Boosting UPR transcriptional activator XBP1 accelerates acute wound healing

Abstract Patients’ suffering from large or deep wounds caused by traumatic and/or thermal injuries have significantly lower chances of recapitulating lost skin function through natural healing. We tested whether enhanced unfolded protein response (UPR) by expression of a UPR transcriptional activator, X-box-binding protein 1 (XBP1) can significantly promote wound repair through stimulating growth factor production and promoting angiogenesis. In mouse models of a second-degree thermal wound, a full-thickness traumatic wound, and a full-thickness diabetic wound, the topical gene transfer of the activated form of XBP1 (spliced XBP1, XBP1s) can significantly enhance re-epithelialization and increase angiogenesis, leading to rapid, nearly complete wound closure with intact regenerated epidermis and dermis. Overexpression of XBP1s stimulated the transcription of growth factors in fibroblasts critical to proliferation and remodeling during wound repair, including platelet-derived growth factor BB, basic fibroblast growth factor, and transforming growth factor beta 3. Meanwhile, the overexpression of XBP1s boosted the migration and tube formation of dermal microvascular endothelial cells in vitro. Our functional and mechanistic investigations of XBP1-mediated regulation of wound healing processes provide novel insights into the previously undermined physiological role of the UPR in skin injuries. The finding opens an avenue to developing potential XBP1-based therapeutic strategies in clinical wound care protocols.


Introduction
Skin wound repair needs an orchestra of multiple molecular and cellular components in three sequential but overlapping phases, inflammation, proliferation, and remodeling (1). In adults, skin wounds caused by traumatic and/or thermal injuries, particularly large or deep injuries, often fail to recapitulate skin structure and heal by scar, many of which can cause cosmetic or functional defects. When skin wounds occur in patients with diabetes or atherosclerosis, these wounds can easily progress to nonhealing ulcers despite standard wound care protocols (2). Therefore, there is an unmet need for effective regimens conducive to a favorable microenvironment for wound healing.
The high demand for protein folding and clearance of malfunctional proteins and the increase in new protein synthesis in wound repair and regeneration increases the burden on the endoplasmic reticulum (ER), disrupting ER homeostasis and causing ER stress (3). The ER stress response, or unfolded protein response (UPR), classically encompasses the activation of three separate ER transmembrane stress sensors, with the inositol-requiring enzyme 1α (IRE1α)/X-box-binding protein 1 (XBP1) pathway being the most conserved UPR pathway among species (4). Once activated, IRE1α cleaves a 26-base pair segment from XBP1 transcripts and produces the active, spliced isoform of XBP1 (XBP1s), functioning as a transcriptional factor to induce multiple genes involved in the restoration of ER homeostasis (4). XBP1 is known to exert substantial effects in augmenting ER protein folding capacity and remodeling the secretory pathway (4,5). In addition, XBP1 might carry out distinct roles in vascular development in the fetal stage (6). However, the potential roles of XBP1 in regulating the adult tissue repair process are unclear to date.
In this study, we evaluated the efficacy of XBP1 gene transfer on burn and diabetic wounds. Downstream pathways in fibroblasts and endothelial precursor cells have been explored to dissect the XBP1-induced skin repair and regeneration capacity. Our study has provided novel discoveries implicating a potentially high-effective therapeutic approach for acute burn or traumatic injury-associated wound healing through enhanced XBP1 expression.

XBP1 improves burn wound healing in vivo
In a mouse model of the deep partial-thickness burn wound, we infected the wounds with the adenovirus carrying activated/ spliced XBP1 (Ad-XBP1s) sequence 2 days after wounding, using adenovirus carrying egfp (Ad-GFP) as a control. The elevated mRNA levels of activated/spliced XBP1 (XBP1s) in wound tissue were confirmed by real-time PCR at the end of the experiment (Fig. 1A). Our results showed that Ad-XBP1s healed the wounds significantly faster than Ad-GFP ( Fig. 1B and C), starting from day 14. At the end, wounds receiving Ad-XBP1s were considered clinically healed, while wounds receiving Ad-GFP were covered by large scabs. The keratin 14 staining suggested that XBP1s-expressing wounds processed a continuous, thick, and slightly keratinized epidermis (Fig. 1D), and a decent depth of dermis (Fig. 1E). In contrast, GFP-expressing wounds had no or partially formed epidermis and thin dermis with extensive inflammatory cells ( Fig. 1D and E). The Masson's trichrome staining suggested that the collagen fibers were moderately increased (P = 0.0571) and better organized and corrugated structured at the wound edge of XBP1s-expressing wounds, although they were comparable in the wound center area ( Fig. 1F and G). These data indicate that XPB1s is potent in promoting the regeneration of functional dermal and epidermal layers for burn-injured skin and refining the inflammation.

