MiR equal than others: MicroRNA enhancement for cutaneous wound healing

Keratinocyte migration is vital in the re‐epithelialisation of the skin during wound healing. Multiple factors conspire to impair closure of chronic wounds such as diabetic foot ulcers, venous leg ulcers and pressure wounds. Despite deep mechanistic understanding of microRNA (miRNA) biogenesis and function, the translational potential of these small genetic molecules has not been exploited to promote wound repair. In this review, I focus on miRNAs whose importance for wound healing stems from their impact on epidermal keratinocyte behaviour. These include miR‐21‐5p, miR‐31‐5p, miR‐132‐3p, miR‐19b, miR‐20a, miR‐184, miR‐129‐5p and miR‐335‐5p which regulate diverse aspect of keratinocyte biology such as migration, proliferation, differentiation, inflammation and wound closure. A combinatorial approach where two or more miRNA mimics targeting distinct but complementary wound healing processes is proposed as this may enhance wound repair more effectively than any single miRNA mimic alone.

protein output (Bartel, 2018). A given miRNA can regulate numerous mRNAs, and consequently miRNAs function as master regulators of the molecular status of the cell. As a result, miRNAs represent attractive potential therapeutic targets for complex diseases with multiple pathological features, such as chronic wounds.
As the skin is constantly exposed to potential injury, wound healing is a fundamental physiological process required to maintain the integrity of the skin after trauma. It consists of a series of successive overlapping phases spanning haemostasis, inflammatory, proliferative and remodelling phases (Baltzis et al., 2014;Gonzalez et al., 2016). These stages are driven co-ordinated activity of diverse cell types including keratinocytes, fibroblasts, endothelial cells and infiltrating immune cells, of which keratinocyte replication and migration during the proliferative phase drive the re-formation of the epidermis to secure wound closure (Gonzalez et al., 2016;Martin & Nunan, 2015).
Recent years have seen significant growth in our understanding of miRNA function in keratinocyte migration and as candidate targets for the development of novel therapies for wound healing (Mulholland et al., 2017). In addition, long noncoding RNAs (lncRNA), which are broadly defined as >200 nucleotides, have recently been implicated in wound repair, wound and keratinocyte migrationassociated long noncoding RNA 1 (WAKMAR1) and WAKMAR2 being notable examples . Here, I introduce mechanisms of miRNA expression and function briefly, then focus on miRNAs that have translational promise for wound repair through elevated expression, for instance through miRNA mimics that stimulate keratinocyte migration and other aspects of wound healing.

