Polyamine Regulator AMD1 Promotes Cell Migration in Epidermal Wound Healing

ORIGINAL ARTICLE

Polyamine Regulator AMD1 Promotes Cell Migration in Epidermal Wound Healing Hui Kheng Lim1,2, Anisa B. Rahim1,2, Vonny Ivon Leo1,2, Shaptura Das1,2, Thiam Chye Lim3, Takeshi Uemura4, Kazuei Igarashi4, John Common1,2 and Leah A. Vardy1,2,5 Wound healing is a dynamic process involving gene-expression changes that drive re-epithelialization. Here, we describe an essential role for polyamine regulator AMD1 in driving cell migration at the wound edge. The polyamines, putrescine, spermidine, and spermine are small cationic molecules that play essential roles in many cellular processes. We demonstrate that AMD1 is rapidly upregulated following wounding in human skin biopsies. Knockdown of AMD1 with small hairpin RNAs causes a delay in cell migration that is rescued by the addition of spermine. We further show that spermine can promote cell migration in keratinocytes and in human ex vivo wounds, where it significantly increases epithelial tongue migration. Knockdown of AMD1 prevents the upregulation of urokinase-type plasminogen activator/urokinase-type plasminogen activator receptor on wounding and results in a failure in actin cytoskeletal reorganization at the wound edge. We demonstrate that keratinocytes respond to wounding by modulating polyamine regulator AMD1 in order to regulate downstream gene expression and promote cell migration. This article highlights a previously unreported role for the regulation of polyamine levels and ratios in cellular behavior and fate. Journal of Investigative Dermatology (2018) -, -e-; doi:10.1016/j.jid.2018.05.029

INTRODUCTION The process of wound healing is an elaborate and regulated sequence of events involving a multitude of different cell types and signaling pathways (Maddaluno et al., 2017). In normal skin, keratinocytes in the basal layer of the epidermis are in contact with the basement membrane and are proliferative. On epidermal wounding, keratinocytes at the wound edge undergo a transition from a non-motile epithelial state to a mesenchymal-like state where they lose cellecell contacts and become motile. Migrating cells reorganize their actin cytoskeleton and secrete proteases to remodel the extracellular matrix and enable migration across the wound (Leopold et al., 2012; Pastar et al., 2014; Shaw and Martin, 2016; Stone et al., 2016). Directly behind the migrating cells, keratinocytes rapidly proliferate to provide enough cells to cover the wound. The urokinase-type plasminogen activator (uPA) is upregulated in migrating keratinocytes at the wound edge and, along with its receptor, uPAR, is important for cell migration (Romer et al., 1994; Solberg et al., 2001). uPA/uPAR plays a role in promoting degradation of the 1

Institute of Medical Biology, Agency for Science, Technology and Research, Immunos, Singapore; 2Skin Research Institute of Singapore, Agency for Science, Technology and Research, Immunos, Singapore; 3 Division of Plastic, Reconstructive and Aesthetic Surgery, Department of Surgery, National University Hospital and National University of Singapore, Kent Ridge Wing, Singapore; 4Graduate School of Pharmaceutical Sciences, Chiba, University, Chiba, Japan; and 5School of Biological Sciences, Nanyang Technological University, Singapore Correspondence: Leah A. Vardy, Institute of Medical Biology, Agency for Science, Technology and Research, 8A Biomedical Grove, Immunos, Singapore 138648, Singapore. E-mail: [email protected] Abbreviations: uPA, urokinaseetype plasminogen activator; sh, small hairpin; uPAR, urokinaseetype plasminogen activator receptor Received 10 January 2018; revised 29 April 2018; accepted 23 May 2018; accepted manuscript published online 12 June 2018; corrected proof published online XXX

extracellular matrix at the wound edge. In addition, it signals through the RHO family of small GTPase RAC to promote actin cytoskeleton reorganization at the leading edge of the migrating cells (Smith and Marshall, 2010). Defects in keratinocyte migration are characteristic of chronic non-healing wounds where re-epithelialization fails to occur (Lindley et al., 2016). The polyamines, putrescine, spermidine, and spermine are naturally occurring cations present in all cells, and are essential for proliferation and a wide range of cellular events. Putrescine is synthesized within cells by decarboxylation of ornithine by the rate-limiting enzyme, ODC1. Putrescine is then sequentially converted into spermidine and spermine by the addition of an aminopropyl group, from dcAdoMet. dcAdoMet is generated by the decarboxylation of AdoMet by AMD1, a second rate-limiting enzyme in the polyamine pathway. The levels and activity of ODC1 and AMD1 thus directly influence the levels and ratios of the three polyamines produced (Pegg, 2016). Polyamine catabolism also contributes to the regulation of polyamine levels and is tightly regulated (Casero and Pegg, 2009). While tightly controlled, the levels of the polyamines are altered in different cellular contexts. High levels are present in proliferative cells, and many cancers show elevated levels of ODC1, which drives increased polyamine levels (Nowotarski et al., 2013). Controlled fluctuations of the levels of the three polyamines within a physiological range can promote different cellular phenotypes and have been suggested to play a regulatory role (Pegg, 2016; Zhang et al., 2012). Within the cell, the three polyamines can modulate gene transcription and translation, kinase function, cytoskeleton assembly, and ion channel function (Igarashi and Kashiwagi, 2010, Lightfoot and Hall, 2014). Intracellular polyamine levels are upregulated in response to injury and cellular damage (Casero and Pegg; 2009, Gao et al., 2013; Gilad and Gilad, 2003; Schimchowitsch and

ª 2018 The Authors. Published by Elsevier, Inc. on behalf of the Society for Investigative Dermatology.

www.jidonline.org

1

HK Lim et al.

