Bone Marrow-Derived Mesenchymal Stem Cells Repair Necrotic Pancreatic Tissue and Promote Angiogenesis by Secreting Cellular Growth Factors Involved in the SDF-1α/CXCR4 Axis in Rats

Acute pancreatitis (AP), a common acute abdominal disease, 10%–20% of which can evolve into severe acute pancreatitis (SAP), is of significant morbidity and mortality. Bone marrow-derived mesenchymal stem cells (BMSCs) have been reported to have a potential therapeutic role on SAP, but the specific mechanism is unclear. Therefore, we conducted this experiment to shed light on the probable mechanism. We validated that SDF-1α significantly stimulated the expressions of VEGF, ANG-1, HGF, TGF-β, and CXCR4 in BMSCs, which were inhibited by its receptor agonist, AMD3100. The capacities of proliferation, migration, and repair of human umbilical vein endothelial cells were enhanced by BMSCs supernatant. Meanwhile, BMSCs supernatant could also promote angiogenesis, especially after the stimulation with SDF-1α. In vivo, the migration of BMSCs was regulated by SDF-1α/CXCR4 axis. Moreover, transplanted BMSCs could significantly alleviate SAP, reduce the systematic inflammation (TNF-α↓, IL-1β↓, IL-6↓, IL-4↑, IL-10↑, and TGF-β↑), and promote tissue repair and angiogenesis (VEGF↑, ANG-1↑, HGF↑, TGF-β↑, and CD31↑), compared with the SAP and anti-CXCR4 groups. Taken together, the results showed that BMSCs ameliorated SAP and the SDF-1α/CXCR4 axis was involved in the repair and regeneration process.


Introduction
Though acute pancreatitis (AP) is considered one of the commonest acute abdominal diseases, no effective treatment has yet been available. The incidence of AP is rising continually with 10%-20% of patients progressing to severe acute pancreatitis (SAP) [1], which is associated with significant morbidity and mortality. AP can also transform into chronic pancreatitis and even pancreatic cancer and the overall mortality rate is reported to be about 15%-20% [2]. However, the pathogenic mechanisms of AP have still not been understood. It is well recognized that AP begins with pancreatic acinar injury, attributed to the premature activation of trypsin within the pancreatic acinar cells, after which exudation, edema, and a local inflammatory response are observed [3,4]. Further pathological deterioration can be prevented by the self-regulation of the bodies in most patients. But some AP patients can still progress to SAP with hemorrhage, necrosis, and a systematic inflammatory response, further leading to shock, multiple organ failure, and even death. Consequently, the treatment of SAP focuses predominantly on inhibiting the synthesis and secretion of trypsin and averting the systemic inflammatory response. Most patients with SAP require drug treatment, whereas some choose to receive surgery. However, no satisfactory therapeutic effect has been discovered whichever method is adopted.
Mesenchymal stem cells (MSCs), first described by Friedenstein et al. in 1976 [5], are multipotent adult stem cells that can be derived from many different organs and tissues. MSCs are characterized by the expression of cellsurface molecules, including CD44, CD73, CD90, and CD105 2 Stem Cells International [6]. MSCs can differentiate into bone, cartilage, muscle, adipose tissue, and so forth in vitro, depending on the culture conditions [7]. MSCs also have low or no immunogenicity which make them ideal choices for cell transplantation. Many studies have also demonstrated that MSCs have therapeutic effects on several autoimmune, ischemic, and inflammatory diseases [8]. Therefore, they are widely valued and studied. MSCs have been isolated initially from bone marrow and much of the literature uses these cells [9][10][11], though other sources of MSCs have been described [12]. Recent researches have also shown that bone marrow-derived mesenchymal stem cells MSCs (BMSCs) are involved in the pathogenesis of AP and can relieve the severity and improve the prognosis of SAP [9][10][11]. Thus, we selected these cells as our therapeutic tool for SAP. Jung et al. [9] found that BMSCs reduced the levels of proinflammatory cytokines and increased the numbers of Foxp3+ regulatory T cells to improve SAP repair. Tu et al. [10] believed that BMSCs attenuated systemic inflammation through relieving injury to pancreatic acinar cells and small intestinal epithelium. However, the specific mechanism has been a controversy so far. In our previous work, we have firstly demonstrated that stromal-cell-derived factor 1 (SDF-1 ) and its receptor, chemokine receptor 4(CXCR4), played an important role in the process of BMSCs therapy for SAP [11].
SDF-1 is a chemokine that regulates the migration of BMSCs by interacting with CXCR4 [13,14]. However, recent studies have also shown that the SDF-1 /CXCR4 axis not only regulates the migration of cells, but also repairs damaged tissue and promotes angiogenesis [15,16]. SDF-1 /CXCR4 axis has been reported to induce neovascularization in ischemic, tumor, and wounded tissues [17][18][19]. Besides, SDF-1 combines with vascular endothelial growth factor (VEGF) to enhance angiogenesis [20,21]. Therefore, we come to the conclusion that the SDF-1 /CXCR4 axis is probably involved in the repair of acute necrotizing pancreatitis through promoting the formation of new vessels.
To demonstrate our hypothesis, we conducted this experiment, further investigating whether SDF-1 /CXCR4 axis can regulate the migration of BMSCs to injured pancreas and validating whether it can also enhance tissue repair and regeneration to recover SAP by inducing angiogenesis in rats.
In the current study, we used superparamagnetic iron oxide nanoparticles (SPION) to label BMSCs in order to track their distribution, as previously described [22,23]. SPION have frequently been used to trace BMSCs in vivo. Recent studies have also indicated that the proliferation, migration, and differentiation capacities of SPION-labeled BMSCs are not affected by labeling (maintaining their "stemness") [24,25].