| MicroRNA biogenesis and function
The canonical pathway for miRNA biogenesis involves transcription from diverse genomic loci including introns, exons and intergenic regions to generate primary (pri-miRNA) (Finnegan & Pasquinelli, 2013). Subsequent processing of the pri-miRNA transcripts into~70 nt precursor miRNA (pre-miRNA) hairpin loop structures appears to be almost completely dependent on DROSHA, a nuclear ribonuclease (RNAse) III enzyme that functions as a complex with the protein product of DiGeorge syndrome critical region gene 8 (DGCR8) (Denli et al., 2004;Gregory et al., 2004;Han et al., 2004;Landthaler et al., 2004;Lee et al., 2003). The pre-miRNAs are transferred from the nucleus into the cytoplasm via the Exportin 5 complex with RanGTP (Bohnsack et al., 2004;Lund et al., 2004;Yi et al., 2003). However, genetic ablation of the XPO5 gene that encodes Exportin-5 only reduced miRNA maturation modestly, indicating the existence of alternative pathways for pre-miRNA translocation (Y. K. Kim et al., 2016). Once in the cytoplasm, each pre-miRNA is processed into a mature miRNA duplex by the RNAse III enzyme DICER, which contributes to the biogenesis of several small regulatory RNAs (Song & Rossi, 2017). One strand of the duplex is loaded into Argonaute (AGO) proteins to form the RNA-induced silencing complex (RISC), strand selection by AGO depending on 5′ nucleotide identity and the relative thermodynamic stabilities of the two ends of the miRNA duplex (Meijer et al., 2014;Sheu-Gruttadauria & MacRae, 2017). The guide stand is stabilised within AGO and targets the RISC to mRNA transcripts, while the other, minor species (the passenger strand, or miRNA*) is usually degraded (Bartel, 2018). However, both strands can accumulate F I G U R E 1 Schematic depiction of the canonical pathway of miRNA biogenesis. After transcription from diverse genomic loci, the primary miRNA transcript (pri-miRNA) is cleaved by the DROSHA/ DGCR8 microprocessor complex, yielding pre-miRNA which is exported from the nucleus to the cytoplasm predominantly by the Exportin-5 Ran-GTP complex. After processing by DICER to form a mature miRNA duplex one strand from the duplex is loaded onto the AGO family of proteins to form the RISC, while the other strand is largely degraded. The RISC can repress translation but the dominant mechanism of silencing involves mRNA destabilization. AGO, Argonaute; mRNA, messenger RNA; miRNA, microRNA; RISC, RNA-induced silencing complex ROSS | 8051 to detectable levels and mediate RISC function, hence mature miRNAs are designated miR-#−5p or miR-#−3p according to the precursor hairpin arm from which they arise (Guo & Lu, 2010;Marco et al., 2012;Okamura et al., 2008;J. S. Yang, Phillips, et al., 2011). Such unambiguous nomenclature is also pertinent given that the choice of guide or passenger strand can vary with cell type or under pathological conditions, with very recent work indicating that the uridylation status of the miRNA strands is a key factor driving this choice (H. Kim et al., 2020). In any case, the RISC binds to the 3′ untranslated region (3′ UTR) of target mRNA transcripts and, in animals, results in abrogation of protein expression through complex mechanisms dominated by mRNA destabilization rather than translational repression (whose overall impact is modest) or mRNA cleavage, which dominates RISC function in plants (Eichhorn et al., 2014;Fang & Qi, 2016;Iwakawa & Tomari, 2015). Key elements of the canonical pathway of miRNA biogenesis are summarised in Figure 1, but it should be noted that several pathways of miRNA biogenesis are known to bypass Drosha or Dicer processing, and these noncanonical mechanisms have been reviewed critically elsewhere (Bartel, 2018;Treiber et al., 2019).

| Enhancing miRNA levels for wound healing: A focus on keratinocytes
The formation of new epidermal tissue over the denuded wound surface is fundamental to the completion of wound healing. Keratinocytes migrate from the wound edge to repopulate the exposed extracellular matrix but in chronic wounds, such keratinocyte migration is impaired (Usui et al., 2008). Diverse cytokines and growth factors promote keratinocyte migration, and early work by Woodley and colleagues found transforming growth factor α (TGFα) was the most potent stimulator of keratinocyte migration among a panel of 11 cytokines and growth factors, while transforming growth factor β (TGFβ) was the weakest (Y. Li et al., 2006). Nonetheless, several TGFβ-induced miRNAs, including miR-21-5p, miR-31-5p and miR-132-3p that have emerged as key miRNAs whose elevation drives keratinocyte migration, re-epithelialisation and other elements of wound healing (Table 1), while the impact of TGFα on miRNA expression and function in keratinocytes remains obscure.

