin secretion through Apof, we
   overexpressed Apof in B7-H4 cKO mice via pancreatic intraductal
   infusion with AAV2/8-Apof- 3xFlag or AAV2/8-GFP control virus. As shown
   in [171]sFigure 6C, the overexpression of Apof led to a sharp increase
   in the Apof protein level. Glucose tolerance was markedly improved by
   Apof overexpression in B7-H4 cKO mice, and the GTT results revealed a
   sharp reduction in blood glucose levels in B7-H4 cKO mice with Apof
   overexpression at 15, 30, 60, 90 and 120 min after glucose injection,
   although there was a slight decrease in glucose levels in Flox mice
   ([172]Figure 6H, I). Insulin secretion was also markedly improved in
   Apof-overexpressing B7-H4 cKO mice, and serum insulin concentrations
   increased significantly at every time point after glucose stimulation,
   even during fasting, suggesting that Apof ameliorates the basal and
   glucose-stimulated insulin secretion deficiency induced by B7-H4
   defects ([173]Figure 6J, K). Transmission electron micrographs revealed
   that Apof overexpression in B7-H4 cKO mice reduced the number of
   retained vesicles in β cells, and these vesicles were smaller and
   closer to the plasma membrane ([174]Figure 6L–O), indicating that Apof
   promotes the secretion of insulin secretory vesicles and improves the
   functional status of β cells in B7-H4 cKO mice. ITTs revealed slight
   changes in insulin sensitivity after Apof overexpression in B7-H4 cKO
   mice ([175]sFigure 6D, E). Furthermore, we found that Apof
   significantly reduced the elevated serum and islet cholesterol levels
   in B7-H4 cKO mice ([176]Figure 6P, Q).

3.7. B7-H4 regulates islet cholesterol metabolism through Stat5 signalling

   In addition to β-cell function and insulin secretion markers,
   cytokine–cytokine receptor interactions were also enriched in DEGs of
   B7-H4 cKO and B7-H4-overexpressing mice, and 75 genes whose expression
   was altered by B7-H4 KO but reversed by B7-H4 overexpression were
   enriched in the JAK‒STAT signalling pathway. Previous studies have
   shown that B7-H4 can regulate cytokine expression in T cells [[177]43].
   These findings indicate that B7-H4 may be involved in the regulation of
   the immune microenvironment of islets. We performed Luminex
   multiple-factor assays to measure the serum concentrations of
   cytokines, including interleukins, CC- and CXC- chemokines and
   interferon, in Flox and B7-H4 cKO mice. A heatmap of these cytokines
   presented individually or as the means of each group revealed that
   MDC/CCL22, eotaxin/CCL11, KC/CXCL1, TARC/CCL17, IFN-gamma,
   MIP-1alpha/CCL3 and MCP-3/CCL7 were higher and that CTACK/CCL27,
   IP-10/CXCL10, IL-2, I-309/CCL1, IL-1beta and IL-4 were lower in the
   serum of B7-H4 cKO mice ([178]Figure 7A). Analysis of the differences
   between the groups revealed that TARC/CCL17 expression was
   significantly higher and that CTACK/CCL27 expression was significantly
   lower in the serum of B7-H4 cKO mice ([179]Figure 7B–D). Real-time
   quantitative PCR was used to detect the expression of islet cytokines
   in Flox and B7-H4 cKO mice fed a NCD or HDF. Consistent with the change
   in serum concentrations, the mRNA expression of CCL17 in islets from
   cKO mice was significantly upregulated, but in contrast to the serum
   concentrations, the mRNA levels of CCL27, IL-4 and IL-6 in islets from
   cKO mice were significantly greater than those in islets from Flox mice
   ([180]Figure 7E, F). Moreover, we examined the mRNA levels of chemokine
   receptors in the islets of the four groups of mice and found that CCR4
   and CCR10 were significantly upregulated in the islets of cKO mice
   ([181]Figure 7G). Changes in cytokines and chemokines may alter the
   composition of islet immune cells. Given that the immune cells of
   islets are mainly macrophages, we used the myeloid marker Iba1 to stain
   the islets of Flox and B7-H4 cKO mice via immunofluorescence. Although
   there was no significant difference in the ratio of Iba1-positive
   islets, the intensity of Iba1 was significantly greater in B7-H4 cKO
   mice than in flox mice ([182]Figure 7H–J), indicating that the number
   of infiltrating myeloid cells in B7-H4 cKO islets was significantly
   higher.

