SDF‐1 inhibits the dedifferentiation of islet β cells in hyperglycaemia by up‐regulating FoxO1 via binding to CXCR4

Abstract Islet β cell dedifferentiation is one of the most important mechanisms in the occurrence and development of diabetes. We studied the possible effects of chemokine stromal cell‐derived factor‐1 (SDF‐1) in the dedifferentiation of islet β cells. It was noted that the number of dedifferentiated islet β cells and the expression of SDF‐1 in pancreatic tissues significantly increased with diabetes. In islet β cell experiments, inhibition of SDF‐1 expression resulted in an increase in the number of dedifferentiated cells, while overexpression of SDF‐1 resulted in a decrease. This seemed to be contradicted by the effect of diabetes on the expression of SDF‐1 in pancreatic tissue, but it was concluded that this may be related to the loss of SDF‐1 activity. SDF‐1 binds to CXCR4 to form a complex, which activates and phosphorylates AKT, subsequently increases the expression of forkhead box O1 (FOXO1), and inhibits the dedifferentiation of islet β cells. This suggests that SDF‐1 may be a novel target in the treatment of diabetes.

and mice being as high as 99%. 5 As a member of the chemokine family, it exerts unpredictable effects on various pathophysiological processes including cell differentiation, immune surveillance and inflammation. 6 On one hand, SDF-1 can regulate the differentiation and function of immune cells and play an anti-inflammatory and immunomodulatory role in type 1 diabetes. 7,8 On the other hand, a significantly higher plasma SDF-1 level is common in subjects with type 2 diabetes, 9 and it is associated with diabetic insulitis, nephropathy and adipose tissue inflammation. [10][11][12] SDF-1 is associated with the mediation of islet β cells function.
For example, our previous study found that the DPP-IV inhibitor Saxagliptin improved the function of islet β cells through regulating SDF-1 expression. 13 Another study also revealed the important role of SDF-1 in β cells survival after islet transplantation. 14 Furthermore, during the terminal differentiation stage of mature β cells, SDF-1 has been proven to prevent apoptosis and necrosis through activating PI3K/AKT and WNT/β-catenin pathways. [15][16][17] Furthermore, SDF-1 is also related to tissue damage repair in diabetic patients, including heart repair after acute myocardial infarction, 18 wound vascular healing 19 and skin scar formation. 20 In this study, we investigated the effect of SDF-1 on the dedifferentiation of pancreatic β -cells and investigated one of its possible mechanisms through cell experiments based on pancreatic tissues observation. Specifically, active SDF-1 can bind to CXCR4 and ultimately upregulate FOXO1 expression by phosphorylating AKT, thereby inhibiting the dedifferentiation of pancreatic β cells.
However, hyperglycaemia causes a partial loss of SDF-1activity, which is then unable to bind to CXCR4 or inhibit the dedifferentiation of pancreatic β cells.

| Human pancreatic tissue specimen
Pancreatic samples were collected from subjects undergoing pancreatectomy caused by non-neoplastic proliferations. Our study enrolled 3 participants with type 2 diabetes and 3 age-matched subjects without diabetes, from September 2018 to September 2019. A tissue sample from each subject was immediately fixed in formalin within 240 s after the pancreatectomy, prior to paraffin-embedding. The remaining specimens were frozen at −80°C for future use. (Ethics number:SZX-IRB-ZD-004(F)-002-01).

| Experimental animal
A total of 15 12-week-old male db/db mice (53 ± 4 g) and db/m mice (24 ± 1 g) were purchased from Jiangsu Jicui Yaokang Biotechnology Co., Ltd. Three db/m mice and three db/db mice were randomly selected for the experiment and repeated three times. The mice were housed in an environment at 24 ± 2°C with a 12 h day/night cycle, and had free access to tap water and standard chow diet. The mice were given analgesia immediately before surgery via buprenorphine

| Plasmid transfection
Lipo2000 was used for transient transfection (6-well plate, 4 μg plasmid, 10 μl lipo2000). MIN6 cells were divided into 3 groups: group 1 was supplemented with high glucose, group 2 was supplemented with high glucose and control-plasmid, and high glucose and SDF-1plasmid were added to group 3.

| Small interference RNA transfection
Lipo2000 was used for transient transfection(6-hole plate; 100 pmol siRNA; 5.0 μl Lipo2000). MIN6 cells were divided into 3 groups: group 1 was supplemented with high glucose, group 2 was supplemented with high glucose and control-siRNA, and high glucose and SDF-1-siRNA were added to group 3.

| Tissue immunofluorescence staining
The pancreatic tissue was fixed in 4% paraformaldehyde for 24 h, embedded in paraffin and sectioned into 4 μm slices. Before immunofluorescence staining, the slices were dewaxed in xylene and ethanol, and boiled in 1× TRIS-EDTA to repair the antigens. Next, the slices were covered with 1%Triton X-100, diluted by PBS and sealed with 1% BSA diluted by PBS, for 30 min, before being rinsed twice with PBS for 5 min.

