Homo Sapiens (Hsa)-microRNA (miR)-6727-5p Contributes to the Impact of High-Density Lipoproteins on Fibroblast Wound Healing In Vitro

Chronic, non-healing wounds are a significant cause of global morbidity and mortality, and strategies to improve delayed wound closure represent an unmet clinical need. High-density lipoproteins (HDL) can enhance wound healing, but exploitation of this finding is challenging due to the complexity and instability of these heterogeneous lipoproteins. The responsiveness of primary human neonatal keratinocytes, and neonatal and human dermal fibroblasts (HDF) to HDL was confirmed by cholesterol efflux, but promotion of ‘scrape’ wound healing occurred only in primary human neonatal (HDFn) and adult fibroblasts (HDFa). Treatment of human fibroblasts with HDL induced multiple changes in the expression of small non-coding microRNA sequences, determined by microchip array, including hsa-miR-6727-5p. Intriguingly, levels of hsa-miR-6727-5p increased in HDFn, but decreased in HDFa, after exposure to HDL. Delivery of a hsa-miR-6727-5p mimic elicited repression of different target genes in HDFn (ZNF584) and HDFa (EDEM3, KRAS), and promoted wound closure in HDFn. By contrast, a hsa-miR-6727-5p inhibitor promoted wound closure in HDFa. We conclude that HDL treatment exerts distinct effects on the expression of hsa-miR-6727-5p in neonatal and adult fibroblasts, and that this is a sequence which plays differential roles in wound healing in these cell types, but cannot replicate the myriad effects of HDL.


Introduction
Chronic, non-healing wounds are a significant cause of morbidity and mortality in the United Kingdom; according to The Health Improvement Network (THIN database), there were an estimated 3.8 million patients with a wound managed by the National Health Service (NHS) in 2017/2018, of which 89% and 49% of acute and chronic wounds healed, respectively, within the study year [1]. The impact on the NHS was assessed as £8.3 billion, the majority of this cost incurred in the community, with £5.6 billion spent managing unhealed wounds [1].

In Vitro Wound Healing: Scrape ('Scratch') Wound Assays
A scrape wound was introduced to confluent (90%) monolayers of fibroblasts or keratinocytes, using a sterile 100 µL pipette tip as previously described [37]. Fibroblasts were incubated in serum-free DMEM, or serum-free media supplemented with HDL (5-20 µg mL −1 ) or with 10% (v/v) FBS; keratinocytes were incubated in EpiLife media supplemented with S7 (above) and plated on coating matrix. Cell migration was monitored by capturing triplicate images of wound area (0-72 h) using Image J software (University of Wisconsin) at the intervals indicated in the figure legends, and values normalised by comparison with the corresponding initial wound size.

Total RNA Isolation, Microchip Analysis and Quantitative PCR
Isolation of RNA was performed using either a Directzol RNA MiniPrep Plus kit (Zymo Research, Cambridge Bioscience, Cambridge, UK) or a Nucleospin RNA kit (Machery-Nagel, Germany). Complementary DNA (cDNA) to be used for assessment of gene expression was generated from 250 ng RNA using the High-Capacity cDNA Reverse Transcription kit (ThermoFisher Scientific, Waltham, MA, USA); cDNA for measurement of miRNA and/or mRNA was generated using miScript II RT kit (Qiagen, Manchester, UK) and HiFlex buffer from the same supplier, and performed according to the manufacturer's instructions. Analysis of miRNA expression was carried out by LC Sciences (Houston, TX, USA): 2632 unique mature hsa-miR sequences derived from miRBase (version 22) [38] were assessed using uParaflo microfluidic chip technology (http://www.lcsciences.com/ discovery/applications/transcriptomics/mirna-profiling/mirna/ (accessed on 23 September 2021)). Relative quantitative (q)PCR for miRNA and mRNA expression was performed using HOT FIREPol EvaGreen Q-PCR mix plus (Solid BioDyne), the primer sequences defined in Table 1 and a C1000 thermal cycler with a CFX96 real-time system; no dimer formation was detected using melt curves. Expressions of target miRNA and mRNA were determined using the 2 −∆∆Ct method relative to an invariant control sequence or gene, as indicated in the figure legends. Primer efficiencies were confirmed as 90-110% for each sequence investigated. Statistical analysis for significance (p < 0.05) used the ∆Ct values, compared with the relevant housekeeping sequence.
Delivery of hsa-mir-6727-5p mimic, inhibitor and respective scrambled controls (Qiagen, Manchester, UK) to primary human neonatal and adult fibroblasts was achieved using HiPerFect Transfection Reagent (Qiagen, Manchester, UK), following the manufacturer's instructions. Transfection complexes were added in serum-free DMEM, and incubated at 37 • C and 5% CO 2 at the concentrations, and for the periods of time indicated, in the figure legends.

