Hypoxia Regulates mTORC1-Mediated Keratinocyte Motility and Migration via the AMPK Pathway

Keratinocyte migration, the initial event and rate-limiting step in wound healing, plays a vital role in restoration of the intact skin barrier, also known as re-epithelialization. After acute tissue injury, hypoxic microenvironment gradually develops and acts as an early stimulus to initiate the healing process. Although we have previously found that hypoxia induces keratinocyte migration, the underlying mechanism remains unknown. Here, we first observed that hypoxia increased mTORC1 activity. Recombinant lentivirus vector and Rapamycin were used for silencing mTORC1 in HaCaT cells and primary mouse keratinocytes (MKs). Using cell migration assay and a Zeiss chamber equipped with imaging system, we also demonstrated that mTORC1 downregulation reversed hypoxia-induced keratinocyte motility and lateral migration. Importantly, hypoxia-activated mTORC1 was accompanied by the AMPK downregulation, and we found that the AMPK pathway activators Metformin (Met) and 5-Aminoimidazole-4-carboxamide 1-β-D-ribofuranoside (AICAR) decreased the mTORC1 activity, cell motility and lateral migration. Thus, our results suggest that hypoxia regulates mTORC1-mediated keratinocyte motility and migration via the AMPK pathway.


Introduction
Wound healing is a dynamic and well-ordered biological process that requires the spatial and temporal orchestration of several distinct components, including coagulation, inflammation, re-epithelialization, contraction and remodeling [1]. The essential feature of successful wound healing is the reestablishment of the intact epidermal barrier. Thus, wound epithelialization, also called re-epithelialization, is the key and defining feature of wound repair. The term "reepithelialization" refers to an intricate process that the keratinocytes migrate from wound margins to resurface the wounded area. During this process, keratinocyte migration into the wound is the initial event and rate-limiting step [2], since defects in migration, but not in differentiation or proliferation, are closely related with the non-healing wounds.
Under normal conditions, keratinocytes develop a mechanical barrier against chemical stimulus and microorganism through terminal differentiation. During wound healing, a complex balance of genes and signals are regulated in a temporal and spatial manner to promote PLOS ONE | DOI: 10

Cell culture and hypoxia treatment
HaCaT cells were obtained from the Cell Bank of the Chinese Academy of Sciences (Beijing, China). Cells were cultured in RPMI 1640 medium (Hyclone, USA) supplemented with 100 U/ml penicillin, 100 mg/ml streptomycin, and 10% fetal bovine serum (Gibco, USA) in a 37˚C humidified incubator at 5% CO 2 . New-born BALB/c mice were commercially purchased from the Laboratory Animal Center of the Third Military Medical University and sacrificed by cervical dislocation. Primary MKs were isolated from the skin of new-born BALB/c mice (postnatal day 1-3) as previously described [19]. Briefly, keratinocytes were obtained from the skin by incubation with a 0.25% trypsin/0.04% EDTA solution (Invitrogen, USA) at 4˚C overnight and were then plated into collagen type IV-coated dishes. The cells were then maintained in keratinocyte serum-free medium (K-SFM medium) (Gibco, USA) supplemented with antibiotics in a humidified atmosphere containing 5% CO 2 at 37˚C. A Forma Series II Water Jacket CO 2 incubator (model: 3131; Thermo Scientific), which maintains a desired and precise culturing environment, kept CO 2 at 5% and O 2 at 2% by nitrogen displacement. Rapamycin (100 mM) and corresponding amounts of DMSO, used as an mTORC1 inhibitor and control, were added to the cultures, which were pre-incubated at 37˚C for 4 h. Cells were incubated in the presence of the AMPK activators Met (500 μM) or AICAR (0.5 mM) throughout the experiment, unless otherwise indicated.

Lentiviral transduction
The lentivirus for silencing Raptor (shRaptor) and non-targeting control (shNC) sequence were purchased from Shanghai GeneChem, Co. Ltd (Shanghai, China), and transfections were performed according to the manufacturer's instructions.