Polyamines and Wound Healing

Cassel, 2006). Polyamine levels and the regulator ODC1 have been found to be increased on epidermal wounding, yet it remains unclear what drives these changes or how they impact the wound healing process (Maeno et al., 1990; Mizutani, 1974; Shi et al., 2002). Here, we address the role of the polyamine regulator, AMD1, in the wound healing process in the human epidermis. We find that AMD1 is rapidly upregulated in wound-edge keratinocytes. Upregulation of AMD1 and high spermine levels are required for keratinocyte migration and play a role in the activation of the uPAR/uPA signaling system. Knockdown of AMD1 results in decreased spermine levels and a failure to upregulate uPA/uPAR signaling to promote actin reorganization in the cells at the leading edge of the wound. Our data suggest that on wounding, AMD1 functions to modulate the ratios of polyamines by promoting high levels of spermine at the expense of putrescine and that these changes are required for the promotion of a keratinocyte migratory phenotype to enable wound closure. RESULTS AMD1 is upregulated during keratinocyte differentiation

Polyamine levels are known to be high in the epidermis compared with the dermis (El Baze et al., 1983), but their function is not well understood. Polyamine regulator AMD1 is rate limiting for the conversion of putrescine to spermidine and spermine, therefore, changes in its activity influence the putrescine to spermidine and spermine ratio. To address the role of AMD1 in the human epidermis, we analyzed AMD1 expression in human skin histologic sections by immunofluorescence. AMD1 was highly expressed in the more differentiated granular layer of the epidermis, where it colocalized with involucrin, a granular layer marker. AMD1 was also present in the less differentiated stratum spinosum, but was at low levels in the proliferating basal cell layer (Figure 1a). AMD1 protein levels were also upregulated in the N/TERT-1 human keratinocyte cell lines following 6 days of calciuminduced differentiation. AMD1 starts to increase by day 2 and peaks by day 4 of keratinocyte differentiation (Figure 1b). The terminal differentiation markers, involucrin and filaggrin, were also upregulated, confirming the efficiency of cell differentiation (Figure 1b). AMD1 is upregulated during cutaneous wound healing

To determine whether AMD1 protein is upregulated on wounding, we performed immunofluorescence studies of AMD1 expression in ex vivo wounded human skin biopsies. AMD1 expression was upregulated in keratinocytes at the wound edge at 24 hours post wounding (Figure 2a). Sections were also stained for K6, a marker of activated pro- migratory keratinocytes known to be present at the wound edge (TomicCanic et al., 1998). We quantified the intensity of AMD1 staining in the basal cells from non-wounded tissue and the migrating cells from wound edge tissue using ImageJ software (National Institutes of Health, Bethesda, MD). This confirmed the increase in AMD1 staining at the wound edge keratinocytes compared to basal layer non-migrating keratinocytes (Figure 2b). Ki-67 staining showed that at 24 hours post wounding there was no significant increase in cell proliferation at the wound edge compared to the non-wounded 2

Journal of Investigative Dermatology (2018), Volume

-

tissue (Supplementary Figure S1 online). To model cell migration at the wound edge, we performed scratch assays on cultured human keratinocytes. Western blot and quantitative real-time reverse transcriptaseePCR analysis of keratinocytes before and 6 hours post scratch showed that AMD1 was upregulated at the protein and the RNA level following scratch (Figure 2c, 2d). Immunofluorescence studies showed a strong increase in AMD1 expression at the wound edge cells 6 hours post-scratch (Figure 2e). Rapid upregulation of AMD1 was detected within 10 minutes of scratch wounding in the leading edge cells. By 1 hour post scratch, increased labeling of AMD1 was detected over 18 cells deep from the wound edge. The upregulated AMD1 was then redistributed to wound-edge keratinocytes two to five cells deep by 6 hours post scratch. Interestingly, the distribution and level of ODC1 and AMD1 upregulation in wound-edge keratinocytes was very different, as seen by immunofluorescence (Figure 2e, 2f). AMD1 shows a strong upregulation in woundedge cells, while ODC1 protein is only mildly upregulated many cells back from the wound edge. Upregulation of AMD1 is dependent on the scratch-induced calcium wave

On wounding, there is an immediate increase in intracellular calcium in the cells adjacent to the wound. This calcium flux moves as a wave from the wound edge and is essential for wound healing to proceed (Tran et al., 1999). The increased calcium levels trigger changes in gene expression that are essential for many aspects of wound healing, including reorganization of the actin cytoskeleton (Cordeiro and Jacinto, 2013). We sought to determine whether the upregulation of AMD1 at the wound-edge was dependent on the wound-induced calcium wave. Human keratinocytes were labeled with the Ca2þ probe Fluo-4 for 30 minutes prior to scratching to mimic a wound response. Live-cell imaging of the scratch wound edge confirmed a rapid calcium flash extending outward from the point of scratch as a wave; the signal subsequently decayed to background level after 3 minutes (Figure 3a, 3b). To determine whether the upregulation of AMD1 at wound-edge keratinocytes was dependent on the calcium wave, cells were pretreated for 24 hours with 1 mM thapsigargin and for 1 hour with 2 mM EGTA prior to scratch, to deplete internal calcium stores and extracellular calcium, respectively (Figure 3c). AMD1 upregulation in the scratched cells was completely abrogated following a calcium block, as shown by immunofluorescence and Western blot of cells 6 hours post-scratch (Figure 3d, 3e). This suggests that the calcium wave is required for AMD1 upregulation at the wound-edge keratinocytes. Collectively, these data indicate that AMD1 upregulation on wounding is downstream of the wound-induced calcium wave. AMD1 upregulation is required for cell migration on wounding

To determine the function of AMD1 upregulation in keratinocytes during wound healing, we performed in vitro wound healing assays on small hairpin (sh)AMD1 human keratinocytes. Keratinocytes expressing shRNAs targeting AMD1 showed a 70e80% decrease in AMD1 RNA and protein expression (Figure 4a, Supplementary Figure 2a online). shAMD1 cells failed to upregulate AMD1 RNA or protein

HK Lim et al.