Statistical Analyses
All datasets were deposited in Mendeley Data, V1, doi: 10.17632/mxpjjw4t4g.1. The results are expressed as mean ± SD or SEM of the number of experiments indicated in the figure legends. Statistical analysis was performed by a one-way or two-way ANOVA and Dunnett's post-test, or a Student's t-test, as indicated in the figure legends; please note that normal distribution was assumed, but could not be proven with the number of independent experiments performed in this study. All testing was performed using GraphPad Prism 8.0 software San Diego, CA, USA; * p < 0.05, ** p < 0.01 and *** p < 0.01.

The Effect of HDL on Viability of HKn, HDFn and HDFa
Human neonatal keratinocytes showed positive staining for E-cadherin, and negative staining for vimentin, and the isotype control (data not shown); treatment (24 h) with HDL (0-100 µg mL −1 ) had no significant impact on cell viability ( Figure 1A), as judged by the conversion of MTT to formazan [33]. Neonatal human dermal fibroblasts stained positively for vimentin and negatively for E-cadherin and the isotype control (data not shown), while exposure to the same range of HDL concentrations revealed significant increases in conversion of MTT to formazan at 5 µg mL −1 HDL (21.5 ± 2.43%; p < 0.05) and 10 µg mL −1 HDL (19.6 ± 6.90%; p < 0.05) ( Figure 1B) when compared with the control condition. An equivalent staining pattern was observed for HDFa (data not shown); increased conversion of MTT to formazan was observed at 20 µg mL −1 HDL (32.3 ± 7.14%; p < 0.05) compared to the control ( Figure 1C).

The Impact of HDL on Scrape Wound Healing by HKn, HDFn and HDFa
The effects of HDL (5-20 µg mL −1 ) on closure (0-72 h) of scrape wounds in vitro by HKn, HDFn and HDFa are shown in Figure 3. No significant increase in wound closure was noted in the presence of HDL, compared to control, in HKn ( Figure 3A), and was, therefore, not investigated further. Exposure of HDFn to HDL enhanced wound closure at 24 h, 48 h and 72 h compared with the control (serum-free media); after 24 h, 10 µg mL −1 and 20 µg mL −1 HDL increased scrape closure by 1.50-fold (p < 0.001) and 1.78-fold (p < 0.001), respectively, compared with control ( Figure 3B). By contrast, treatment with HDL (5 µg mL −1 ) significantly enhanced wound closure in HDFa ( Figure 3C) after 6 h (1.38-fold; p < 0.01); this effect was sustained, but not increased, over 72 h (1.32-fold; p < 0.001), compared with the control, but no dose dependency was observed. Complete wound closure was achieved in both HDFn and HDFa at 72 h in the presence of media supplemented with 10% (v/v) FBS (data not shown); analysis of the area under the curve (AUC) revealed no significant differences in wound healing between HDFn and HDFa under basal conditions.  . 'Scrape' wound healing, in the presence or absence of HDL (5-20 µg mL −1 ; 0-72 h), compared with the control (serum-free media) condition, by HKn, HDFn and HDFa, are shown in (A-C), respectively. Data are the mean ± SEM of three independent experiments (each performed using triplicate wells). The data were analysed using two-way ANOVA and Tukey post-hoc test; * p < 0.05; ** p < 0.01; *** p < 0.001. In (C), after 6 h, significance (p < 0.01) from control is noted after treatment with HDL (5 µg mL −1 and 20 µg mL −1 ); at 9 h, significance is noted for 5 µg mL −1 (p < 0.01), 10 µg mL −1 (p < 0.001) and 20 µg mL −1 (p < 0.001); for time points thereafter, significance (p < 0.001) from control is noted for all three concentrations tested.

Changes in microRNA Expression Caused by Exposure to HDL in HDFn and HDFa
The changes in expression of microRNA (miRNA) sequences in HDFn and HDFa, caused by exposure (24 h) to 20 µg mL −1 HDL compared to the relevant controls (serumfree media) are shown in Figure 4. Treatment with HDL induced pleiotropic changes in miRNA expression, in HDFn and HDFa up-regulating ( Figure 4A,D, respectively) and down-regulating ( Figure 4B,E, respectively) distinct sequences, compared with the control, visualised using heatmaps [39]; the volcano plots ( Figure 4C,F) shows the unstandardised signal (log-fold change) vs. standardised signal (t-statistic) [40], indicating sequences that show alterations (+/−) from the control condition with high significance in HDFn and HDFa, respectively.