Animal Model.
Healthy Sprague Dawley rats weighing 200-250 g ( = 99) were purchased from Shanghai Laboratory Animal Co. Ltd (Shanghai, China). All animals were maintained at about 25 ∘ C with an alternating 12 h dark/12 h light cycle, with a free access to standard laboratory water and food. An animal model of SAP was established by retrograde pancreatic duct injection of Na-taurocholate as previously described [14,26]. All animal procedures were conducted according to the Shanghai Laboratory Animal Ordinance and approved by the Ethics Committee of Shanghai Tenth People's Hospital (Tongji University, Shanghai, China).

Cells and Cell
Culture. BMSCs were isolated, cultured, and identified as described in our previous study [14]. A human umbilical-vein endothelial cell line (EA.hy926 cells) was purchased from Shanghai Cell Bank of the Chinese Academy of Sciences. EA.hy926 cells were cultured in DMEM-LG complete medium with 10% FBS, 100 U/mL penicillin and 100 g/mL streptomycin at 37 ∘ C in 5% CO 2 . The cells were digested and passaged by 1 : 3 when they reached >90% confluence.

Sample
Collection. Blood samples and pancreatic tissues were collected from the normal control, sham, SAP, and SAP+BMSCs, SAP+anti-CXCR4 BMSCs groups. The serum was first extracted by centrifugation at 8000 ×g at 4 ∘ C and then stored at −80 ∘ C. The pancreatic tissues were stored in liquid nitrogen or fixed in 4% paraformaldehyde. After pretreatment with SDF-1 for 48 h, the BMSCs supernatants were collected, including the normal control, SDF-1 10 ng/mL, and SDF-1 100 ng/mL groups, and stored at −80 ∘ C. Meanwhile, the normal BMSCs supernatants were also collected at 24 h after cells were given fresh DMEM-LG complete medium.

Hematoxylin-Eosin (H&E)
Staining and the Detection of Serum Amylase Activity. H&E staining and the detection of serum amylase activity were performed as previously described [11].