Polyamines and Wound Healing

Figure 1. AMD1 is expressed in the differentiated layers of the human skin epidermis. (a, b) Detection of AMD1 and epidermal differentiation markers (FLG and IVL) by (a) immunostaining of human skin epidermis (n ¼ 5) and (b) Western blot (n ¼ 3) following induction of keratinocyte differentiation with 1.2 mM calcium for 0, 2, 4, and 6 days. White dashed line delineates boundary between epidermis and dermis. FLG, filaggrin; IVL, involucrin. Scale bar ¼ 25 mm.

expression on wounding compared with shScrambled control cells (Figure 4b, Supplementary Figure S2a). The absence of AMD1 upregulation resulted in a significantly delayed wound closure in a scratch assay (Figure 4c, Supplementary Figure S2b). At 6 hours post scratch, AMD1 knockdown resulted in a 20% increase in relative wound width compared with control cells (Figure 4d, Supplementary Figure 2c). To determine whether the delay in cell wound closure on knockdown of AMD1 was due to a decrease in proliferation, we stained the wound-edge cells with proliferation marker Ki-67. Cells expressing AMD1 shRNAs did not show a decrease in cell proliferation at the wound edge following scratch (Supplementary Figure S2d). This suggests that the decrease in scratch wound closure was due to a delayed cell migration and not decreased cell proliferation. We showed that cells expressing shAMD1 did not show a dramatic reduction in proliferation rate, suggesting that proliferation was not responsible for the delayed wound closure (Supplementary Figure S2e). To further confirm that cell proliferation does not contribute to the delayed scratch wound closure, we incubated cells with and without mitomycin C to block cell proliferation. We observed that the inhibition of cell proliferation did not delay scratch wound closure within the time frame of these studies (10 hours). We observed that the inhibition of cell proliferation with mitomycin C resulted in a more rapid wound closure following scratch (Figure 4e, 4f). Taken together, these data suggest that AMD1 upregulation at the wound edge is required for cell migration. AMD1 promotes high spermine levels to drive cell migration at the wound edge

Our data show that AMD1 is upregulated in keratinocytes at the wound edge post injury. Changes in AMD1 will alter the availability of the aminopropyl donor dcAdoMet required for the generation of spermidine and spermine from putrescine. As a consequence, the knockdown of AMD1 will disrupt the

ratio of putrescine to spermidine and spermine. Thus, we measured the intracellular polyamine levels in control and shAMD1 knockdown keratinocytes and found that putrescine levels were 2.7-fold higher compared with shScrambled cells, while spermine levels were decreased by 2-fold. There was no significant change in spermidine levels (Figure 5a). To determine whether spermine can promote cell migration on wounding, we added spermine to scratched keratinocytes and measured the rate of wound closure. Strikingly, we observed that spermine enhances cell migration and scratch wound closure (Figure 5b, 5c). To determine whether spermine can promote cell migration in a human ex vivo wound model, we added spermine to human ex vivo wounds and examined the length of the migrating tongue after 72 hours. We observed that addition of spermine promoted a striking improvement in epithelial tongue migration (Figure 5d, 5e, Supplementary Figure S3 online), which was consistent with our in vitro model. Our previous data suggest that AMD1 promotes cell migration independent of proliferation, at least during the early stages of wound healing. We stained the ex vivo wounds with proliferation marker Ki-67 and did not see an increase in staining in the migrating tongue (Supplementary Figure S4 online). There was a marginal increase in Ki-67 staining in cells behind the migrating tongue. We next measured the levels of polyamines in keratinocytes before and after scratch and confirmed that spermine levels remained low in the shAMD1 knockdown cells after scratch, while putrescine levels remained high (Supplementary Figure S5 online). To determine whether decreased putrescine levels are required for cell migration, we evaluated whether the topical application of putrescine can influence the migration of keratinocytes following a scratch assay (Figure 5f, 5g). Putrescine treatment caused an inhibition of keratinocyte migration and a delay in closure of the scratch wound. Taken together, these data suggest that AMD1 upregulation on wounding is required to promote high spermine levels, which is important for cell migration. www.jidonline.org

3

HK Lim et al.

Polyamines and Wound Healing

Figure 2. AMD1 is upregulated in wound-edge keratinocytes. (a) Immunofluorescence of AMD1 (green) and K6 (red) expression in skin biopsies from healthy donors (n ¼ 5) before and 24 hours after wounding counterstained with Hoechst (blue, DNA). Right panels show enlarged images from boxed area showing AMD1 staining. White arrowhead indicates site of wounding. White dashed line delineates boundary between epidermis and dermis. Scale bar ¼ 25

4

Journal of Investigative Dermatology (2018), Volume

-

HK Lim et al.