Figure 7.

   [183]Figure 7
   [184]Open in a new tab

   B7-H4 regulates the expression of islet cytokines and chemokine
   receptors.

   (A) Heatmap of serum concentrations of cytokines assessed by luminex
   multiple-factor assays in the indicated groups. n = 5 per group.

   (B–D) Serum concentrations of cytokines in the indicated groups. n = 5
   per group.

   (E–G) mRNA expression of cytokines in islets from the indicated groups.
   n = 9 per group.

   (H) Representative images of pancreatic islet immunofluorescence
   staining in the indicated groups. Green, Insulin; Red, Iba1; Blue,
   DAPI.

   (I and J) The ratio of Iba1-positive islets (I) and Iba-1 indensity (J)
   in the indicated group. n = 3 mouse per group.

   Data are shown as mean ± SEM. ∗: P < 0.05; ∗∗∗∗: P < 0.0001; n.s.: not
   significant. Unpaired two-tailed Student's t-test was used in (B)–(G).
   Three independent experiments were performed.

   Cytokines that activate the JAK-STAT signalling pathway may be involved
   in the regulation of β-cell function and insulin secretion by B7-H4. We
   detected the protein and phosphorylation levels of seven members of the
   STAT family in islets from B7-H4 cKO and B7-H4-overexpressing mice and
   found that the phosphorylation levels of Stat5 were significantly
   higher in islets from B7-H4 cKO mice but reduced in those from
   B7-H4-overexpressing mice ([185]Figure 8A). After tyrosine
   phosphorylation at Y694, Stat5 accumulates in the nucleus [[186]44]. As
   shown in [187]Figure 8B, phosphorylated Stat5 accumulated in the
   nucleus of islets from Flox and B7-H4cKO mice. To investigate whether
   B7-H4 regulates islet cholesterol metabolism and insulin secretion
   through Stat5, we treated B7-H4 cKO mice with the Stat5 inhibitor
   pimozide or vehicle control (VC) for 2 weeks. Pimozide treatment
   decreased Stat5 phosphorylation levels but increased Apof protein
   levels in the islets of B7-H4 cKO mice ([188]Figure 8C). GTTs markedly
   improved glucose clearance in B7-H4 cKO mice treated with pimozide for
   two weeks, which presented significantly lower blood glucose levels at
   15, 30 and 60 min after glucose injection ([189]Figure 8D, E). The
   serum random insulin concentration in B7-H4 cKO mice was significantly
   lower than that in Flox mice, and pimozide treatment for two weeks
   reversed the difference in the serum insulin level between B7-H4 cKO
   mice and Flox mice ([190]Figure 8F). GSIS in vitro revealed that
   pimozide treatment markedly improved glucose-stimulated insulin
   secretion in B7-H4 cKO mice ([191]Figure 8G). Furthermore, we found
   that pimozide treatment significantly reduced the increase in islet
   cholesterol levels in B7-H4 cKO mice ([192]Figure 8H), although there
   was no significant difference in the serum cholesterol concentration
   when the mice were fed a NCD; this trend was consistent with that of
   islet cholesterol ([193]Figure 8I). These data indicate that B7-H4
   regulates islet cholesterol metabolism to enhance β-cell function and
   insulin secretion through Stat5 signalling.

Figure 8.

   [194]Figure 8
   [195]Open in a new tab

   B7-H4 regulates islet cholesterol metabolism through Stat5 signaling.

   (A) Western blot analysis of Stat5 and phosphorylated Stat5 in islets
   from the indicated groups, which showed that phosphorylation levels of
   Stat5 were significantly increased in islets from B7-H4 cKO mice but
   reduced in that from B7-H4 overexpressed mice. n = 3 per group.

   Flox: Mice homozygous for the floxed B7-H4 allele; cKO: β-cell–specific
   B7-H4 knockout mice; OE: Flox mice overexpressed B7-H4 through
   injecting with AAV2/8-Ins2-B7H4; Stat5: Signal transducer and activator
   of transcription 5; p-Stat5: Phospho-signal transducer and activator of
   transcription 5; B7-H4: B7 family member, H4; GAPDH:
   Glyceraldehyde-3-Phosphate Dehydrogenase.