2.7.
ELISAs. The serum levels of IL-1 , IL-4, IL-6, IL-10, TNF-, and TGF-were analyzed with ELISAs. Samples of serum (150 L) were diluted with standard diluent (1 : 1), and then 40 L diluted samples were added to the test sample wells. The test samples (10 L) were added to the wells (a final fivefold dilution of the samples), without touching the well walls as far as possible, and gently mixed. The plate was closed with the closure plate membrane and incubated for 30 min at 37 ∘ C. Wash solution was diluted 30-fold (or 20-fold) with distilled water and reserved. The closure plate membrane was removed, the liquid in the wells was discarded, the wells were dried by swinging the plate, and washing buffer was added to every well. The plate was left to stand for 30 s and then drained, this was repeated five times, and the plate dried by patting. Horseradish peroxidase-(HRP-) conjugated reagent (Agrisera, Sweden) (50 L) was added to each well, except the blank control well. After the plate was closed with the plate closure membrane, it was incubated for 30 min at 37 ∘ C. Wash solution was diluted 30-fold (or 20-fold) with distilled water and reserved. Chromogen solution A (50 L) and chromogen solution B (50 L) were added to each well, and the plate was incubated for 10 min at 37 ∘ C in the dark. Stop solution (50 L) was added to each well to stop the reaction (the solution changed from blue to yellow). Taking the blank well as zero, the absorbance of each well was read at 450 nm 15 min after the stop solution was added. The concentrations of the rat inflammatory cytokines in the samples were determined by comparing the absorbance of the samples with standard curves. Each sample included three repeated measurements.
2.8. Immunohistochemistry. The Histostain-Plus kit was used for the immunohistochemical analysis. Paraffin-embedded pancreatic tissues were dewaxed and rehydrated. Hydrogen peroxide (3%) was used to inactivate any endogenous peroxidase for 20 min at room temperature. Antigen retrieval was performed in a high-pressure cooker for 30 min. The pancreatic slices were incubated overnight with rabbit antirat VEGF (1 : 50) at 4 ∘ C in a wet box, after the endogenous antigens were blocked with 5% bovine serum albumin (BSA) for 30 min at room temperature. On the second day, the slices were incubated with biotin-labeled secondary antibody (goat anti-rabbit, Epitomics, China) for 30 min at 37 ∘ C and then with HRP-conjugated streptavidin for 20 min at 37 ∘ C. After the slices were stained with diaminobenzidine and hematoxylin, they were dehydrated, cleared, and mounted with neutral resin. CD31 and vWF proteins were detected with a similar procedure.
2.9. Immunoblotting Assay. Each sample of pancreatic tissue (approximately 100 g) was crushed to power and 500 L of RIPA lysis buffer containing 5 L of PMSF (100 : 1) was added to extract the total proteins. The concentration of the total proteins was determined with the BCA method. The proteins (20 g) were separated with 12% SDS-PAGE and transferred to 0.45 m nitrocellulose membrane. After the membrane was blocked with 5% skim milk for 1 h, it was incubated overnight with anti-VEGF polyclonal antibody (1 : 1000) at 4 ∘ C. On the following day, the membrane was washed three times with phosphate-buffered saline (PBS) containing Tween 20 for 10 min each time and then incubated with a secondary goat anti-rabbit antibody (1 : 1000) at 37 ∘ C for 1 h and washed again three times. The experiment was repeated five times. Finally, the signal was detected with the Odyssey 3.0 analysis software (LI-COR Biotechnology, USA).

qRT-PCR.
Total RNA was extracted from the cells and frozen pancreatic specimens with TRIzol Reagent. Firststrand cDNA was synthesized with the PrimeScript Reverse Transcriptase Reagent Kit (Takara Biotechnology, Japan) with oligo(dT) and random primer. The gene expression of the BMSCs and EA.hy926 cells was quantified with the KAPA Kit (Kapa Biosystems, USA) and 50 ng of cDNA was amplified in a 10 L reaction using the Applied Biosystems 7500 Real-Time PCR system based on SYBR Green dye (Applied Biosystems). The primers were purchased from Beijing Genomics Institute (Beijing, China). Rat glyceraldehydephosphate dehydrogenase (GAPDH) was used as the endogenous control. The sequences of the primers are listed in Table 1. qRT-PCR was performed with the following cycling conditions: 95 ∘ C for 3 min, followed by 40 cycles of 95 ∘ C for 1 s and 60 ∘ C for 20 s. Quadruplicate cycle threshold (CT) values were analyzed with the SDS software (Applied Table 1: The primer sequences of genes.

Rat Gene
Forward primer Reverse primer PCR amplified products (bp) Biosystems, USA), using the comparative CT method. The procedure was replicated three times. Each measurement was set three repeats.

