Hypoxia Inducible Factor-1α Regulates the Migration of Bone Marrow Mesenchymal Stem Cells via Integrin α 4

Although hypoxic environments have been known to regulate the migratory ability of bone marrow-derived mesenchymal stem cells (BM-MSCs), which is a critical factor for maximizing the therapeutic effect, the underlying mechanisms remain unclear. Therefore, we aimed to confirm the effect of hypoxia-inducible factor-1α (HIF-1α) on the migration of BM-MSCs and to analyze the interaction between HIF-1α and integrin-mediated signals. Hypoxia-activated HIF-1α significantly increased BM-MSC migration. The expression of integrin α 4 was decreased in BM-MSCs by increased HIF-1α under hypoxia, whereas the expression of Rho-associated kinase 1 (ROCK1) and Rac1/2/3 was increased. After downregulation of HIF-1α by YC-1, which is an inhibitor of HIF-1α, BM-MSC migration was decreased via upregulation of integrin α 4 and downregulation of ROCK1 and Rac1/2/3. Knockdown of integrin α 4 by integrin α 4 siRNA (siITGA4) treatment increased BM-MSC migration by upregulation of ROCK1, Rac1/2/3, and matrix metalloproteinase-2 regardless of oxygen tension. Moreover, siITGA4 treatment increased HIF-1α expression and augmented the translocation of HIF-1α into the nucleus under hypoxia. Taken together, the alternative expression of HIF-1α induced by microenvironment factors, such as hypoxia and integrin α 4, may regulate the migration of BM-MSCs. These findings may provide insights to the underlying mechanisms of BM-MSC migration for successful stem cell-based therapy.


Introduction
Mesenchymal stem cells (MSCs) are adult multipotent stem cells capable of differentiation into cells originating from any of the three germ layers, that is, the endoderm, mesoderm, and ectoderm [1]. Bone marrow is an abundant source of MSCs, and bone marrow-derived mesenchymal stem cells (BM-MSCs) have been extensively studied and determined to have the ability to differentiate into osteogenic, chondrogenic, and adipogenic lineages with the capability to migrate toward injured tissues and sites of inflammation [2][3][4][5]. Recently, MSCs have been applied in an increasing number of clinical trials for the treatment of various diseases based on their proven efficacies in preclinical and clinical studies [6,7]. To maximize the therapeutic efficacy of stem cell therapy using BM-MSCs, several requisite characteristics should be established: (1) a high survival rate and a high proliferative potential of transplanted cells; (2) effective homing properties; and (3) sufficient interactions between grafted MSCs and environmental factors at sites where MSCs have migrated [8]. Among these factors, effective homing, which is the capability of MSCs to migrate into target sites, is the most important and challenging characteristic to achieve [9].
Generally, the homing mechanisms of MSCs have been shown to be similar to those of inflammatory cells [10]. Proinflammatory factors released from damaged tissues induce the production of chemokines, cytokines, and the expression of adhesion molecules [11,12]. Kim et al. reported that pretreatment with tumor necrosis factor-, a cytokine involved in acute inflammation, enhanced the adhesiveness 2 Stem Cells International and migration of MSCs through the overexpression of bone morphogenetic protein-2 [13]. Furthermore, stromal cellderived factor 1 (SDF-1 ), which is also known as C-X-C motif chemokine 12, has been shown to play a crucial role in cell-cell adhesion, adhesion to extracellular matrix (ECM), and cell migration [14,15].
Hypoxia-inducible factor-1 (HIF-1 ) is a key mediator of the adaptive cellular response to hypoxia and is upregulated under hypoxic conditions, modulating the expression of numerous genes that affect cellular survival and metabolism [16]. Recently, it was determined that MSC migration was enhanced under hypoxic conditions through the increased expression of chemokine receptors, such as CX3C chemokine receptor 1, C-X-C chemokine receptor type 4 (CXCR4), and SDF-1 [17]. Therefore, hypoxia preconditioning was tried and revealed to enhance the survival and engraftment of MSCs in previous in vitro and in vivo studies [18][19][20].
The homing capabilities of MSCs are also affected by the dynamic expression of integrins, which are heterodimeric transmembrane proteins composed of and subunits that regulate cell-cell adhesion, cell differentiation, and migration [21]. Saller et al. suggested that low oxygen concentration augments the stemness and migration of MSCs and alters integrin expression [22]. However, the underlying mechanisms through which altered integrin expression under hypoxia affects the migration of MSCs have not been fully elucidated.
The Rho GTPase family consists of three members: Rac1, RhoA, and Cdc42 [23]. Rho GTPases have been determined to be associated with a variety of cellular processes, especially in the regulation of cytoskeletal dynamics and cell migration [24]. The overexpression of RhoA in noninvasive cells gives rise to the increased invasiveness, whereas the inhibition of RhoA suppresses cellular invasiveness by modulating downstream signals, such as Rho-associated kinase (ROCK) and myosin light chain [25]. However, it is still unclear whether the Rho family of GTPases influences the migration and homing of MSCs.
Therefore, we aimed to analyze the effect of increased expression of HIF-1 on the migration of BM-MSCs and to assess the alterations of integrins and Rho GTPases in BM-MSCs under hypoxia. Moreover, we evaluated the correlation between HIF-1 and integrin 4 in BM-MSCs.