| MicroRNA-21-5p in keratinocytes
Early studies demonstrated miR-21-5p induction by TGFβ in HaCaT keratinocytes (X. Yang, Wang, et al., 2011). Antisense inhibition of miR-21-5p slowed TGFβ-dependent migration of these cells in scratch assays, while a miR-21 mimic significantly enhanced HaCaT keratinocyte migration (Ahmed et al., 2011;X. Yang, Wang, et al., 2011). In mouse wounds, miR-21-5p appears as one of the most elevated miRNAs during the proliferative granulation formation stage, and antisense inhibition of miR-21-5p impaired reepithelialisation of murine skin wounds while a pre-miR-21 plasmid enhanced granulation tissue formation and wound contraction (T. Wang, Feng, et al., 2012). A miR-21 mimic was also recently shown to accelerate wound closure . Interestingly, downregulation of miR-21-5p in cutaneous diabetic mouse wounds was associated with delayed wound healing, although a causal relationship was not established (Madhyastha et al., 2012). Together, these rodent studies suggest abrogation of miR-21-5p activity impairs wound healing while enhancement of pre-miR-21 accelerates wound repair. On the other hand, topical application of a miR-21-5p mimic inhibited re-epithelialisation in ex vivo human skin wounds and reduced granulation tissue formation and re-epithelialisation in a rat model (Pastar et al., 2012). Thus, while a very recent study showed that keratinocyte-derived microvesicles carrying miR-21-5p po- Furthermore, unlike miR-31-5p and miR-132-3p (see below), miR-21-5p does not appear to stimulate keratinocyte proliferation, at least as measured in HaCaT cells (Ahmed et al., 2011;X. Yang, Wang, et al., 2011), although a recent study suggests otherwise .
Recently, Wang and colleagues also observed that miR-21-5p was reduced in aged mouse skin compared to their younger counterparts, and this was associated with impaired wound healing in the aged mice (Long et al., 2018). The delayed wound closure was reversed in aged miR-21 knock-in mice or by intradermal injection of a pre-miR-21 plasmid.
However, although miR-21 was depleted in aged mouse skin (12-month vs. 2-month-old mice) in the Wang study, Botchkarev and co-workers found miR-21-5p was elevated in aged mice (2-year vs. 8-week-old mice) (Ahmed et al., 2019). Hence, the dynamics of miR-21 expression in aged murine skin require further clarification, as do the mechanisms underpinning alterations in miR-21 levels in aging skin.
Some aspects of keratinocyte migration during wound healing, such as the reduction of cell-cell and cell-matrix adhesion overlap with features of epithelial mesenchymal transition (EMT) (Haensel & Dai, 2018). Conflicting observations have been reported in relation to the impact of miR-21-5p on EMT in HaCaT keratinocytes: work by Su and colleagues suggest modulation of miR-21-5p had limited impact on EMT whereas recent studies by Qian and co-workers found miR-21 mediates TGF-β1-dependent mesenchymal transition (J. Wang et al., 2016;. In any case, it will be important to establish the impact of miR-21-5p on EMT in primary keratinocytes and ex vivo human skin to verify the physiological significance of these observations.

| MicroRNA-132-3p in keratinocytes
Inflammation is an essential early stage of wound healing that supports the generation of a provisional extracellular matrix for subsequent phases and helps neutralise infectious agents and remove debris . However, prolonged inflammation is detrimental to wound healing, which is where the translational utility of miR-21-5p and miR-31- The other key finding from the gene ontology analysis was that genes whose expression was elevated in the pre-miR-132 transfected cells were enriched for processes associated with the cell cycle (D. . Indeed, pre-miR-132 enhanced keratinocyte proliferation by increasing the activation of EGFR and its downstream targets STAT3 and ERK. leading to impaired NF-κB activation Shaked et al., 2009) and epidermal keratinocytes have long been known to express AChE (Grando et al., 1993).
More importantly, depletion of miR-132 delayed wound closure in mouse skin while a miR-132-3p mimic promoted wound healing in leptin receptor-deficient diabetic (db/db) mice (D.  X. Li, D. Li, Wang, et al., 2017). Similarly, in human ex vivo skin wounds, inhibition of miR-132 abrogated re-epithelialisation, while a miR-132-3p mimic enhanced this process (D. X. Li, D. Li, Wang, et al., 2017). Because miR-132 was under-expressed in diabetic foot ulcers compared with wounded healthy skin, these findings together suggest elevation of miR-132 holds strong translational potential for treatment of chronic wounds.

| MicroRNA-19a/b and microRNA-20a in keratinocytes
Very recently, the Landén group has also identified a further set of miRNAs that dampen inflammation in wound healing. The six miRNAs encoded by the miR-17~92 cluster (miR-17, miR-18a, miR-19a, and miR-19b, miR-20a, miR-92) were depleted in chronic wounds compared to wounded healthy skin, with miR-19a, miR-19b and miR-20a specifically downregulated in the epidermis . Wound closure was delayed in mice with keratinocytespecific miR-17~92 conditional knockout mice, especially when diabetes was induced experimentally by streptozocin injection.
Conversely, wound repair was accelerated in diabetic mice with KC-specific conditional knock-in (cKI) of the miR-17~92 cluster or miR-19b alone (D. Li et al., 2020). Based on observations that