Labeling BMSCs.
BMSCs at >90% confluence were pretreated with or without AMD3100 (10 g/mL) and then incubated with labeling medium containing 25 g/mL Fe 3+ and 0.75 ng/mL poly-l-lysine for 24 h. After they were labeled, the cells were washed three times with PBS and harvested with 0.25% trypsin-EDTA. The viability of the cells was assessed with Trypan Blue exclusion before cell transplantation. The functions of the different concentrations of Fe 3+labeled BMSCs were analyzed in the following experiments.

Cell Migration Assay.
In this experiment, 5 × 10 4 EA.hy926 cells were added to the upper chamber of the Transwell apparatus and supplemented with 180 L of DMEM and 20 L of 1% BSA, and the lower chamber was filled with 500 L of DMEM-LG complete medium with or without BMSCS supernatant (100 L). After incubation for 12 h, the upper chamber was removed and the cells were fixed with 4% paraformaldehyde for 30 min. The lower surface of the filter membrane was stained with 0.1% crystal violet for 10 min in the dark. After the membrane was thoroughly washed, it was observed with a digital camera and light microscope. The experiment was repeated three times. Finally, the crystal violet was dissolved in 300 L of 33% acetic acid and the absorbance of the solution was measured at 573 nm with an ELISA plate reader (Gene Company Limited, HK, China). After the slices were fixed, dehydrated, embedded, and sectioned, the tissues were stained with Prussian blue, as described above. The blue granules were counted by a person, who did not know the experiment, from ten areas of each histologic section. Experimental data are shown as means ± standard deviations (SD) and compared with Student's or a paired test or oneway ANOVA. A value of < 0.05 was deemed to indicate significant differences. significantly stimulated the expressions of VEGF, ANG-1, and CXCR4 in the BMSCs, which was also positively correlated with the concentration of SDF-1 . This effect was blocked by the CXCR4 agonist, AMD3100 (Figures 1(a)-1(b) and 1(d)-1(g)). The mRNA levels of HGF and TGF-were both upregulated after stimulation with SDF-1 but were inhibited by AMD3100 (Figure 1(c)).

BMSCs Supernatant
Promotes the Proliferation, Migration, and Repair of EA.hy926 Cells. The BMSCs supernatant promoted the growth and migration of EA.hy926 cells significantly more than DMEM-LG complete medium (Figures 2(a)  and 2(b)). Moreover, EA.hy926 cells impaired by trypsin were more effectively repaired by the BMSCs supernatant than by DMEM-LG complete medium (Figure 2(c)).

SDF-1 /CXCR4 Axis Induces Angiogenesis In Vitro, Being
Related to the Expressions of VEGF and ANG-1. The BMSCs supernatant pretreated with SDF-1 significantly promoted angiogenesis compared with the normal BMSCs supernatant in vitro, which was also positively related to the concentration of SDF-1 . Furthermore, the angiogenesis was reduced in both VEGF siRNA and ANG-1 siRNA compared with normal BMSCs supernatant (Figure 1(h)).
positive and the result showed that the high signals (white points) in T1WI became low signals (dark points) in T2WI, whereas the low signals (dark points) in T1WI became high signals (white points) as shown in Figure 5(c). Furthermore, the high signals in T1WI decreased obviously in anti-CXCR4 group compared with BMSCs group on posttransplant day 5 ( Figure 5(b)).

Transplanted BMSCs Reduced Pancreatic Edema, Hemorrhage, and Necrosis, Inhibited Systematic Inflammation, and Promoted the Formation of TCs Involved in the SDF-1 /CXCR4
Axis. Pancreatic edema, hemorrhage, and necrosis were markedly reduced and the levels of serum amylase were significantly lower in the SAP+BMSCs group than in the SAP and SAP+anti-CXCR4 BMSCs groups (Figures 6(a) and 6(b)). Inflammatory cytokines were detected with ELISAs.
The results show that the levels of serum proinflammatory cytokines (IL-1 , IL-6, and TNF-) were significantly downregulated in the SAP+BMSCs group compared with the SAP and SAP+anti-CXCR4 BMSCs groups. In contrast, the levels of serum anti-inflammatory cytokines (IL-4, IL-10, and TGF-) were significantly upregulated in the SAP+BMSCs group compared with the SAP and SAP+anti-CXCR4 BMSCs groups (Figure 6(c)). We also found that a large number of TCs appeared in the BMSCs group (as indicated by red arrow and a magnified picture in Figure 6(a)).