Downregulation of Integrin 4 in BM-MSCs Using Short
Interfering RNA (siRNA). BM-MSCs were treated with integrin 4 siRNA (siITGA4) (Santa Cruz Biotechnology, Santa Cruz, CA, USA) or mock to inhibit integrin 4 expression according to the manufacturer's protocol under normoxic or hypoxic conditions. Briefly, 4 L of Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) containing 100 pM of siITGA4 was added to the cells. The cells were cultured for 24 hours under normoxic or hypoxic conditions.

Immunofluorescence.
To investigate the expression of HIF-1 and integrin 4 , BM-MSCs (5 × 10 4 cells/well) seeded onto coverslips were pretreated with 20 M of YC-1, which is a specific inhibitor of HIF-1 , and cultured under normoxic or hypoxic conditions for 24 hours. The cultured cells were washed with cold phosphate-buffered saline (PBS) and incubated with serum-free protein blocking buffer (Dako, Glostrup, Denmark) at 37 ∘ C for 1 hour followed by incubation with mouse anti-HIF-1 antibody (1 : 100, Santa Cruz Biotechnology) or mouse anti-integrin 4 (1 : 100, Novus Biologicals, Littleton, CO, USA) overnight at 4 ∘ C. After washing with PBS, the cells were incubated with a secondary antibody conjugated to Alexa 488 or Alexa 588 (1 : 500, Invitrogen) for 1 hour at room temperature. Then, the cells were stained with 4 ,6-diamidino-2-phenylindole (DAPI) for nuclear counterstaining and observed under a fluorescence microscope at a magnification of 400x (Axioskop2, Carl Zeiss Microimaging, Germany).

Invasion Assay.
The invasiveness of BM-MSCs was analyzed using 24-well filtered inserts with membranes (8 m pore size; Thermo Fisher Scientific, Rockford, IL, USA) on 24-well plates. In the inserts, 3 × 10 4 cells were seeded and treated with serum-free Opti-MEM (Gibco-BRL) containing siITGA4 or YC-1. Culture medium containing FBS was added in the lower well and incubated for 24 hours under normoxic or hypoxic conditions. After incubation, the cells in the upper wells were completely removed with a cotton swab. The invading cells that had attached to the bottom side of the filter were fixed with methanol for 20 minutes and stained with Mayer's hematoxylin (Sigma-Aldrich) for 20 minutes. The cell invasion ability was determined by counting the number of stained cells attached to the other side of the filter in seven randomly selected fields on the membranes at a magnification of 100x. Cell invasion under different treatments was normalized to controls and expressed as the mean invasion (% invasion ± SEM).