| Triple staining
Triple staining of the pancreatic tissue slices was conducted using a fourcolour multi-labelled immunofluorescence staining kit. According to the instructions, diluted primary antibody solution was applied on top of the sample area to incubate the tissue at RT for 1 h. The slices were then twice washed with TBST for 3 min. Following that, the tissues were incubated with HRP secondary antibody working solution at RT for 10 min, and the slices were then twice washed with TBST for 3 min. Next, the tissue was covered with 1× 100 ul dye working solution (diluted with a signal magnifier at 100%) at RT for 10 min, and the slices were washed twice with TBST for 3 min. The process was then repeated, and the same slices were stained with different colour immunofluorescence dyes.
Finally, the slices were observed via a fluorescence microscope.

| Western blot
Equal amounts of total protein extracted from pancreatic tissue or MIN6 cells were set with 12.5% SDS-PAGE and transferred to a nitrocellulose membrane. After sealing with 5% skimmed milk, the mem- (1:1000). Finally, chemiluminescence was detected using an ECL imaging system, and the grey value was analysed via Image J software.

| Co-immunoprecipitation
Firstly, Dynabeads were incubated with a CXCR4 antibody (diluted at 1:50 with primary antibody dilutant) at RT for 10 min and then gently washed with PBS on the magnetic rack for 5 min to remove loose antibodies. Next, protein samples extracted from the cells were added and incubated with the Dynabeads for 10 min. After washing with PBS, the Dynabeads were heated in 1×loading to elute the proteins.

| Statistical analysis
GraphPad Prism software (version 8.0) was used for statistical analysis. A Student's t-test was used for comparison between the two groups, and a one-way ANOVA was used to compare three or more conditions. p < 0.05 was considered statistically significant.

| Phenotypic observation of the relationship between SDF-1 and dedifferentiation of islet β cells
In order to identify the dedifferentiated islet β cells, we selected SRYbox transcription factor 9 (SOX9). Its specific inactivation in mouse organs can result in exhaustion of the progenitor cell pool, leading to severe pancreatic hypoplasia. SOX9 can also maintain pancreatic progenitor cells by stimulating the proliferation, survival and persistence of pancreatic cancer cells. It can be used as an islet β-cell dedifferentiation marker, 21 as when dedifferentiation occurs, its expression level increases. Another dedifferentiation protein marker is Neurogenin 3 (NGN3), 22 which is highly expressed in pancreatic progenitor cells, and is very important in endocrine differentiation.
When dedifferentiation occurs, its expression level increases.

| The number of dedifferentiated islet β cells in diabetic patients and diabetic mice is significantly increased
In the first set of experiments, we investigated the differences in islet β-cell dedifferentiation between diabetic and non-diabetic states. Firstly, paraffin-treated sections of non-diabetic and diabetic pancreatic tissues were selected for the double immunofluorescence staining of insulin and SOX9. Red denoted Insulin, green was SOX9, and the insulin-labelled area is the islet. As can be seen in Figure 1A,B, there was a significantly higher expression of SOX9 in the islets of diabetic patients. In the Western blot, it can be seen that the expression of the dedifferentiation marker proteins SOX9 and NGN3 is increased in the pancreatic tissue of diabetic patients ( Figure 1E,F). Thus, it can be considered that more undifferentiated islet β cells appear in diabetic patients. Consistent results were obtained in the db/m and db/db mice experiments (Figure 1C,D,G,H).

| SDF-1 is significantly related to the dedifferentiation of islet β cells
Paraffin-treated sections of pancreatic tissue from diabetic and nondiabetic patients were double immunofluorescence stained. Insulin labelled with red fluorescence was used to mark the outline of the islet. As shown in Figure 2A,B, the green fluorescence representing SDF-1 mainly appeared in the islet area and was much denser in the diabetic group. In Figure 2E, SOX9 was used to mark the dedif- value appeared at 60 h ( Figure S1D). Therefore, the optimal intervention condition for dedifferentiation of Min6 cells occurred when the glucose concentration was 35 mmol/L with a treatment time of 60 h. However, as shown in Figure 3A,B, under these conditions, there was no significant change in SDF-1 expression before or after intervention. The interaction between SDF-1 and FOXO1 can be found in the STRING database ( Figure S1E). Currently, the most in-depth study focuses on the SDF-1/CXCR4 pathway. Therefore, we chose to investigate whether SDF-1 could affect the dedifferentiation of islet β cells through the SDF-1/CXCR4/AKT/FOXO1 pathway.

|
To verify this hypothesis, we detected the effect of diabetes on FOXO1 expression in pancreatic tissue via Western blot ( Figure 4A). Expression of FOXO1 in pancreatic tissue in a diabetic state was significantly reduced compared to patients/mice without diabetes ( Figure 4B). Consequently, we examined the effect 2. There was no significant change in SDF-1 expression before or after Min6 cell dedifferentiation induced by high glucose content.