SDF-1 /CXCR4 Axis Enhances the Expressions of VEGF and ANG-1 in Damaged Pancreatic
Tissues. VEGF expression in injured pancreatic tissues was measured with qRT-PCR and immunoblotting. VEGF expression was higher in the SAP+BMSCS group than in the normal control, SAP and SAP+anti-CXCR4 BMSCs groups on postoperative days 1 and 4 but had decreased on postoperative day 7 (Figures 7(b), 7(c), and 7(e)). The expression of ANG-1 was detected with qRT-PCR and immunoblotting and its expression was higher in the SAP+BMSCS group than in the normal control, SAP, and SAP+anti-CXCR4 BMSCs groups (Figures 7(a), 7(c),  The Prussian blue staining of pancreatic tissue indicates that the cells were stained blue (as indicated by black arrows), gradually increased, and peaked on postoperative days 5-7, when the formation of tubular complexes was also maximal (as indicated by red arrows). However, the migration was partly inhibited and the trend was not obviously investigated in anti-CXCR4 group. (b) The lung tissues were stained by Prussian blue and the blue particles (as indicated by black arrows) were gradually decreasing in both BMSCs and anti-CXCR4 groups. (c) The blue particles in lung, pancreas, liver, spleen, and small intestine were counted and analyzed between BMSCs and anti-CXCR4 groups. The result showed that the number of blue particles of pancreas in BMSCs group was significantly more than in anti-CXCR4 group at postoperative days 1, 3, 5, and 7. Data are expressed as mean ± SD ( * * < 0.01 and * * * < 0.001 for BMSCs versus anti-CXCR4 at each corresponding time point). and 7(d)). The mRNA levels of HGF and TGF-were also detected with qRT-PCR and were significantly higher in the SAP+BMSCs group than in the normal, SAP, and SAP+anti-CXCR4 BMSCs groups (Figures 7(f) and 7(g)).

SDF-1 /CXCR4 Axis Induces the Neovascularization In
Vivo. To assess the neovascularization, we conducted qRT-PCR or immunohistochemistry for detecting these angiogenesis markers (CD31, VEGF, and vWF). The results indicated that the expressions of CD31, VEGF, and vWF in necrotic pancreatic tissue were higher in the SAP+BMSCs group than in the normal control, SAP, and SAP+anti-CXCR4 BMSCs groups, especially in early phase (Figures 7(h) and 8(a)-8(d)).