Hsa-miR-6727-5p Mimic and Inhibitor: Impact on Cell Viability and Scrape Wound Healing in HDFn and HDFa
The impacts of transient delivery of hsa-miR-6727-5p mimic and inhibitor on the viability of HDFn and HDFa, as judged by conversion of MTT to formazan, are shown in Figure 6A,D. Treatment with mimic (10 nM; 24 h) caused a minor, but statistically significant (4.6%; p < 0.05) increase in the production of formazan in HDFn, but not HDFa, compared with the scrambled siRNA control; no changes in viability were noted after treatment with the inhibitor sequence (50 nM; 24 h) compared with the same concentration of its control. The effects of treatment with hsa-mir-6727-5p mimic in the scrape wound healing assay in HDFn and HDFa are shown in Figure 6B,E. In the neonatal HDF addition of the mimic maximally increased wound healing after incubation for 9 h (1.47-fold; p < 0.05) with stimulation at time points examined thereafter ( Figure 6B). By contrast, the mimic significantly reduced wound closure in HDFa, maximally at 9 h (44%; p < 0.01), with inhibition sustained after 72 h ( Figure 6D). The same inverse relationship was noted in response to hsa-miR-6727-5p inhibitor in HDFn and HDFa ( Figure 6C,E): the inhibitor reduced wound healing from 24-72 h by around 25% in HDFn, but significantly (p < 0.001) promoted wound closure in HDFa at each time point tested, maximally at 3 h (2.1-fold).   F). Data are the mean ± SD of four independent experiments, each performed using duplicate wells; SD was utilised as error bars were too small to indicate the variance of the data using SEM. The data were analysed using two-way ANOVA and Tukey post-hoc test; * p < 0.05, ** p < 0.01; *** p < 0.001.