   (B) Western blot showed that phosphorylated Stat5 accumulated in the
   nucleus of islets from Flox and B7-H4cKO mice.

   Flox: Mice homozygous for the floxed B7-H4 allele; Lamb1: Laminin beta
   1; Stat5: Signal transducer and activator of transcription 5; p-Stat5:
   Phospho-signal transducer and activator of transcription 5.

   (C) Western blot showed that pimozide treatment decreased the Stat5
   phosphorylation levels and cholesterol content, while upregulate Apof
   protein levels in islets of B7-H4 cKO mice. n = 3 per group.

   Flox: Mice homozygous for the floxed B7-H4 allele; VC: cKO mice
   injected with vehicle control for 2 weeks; Piz: cKO mice injected with
   pimozide for 2 weeks; Stat5: Signal transducer and activator of
   transcription 5; p-Stat5: Phospho-signal transducer and activator of
   transcription 5; Apof: Apolipoprotein F.

   (D and E) GTT results of the mice treated by pimozide. n = 5–7 per
   group.

   (F) Serum random insulin concentration in the indicated group. n = 6
   per group.

   (G) Results of in vitro GSIS in the indicated group. n = 6 per group.

   (H and I) Islet and serum cholesterol content in the indicated group.
   n = 6 per group.

   Data are shown as mean ± SEM. ∗: P < 0.05; ∗∗: P < 0.01; ∗∗∗:
   P < 0.001; ∗∗∗∗: P < 0.0001; n.s.: not significant. Unpaired two-tailed
   Student's t-test was used in (D)–(H). Three independent experiments
   were performed.

4. Discussion

   Pancreatic islet inflammation is a pathological mechanism of T2D, and
   therapies targeting islet inflammation are promising for the prevention
   and treatment of T2D and its complications. However, the molecules
   regulating islet inflammation in the pathological process of T2D are
   still unclear. Through islet transcriptome data analysis of T2D
   patients, we identified six highly expressed immune checkpoint genes
   that are important for immune homeostasis. Here, we elucidate the
   physiological role of one of these molecules, B7-H4, in β-cell function
   and its pathological effects on islet inflammation and T2D.

   As an inhibitory costimulatory molecule of T cells, B7-H4 has been
   implicated in many diseases, such as cancer, allograft rejection, and
   autoimmune diseases, due to its immunosuppressive function. B7-H4 is
   also expressed in nonhaematopoietic cells and contributes to immune
   tolerance, such as maternal–foetal immune tolerance [[196]19]. In the
   pancreas, Jun Sik Lee et al. [[197]20] revealed that, unlike other B7
   family members, the B7-H4 protein is not expressed on immune cells but
   rather on pancreatic islet cells. The B7-H4 protein is expressed and
   colocalized with insulin [[198]25], and its polymorphism influences the
   prevalence of diabetes [[199]26]. Although it has been speculated that
   B7-H4 may play a potential physiological role in the function of β
   cells and may have pathological effects on T2D [[200]23], there have
   been no relevant studies to date. For the first time, we used
   β-cell-specific B7-H4 knockout mice to explore the in vivo functions of
   B7-H4. We found that B7-H4 affects not only the expression of islet
   cytokines but also the secretion of insulin vesicles. Unlike the
   reduction in islet B7-H4 protein expression in T1D through the
   proteolytic cleavage of the B7-H4 protein from the cell membrane by the
   metalloproteinase nardilysin (NRD1) [[201]21,[202]22], we found that
   the expression of the islet B7-H4 protein was increased in HFD-induced
   obese and genetically modified T2D mice. This upregulation may be a
   protective mechanism against metabolic stress. Indeed, the upregulation
   of Nrd1 and elevated soluble B7-H4 are characteristic of patients with
   T1D but not T2D [[203]22]. Immunostaining revealed that upregulated
   B7-H4 expression occurred mainly in islet β cells but not in pancreatic
   macrophages. In tumour microenvironments, cancer cells increase B7-H4
   expression in response to hypoxia and transforming growth factor β1
   (TGFβ1) [[204]45,[205]46]; IL-6 and IL-10 can also stimulate B7-H4 mRNA
   expression in macrophages [[206]47]. Here, we found that in the
   diabetic microenvironment, high glucose and IL-1β can stimulate B7-H4
   mRNA expression in the β-cell line MIN6 cells.