Discussion
Several studies have shown that MSCs from the bone or umbilical cord can attenuate acute necrotic pancreatitis by inhibiting systematic inflammation, regulating the immune responses, reducing the apoptosis of pancreatic acinar cells, and repairing damaged the small intestinal epithelium [9][10][11][12]. However, these functions do not account for how to repair and regenerate the necrotic pancreatic tissues. In the study, we found that BMSCs can reduce severe acute pancreatitis by alleviating the systematic inflammation, promoting the formation of tubular complexes (TCs), and inducing angiogenesis involving the SDF-1 /CXCR4 axis by enhancing the expression of cell growth factors in a rat model of SAP. Large previous studies have found that MSCs can not only differentiate into different kinds of functional adult cells but also secrete various soluble factors, including VEGF, HGF, and TGF- [27]. These factors which promote angiogenesis and mitosis and reduce apoptosis can explain tissue repair and regeneration after MSCs transplanted [28,29]. Moreover, the regeneration of the pancreas begins with the TCs that consist of a cluster of epithelia surrounded by the mesenchymal cells [30,31]. Thus, we inferred that SDF-1 /CXCR4 axis was involved in the repair and regeneration of SAP. SDF-1 and its receptor CXCR4 have been studied extensively, together with their roles in migration, homing, proliferation, angiogenesis, and so forth [11, 13-15, 18, 32, 33]. Many researches have shown that BMSCs can migrate to injured tissues via the SDF-1 /CXCR4 axis and relieve myocardial infraction [13,14], promote wound healing [19], heal bone fracture [34], reduce cerebral ischemia [35], and ameliorate acute necrotizing pancreatitis [11]. SDF-1 expression is upregulated by hypoxia-inducible factor-1 (HIF-1) [36]. In the study, we successfully established a rat model of acute necrotizing pancreatitis with the retrograde injection of Na-taurocholate, which causes acinar cell injury/necrosis by triggering transient pathological intra-acinar-cell calcium ions and activating digestive zymogens [37][38][39]. The injured/necrotic pancreatic tissue causes HIF-1 expression and thus induces SDF-1 expression, which attracts BMSCs to injured pancreas through the interaction with CXCR4. For further investigating whether BMSCs could migrate to injured pancreas and the migration was regulated by SDF-1 /CXCR4 axis in a SAP rat model, SPION was used to label BMSCs. SPION were initially used as a unique MR contrast agent in the clinical context. Recently, their two uses have been discovered-the labeling of live cells and tissues, disease diagnosis, and therapy [22,23,[40][41][42]. Studies have demonstrated that SPION-labeling does not deleteriously affect the functions of the target cells [25]. In our study, we also demonstrated that there was no difference in the proliferation capacities between SPION-labeled (Fe = 25 g/mL, labeling efficiency > 95%) and unlabeled BMSCs. Moreover, the expression of CXCR4 was also similar between the labeled and unlabeled cells. SPION-labeled BMSCs could be detected by both MRI and Prussian blue staining because SPION is the nanometer particles of ferric oxide. The dynamic movement of labeling cells could be observed at the photos of MRI so that we can intuitively assess the proportions of SPION-labeled BMSCs in vivo. The result showed that these cells could migrate to the damaged pancreatic tissues and there was also a gradual increase in migration, peaking on postoperative days 5-7, with the formation of large numbers of TCs. SDF-1 /CXCR4 axis could regulate the migration of BMSCs as our previously described [11]. Meanwhile, the † † † *   ×100) is showing that the edema, infiltration, and necrosis were significantly reduced in the SAP+BMSCs group compared with SAP and SAP+anti-CXCR4 BMSCs groups at postoperative days 1, 4, and 7, respectively. A large number of tubular complexes were also investigated in SAP+BMSCs group (as indicated by red arrow in a 400x magnified picture). Serum amylase activity was also significantly reduced in SAP+BMSCs group than in normal, SAP, and anti-CXCR4 BMSCs groups at postoperative days 1, 4, and 7, respectively. (c) The levels of serum proinflammatory cytokines are significantly lower in SAP+BMSCs than in SAP and SAP+anti-CXCR4 BMSCs groups. In contrast, the levels of serum anti-inflammatory cytokines are significantly higher in SAP+BMSCs than in SAP and SAP+anti-CXCR4 BMSCs groups. Data are expressed as mean ± SD ( # > 0.05 for sham versus normal, * < 0.05, * * < 0.01, and * * * < 0.001 for SAP versus Normal, † < 0.05, † † < 0.01, and † † † < 0.001 for SAP versus SAP+BMSCs, $ < 0.05, $$ < 0.01, and $$$ < 0.001 for SAP+BMSCs versus SAP+anti-CXCR4 BMSCs). Data analysis was performed by Student's test (BMSCs, bone marrow-derived mesenchymal stem cells, H&E, hematoxylin-eosin).