Western blotting
Cells were washed with ice-cold phosphate-buffered saline (PBS) and harvested on ice. The protein concentrations of the lysates were determined using an RCDC protein assay kit (Sigma, USA). A prestained standard protein molecular weight marker and the samples were loaded into the wells of 8% SDS-PAGE gels and transferred by electroblotting to polyvinylidene difluoride (PVDF) membranes. After blocking the membranes with 5% non-fat dried milk or bovine serum albumin (BSA), the desired bands were incubated overnight at 4˚C with the corresponding primary antibodies. Horseradish peroxidase-conjugated secondary IgG (1:5,000, Proteintech, USA) was subsequently used as a secondary antibody. The results were analyzed using a ChemiDoc imaging system (Bio-Rad, USA). The primary antibodies used in this study were as follows: phospho-p70S6K (Thr389), p70S6K, phospho-4E-BP1 (Thr70), 4E-BP1, phospho-AMPKα (Thr172), AMPK, phospho-ACC (Ser79), ACC (1:1,000, Cell Signaling, USA). acquired every 3 min for 3 h and were analyzed using NIH Image J software (http://rsb.info. nih.gov/ij/). The cell trajectories were imaged based on the positions of the cell nuclei at frame intervals of 3 min, and their trajectory speed (μm/min) was calculated as the total length (μm) of the trajectories divided by the time (minute), reflecting cell motility.

Wound migration assay
A wound migration model of cultured cells was used as described in previous studies, with some modifications. Briefly, Culture-Inserts (Ibidi) were used to measure cell migration. Cell suspension at a density of 7×10 4 /mL (70 μL volume) was applied to each well of the Culture-Inserts. After the appropriate duration for cell attachment (24 h), the cells were incubated at 37˚C for additional 2 h in the presence of mitomycin-C (5 μg/mL) [20] to inhibit cell proliferation. A cell-free gap of 500μm was created by removal of the Culture-Insert. Images were captured every 6 h using an inverted phase-contrast microscope. The percent of wound closure in five randomly chosen fields was calculated with NIH ImageJ software.

Statistical analysis
The data are presented as the mean ± standard deviation (SD). SPSS 13.0 was used for statistical analysis, and statistical significance among multiple groups was evaluated by one-way ANOVA. P values < 0.05 were considered statistically significant.

Hypoxia increases the activities of distinct mTORC1 downstream targets and their expression in the nucleus
The activity of mTORC1 can be measured by analyzing phosphorylation of the direct downstream target of ribosomal protein kinase S6 (p70S6K) at Thr389 and eukaryotic initiation factor 4E binding protein 1 (4E-BP1) at Thr70. Thus, we performed time-course experiments in which MKs and HaCaT cells were exposed to hypoxia (2% O 2 ) for different durations. The phosphorylation of p70S6K-Thr389 and 4E-BP1-Thr70 were then analyzed. As shown in Fig  1A, the phosphorylation of p70S6K (Thr389) and 4E-BP1 (Thr70) in MKs and HaCaT cells rapidly increased after exposure to hypoxic conditions and remained elevated compared with normoxic cells. The total p70S6K and 4E-BP1 levels remained unchanged under hypoxia. The average quantifications from cumulative experiments shown in Fig 1B and 1C indicated that the p70S6K and 4E-BP1 activities were significantly increased. Subcellular localization, constituting the environment in which proteins act, is a crucial determinant of protein function and regulation. We examined whether hypoxia (2% O 2 ) influences the cellular localization of mTORC1 downstream substrates. Under all tested conditions, phospho-p70S6K was localized to only distinct nuclear structures, whereas phospho-4E-BP1 was detected predominantly in nuclei, as shown by immunostaining in S1 Fig. Exposure to hypoxic conditions for 6 h clearly increased the phospho-p70S6K and phospho-4E-BP1 expression levels in the nuclei. Thus, we found that hypoxia rapidly promoted the phosphorylation of p70S6K and 4E-BP1 and increased their expression levels in nuclei. These results indicated that mTORC1 is activated by hypoxia.
The mTORC1 pathway is required for hypoxia-induced keratinocyte motility and migration A tremendous amount of data has established how mTORC1 controls different processes involved in disease, such as the migration and invasion of multiple types of tumour cells.  Therefore, we hypothesized that mTORC1 might regulate single-keratinocyte motility and lateral migration under hypoxia.
To determine the role of mTORC1 in keratinocyte migration, we first used rapamycin, a classical inhibitor of mTORC1, to block its activity. As previously shown, exposure to hypoxic conditions for 6 h resulted in markedly increased phosphorylation of p70S6K and 4E-BP1, which was inhibited by rapamycin (Fig 2A and 2B). Next, Raptor, an essential component of mTORC1, was selectively silenced. Keratinocytes with stable shRNA-mediated knockdown of Raptor was generated and exhibited > 90% reduction in expression of the targeted protein (S2 Fig). Cells expressing shRaptor had significantly reduced levels of phospho-p70S6K and phospho-4E-BP1 compared with hypoxic cells (with or without shNC), whereas total p70S6K and 4E-BP1 levels were not affected by hypoxia or shRNA (Fig 2A and 2C).
Migration is a critical step in the initial process of wound repair. A Zeiss imaging system was used to study the role of mTORC1 on single-keratinocyte motility independent of cell proliferation and survival. Wound migration assay was performed to observe keratinocytes lateral migration. Our results showed that the movement range of MKs exposed to hypoxia was greatly increased, thus indicating better motility. Both allosteric inhibition by rapamycin and the knockdown of Raptor markedly attenuated the hypoxia-induced cell motility ( Fig 3A and  3B, S1 Movie). Similar results were also found in HaCaT cells ( Fig 3B). The positive role of mTORC1 in hypoxia-induced lateral migration was further confirmed in an in vitro wound migration assay. After 24 h, hypoxic keratinocytes (H + DMSO, H + shNC) migrated into nearly 100% of the original non-cell-covered areas, whereas mTORC1-inhibited keratinocytes (H + rapamycin, H + shRaptor) migrated into only 25% of the original cell-free areas under hypoxic conditions (Fig 3C and 3D). Similar results were observed in treated HaCaT cells ( Fig  3D). In summary, the pharmacological blockade of mTOR signaling using rapamycin and the genetic modification of an essential component of mTORC1 significantly inhibited single-keratinocyte motility and lateral migration. The above data have revealed that mTORC1 activation is essential for hypoxia-enhanced keratinocyte motility and migration, although the underlying molecular mechanism remains to be elucidated.