   B7-H4 has been proposed to have a direct effect on cell functions such
   as the proliferation and differentiation of expressed cells [[207]48].
   Here, we revealed that B7-H4 knockout in β cells results in decreased
   β-cell mass and pancreatic areas in mice. Β-cell mass is determined by
   prenatal development of islets, postnatal compensatory growth and
   regeneration of β cells [[208]49]. Dynamic changes in the β-cell mass
   also occur during pancreatic development and are triggered by multiple
   signalling pathways. Apoptosis, dedifferentiation and
   transdifferentiation are the main causes of reduced β-cell mass in
   diabetes mellitus patients. Although the decreased β-cell mass in B7-H4
   cKO mice can be partially explained by an increase in β-cell apoptosis,
   the number of islets was significantly lower in cKO mice than in
   control mice, suggesting a developmental defect in β cells upon B7-H4
   knockout. This might be related to altered cytokines and infiltrated
   immune cells in islets due to the depletion of B7-H4 in β cells.
   Indeed, some immune signals, such as TGF-β and IL-18, have been
   reported to play key roles in the development of pancreatic islets
   [[209]50,[210]51]. The altered identity of β cells has also been
   proposed as a mechanism for the loss of β cells in diabetes [[211]52].
   RNA-seq of islets revealed that B7-H4 knockdown in β cells
   downregulated the expression of β-cell maturation-related genes such as
   Nkx6.1, Mafa, Mafb, Pax6, and Neurod1, whereas B7-H4 overexpression
   rescued the expression of these genes in HFD-fed obese mice. Although
   the expression of β-cell disallowed genes, which are typically
   repressed in mature adult β cells, did not change in B7-H4 cKO mice,
   B7-H4 overexpression rescued the highly expressed disallowed genes
   caused by metabolic stress in HFD-induced obesity. These findings
   suggest that B7-H4 may play a promising role in reversing β-cell
   trans-differentiation and maintaining β-cell identity in individuals
   with obesity and T2D.

   Another important observation of our study is that B7-H4 regulates
   insulin secretion by modulating cholesterol metabolism in β cells. The
   insufficiency of serum insulin in B7-H4 cKO mice after glucose
   injection was partly due to decreased β-cell mass. However, when we
   took the same number of islets ex vivo and stimulated them with
   glucose, insulin secretion from the islets of B7-H4 cKO mice decreased
   significantly, indicating that the defect in insulin secretion was also
   a crucial factor. The SEM results revealed that the vesicles in β cells
   from B7-H4 cKo mice were numerous but immature and further away from
   the plasma membrane. To investigate the molecular mechanism by which
   B7-H4 affects β-cell function, we focused on the 75 differential genes
   altered by B7-H4 KO but reversed by B7-H4 overexpression. To our
   surprise, these 75 genes were enriched in cholesterol metabolism.
   Cholesterol metabolism has been reported to be involved in the
   maintenance of β-cell function and insulin secretion. Excess
   cholesterol in β cells impairs β-cell function, causing insulin
   secretory vesicle enlargement and reduced glucose-stimulated secretion
   [[212][28], [213][29], [214][30], [215][31], [216][32]]. Indeed, islet
   cholesterol content detection revealed that the islet cholesterol
   content of B7-H4 cKO mice was significantly increased, whereas B7-H4
   overexpression significantly reduced the islet cholesterol content. By
   comparing the 75 genes in mice with the DEGs in healthy humans with
   high and low expression of B7-H4 in the GTEx database, we found that
   Apof is a common gene regulated by B7-H4 in both humans and mice. We
   verified the downregulated protein levels in islets from B7-H4 cKO mice
   and upregulated levels in those from B7-H4-overexpressing mice.
   Moreover, LTIP (lipid transfer inhibitor protein) inhibits the activity
   of CETP (cholesteryl ester transfer protein) and, via this inhibition,
   functions to regulate cholesterol transport [[217]53]. The expression
   and function of Apof in β cells are still unknown. However, other
   apolipoproteins, such as apolipoprotein A-I, reportedly improve
   pancreatic β-cell function [[218]33] and protect them from
   cholesterol-induced mitochondrial damage, restoring their ability to
   secrete insulin [[219]34]. The overexpression of Apof in the β cells of
   B7-H4 cKO mice, which have reduced Apof expression, markedly rescued
   the intolerance of glucose nearly to normal levels, indicating that the
   downregulation of Apof expression was the main cause of β-cell
   dysfunction caused by B7-H4 KO.