| MicroRNA-184 in keratinocytes
While miR-31-5p and miR-132-3p drive keratinocyte migration and proliferation, miR-184 is distinguished by its ability to stimulate keratinocyte migration and differentiation (Nagosa et al., 2017;Richardson et al., 2019Richardson et al., , 2020. Mature miR-184 arises from the 3p arm of the pre-miR-184 duplex and there is no evidence for a minor miRNA from the 5p arm (see http://www.mirbase.org/). Although early studies did not detect miR-184 in proliferating epidermal keratinocytes maintained in monolayer culture, we observed miR-184 expression in reconstituted human epidermis, which comprises proliferating keratinocytes and differentiating suprabasal cells (Roberts et al., 2013). This led us to suspect that miR-184 may play a role in keratinocyte differentiation. Consistent with this, we and others recently showed that elevation of extracellular Ca 2+ , a major inducer of keratinocyte differentiation, induces miR-184 in these cells (Nagosa et al., 2017;Richardson et al., 2020). In addition, we found that the induction of miR-184 required Ca 2+ entry through the storeoperated Ca 2+ entry (SOCE) channel ORAI1 though strict dependence on the SOCE trigger STIM1 (stromal interaction molecule 1) has not been verified. Elevation of miR-184 was associated with a reduction in keratinocyte proliferation and enhancement of keratinocyte differentiation through the cyclin E:DNA damage and NOTCH pathways (Nagosa et al., 2017;Richardson et al., 2020).
Given that migration through the suprabasal layers is inherent to epidermal differentiation, we examined the impact of miR-184 on keratinocyte migration (Richardson et al., 2019(Richardson et al., , 2020. High-density scratch wounding of keratinocyte monolayers led to a 50-fold induction of miR-184 after 5 days, suggesting miR-184 may function critically during the latter phases of re-epithelialisation. Exogenous miR-184 accelerated keratinocyte migration threefold, while a miR-184 inhibitor dampened keratinocyte migration threefold (Richardson et al., 2019(Richardson et al., , 2020. However, the targets of miR-184 required for epidermal keratinocyte migration have not been defined and the impact of miR-184 on cutaneous wound healing in vivo is not known. Hence, studies of re-epithelialisation in miR-184-deficient and miR-184 transgenic mice, such as those generated by Shalom-Feuerstein and colleagues (Nagosa et al., 2017), will be crucial to deepen our understanding of miR-184 function in reepithelialisation, along with studies in diabetic mouse models. Further studies are also required to establish whether miR-184 can promote tissue regeneration in normal human or diabetic wounds through its effects on keratinocyte differentiation and migration.
This is important because it is keratinocyte differentiation rather than proliferation that appears to be impaired at the edges of chronic ulcers (Stojadinovic et al., 2008;Usui et al., 2008;Wikramanayake et al., 2014). Given that terminal differentiation is the ultimate destiny of epidermal keratinocytes, the ability of miR-184 to mobilise differentiation pathways in conjunction with migration may underpin its translational potential.
1.8 | MicroRNA-129-5p and microRNA-335-5p in keratinocytes The diabetic wound is a highly proteolytic environment where sustained elevation of matrix metalloproteinases (MMPs) contributes to extracellular matrix degradation, dysregulated inflammation and impaired wound closure (Ayuk et al., 2016). Increasing evidence points to MMP-9 in particular as a key driver of non-healing wound pathology Gooyit et al., 2014;C. Yang et al., 2009).
Ren and colleagues found that the specificity protein 1 (Sp1) regulates MMP-9 expression in HaCaT keratinocytes, and that exposure to glycated albumin to reproduce the advanced glycation end product (AGE)-enriched microenvironment of diabetic wounds raised expression of both Sp1 and MMP-9 in HaCaT and primary keratinocytes (W. Wang et al., 2018). Profiling diabetic patient serum revealed 58 downregulated miRNAs, among which miR-129-5p and miR-335-5p stood out as predicted regulators of Sp1, which was confirmed by luciferase reporter assays and western blot analysis.
More importantly, both miR-129-5p and miR-335 were underexpressed in diabetic skin wounds from patients and rats, and decreased in HaCaT keratinocytes treated with glycated serum ROSS | 8055 albumin, establishing a link between miR-129-5p and miR-335-5p depletion and elevation of their target Sp1 (W. Wang et al., 2018).

| CONCLUSION
It is tempting to speculate that a concomitant or sequential combinatorial approach in which miRNA mimics targeting different aspects of wound repair may yield the best patient outcomes for miRNA- healthy and diabetic skin would also be of interest given recent advances in image processing tools (Driscoll et al., 2019;Wu et al., 2020). Finally, given the asymmetric distribution of mRNA and proteins in polarised migrating cells (Liao et al., 2015), it would be interesting to assess the subcellular localisation of miRNAs during keratinocyte migration.