Hypoxia mediates time-dependent, sustained suppression of the AMPK signaling pathway in keratinocytes
Our finding demonstrated that mTORC1 activation plays crucial roles in the regulation of single-cell motility and lateral migration during hypoxia, we then study the mechanism of mTORC1 activation in hypoxic keratinocytes. The phosphorylation statuses of AMPKα and acetyl-CoA carboxylase (ACC) were examined by immunoblotting analysis using commercially available antibodies to assess AMPK activity. Exposure of keratinocytes to hypoxia resulted in the dephosphorylation (inactivation) of AMPKα and ACC in a time-dependent manner ( Fig 4A). The quantification from cumulative experiments confirmed that the expression levels of phosphor-AMPKα and phosphor-ACC inversely correlated with the duration of hypoxic treatment in MKs and HaCaT cells (Fig 4B and 4C). The total AMPK and ACC levels remained unchanged under hypoxia. Considering the early reports demonstrating that AMPK can suppress mTORC1, we reasoned that there might be a causal relationship between AMPK and mTORC1.

AMPK activation reverses hypoxia-induced mTORC1 activation
To study the possible causal relationship between AMPK suppression and mTORC1 activation, keratinocytes were treated with the AMPK activators Met and AICAR. They activated AMPK effectively in MKs, as shown by the increased phosphorylation of AMPKα and ACC   after exposure to hypoxic conditions for 6 h. Meanwhile, the expression levels of phosphorylated p70S6K and 4E-BP1 were consistently decreased (up to 90%) by Met and AICAR, indicating the suppression of mTORC1 activity. The total levels of AMPK, ACC, p70S6K, and 4E-BP1 were not affected by either hypoxia or the AMPK activators (Fig 5A and 5B). Met and AICAR had similar effects on mTORC1 in HaCaT cells (Fig 5A and 5C). These results suggest that the AMPK pathway negatively regulates the mTORC1 activity in hypoxic keratinocytes.