   The initial cephalic phase and postprandial insulin secretion have been
   reported to be modulated by cytokines such as IL-β and IL-33
   [[220][54], [221][55], [222][56]]. Here, we found that many types of
   cytokines and chemokine receptors were upregulated or downregulated in
   islets from B7-H4 cKO mice. Cytokines typically function through the
   downstream JAK-STAT signalling pathway. We confirmed that Stat5 plays a
   key role in the regulation of insulin secretion by B7-H4 and that
   inhibiting Stat5 can reverse impaired glucose tolerance and insulin
   secretion in B7-H4 cKO mice. This effect of Stat5 is achieved by
   regulating the concentration of islet cholesterol, and Stat5 inhibitors
   significantly increase the expression of Apof in islets from B7-H4 cKO
   mice. Stat5 has been reported to be involved in the regulation of
   cholesterol metabolism [[223]35,[224]57], although the molecular
   mechanism is not clear. As a transcription factor, Stat5 usually
   activates the expression of target genes. However, we found that the
   expression of Apof was upregulated after Stat5 inhibition, which may be
   related to the indirect role of Stat5 in acting as a cytoplasmic
   platform for other proteins, as reported previously [[225]57]. Indeed,
   specific knockout of Stat5 in the liver can upregulate bile acid
   conversion genes [[226]35].

   In this study, we identified the role of B7-H4 in regulating β-cell
   mass and insulin processing and secretion through cholesterol
   metabolism in β-cell-specific B7-H4 knockout mice and HFD-fed mice with
   β-cell-specific B7-H4 overexpression. This is the first in vivo
   evidence showing that B7-H4 plays a physiological role in regulating
   β-cell mass and insulin secretion, which are modulated by Stat5
   signalling. Our findings identify B7-H4 as a regulator of β-cell mass
   and insulin secretion, and its manipulation in β cells offers a
   therapeutic opportunity for T2D.

5. Limitations of this study

   There are still some unanswered questions in our study. First, how does
   B7-H4 affect Stat5 signaling, and is it related to cytokines? What is
   the cellular source of these cytokines? Second, how does Stat5 regulate
   the expression of Apof? Third, although we used two kinds of Cre mice
   (RIP Cre and Pdx-1Cre) to exclude the effects of transgene and
   non-specific roles, both Ins2 and Pdx-1genes may be expressed in the
   hypothalamus, so the influence of B7-H4 in the hypothalamus cannot be
   excluded. Overall, our study elucidates, for the first time, the
   physiological function of B7-H4 in regulating insulin secretion through
   cholesterol metabolism.

CRediT authorship contribution statement

   Fangzhen Xia: Writing – review & editing, Writing – original draft,
   Visualization, Validation, Supervision, Project administration,
   Methodology, Investigation, Funding acquisition, Formal analysis, Data
   curation, Conceptualization. Ziteng Zhang: Writing – review & editing,
   Writing – original draft, Visualization, Methodology, Investigation,
   Formal analysis. Zhen Qian: Writing – review & editing, Writing –
   original draft, Visualization, Validation, Methodology, Investigation,
   Formal analysis. Xiaoyu Fang: Writing – review & editing,
   Visualization, Formal analysis. Junxue Wang: Writing – review &
   editing. Yan Wang: Writing – review & editing, Visualization. Guoting
   Sun: Writing – review & editing. Yuefeng Yu: Writing – review &
   editing, Visualization. Ninjian Wang: Writing – review & editing. Junke
   Zhen: Writing – review & editing, Validation, Methodology,
   Investigation, Data curation, Conceptualization. Yan Liu: Writing –
   review & editing, Validation, Methodology, Investigation, Data
   curation, Conceptualization. Yingli Lu: Writing – review & editing,
   Validation, Supervision, Project administration, Methodology,
   Investigation, Funding acquisition, Data curation, Conceptualization.

Declaration of competing interest

   The authors declare no competing interests.

Acknowledgments