Gelatin Zymography.
To analyze the activities of matrix metalloproteinase-(MMP-) 2 and MMP-9, BM-MSCs were treated with YC-1 or siITGA4 and cultured under hypoxia. Then, the conditioned medium was analyzed by zymography. The conditioned medium was separated by 12% SDS-PAGE supplemented with 1 mg/mL of gelatin (Bio-Rad Laboratories). The separated proteins were incubated for 30 minutes using a renaturation buffer (Bio-Rad Laboratories), rinsed, and incubated in a development buffer (Bio-Rad Laboratories) at 37 ∘ C for 24 hours. The gels were stained with Coomassie Brilliant Blue R-250 solution for 2 hours at room temperature and then destained with a buffer comprised of 10% acetic acid, 30% methanol, and 60% deionized water until the zymogen bands were visualized. The activities of MMP-2 and MMP-9 were analyzed by the density of unstained bands. All experiments were performed in triplicate.

Statistical
Analysis. Student's t-tests were performed for group-wise comparisons and a P value less than 0.05 was considered statistically significant. All experiments were performed in triplicate.

HIF-1 Induced by Hypoxia Enhances the Migration of BM-MSCs through Activation of MMP-2.
To confirm the effect of hypoxia on the migration of BM-MSCs, we analyzed the expression of HIF-1 . When the cells were cultured under hypoxia, the transcription and expression of HIF-1 were significantly higher than those of cells cultured under normoxia ( < 0.05; Figures 1(a) and 1(b)). After the pretreatment with YC-1, an inhibitor of HIF-1 , the expression of HIF-1 significantly decreased under both normoxia and hypoxia (Figure 1(c)). The localization of HIF-1 into the nucleus, which indicates activation of HIF-1 , was more pronounced when the BM-MSCs were cultured under hypoxia. However, the translocation of HIF-1 in BM-MSCs was decreased under both normoxia and hypoxia when the cells were pretreated with YC-1 (Figure 1(d)). While the migration of BM-MSCs was significantly augmented under hypoxic conditions compared with normoxic conditions, the migration was suppressed by YC-1 pretreatment under both normoxic and hypoxic conditions ( < 0.05; Figure 1(e)). In a previous study, the increased activities of MMP-2 and MMP-9 were determined to contribute to BM-MSCs [26]; we therefore verified the changes in the activities of MMP-2 and MMP-9 by zymography. As shown in Figure 1(f), the activity of MMP-2 significantly increased under hypoxia compared with normoxia ( < 0.05), whereas the activity of MMP-9 did not significantly change. Collectively, these results suggest that the migration of BM-MSCs was directly regulated by HIF-1 through MMP-2 activation under hypoxic conditions.

Expression of Integrins and Rho GTPases in BM-
MSCs under Hypoxia. Generally, the dynamic expression of molecules involved in cell adhesion and cytoskeletal remodeling, such as integrins and Rho family proteins, through environmental signals is required for cellular migration. Therefore, we explored the effect of hypoxia on the migration ability of BM-MSCs through the alternative expression of adhesion molecules, including integrins, phosphorylated FAK, and Rho GTPases. Interestingly, the expression of integrin 4 was significantly lower when BM-MSCs were exposed to hypoxia, whereas the expression of integrin 5 was significantly higher under hypoxia ( < 0.05; Figure 2(a)). However, the expression of integrin 7 and phosphorylated FAK was not significantly different between cells cultured under normoxia and hypoxia (Figure 2(a)). Next, we analyzed the expression of Rho proteins. Although the expression of RhoA and phosphorylated Rac1/cdc42 was not changed by hypoxia, the expression of Rac1/2/3 and ROCK1, which is a downstream signal of integrins and Rho GTPase family, was augmented by hypoxia ( < 0.05; Figure 2(b)). These findings indicate that the alternative expression of integrin 4 and ROCK1 in BM-MSCs is induced by hypoxia. Moreover, integrin 4 could be negatively correlated with migration of BM-MSCs.    (Figure 3(a)). However, the levels of mRNA transcription and protein expression of integrin 4 in BM-MSCs were increased when the cells were pretreated with YC-1 compared with controls regardless of oxygen concentration (Figures 3(b) and 3(c)). To verify the effect of HIF-1 on Rho GTPases, we analyzed the expression of ROCK1 and Rac1/2/3 in BM-MSCs after YC-1 pretreatment. The augmented expression of ROCK1 and Rac1/2/3 in BM-MSCs under hypoxia significantly declined with YC-1 pretreatment (Figure 3(d)). Therefore, these findings indicate that HIF-1 reduces the transcription and expression of integrin 4 and enhances the expression of ROCK1 and Rac1/2/3 in BM-MSCs under hypoxia. The level of mRNA integrin 4 was significantly decreased by the transfection of siITGA4 under both hypoxic and normoxic conditions (Figure 4(a)). After siITGA4 treatment, the migration of BM-MSCs was revealed to be significantly increased under both normoxia and hypoxia ( < 0.05; Figure 4(b)). Moreover, enhanced migration induced by siITGA4 treatment was more evident under hypoxia compared with normoxia ( < 0.05; Figure 4 the MMP-2 activity in BM-MSCs, which was elevated by hypoxia, significantly increased when the cells were transfected with siITGA4 (Figure 4(c)). While the MMP-9 activity was not altered by hypoxia, downregulation of integrin 4 by siITGA4 induced activation of MMP-9 under hypoxic conditions (Figure 4(c)). Taken together, downregulated integrin 4 enhances the migration of BM-MSCs through the stimulation of MMP-2 activity under normoxia as well as hypoxia.