Control
number of cells, which migrated to the necrotic pancreas, was positively related to the therapeutic effect. However, SPIONlabeled BMSCs could be rarely seen on postoperative day 10, perhaps because the SPION were degraded or the BMSCs had differentiated into other cells, including pancreatic exocrine cells.
To validate whether the SDF-1 /CXCR4 axis also plays an important role in the repair and regeneration of SAP, we conducted the study in vitro and vivo. We found that the secretions of BMSCs could promote the proliferation, migration, and restoration of EA.hy926 cells. The BMSCs supernatant induced angiogenesis and the SDF-1 /CXCR4 axis stimulated BMSCs to express VEGF, ANG-1, HGF, and TGF-, promoting further angiogenesis in vitro. Both VEGF and ANG-1 play important roles in angiogenesis. Moreover, the expressions of VEGF, ANG-1, HGF, and TGFin the damaged pancreatic tissues were higher in the BMSCs transplantation group than in the normal control, SAP, and SAP+anti-CXCR4 BMSCs groups. Taken all together, the result showed that transplanted BMSCs repaired the damaged pancreas through secreting large cellular growth factors and the effect was regulated by SDF-1 /CXCR4 axis.
As we all know, VEGF, a potent inducer of angiogenesis, promotes the repair of injured vascular endothelial cells and neovascularization. Studies have shown that the SDF-1 /CXCR4 axis promotes the synthesis and secretion of VEGF [17,43] and SDF-1 , combined with VEGF, enhances ischemic angiogenesis [20]. In our study, we confirmed that  The mRNA level of HGF, TGF-, and CD31 was significantly higher in SAP+BMSCs group than in normal, SAP, and SAP+anti-CXCR4 BMSCs groups. Data are expressed as mean ± SD ( # > 0.05 for sham versus NC, * < 0.05 and * * < 0.01 for SAP+BMSCs versus NC, † < 0.05, † † < 0.01, and † † † < 0.001 for SAP+BMSCs versus SAP, $ < 0.05, $$ < 0.01, and $$$ < 0.001 for SAP+BMSCs versus SAP+anti-CXCR4 BMSCs at each corresponding time point) (NC, normal control, ANG-1, angiopoietin-1, VEGF, vascular endothelial growth factor, SAP, severe acute pancreatitis, and BMSCs, bone marrow-derived mesenchymal stem cells). the SDF-1 /CXCR4 axis correlates positively with VEGF expression in BMSCs. In vivo, VEGF expression in pancreatic tissues in the early phase of BMSCS transplantation was significantly higher than that in the normal, SAP, and SAP+anti-CXCR4 BMSCs groups. Therefore, we confirmed that the SDF-1 /CXCR4 axis is involved in the upregulation of VEGF expression in pancreatitis tissues. On the one hand, the upregulation of SDF-1 attracted BMSCs to damaged pancreatic tissues. On the other hand, the interaction of SDF-1 with CXCR4 increased the expression of VEGF, which repaired the damaged vascular endothelial cells and further enhanced angiogenesis. SDF-1 also increased the expression of the CXCR4 receptor as shown in our experiment and further promoted the migration of BMSCs to injured pancreatic tissues. Meanwhile, VEGF can also enhance the expression of CXCR4 [15]. Thus, the SDF-1 /CXCR4 axis and VEGF form a virtuous circle, producing a cascade effect. As a result, the association between the SDF-1 /CXCR4 axis and VEGF could account for the migration of BMSCs, the restoration of damaged vascular endothelial cells, neovascularization, and the ultimate repair and regeneration of necrotic pancreatic tissues. However, we also found that the expression of VEGF had decreased at posttransplant day 7 because the ischemia and hypoxia in the pancreatic tissues were significantly reduced after cell transplantation.
ANG-1, also involved in the process of angiogenesis, was upregulated by the SDF-1 /CXCR4 axis, and that ANG-1 expression was higher in the SAP+BMSCS group than in the normal, SAP, and SAP+anti-CXCR4 BMSCs groups. Large studies have also demonstrated that VEGF, combined with ANG-1, promotes angiogenesis more effectively [44]. Thus, the SDF-1 /CXCR4 axis, VEGF, and ANG-1 are jointly involved in the process of neovascularization of necrotic pancreatic tissues. In addition, HGF was significantly expressed after BMSCs were stimulated by the SDF-1 /CXCR4 axis. HGF expression was also elevated in damaged pancreatic tissues after BMSCS transplantation. HGF can stimulate epithelial cell proliferation, motility, morphogenesis, and angiogenesis in various organs via the tyrosine phosphorylation of its receptor, c-MET [45]. Bai et al. [46] found that HGF mediated MSC-induced recovery in a model of multiple sclerosis. Wang et al. [31] found that HGF/c-Met are expressed inside pancreatic tubular complexes and are possibly involved in the regeneration of islet cells. In our study, we have also confirmed that HGF participates in the process of repairing and regenerating necrotic pancreatic tissue.
As for TGF-, it is a cytokine with a wide range of diversity and often has contradictory functions. Recent studies lay emphasis on its anti-inflammatory and antiatherogenic