Polyamines and Wound Healing

Figure 2. Continued

To confirm the role of spermine in cell migration, we attempted to rescue the AMD1 knockdown migration phenotype with supplementation of spermine. Replacement of spermine restored impaired cell migration in AMD1 knockdown keratinocytes (Figure 5h, 5i). This was further confirmed with the use of EGBG (ethylglyoxal bis(guanylhydrazone)), an inhibitor of AMD1 (Igarashi et al., 1984). EBGB prevented cell migration following scratch wound assay in a manner very similar to the shAMD1 and this was fully rescued through the supplementation with spermidine and spermine (Figure 5j, 5k). AMD1 upregulation promotes actin cytoskeleton reorganization through the uPA/uPAR pathway at the wound edge

To determine the mechanism by which spermine and AMD1 promote cell migration, we determined the effect of the knockdown on expression of proteins known to be required

for cell migration. The uPAR and uPA are upregulated on wounding and function to promote reorganization of the actin cytoskeleton at the leading edge of migrating cells and to degrade the extracellular matrix, allowing the migrating cells to move freely (Smith and Marshall, 2010). Western blot analysis of keratinocytes before and 6 hours post scratch indicated that uPA and uPAR proteins were rapidly upregulated on wounding in a manner similar to AMD1 (Figure 6a). There was also marked upregulation of uPAR and uPA mRNA levels (6.8-fold and 3.6-fold, respectively) in scratched control cells (Supplementary Figure S6a online). Strikingly, uPA and uPAR upregulation post scratch was impeded when AMD1 was knocked down (Figure 6a, Supplementary Figure S6a, S6b). Western blot analyses on scratched cells are subject to a significant dilution effect, as only the cells at the wound edge are migrating. To confirm the AMD1dependent upregulation of uPA in wounded keratinocytes, we performed immunofluorescence staining in scratched

= mm. (b) Quantitative analysis of AMD1 fluorescence intensity in (a) using ImageJ software. (c, d) Increased expression of AMD1 observed in scratched keratinocytes by Western blot (c) and reverse transcriptaseePCR (d). (e, f) Immunofluorescence staining of AMD1 (e) and ODC1 (f) in scratched keratinocyte monolayer (red). Hoechst DNA staining is in blue. NS, no scratch. White dashed line indicates site of scratch wound. Scale bar ¼ 25 mm. All in vitro experiments were repeated in biological triplicates and the data are presented as means  standard error of mean (n ¼ 3). *P < 0.05.

www.jidonline.org

5

HK Lim et al.

Polyamines and Wound Healing

Figure 3. AMD1 upregulation is dependent on the scratch-induced calcium wave. (a) Ca2þ probe Fluo-4 labeling (1 mg/ml) showing a calcium wave spreading outward from the scratch-wound margin. The time labels (minutes) reflect the time points post scratch. (b) Kinetics of calcium wave propagation after scratch from images taken in (a). (c) Pretreatment of keratinocytes with 1 mM thapsigargin (24 hours) and 2 mM EGTA (1 hour) abrogates the propagation of calcium wave upon scratch. White arrow indicates site of scratch wound. Scale bar ¼ 50 mm. (d, e) Scratches were made in confluent keratinocytes with and without thapsigargin and EGTA pretreatment. Immunofluorescence staining (d) and Western blot (e) demonstrating impeded AMD1 upregulation post scratch following thapsigargin and EGTA treatment. Scale bar ¼ 25 mm. NS, no scratch. White dashed line indicates site of scratch wound. All experiments were repeated in biological triplicates.

6

Journal of Investigative Dermatology (2018), Volume

-

HK Lim et al.

Polyamines and Wound Healing Figure 4. AMD1 upregulation is required for cell migration on wounding. (a) Western blot and RTPCR showing AMD1 levels in keratinocytes expressing scrambled and AMD1 shRNA. RT-PCR data are expressed in relative level and compared to N/TERT-1 untransduced control. (b) Western blot and RT-PCR showing increase in AMD1 expression 6 hours post scratch in scrambled cells compared with shAMD1 cells. qRTPCR data is expressed in relative level and compared to scrambled nonscratched (NS) cells. (c, d) Cell migration of scrambled and shAMD1 keratinocytes were monitored by in vitro scratch assay. Light micrograph of wounded cell monolayer showing impaired wound closure at 6 hours in shAMD1 keratinocytes using the IncuCyte Live Cell Analysis system. Relative wound width was quantified using IncuCyte software by measuring wound width at various time points after scratch. Data were normalized to initial wound width made at time 0. (e, f) Scratch assay showing cell migration in mitomycin Cetreated keratinocytes (5 mg/ml) compared with untreated control. Mitomycin Cetreated keratinocytes show no delay in cell migration. All experiments were repeated in biological triplicate. qRTPCR, quantitative real-time reverse transcriptaseePCR; RT-PCR, reverse transcriptaseePCR; sh, small hairpin. The data are presented as mean  standard error of mean (n ¼ 3). *P < 0.05, **P < 0.01, and ***P < 0.001.

keratinocytes. uPA expression was upregulated at the wound edge in control but not in the AMD1 knockdown cells post wounding (Figure 6b), confirming its dependence on AMD1. Analysis of ex vivo human skin histologic sections showed the upregulation of uPAR 24 hours post wounding (Figure 6c), confirming what has previously been reported in mice (Romer et al., 1994; Solberg et al., 2001). Activation of uPAR binding has been shown to function through b3 integrins to promote activation of RAC, which then signals through the WAVE complex to promote F-actin assembly at the leading edge of cells to enable cells to migrate (Smith and Marshall, 2010). To assess the effect of AMD1 depletion on F-actin reorganization upon wounding, F-actin was visualized using fluorescein-labeled phalloidin in control and AMD1 knockdown cells (Figure 6d). In control keratinocytes 6 hours after scratch, the cells at the wound edge displayed

characteristic actin reorganization, with the extension of filopodia at the leading edge. In contrast, AMD1 knockdown keratinocytes displayed reduced actin reorganization with significantly fewer filopodia. These data suggest that high AMD1 levels are required for actin reorganization in cells at the wound edge. Taken together, our data show that on epidermal wounding, AMD1 is rapidly upregulated at the wound edge and functions to drive cell migration at least in part through the promotion of uPA/uPAR signaling and actin cytoskeletal reorganization (Figure 7). DISCUSSION We show that the polyamine regulator, AMD1, is a critical upstream regulator of wound healing. AMD1 is upregulated at the wound edge within 10 minutes of wounding, where it www.jidonline.org

7

HK Lim et al.