Interaction between Integrin 4 and Hypoxia and Its
Effect on the Rho GTPase Family. To explore the possible interaction between integrin 4 and HIF-1 under hypoxia and its effect on the migration of BM-MSCs via Rho GTPases, we analyzed the expression levels of ROCK1, Rac1/2/3, and HIF-1 after siITGA4 transfection under normoxia and hypoxia. The expression of HIF-1 , ROCK1, and Rac1/2/3 was significantly increased by downregulated integrin 4 expression under both normoxia and hypoxia. Specifically, augmentation of their expression by siITGA4 was marked under hypoxic conditions compared with normoxic conditions ( Figure 5(a)). HIF-1 was revealed to regulate the expression of integrin 4 and to be modulated by integrin 4 , indicating the possible positive feedback between HIF-1 and integrin 4 . Therefore, we assessed the translocation of HIF-1 in BM-MSCs under hypoxia using immunofluorescence. HIF-1 translocated into the nucleus of the cells under hypoxia, whereas it was located in cytoplasm under normoxia. Furthermore, the translocation of HIF-1 into the nucleus became obvious with the knockdown of integrin 4 by siRNA transfection (Figure 5(b)). These results suggest that HIF-1 and integrin 4 interact and regulate each other and modulate the dynamic expression of the Rho GTPase family leading to the regulation of BM-MSC migration.