SDF-1 expression in Min6 cells.
To attempt to solve these contradictions, several experiments were conducted.

| SDF-1 inhibited the dedifferentiation of islet β cells after binding with CXCR4 to form a complex
In some cases, SDF-1 alone did not perform any biological function.
However, upon binding with CXCR4 to form a SDF-1/CXCR4 complex, various effects occur. 25 Consequently, we explored whether, during the dedifferentiation of islet β cells, SDF-1 also bound to CXCR4 before it could function.
DPP-IV can cleave and inactivate SDF-1, which can then no longer activate CXCR4. In our previous research group studies, we found that compared with non-diabetic mice, the expression of DPP-IV in the islets of diabetic mice was enhanced. 13  To investigate whether SDF-1 affected the dedifferentiation of islet β cells by binding to CXCR4, Min6 cells were treated with different concentrations of DPP-IV including 250 ng/ml, 500 ng/ml, 750 ng/ml and 1,000 ng/ml for 48 h ( Figure 5A). It can be seen that DPP-IV did not affect the expression of SDF-1 and CXCR4 but did reduce the expression of FOXO1( Figure 5B). The optimal concentration affecting FOXO1 expression was determined as 1,000 ng/ ml; therefore, this intervention concentration was adopted in subsequent experiments.
The Western blot results demonstrated that after DPP-IV intervention, the expression of SOX9 and NGN3 in Min6 cells was upregulated compared with the non-intervention group, suggesting that more cells underwent dedifferentiation. At this time, the expression of p-AKT and FOXO1 was downregulated, while the expressions of SDF-1 and CXCR4 were not significantly changed ( Figure 5C,D). However, CO-IP results showed that binding of SDF-1 and CXCR4 was reduced after DPP-IV intervention ( Figure 5E,F).
So, we propose that DPP-IV did not change the overall expression of SDF-1 and CXCR4 but did reduce the binding amount between them, which further affected the downstream p-AKT and FOXO1 pathways, reducing their expression and increasing the number of dedifferentiated β cells.

| Verification that SDF-1 in combination with CXCR4 can upregulate FOXO1expression
It has been reported that phosphorylated FOXO1 expression in Min6 cells increases after 24 h of 25 mmol/L glucose treatment. 26 In order to verify this, we examined min6 cells under the same conditions.
After glucose intervention, Western blot results showed that the expression of SDF-1 and CXCR4 had not significantly changed, but the expression of p-AKT and FOXO1 had increased ( Figure 6A,B). The CO-IP results showed that the relative amount of binding of SDF-1 to CXCR4 increased ( Figure 6C,D). At the same time, we also verified the binding of SDF-1 and CXCR4 before and after the transfection of SDF-1-plasmid and SDF-1-siRNA in Min6 cells ( Figure 6E,G).
Therefore, we propose that a combination of SDF-1 and CXCR4 can increase FOXO1expression.

| DISCUSS ION
In this study, we discussed the relationship between SDF-1 and the dedifferentiation of islet β cells. The results indicated that SDF-1 can block the dedifferentiation of islet β cells by combining with CXCR4 to form the SDF-1/CXCR4 complex. This is then For instance, insulin secretion in diabetic patients and diabetic rats was significantly increased after bariatric surgery. 30 However, longterm exposure to hyperglycaemia/stress will cause β cell degeneration and will destroy the plasticity. Some studies have revealed that dedifferentiation occurred before apoptosis. 31 Therefore, the early recovery of dedifferentiated β cells function is particularly important.
We hope that our study on SDF-1 inhibition of islet β cell dedifferentiation can play a role in the treatment of diabetes.
Islet β cells firstly dedifferentiate into progenitor cells, then transform into glucagon-producing α cells 32 and lose their insulin secretion function, which leads to the onset of diabetes. Islet β cell dedifferentiation occurs in both humans and mice. 33 At this time, functional β cells will be greatly reduced, but not all of them will die.
Some of them dedifferentiate into endocrine progenitor cells; others are transdifferentiated into other endocrine cells such as α, δ and Pp. 4 Consequently, β cells 'escape from responsibility' by avoiding excessive metabolic pressure and protect themselves through adaptive mechanisms that avoid cell death under stress conditions. 34 The dedifferentiation of β cells is a 'selfish' protective, and adaptive, mechanism to minimize cell damage. 35,36 It has been reported that in type 1 diabetic NOD mice, β cell dedifferentiation can be achieved by knocking out the IRE1 α gene, while IRE1α-deficient NOD mice can be protected from autoimmune damage and the effects of diabetes. Consequently, it remains to be seen whether treating diabetes by inhibiting the dedifferentiation of β cells is only a temporary fix that is bad in the long-term.
In conclusion, SDF-1 can inhibit the dedifferentiation of islet β cells, but for a single gene, a single metabolic pathway to inhibit or even reverse dedifferentiated β cells in diabetes treatment, the role of SDF-1 still has a long way to go.

ACK N OWLED G EM ENT
This work was supported by the Science &Technology Development Fund of Tianjin Education Commission for Higher Education (grant numbers 2018KJ047).

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.  Writing -review & editing (equal).

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.