Discussion
There is an urgent need for the development of new therapeutics to aid chronic wound healing, based on understanding of the mechanisms which promote progression of this process. This study demonstrates the impact of HDL, at concentrations which induce cholesterol efflux (Figure 2), on scrape wound healing in vitro by primary human neonatal keratinocytes, neonatal human dermal fibroblasts and adult human dermal fibroblasts (Figure 3). Improvements in wound closure were noted in cultures of HDFn and HDFa, with the former proving more dose responsive to HDL treatment. Incubation with HDL induced multiple changes in miRNA sequences in both HDFn and HDFa, as determined by microchip array (Figure 4), although some of these changes could not be validated by qPCR ( Figure 5). Intriguingly, the expression of hsa-miR-6727-5p was significantly enhanced in HDFn, but decreased in HDFa exposed to HDL; delivery of a hsa-miR-6727-5p mimic elicited differential repression of target genes in HDFn (ZNF584) and HDFa (EDEM3, KRAS). No marked changes in cellular viability were noted in the presence of hsa-miR-6727-5p mimic or inhibitor; notably, promotion of wound closure was observed in the presence of the hsa-miR-6727-5p mimic in HDFn, and by the inhibitor in HDFa (Figure 6).
Effective cholesterol homeostasis in keratinocytes and dermal fibroblasts is needed to sustain the integrity of cutaneous permeability barrier function [41][42][43][44] and to provide one of the key membrane components needed for cellular proliferation and migration [45][46][47][48]. Here, cholesterol efflux was used primarily as confirmation of cellular responsiveness to HDL (Figure 2), but some interesting findings emerge, most notably that efflux from neonatal dermal fibroblasts was lower than that observed in adult dermal fibroblasts and keratinocytes. Differential expression of many genes is found in neonatal and adult fibroblasts [49], including those encoding proteins involved in the organisation and structure of extracellular matrix, cell adhesion, proliferation and migration; our own work has established differential roles for connexin 43 in mediating wound closure in neonatal, juvenile and adult dermal cells [37]. Changes in cholesterol metabolism in dermal fibroblasts are known to occur in genetic conditions such as mitochondrial disorders [50], and inherited disorders of cholesterol metabolism [51,52], but ageing also decreases the cholesterol content within plasma membrane lipid rafts [53], and enhances the cholesteryl ester mass, of human fibroblasts [54], changes which can induce compensatory increases in cholesterol efflux [55].
Treatment with HDL promoted scrape wound closure in HDFn and HDFa, but not keratinocytes. This implies, but does not prove, that the impact of HDL on wound healing [12] is primarily an event mediated by fibroblasts. It is possible that one of the serum-free media components used to sustain keratinocyte growth in culture may mimic the effect of HDL on wound closure, but not cholesterol efflux (above), or that removal of cholesterol via the efflux pathway may limit the source of this material for cell division. Treatment with HDL does not affect keratinocyte viability, but does increase conversion of MTT to formazan in HDFn and HDFa, a finding which implies increased cell metabolism, mitochondrial function, or cell numbers [33]. The concentration of HDL required to increase wound healing in vivo [12] was substantively higher than those used here: 800 µg protein mL −1 of HDL in 20% pluronic F-127 gel, every 2 days for 10 days, increased wound healing in apoE −/− mice by nearly 50%, compared with the pluronic gel control. Yu et al. (2017) also utilised high levels of HDL (100 mg mL −1 apoA-I concentration) to achieve increases in proliferation and migration of type II alveolar epithelial cells [23], but Zhang et al. (2010) employed equivalent concentrations to those used here to stimulate proliferation of endothelial progenitor cells and promote wound healing [13], via a pathway which involved activation of phosphoinositide -3-kinase (PI3K), protein kinase B (PKB; Akt1) and cyclin D.
An initial screen of the microRNA sequences modified by HDL treatment in HDFn and HDFa revealed multiple sequences which could potentially impact on the complex process of wound closure. However, not all of the regulated sequences identified in this screen could be validated by qPCR in three independent experiments: this discrepancy may be explained in part by signal intensities in the microarray, as lower threshold levels may elicit false positives [56][57][58]. Use of an intercalating dye for detection of poorly expressed miRNA sequences by qPCR may also result in loss of specificity [59][60][61]. Furthermore, the commercial miRNA detection platform uses the miRBase repository for sequence information, which is regularly updated; this could result in non-alignment of the microarray probe sequences and the corresponding miScript primer sequences used here, which were not publicly available.
However, hsa-miR-6727-5p was confirmed to be up-regulated in HDFn, and downregulated in HDFa, by treatment with HDL ( Figure 5); intriguingly, this was not one of the sequences identified as regulated by HDL or apoA-I in human PANC-1 hybrid 1.1B4 pancreatic beta cells under equivalent conditions [28]. Only one published report exists on this sequence, indicating that it promotes the proliferation, migration and invasion of cervical cancer cell lines [60] suggesting a possible role in mediating the impact of HDL in wound healing. This proved to be the case, despite the fact that multiple miRNA sequences are regulated by exposure to HDL: a rise in this sequence in HDFn promoted wound closure, while inhibition of this sequence achieved the same goal in HDFa. The mechanism by which this is achieved remains unknown. Both cell types exhibit repression of PRX, the gene encoding periaxin ( Figure 6); while periaxin is expressed in fibroblasts [61,62], this protein is more usually associated with myelination in Schwann cells [63], where it interacts with the dystroglycan-dystrophin-related protein -2 complex that links the cytoskeleton to the extracellular matrix. By contrast, KRAS, the most frequently mutated isoform of the RAS proto-oncogene, is modestly increased in HDFn, and repressed in HDFa by the hsa-miR-6727-6p mimic. The RAS/mitogen-activated protein kinase (MAP) signalling pathway has been implicated in metabolic reprogramming [64], and cancerassociated fibroblasts are a key element in the tumour microenvironment, influencing cellular proliferation, matrix deposition and remodelling and crosstalk with infiltrating leucocytes [65]. The only other biologically significant finding was the repression of EDEM3 in HDFa, but not HDFn: EDEM3 is a mannosidase protein involved in ER quality-control, which recognises misfolded proteins targeted for proteasomal degradation [66]. Fibroblasts with a missense variant of EDEM3 exhibit decreased trimming of Man 8 GlcNAc 2 isomer B to Man7LCNAc2, a decrease in EIF2AK3 (Eukaryotic Translation Initiation Factor 2 alpha kinase; protein kinase R (PKR)-like endoplasmic reticulum kinase (PERK)) expression in response to tunicamycin, and an impaired unfolded protein response [66]. At present, it is merely speculation as to whether any of these changes in gene expression contribute to the differing responses to hsa-mir-6727-5p in HDFn and HDFa: it is an obvious limitation of this study that overexpression and knockdown of these genes has not been performed to confirm a functional role in wound closure, remaining an interesting avenue for further exploration.

Conclusions
High-density lipoproteins improve wound healing in both neonatal and adult dermal fibroblasts (human), but not keratinocytes, in a manner which does not seem to be related to the ability of this lipoprotein to facilitate cholesterol efflux. Intriguingly, levels of hsa-miR-6727-5p increased in HDFn, but decreased in HDFa, after exposure to HDL. Delivery of a hsa-miR-6727-5p mimic elicited repression of different target genes in HDFn (ZNF584) and HDFa (EDEM3, KRAS), and promoted wound closure in HDFn. By contrast, a hsa-miR-6727-5p inhibitor promoted wound closure in HDFa. We conclude that HDL treatment exerts distinct effects on the expression of hsa-miR-6727-5p in neonatal and adult fibroblasts, and that this sequence plays differential roles in wound healing in these cell types, but cannot replicate the myriad effects of HDL Acknowledgments: The authors acknowledge the excellent technical expertise within the Department of Biological and Biomedical Sciences. We would wish to acknowledge the award of a Kuwaiti PhD scholarship to Khaled Bastaki.

Conflicts of Interest:
The authors declare no conflict of interest.