XBP1 induces growth factor and cytokine expressions
Next, we sought to investigate whether XBP1 boosts growth factors. We overexpressed XBP1s or GFP control in human embryonic kidney cells (consisting of fibroblasts, endothelial cells, and epithelial cells). Real-time PCRs showed that mRNAs encoding the growth factors ( Fig. 2A), including platelet-derived growth factor (PDGF) BB, basic fibroblast growth factor (FGF2), and transforming growth factor beta 3 (TGFβ3), were significantly elevated in XBP1s-expressing cells. The protein levels of these three growth factors were increased by XBP1s gene transfer as measured by Western blot (Fig. 2B). Furthermore, immunofluorescent staining of PDGF BB (Fig. 2C), FGF2 (Fig. 2D), and TGFβ3 (Fig. 2E) in the wound sections suggested that XBP1s gene transfer increased these growth factors in wound healing in vivo. As shown in Fig. 1F, tissue remodeling factors such as collagen type III (COL III) were moderately increased (P = 0.0689), while collagen type I (COL I) moderately decreased (P = 0.0571).

XBP1 accelerates wound healing in healthy and diabetic animals in vivo
In traumatic wounds (full-thickness excisional) in either type 2 diabetic (db/db) or normal (db/+) mice, Ad-XBP1s was applied to in the wound bed immediately after wounding using Ad-GFP as control. In healthy mice (db/+), Ad-XBP1s accelerated wound healing starting from day 8 and reached full closure on day 10 when the control wounds were healed by <80% (Fig. 3A and B). More importantly, in type 2 diabetic mice (db/db), the Ad-XBP1s-treated wounds demonstrated significantly faster closure from day 2 ( Fig. 3C and D). In the end, Ad-XBP1s-treated wounds were ∼80% closed, while control wounds were healed only by ∼50%. These results indicated that enhancing XBP1 accelerates tissue repair in both healthy and diabetic animals with more remarkable efficacy in diabetic animals. Keratin 14 staining suggested that XBP1s gene transfer augmented wound re-epithelialization in healthy and diabetic animals (Fig. 3E). The epidermal layers and dermal layers were thicker in XBP1s group (Fig. 3F and G). The collagen deposition was enhanced by XBP1s in wound dermal layers in healthy animals and a similar trend in diabetic animals ( Fig. 3H and I).
In both normal and diabetic wounds, XBP1s also stimulated capillary formation at the wound edge where the most robust angiogenesis occurred, as stained by CD31 (Fig. 3J and K). Meanwhile, we also found increased levels of PDGF BB (Fig. 3L), FGF2 (Fig. 3M), and TGFβ3 (Fig. 3N) in wound beds by immunofluorescent staining.

XBP1 promotes angiogenesis in vitro
Human dermal microvascular endothelial cells (HDMVECs) were transfected with Ad-XBP1s or Ad-GFP. The functional assays showed that while their proliferation stayed comparable (Fig. 4A), HDMVECs transfected with Ad-XBP1s displayed significantly higher migration (Fig. 4B) and tube formation activity (Fig. 4C) compared with their controls, suggesting the permissive role of XBP1in angiogenesis.