Polyamines and Wound Healing Figure 5. AMD1 promotes high spermine levels to drive cell migration during wound repair. (a) High-performance liquid chromatography analysis showing accumulation of putrescine and reduction of spermine levels in shAMD1 keratinocytes. (b, c) In vitro scratch assay of keratinocytes with and without 5 mM spermine pretreatment. Cells were monitored by IncuCyte Live Cell Analysis system. Spermine promotes scratch wound closure compared with untreated cells. Relative wound width was quantified using IncuCyte software by measuring wound width at various time points after scratch. Data were normalized to initial wound width made at time 0. Scale bar ¼ 50 mm. (d) Hematoxylin and eosinestained images of human skin at 0 and 72 hours after wounding. Upon wounding, wounds were treated topically with 500 mM spermine. Control wounds were treated with 1 PBS. Treatment was administered daily for 3 consecutive days. New epithelial tongues extended from the wound margin are indicated by black hatched lines. Scale bar ¼ 250 mm. (e) Epithelial tongue length was quantified following spermine or 1 PBS treatment (n ¼ 5). (f, g) Scratch assay as in (b, c) showing addition of putrescine significantly impedes cell migration. (h, i) The effect of spermine on migration of scratched keratinocytes. Spermine restores impaired cell migration in shAMD1 keratinocytes. (j, k) Pharmacologic inhibition of AMD1 with specific inhibitor EGBG (25 mM) alone or EGBG with supplementation of 15 mM spermidine/spermine on scratched keratinocytes. EGBG treated keratinocytes show impaired cell migration. Supplementation with spermidine/spermine restored impaired cell migration in EGBGtreated cells. All in vitro experiments were performed in at least biological triplicate. EGBG, ethylglyoxal bis(guanylhydrazone); PBS, phosphate buffered saline; sh, small hairpin. Data are presented as mean  standard error of mean. *P < 0.05.

is required to promote high spermine levels. Upregulation of AMD1 is required for cell migration during wound healing and functions in part to promote uPA/uPAR levels to enable reorganization of the actin cytoskeleton required for cell migration. Supplementation with spermine rescues the AMD1 knockdown phenotype and addition of spermine alone is sufficient to drive increased cell migration (Figure 7). 8

Journal of Investigative Dermatology (2018), Volume

-

The absolute requirement for polyamines in normal cell function is well established. However, less is known about their regulatory role, where fluctuations in the levels and ratios of the three polyamines influence cellular behavior and function (Pegg, 2016). While ODC1 is the major rate-limiting enzyme for the synthesis of putrescine, AMD1 is rate limiting for the conversion of putrescine to spermidine and spermine

HK Lim et al.

Polyamines and Wound Healing Figure 5. Continued

and thus can regulate the ratio of putrescine to spermidine and spermine. Here, we show that keratinocytes at the epidermal wound edge rapidly upregulate AMD1 on wounding. This occurs in the presence of only mild ODC1 upregulation. As a result, putrescine levels are decreased due to its conversion to spermidine and spermine at the wound edge. We propose that wound-edge keratinocytes are sensitive to modulation of polyamine ratios, which leads to regulation of gene expression changes required for cell migration. Our data show that the knockdown of AMD1 results in decreased spermine levels and an inhibition of migration. This phenotype can be rescued with the addition of spermine and, in addition, spermine alone can drive

migration of cultured keratinocytes and re-epithelialization in human ex vivo wounds. Wound healing is a complex process that involves crosstalk between multiple cell types and is challenging to model (Elliot et al., 2018). Here, we have used cultured keratinocytes to model the re-epithelialization phase of wound healing. This allows the genetic manipulation of target genes to gain insight into how cell migration is controlled. We have validated the key findings in human ex vivo wounds and confirmed that AMD1 protein is upregulated in human ex vivo wound edges and that spermine promotes re-epithelialization in a human ex vivo wound model. We believe that these data together support a role for www.jidonline.org

9

HK Lim et al.

Polyamines and Wound Healing

Figure 6. AMD1 is required for uPA/uPAR signaling and actin cytoskeletal reorganization. (a) Western blot showing upregulation of uPAR and uPA in scrambled and shAMD1 scratched keratinocyte monolayers. (b) Immunostaining of uPA in scrambled and shAMD1 scratched keratinocyte monolayers (red). White dashed line indicates site of scratch wound. NS, no scratch. Scale bar ¼ 50 mm. (c) Fluorescence imaging of human wound sections stained with uPAR (red/gray), showing increased expression at the wound edge at 24 hours post wounding. Hoechst DNA stain is in blue. White dashed line delineates boundary between epidermis and upper dermis. Scale bar ¼ 50 mm. White arrowhead indicates site of initial wound. (d) Immunostaining of F-actin (green) with Hoechst DNA stain (blue) showing increased expression of filopodia (green arrow) at the wound edge of scrambled control cells. shAMD1 knockdown cells show reduced filopodia formation at the wound edge. sh, small hairpin; UPA, urokinaseetype plasminogen activator; uPAR, urokinaseetype plasminogen activator receptor. Scale bar ¼ 10 mm. Enlarged images are shown in the right panels. All in vitro experiments were performed in at least biological triplicate.

10

Journal of Investigative Dermatology (2018), Volume

-

HK Lim et al.