The AMPK suppression is involved in hypoxia-enhanced motility and migration
To further study the role of AMPK in the augmented motility and migration of hypoxic keratinocytes, we analyzed the movement of these cells using a Zeiss imaging system and wound migration assays with or without an AMPK pharmacological activator. Images recording cell trajectories showed that hypoxia greatly increased movement range of keratinocytes, whereas the pharmacological activator of AMPK significantly reduced cell motility (Fig 6A, S2 Movie). The statistical analysis of the velocity confirmed the inhibitory effect of AMPK activation on the keratinocyte motility, because the corresponding histograms for the MKs and HaCaT cells showed 2-fold reductions in their trajectory speeds in the presence of Met and AICAR ( Fig  6B). The similar effect of AMPK activation on migratory capacity was observed using wound healing assays. Hypoxic keratinocytes covered nearly 90% of the original cell-free area, while the areas covered by cells treated with Met and AICAR were only 30% and 50%, respectively; the statistical analysis of the monolayer migration assays showed the similar result (P< 0.05) (Fig 6C and 6D). Taken together, these results demonstrate that the AMPK pathway is negatively correlated with hypoxia-induced keratinocyte migration and that this correlation contributes to the crucial role of mTORC1 in regulating keratinocyte migration after hypoxia.

Discussion
Re-epithelialization, the ultimate goal and defining sign of a healed wound, includes three overlapping keratinocyte functions: migration, proliferation, and differentiation. Keratinocytes migration, the most limiting process, has received considerable attention. Hypoxic microenvironment is vital and significant occurrence to which keratinocytes are exposed. In this study, we investigated the mechanisms responsible for the augmented motility and migration of keratinocytes in the mild hypoxic microenvironment and found that mTORC1 activation, which was mediated by AMPK suppression, played a vital role in hypoxia-induced cell migration. First, we demonstrated that hypoxia-induced mTORC1 activation was greatly involved in keratinocytes motility and migration (Figs 1-3, S1 Fig, S1 Movie). Second, we found that the hypoxia suppressed the AMPK signaling pathway in a time-dependent manner (Fig 4). Third, we showed that AMPK activation via Met and AICAR reversed hypoxia-induced mTORC1 activation, keratinocytes motility, and migration (Figs 5 and 6   The disruption of skin integrity with the loss of epithelial cells continuity results in the activation of an intricate process aiming to restore an intact epidermal barrier, and keratinocyte migration is a critical step in this process. Migrating cells at the epithelial tongue are likely to encounter mild hypoxia due to decreased oxygen supply and increased oxygen demands. After acute injury, hypoxic stress induces or suppresses certain genes and protein synthesis, which is the essential strategy for adaptation to hypoxic environment. Pimonidazole, a specific tissue hypoxia marker, is used to detect moderate hypoxia at the leading edge of the healing area of epidermis [26]. Unlike complete anoxia, exposure of cells to mild hypoxia (1-2%O 2 , PO 2 = 7-14 mmHg) can result in various consequences for metabolism, intracellular signaling, and survival. For instance, 1-2% O 2 does not cause cells to undergo growth arrest or apoptosis [27], whereas cells undergo apoptosis at oxygen concentrations of closer to 0% [28]. In addition, an oxygen concentration of 0.5% or lower will limit the respiratory rate and decrease the ATP levels [29,30]. In contrast, mild hypoxia elicits a well-coordinated series of events that influence gene expression and the phenotype. The early application of semi-occlusive dressings after injury has been shown to promote re-epithelialization substantially when compared with wounds allowed to air-dry [31,32]. In conclusion, the moderate hypoxia microenvironment has important effects on the processes that occur during wound healing. mTORC1 is critical in cell molecular activities. In the last years, many studies have established how the mTORC1 integrates the signals from extracellular environment in the regulation of cellular physiology and molecular biology, for instance, the mTORC1 pathway is pivotal in cancer cell invasion. During wound healing, upregulation of mTORC1 downstream targets are found in the epithelial tongue correspondent to the migratory areas of normal wounds. Coincidentally, hypoxia found in the healing margin of normal wounds was related to cell migration, but not to proliferation or inflammatory infiltration. These results prompted us to study the possibility that the hypoxic microenvironment promotes mTORC1 activation in migrating keratinocytes. In this study, we have demonstrated that hypoxia induced phosphorylation of the mTORC1 substrates p70S6K (Thr389) and 4E-BP1 (Thr70), and their expression in nuclear structures markedly increased. The phosphorylation of aforementioned proteins was inhibited by rapamycin pre-treatment and Raptor silencing. In line with these phosphorylation studies, Raptor silencing and rapamycin, thus