Discussion
In this study, we have obtained strong evidence that exogenous expression of the UPR transcriptional factor XBP1 boosts angiogenesis and epithelialization, leading to rapid, nearly complete wound healing in burn wounds in healthy animals or traumatic wounds in diabetic animals. Potential mechanisms involved include cytokine/growth factor expression and the augmentation of angiogenesis.
PDGF BB, FGF2, and TGFβ3 were significantly upregulated in XBP1s-expressing cells and wounds (Figs. 2C-E and 3L-N), which supports XBP1's action in modulating the proliferative phase. Angiogenesis and growth factor production is essential for diabetic wound healing. PDGF is currently the only growth factor treatment for diabetic wounds approved by FDA (7). FGF2 has been reported to contribute critically to the modulation of scar formation (8). TGFβ family members promote re-epithelialization and angiogenesis (9), with TGFβ3 uniquely inhibiting myofibroblast formation and scarring in adult wounds (10). These growth factors were chosen because of their significant roles in the skin healing process, but it is worth investigating many more factors under the control of XBP1. For example, previous reports have shown mutually positive interactions between VEGF and XPB1 splicing in various settings (6,11), serving as a potential mechanism for improved angiogenesis in diabetic wounds. In fact, we observed an overall moderately increased response of inflammatory cytokine expression upon XBP1 expression (Supplemental Fig. S1). Among them, the increase of anti-inflammatory, pro-angiogenic IL-8 (12) reached statistical significance, and there was a trend that pro-inflammatory IL-1β was also increased (P = 0.165). While inflammatory cytokines are required for the wound-healing process, excessive levels of the pro-inflammatory cytokines might harm the angiogenesis or wound-healing process (13). However, our in vivo data have shown that XBP1s facilitates the healing process and promotes wound angiogenesis. These observations support that XBP1s stimulates the production of select antiinflammatory cytokines while not acting as a negative factor in angiogenesis.
The healing outcome of burn wounds largely relies on scar formation, matrix deposition, and reorganization, which are critical events in wound proliferation and remodeling. Our histological analyses suggested that XBP1s gene transfer increased collagen fibers in dermal layers of the wounds (Fig. 1F and G), indicating enhanced skin strength, therefore, potentially improved wound quality. However, there was a concern that prolonged or extensive XBP1 activation might increase hypertrophic scar formation because a previous study found that the inhibition of XBP1 upstream activator IRE1α decreased scar formation in a mouse excisional wound model (14). The collagen turnover in dermal layers can occur at a rapid pace. While Col I is always the primary collagen type, in a healing wound, Col III is the first to be synthesized in the early stage and is replaced by Col I (15). Collagen in post-burn hypertrophic scarring consists primarily of Col I and less Col III, compared with the uninjured dermis (15). Our in vitro data suggested that XBP1 activation induced a moderate decrease in Col I but an increase in Col III (Fig. 2F), suggesting that Col III might be the dominant type in the XBP1s-treated wounds. In addition, persistent activation of XBP1 may trigger carcinogenic responses due to boosting growth factors. While it is beyond the scope of the current study to examine the long-term safety concerns of XBP1 gene transfer, the abovementioned observations have constituted a valid concern needed to be addressed in the future.
In conclusion, our study uncovered a previously undetermined role of the key UPR activator, XBP1, in promoting reepithelialization and facilitating tissue repair and regeneration in acute burn, traumatic, and diabetic wounds. Applying the XBP1-based treatment protocol might have a high potential to be translated into the clinical arena.

Animals
Male C57BL/6 mice at the age of 12 weeks, type 2 diabetic mice (BKS.db, age of 12 weeks), and their age-and gender-matched nondiabetic healthy littermates (db/+) were purchased from Jackson Laboratory (Bar Harbor, ME, USA). All animal procedures were performed according to Wayne State University Institutional Animal Care and Use Committee (IACUC) guidelines.

Gene therapy for burn wound healing in vivo
Wounds of a second-degree deep contact thermal injury were created to dorsal skin by placing a 1.2-cm diameter stainless steel rod (Fisher Scientific) rod heated to 95°C and applied to the prepped area for 10 s, causing a wound of ∼3.0 cm 2 (16). On day 2, a square of 1.5 by 1.5 cm per area was injected with 2 × 10 8 particle forming units (PFUs) of either Ad-XBP1s (kindly provided by Dr Umut Ozcan from Harvard University) or Ad-GFP (Purchased from Vector Biolabs) solved in 100 μL PBS (17). The XBP1s gene transfer was expected to be effective on day 4 when the wounds were at the peak of denatured protein cleavage and the beginning of new tissue generation. Wound closure rates were calculated as Percentage Closed (y%) = [(Area on Day 0 -Open Area on Day x )/ Area on Day 0 ] × 100.