Polyamines and Wound Healing

Figure 6. Continued

AMD1 and spermine in the promotion of cell migration during wound healing. One of the earliest events detected post-wounding is the release of an intracellular calcium wave (Tran et al., 1999). Given that AMD1 starts to be upregulated within 10 minutes of wounding, we propose that the upregulation of AMD1 is an early downstream effector of calcium signaling required to enable wound closure. AMD1 then functions to drive gene expression changes through the modulation of polyamine levels. Polyamine binding to DNA, RNA, or protein can influence RNA function, promoter activity, and protein function (Igarashi and Kashiwagi, 2010; Lightfoot and Hall, 2014). It is likely that, on wounding, the shift in polyamines impacts gene expression of many downstream targets required to drive the wound healing process. While polyamines have the potential to bind many different RNAs, DNA elements, and proteins, there is, in fact, specificity in their binding and the three polyamines have different targets (Igarashi and Kashiwagi, 2010; Lightfoot and Hall, 2014). The ability of the polyamines to target and influence specific RNAs or promoters is also likely to be highly dependent on cellular context. We propose that, on wounding, polyamine levels are modulated within a physiological range and this shift plays a regulatory role in controlling gene expression to influence cellular behavior to enable cell migration. Migration of keratinocytes at the wound edge is supported by an increase in keratinocyte proliferation behind the migrating cells to provide sufficient cells to cover the wound.

We and others have shown that ODC1 is upregulated at the wound edge and this likely drives an increase in all three polyamines to drive increased proliferation (Mizutani, 1974). We see strong AMD1 upregulation restricted to the keratinocytes 5e6 cells deep at the scratch wound edge. Our data suggest that high spermine along with decreased putrescine is required specifically for leading-edge cells to migrate across the wound. A characteristic of chronic non-healing wounds is a failure of keratinocyte migration with increased proliferation at the wound edge. It is possible that the regulation of polyamines we see in normal wounds is defective in nonhealing wounds, and that the correct balance in the ratios and levels of the polyamines is required for efficient wound closure. It will be interesting in the future to address a potential role for inappropriate polyamine levels in non-healing wounds. Our data demonstrate an essential role for polyamines in regulating and driving wound healing. Further understanding of this regulation may lead to innovative treatments for chronic or slow-healing wounds. MATERIALS AND METHODS Cell culture Human immortalized keratinocytes (Dickson et al., 2000), N/TERT-1, were obtained and cultured in keratinocyte serum-free media (Life Technologies; Thermo Fisher Scientific, Waltham, MA) supplemented with 25 mg/ml bovine pituitary extract (BPE), 0.2 ng/ml epidermal growth factor, and 0.3 mM CaCl2, as described previously (Rheinwald et al., 2002) N/TERT-1 cells at 100% confluence were scratched in www.jidonline.org

11

HK Lim et al.

Polyamines and Wound Healing Figure 7. Model showing how AMD1 promotes cell migration during wound healing. On epidermal wounding, a calcium wave triggers rapid AMD1 upregulation in the wound-edge tissue. This results in a decrease in putrescine, which is converted to spermidine and spermine. This, in turn, drives uPA/ uPAR signaling and actin cytoskeleton reorganization to promote cell migration and wound closure. UPA, urokinaseetype plasminogen activator; uPAR, urokinaseetype plasminogen activator receptor.

media with a 1:1 ratio mixture of supplemented keratinocyte serumfree media and DF-K (as described by Dickson et al., 2000). Lentivirus containing the AMD1 shRNAs were made according to the manufacturer’s protocol (GIPZ; Thermo Fisher Scientific) and used to make stable N/TERT-1 shAMD1 cell lines. Target sequences for AMD1 silencing are 5ʹ-AGACTTCTACAACTTTCCT-3ʹ and 5ʹ- TTAATAGAACAGTCCTAGA-3ʹ.

In vitro scratch cell migration assay Confluent keratinocytes on ImageLock 96-well microplates (Essen BioScience, Inc, Ann Arbor, MI) were scratched using a 96-pin wound maker (Essen BioScience, Inc) to create homogeneous, 700- to 800-mm wide scratch wounds in cell monolayers. Scratched cells were imaged using a real-time cell imaging system, IncuCyte (Essen BioScience, Inc). Putrescine (10 mM), spermidine (15 mM), and spermine (5 mM; Sigma-Aldrich, St Louis, MO) were added to cells and incubated for 48 hours prior to scratch. Cells were pretreated with ethylglyoxal bis(guanylhydrazone), an AMD1 inhibitor (25 mM, provided by Kazuei Igarashi) for 24 hours prior to scratch. Cells were incubated with mitomycin C (M4287; Sigma-Aldrich) for 3 hours before cells were scratched. For analysis of protein and RNA expression in migrating cells, keratinocytes were cultured to full confluency in a six-well culture dish (Corning Costar; Corning, Tewksbury, MA) and scratched with the large end of a 1-ml pipette tip. At specific time points, the cells were either harvested in RIPA buffer for Western blot or TRIzol reagent (Invitrogen; Thermo Fisher Scientific) for RNA analysis.

Calcium (Ca2D) wave imaging To deplete intracellular and extracellular Ca2þ sources, cells were treated with 1 mM thapsigargin (Sigma-Aldrich) for 24 hours and 2 mM EGTA (Sigma-Aldrich) for 1 hour prior to scratch. For Ca2þ flux imaging, cells were incubated with Fluo-4,AM (Life Technologies; Thermo Fisher Scientific) and imaged with an automated confocal microscope (Spinning disk 3D-FRAP; Nikon Ti, Nikon, Tokyo, Japan) at 0.2-second intervals for 5 minutes. Image analysis was performed using ImageJ software.