Discussion
Stem cell therapy using MSCs has been explored for the treatment of various degenerative diseases. However, the mechanisms of action are poorly understood. MSC migration is a critical factor determining the efficacy of stem cell therapy because the therapeutic effect of MSCs can only be expected after the proper engraftment of transplanted MSCs to the damaged tissues. Ceradini and colleagues reported that the recruitment of progenitor cells into the regenerating tissues was regulated by hypoxic gradients via the HIF-1 induction of SDF-1, which binds to CXCR4 on circulating progenitor cells [27]. In a previous clinical study, it was demonstrated that MSC coinfusion improved hematopoietic stem cell engraftment through restoration of a normal level of SDF-1 in 8 patients with acute myeloid leukemia undergoing hematopoietic stem cell transplantation [28]. However, the interactions between HIF-1 and the signaling molecules, such as integrins, MMPs, and Rho GTPases, under hypoxia and their influences on MSC migration have not been fully elucidated.
In the present study, we verified that hypoxia-activated HIF-1 enhanced the migration of BM-MSCs. Furthermore, the hypoxia-induced enhancement of BM-MSC migration was diminished after inhibition of HIF-1 . These results suggest that HIF-1 may be a crucial regulatory signal in facilitating BM-MSC migration under hypoxic conditions. If HIF-1 is activated under hypoxic conditions, HIF-1 translocates into the nucleus and activates the transcription of genes involved in cell survival, differentiation, and migration [29]. In the present study, the translocation of HIF-1 into the nucleus of BM-MSCs was demonstrated under hypoxia and additionally became marked after the knockdown of integrin 4 . We also identified that the expression of integrin 4 was decreased by HIF-1 under hypoxic conditions. Therefore, the results of our study implicate the possible existence of a positive feedback loop between HIF-1 and integrin 4 . Integrins are transmembrane proteins known to regulate a variety of physiologic events, such as cell growth, differentiation, and migration [30,31]. In a previous study, the expression of integrin 5 was determined to increase under hypoxic conditions, resulting in the enhancement of extravillous trophoblast cell migration during early pregnancy [32]. Moreover, it was demonstrated that HIF-1induced upregulation of integrin 5 mediated cancer cell invasion [33]. Although we previously reported the effect of HIF-1 -induced integrin 4 suppression on trophoblast invasion [34], HIF-1 -induced downregulation of integrin 4 and the positive feedback loop between HIF-1 and integrin 4 in BM-MSCs exposed to hypoxia have not been reported so far.
MMPs are proteases that degrade the ECM proteins and play a major role in cancer invasion and metastasis [35]. It was reported that hypoxia-activated HIF-1 reduced E-cadherin expression and augmented MMP-2 expression during cancer cell migration [36]. Therefore, according to the results of the present study, BM-MSC behavior may resemble those of cancer cells with respect to augmenting cellular invasiveness via MMP-2 activation under hypoxia. The activities of MMP-9 as well as MMP-2 in trophoblast cells increased under hypoxic conditions in the aforementioned study [34]. In our study, MMP-2 activity was significantly enhanced under hypoxia contributing to increased BM-MSC migration as a result of suppressing integrin 4 , whereas MMP-9 activity was not altered. This finding suggests that enhancement of BM-MSC migration under hypoxia is attributed to augmented intratissue migration via MMP activity, especially MMP-2, as well as cell adhesion via adhesion molecules.
FAK, a cellular focal adhesion-associated protein kinase, is known to be involved in cell adhesion and migration [37,38]. Skuli et al. showed that HIF and reduced oxygen tension increased the expression of integrin 5 3 leading to increased migration of trophoblast stem cells through FAK activation [39]. In addition, integrin 5 1 was identified to mediate MSC migration during vascular remodeling by inducing FAK activity and platelet-derived growth factor receptor-phosphorylation [40]. Tyrosine phosphorylation of FAK in response to growth factor stimulation and integrin engagement was shown to trigger phosphorylation of paxillin, which regulates Rho GTPases [41]. Based on this evidence, we hypothesized that FAK and Rho GTPases may be negatively regulated by integrin 4 . The knockdown of integrin 4 by siRNA transfection induced upregulation of Rac1/2/3 and ROCK1 and changes in their expression levels were marked under hypoxic conditions. However, FAK and RhoA, which are upstream signals of ROCK1, were not significantly changed by inhibition of integrin 4 (data not shown). Therefore, these findings indicate that increased expression of Rac1/2/3 and ROCK1 as a result of integrin 4 suppression under hypoxia stimulates the migration of BM-MSCs, whereas tyrosine phosphorylation of FAK was not induced.

Conclusions
In summary, our study demonstrated that HIF-1 -mediated downregulation of integrin 4 facilitated the migration of BM-MSCs under hypoxia via MMP-2 activation and enhanced expression of Rac1/2/3, which belongs to the Rho GTPase family, and ROCK1, which is a downstream signal of integrins and Rho GTPases. Moreover, the possible existence of a positive feedback loop between HIF-1 and integrin 4 was revealed. Although further in vivo studies are needed to confirm the influences of the interactions between HIF-1 , integrin 4 , and the Rho family of GTPases on BM-MSC homing and migration, stabilization of HIF-1 or knockdown of integrin 4 in BM-MSCs by genetic manipulation may be a potential therapeutic approach to enhancing the efficacy of stem cell therapy using BM-MSCs.