the inhibition of mTORC1 signaling, all specifically and significantly inhibited single-cell motility and the lateral migration of keratinocytes under hypoxic conditions. These findings indicate that mTORC1 activity is very important in hypoxia-induced migration of keratinocytes. The role of mTORC1 in hypoxic keratinocytes is similar to that observed in hypoxic carcinoma cells whose mTORC1 activation is also associated with carcinoma progression and invasion. In our study, pre-treatment with rapamycin for 4 h, rather than prolonged treatment, did not inhibit mTORC2 assembly, in agreement with previous studies showing that mTORC2 (the RICTOR/mTOR complex) is insensitive to short-term treatment with rapamycin [33]. mTORC1 is traditionally believed to regulate cell growth and survival. Although our results highlight the importance of mTORC1 in migrating keratinocytes under hypoxia, the mechanisms that mTORC1 regulate hypoxic keratinocytes migration remain unclear. During the process of re-epithelialization, a series of orchestrated events transform a mature, differentiated keratinocyte into a migrating keratinocyte. The change from a sedentary cell to a motile cell involves changes in the cell attachments, keratins, matrix metalloproteinase, and tight junctions. Our study is just a preliminary exploration of the role of mTORC1 in migrating keratinocytes, and this finding raises the question that how mTORC1 links the expression of transcription factors and changes in migratory ability. Future studies may further elucidate the roles of mTORC1 in migrating keratinocytes.
In our study, we found that hypoxia mediated time-dependent, sustained suppression of the AMPK signaling pathway in keratinocytes and that this suppression was negatively correlated with the mTORC1 activation in hypoxic cells (Fig 4). The suppression of AMPK was verified by the phosphorylation of Thr172 in its α subunit and ACC, the rate-limiting enzyme in fatty acid synthesis. Accumulating data indicate that the AMPK pathway is involved in wellorchestrated signaling events triggered by hypoxia, and it is believed to sense energy stress and to be activated during sudden periods of acute stress. The different situation in hypoxic keratinocytes and the reasons for this phenomenon are quite intriguing. This discrepancy may potentially be explained by complex regulation of the AMPK network and the specific conditions. There are at least three (but probably more) kinases that phosphorylate AMPK at Thr172, thereby affecting its activity. The AMPK holoenzyme is also activated by allosteric regulation of the β subunit, with or without the phosphorylation of Thr172 in the α subunit [34]. Therefore, the functions of AMPK should be carefully considered depending on the specific context. In our situation, other factors instead of AMP/ATP play a leading role in suppressing the AMPK pathway, which remains to be explored in our further studies. Additionally, previous studies have shown that AMPK is involved in regulating the migration of different types of cells [14,35]. However, there is no evidence showing that AMPK signaling is involved in hypoxic keratinocyte migration. Our results demonstrated that suppression of the AMPK pathway accelerated keratinocyte migration under hypoxia, whereas its pharmacological activation led to decreased cell movement and migration. Thus, this study has revealed that mild hypoxic microenvironment induces keratinocyte migration through the suppression of AMPK during the early stage of wound repair. New studies will continue to uncover the elegant network of events that result in the balance between AMPK suppression related migration and the energy stress.
Activation of mTORC1 should be fine regulated in consideration of energy level in hypoxic cells, so that cells can survive and maintain physiological function. As a biomarker for mTOR activation, the phosphorylation of mTOR-Ser2448 has been suggested to be an important part of a feedback mechanism that regulates its activity [36]. The p70S6 kinase is considered to be the dominant kinase responsible for the phosphorylation of mTOR-Ser2448 in cells [37]. Interestingly, we found that moderately hypoxic condition (2% O 2 ) promoted the sustained and rapamycin-sensitive phosphorylation of mTOR (Ser2448) in keratinocytes, whereas total mTOR expression was not affected by hypoxia. These findings are consistent with other cell responses to hypoxia, preventing or delaying the onset of more severe hypoxia [38]. However, the functional significance of phosphorylated mTOR at Ser2448 is still unknown. Additionally, it is still unclear whether this feedback mechanism is positive or not and to what extent it affects mTORC1 [37,39]. Further studies are necessary to decipher the relationship between hypoxia-induced mTOR phosphorylation (Ser2448) and mTOR function.
In conclusion, our results have revealed that mTORC1 activation caused by the AMPK suppression is closely involved in mild hypoxia-induced single-cell motility and lateral migration. These findings provide a new perspective on the influences of hypoxic microenvironment on keratinocyte signaling pathways, which differ from its effects on signalling in other cells.