Human wound biopsies Human skin samples were sourced from T.C. Lim at the Yong Loo Lin School of Medicine, National University of Singapore. All participants gave written, informed consent. The study was done in accordance with the Declaration of Helsinki and was approved by a local scientific ethics review board. The skin was further dissected into 6-mm cubes or circles. Excisional skin wounds of approximately 3e4 mm were made with surgical scissors. The ex vivo wounded skin explants were cultured at the aireliquid interface and treated 12

Journal of Investigative Dermatology (2018), Volume

-

topically with 500 mM of spermine or 1 phosphate-buffered saline every 24 hours for 3 days. Non-wounded skin and wounded biopsies (day 0, day 1, and day 3) were bisected, with half, snap-frozen in Optimal Cutting Temperature Compound and the other half, fixed in 10% buffered formalin saline and embedded in paraffin for histologic analysis. For AMD1 staining and ex vivo epithelial tongue length analysis, five biologic replicates from different donors were analyzed with three wounds per replicate.

Histology, immunohistochemistry, and immunofluorescence Histologic sections were prepared from wound tissue embedded in paraffin or snap-frozen in Optimal Cutting Temperature Compound. Seven-micrometer-thick sections were stained with hematoxylin and eosin. For immunofluorescence, sections or in vitro scratched cells on cover slips were incubated with primary antibodies against AMD1 and AMD1 blocking peptide (sc-390037; sc-390073 P; Santa Cruz Biotechnology, Inc, Santa Cruz, CA), ODC1 (MA1-25811; Thermo Fisher Scientific), involucrin (Ab53112; Abcam, Cambridge, MA), uPA and uPAR (GTX89445; GTX59605; GeneTex, Inc, Irvine, TX), K6 (a gift from Birgit Lane), and KI-67 (Ab9260, Merck Millipore, Billerica, MA), followed by appropriate Alexa Fluor 488e or 594econjugated secondary antibodies (Molecular Probes; Thermo Fisher Scientific) and counterstained with Hoechst 33342 (Molecular Probes; Thermo Fisher Scientific). To visualize F-actin filaments, the cells were incubated with Alexa Fluor 488elabeled phalloidin (A12379; Life Technologies; Thermo Fisher Scientific) and counterstained with Hoechst 33342. Images were acquired using either a fluorescence microscope (Zeiss AxioImager Z1; Carl Zeiss, Oberkochen, Germany) or confocal microscope (Olympus FV1000). All images were taken at room temperature with a Texas red, GFP, or DAPI filter. Image analysis was performed using Zeiss image analysis software, ZEN Pro.

Quantitation of AMD1 level using ImageJ To measure and compare AMD1 levels in non-wounded and 24hour post-wounding human skin biopsies, at least three immunofluorescence images were captured at random along the length of skin epidermis from each of three biologic replicates. For quantitation, 50 cells at the basal layer of the epidermis were selected at random from non-wounded skin and compared with 50 cells selected from wound edge skin. The nuclei were masked and the cells were numbered and analyzed using ImageJ.

Quantitative real-time PCR Total RNA was reverse transcribed to cDNA (RevertAid; Thermo Fisher Scientific). Luminaris Colour HiGreen Hi ROX master mix (Thermo Fisher Scientific) was used with transcript-specific primers

HK Lim et al.

Polyamines and Wound Healing for quantitative real-time reverse transcriptaseePCR on an ABI PRISM 7900 sequence detection system. CT values were normalized to the glyceraldehyde-3-phosphate dehydrogenase or ribosomal protein LP0 or L13A mRNA.

Western blot and measurement of polyamine levels Ninety micrograms of protein was resolved on a 4e12% Bis-Tris gel (NuPAGE) and transferred onto nitrocellulose membranes. Primary antibodies against filaggrin, involucrin and glyceraldehyde-3phosphate dehydrogenase (Ab3137; Ab53112; Ab9484; Abcam), AMD1 (sc-390037; Santa Cruz Biotechnology, Inc), uPAR, uPA (GTX59605 and GTX89445; GeneTex, Inc) and b-actin (A5441; Sigma-Aldrich) were used followed by incubation with horseradish peroxidaseelabeled secondary antibodies. High-performance liquid chromatography was used to measure the levels of polyamines as described previously (Igarashi et al., 1986).

Statistical analysis All data are presented as means  standard of mean. Two-tailed Student t test was used to determine the significance between two groups. P < 0.05 were considered to be statistically significant. All experiments were performed in at least biological triplicate as indicated in the legends. ORCID Leah A. Vardy: http://orcid.org/0000-0003-4186-684X.

CONFLICT OF INTEREST

Igarashi K, Kashiwagi K. Modulation of cellular function by polyamines. Int J Biochem Cell Biol 2010;42:39e51. Igarashi K, Kashiwagi K, Hamasaki H, Miura A, Kakegawa T, Hirose S, et al. Formation of a compensatory polyamine by Escherichia coli polyaminerequiring mutants during growth in the absence of polyamines. J Bacteriol 1986;166:128e34. Igarashi K, Porter CW, Morris DR. Comparison of specificity of inhibition of polyamine synthesis in bovine lymphocytes by ethylglyoxal bis(guanylhydrazone) and methylglyoxal bis(guanylhydrazone). Cancer Res 1984;44: 5326e31. Leopold PL, Vincent J, Wang H. A comparison of epithelial-to-mesenchymal transition and re-epithelialization. Semin Cancer Biol 2012;22:471e83. Lightfoot HL, Hall J. Endogenous polyamine function—the RNA perspective. Nucleic Acids Res 2014;42:11275e90. Lindley LE, Stojadinovic O, Pastar I, Tomic-Canic M. Biology and biomarkers for wound healing. Plast Reconstr Surg 2016;138(3 Suppl):18Se28S. Maddaluno L, Urwyler C, Werner S. Fibroblast growth factors: key players in regeneration and tissue repair. Development 2017;144:4047e60. Maeno Y, Takabe F, Inoue H, Iwasa M. A study on the vital reaction in wounded skin: simultaneous determination of histamine and polyamines in injured rat skin by high-performance liquid chromatography. Forensic Sci Int 1990;46:255e68. Mizutani AT, y. Changes in polyamine metabolism during wound healing in rat skin. Biochim Biophys Acta 1974;338:183e90. Nowotarski SL, Woster PM, Casero RA Jr. Polyamines and cancer: implications for chemotherapy and chemoprevention. Expert Rev Mol Med 2013;15:e3. Pastar I, Stojadinovic O, Yin NC, Ramirez H, Nusbaum AG, Sawaya A, et al. Epithelialization in wound healing: a comprehensive review. Adv Wound Care (New Rochelle) 2014;3:445e64.

The authors state no conflict of interest.

Pegg AE. Functions of polyamines in mammals. J Biol Chem 2016;291(29): 14904e12.

ACKNOWLEDGMENTS

Rheinwald JG, Hahn WC, Ramsey MR, Wu JY, Guo Z, Tsao H, et al. A twostage, p16(INK4A)- and p53-dependent keratinocyte senescence mechanism that limits replicative potential independent of telomere status. Mol Cell Biol 2002;22:5157e72.

We thank the Agency for Science, Technology and Research for funding. The N/TERT-1 cell line was kindly donated by Jim Rheinwald. We are very grateful to John Lim and Graham Wright at the IMB microscopy suite, Agency for Science, Technology and Research. We would like to thank Kerry McLaughlin of Insight Editing London for critical review of the manuscript.

SUPPLEMENTARY MATERIAL

Romer J, Lund LR, Eriksen J, Pyke C, Kristensen P, Dano K. The receptor for urokinase-type plasminogen activator is expressed by keratinocytes at the leading edge during re-epithelialization of mouse skin wounds. J Invest Dermatol 1994;102:519e22.

REFERENCES

Schimchowitsch S, Cassel JC. Polyamine and aminoguanidine treatments to promote structural and functional recovery in the adult mammalian brain after injury: a brief literature review and preliminary data about their combined administration. J Physiol Paris 2006;99:221e31.

Casero RA, Pegg AE. Polyamine catabolism and disease. Biochem J 2009;421: 323e38.

Shaw TJ, Martin P. Wound repair: a showcase for cell plasticity and migration. Curr Opin Cell Biol 2016;42:29e37.

Cordeiro JV, Jacinto A. The role of transcription-independent damage signals in the initiation of epithelial wound healing. Nat Rev Mol Cell Biol 2013;14:249e62.

Shi HP, Fishel RS, Efron DT, Williams JZ, Fishel MH, Barbul A. Effect of supplemental ornithine on wound healing. J Surg Res 2002;106:299e302.

Supplementary material is linked to the online version of the paper at www. jidonline.org, and at https://doi.org/10.1016/j.jid.2018.05.029.

Dickson MA, Hahn WC, Ino Y, Ronfard V, Wu JY, Weinberg RA, et al. Human keratinocytes that express hTERT and also bypass a p16(INK4a)-enforced mechanism that limits life span become immortal yet retain normal growth and differentiation characteristics. Mol Cell Biol 2000;20:1436e47.

Smith HW, Marshall CJ. Regulation of cell signalling by uPAR. Nat Rev Mol Cell Biol 2010;11:23e36. Solberg H, Ploug M, Hoyer-Hansen G, Nielsen BS, Lund LR. The murine receptor for urokinase-type plasminogen activator is primarily expressed in tissues actively undergoing remodeling. J Histochem Cytochem 2001;49:237e46.

El Baze P, Milano G, Verrando P, Renee N, Ortonne JP. Polyamine levels in normal human skin. A comparative study of pure epidermis, pure dermis, and suction blister fluid. Arch Dermatol Res 1983;275:218e21.

Stone RC, Pastar I, Ojeh N, Chen V, Liu S, Garzon KI, et al. Epithelialmesenchymal transition in tissue repair and fibrosis. Cell Tissue Res 2016;365:495e506.

Elliot S, Wikramanayake TC, Jozic I, Tomic-Canic M. A modeling conundrum: murine models for cutaneous wound healing. J Invest Dermatol 2018;138: 736e40.

Tomic-Canic M, Komine M, Freedberg IM, Blumenberg M. Epidermal signal transduction and transcription factor activation in activated keratinocytes. J Dermatol Sci 1998;17:167e81.

Gao JH, Guo LJ, Huang ZY, Rao JN, Tang CW. Roles of cellular polyamines in mucosal healing in the gastrointestinal tract. J Physiol Pharmacol 2013;64: 681e93.

Tran PO, Hinman LE, Unger GM, Sammak PJ. A wound-induced [Ca2þ]i increase and its transcriptional activation of immediate early genes is important in the regulation of motility. Exp Cell Res 1999;246:319e26.

Gilad GM, Gilad VH. Overview of the brain polyamine-stress-response: regulation, development, and modulation by lithium and role in cell survival. Cell Mol Neurobiol 2003;23:637e49.

Zhang D, Zhao T, Ang HS, Chong P, Saiki R, Igarashi K, et al. AMD1 is essential for ESC self-renewal and is translationally down-regulated on differentiation to neural precursor cells. Genes Dev 2012;26:461e73.

www.jidonline.org

13