Regulatory roles for long noncoding RNAs and circular RNAs in human skin wound repair

Abstract

Human skin wound repair is a multiphase process, which includes inflammation, growth and remodeling, and requires the concerted communication of many cell types. During the course of healing, skin cells find themselves in a hostile environment where processes like migration, differentiation, apoptosis and proliferation must be correctly executed in order to restore the integrity of the injured tissue. These processes are tightly regulated by the interplay between signaling pathways and noncoding transcripts, which are essential for skin wound repair. When the healing phases do not progress in a timely and orderly manner, this results in healing defects like chronic skin ulcers or pathological scars. Due to the complex pathophysiology, these wound impairments are not fully understood and suffer from a lack of effective treatments.

This doctoral thesis aims to investigate the regulatory roles of different classes of long noncoding RNAs (linear – lncRNAs. and circular isoforms – circRNAs) during normal skin wound healing and in chronic ulcers. For this, we recruited healthy volunteers to donate skin and wound biopsies that were collected at different healing stages, and chronic nonhealing ulcers were obtained from patients with diabetic foot ulcers or venous ulcers.

In Paper I, we show that WAKMAR2 – a previously uncharacterized lncRNA – has a dynamic expression and function during wound repair. WAKMAR2 is downregulated during the early stages of healing to stimulate the production of inflammatory chemokines needed for the inflammatory phase, and later its levels increase to promote keratinocyte migration, essential for the proliferative phase. In a human ex vivo wound model, WAKMAR2 depletion delayed wound re-epithelialization, reinforcing its role in keratinocyte migration. We demonstrate that WAKMAR2 expression is directly affected by TGF-b signaling and its function is mediated by NF-kB signaling in keratinocytes. Because we find WAKMAR2 expression to be lower in keratinocytes at the wound edge of chronic wounds compared to normal wounds, we highlight its potential role in this pathology.

In Paper II, we switch the focus to an underexplored class of noncoding RNAs known as circRNAs and, by using microarray profiling, we demonstrate that they are dysregulated in chronic diabetic foot ulcers compared to normal wounds. Our findings support a functional role for hsa_circ_0084443, a circRNA that was upregulated in diabetic wounds, that is related to keratinocyte motility and growth. We show that hsa_circ_0084443 has a longer half-life than linear mRNAs and is resistant to degradation by exoribonuclease R, which strengthens its therapeutic potential. We also uncover a molecular gene network that is regulated by hsa_circ_0084443 and may explain its functions.

Paper III provides a robust catalogue of circRNA expression in human skin wounds at early healing stages and in nonhealing venous ulcers. By using RNA-seq on ribosomal RNAdepleted total RNA, we detected circRNAs and mRNAs in the same set of clinical samples. We show that circRNAs have dynamic expression patterns during wound healing and an aberrant signature in venous ulcers, that is largely independent from their parental mRNAs. Based on circRNA and mRNA co-expression network analysis, we infer putative functions of circRNAs in human wounds. We present this data at a browsable web resource www.xulandenlab.com/humanwounds-circrna that can facilitate further studies. By further integrating the microRNA deregulation data, we also reveal potential circRNA-miRNAmRNA regulatory networks, pinpointing the circRNAs that may act as microRNA sponges in human wounds. We validate the expression of circRNA candidates with experimental approaches and show the functional significance of two circRNAs upregulated in VU.

In Paper IV, we characterize the role of a circRNA – circGLIS3, which was identified in paper III. CircGLIS3 is transiently upregulated in the dermis during the early phases of wound healing and it shows no expression change in the epidermis. We show that circGLIS3 is important for wound fibroblasts to increase their responsiveness to TGF-b, enhance the production of extracellular matrix proteins (collagen type I and fibronectin I) and to promote their differentiation into myofibroblasts – requirements for granulation tissue formation during the early phases of repair. We demonstrate that circGLIS3 is regulated in an IL-1a-dependent manner in human wounds. We unravel a mechanism of action through which circGLIS3 interacts with PCPE-1 protein in the cytosol of fibroblasts to increase PCPE-1 stability and to cooperatively enhance the TGF-b-induced fibroblast activation. Lastly, we show that circGLIS3 knockdown in human ex vivo wound dermis reduced re-epithelialization and contracture of the wounds, reinforcing its beneficial role for wound closure.

Taken together, the findings presented in this thesis shed light on the regulatory roles of long noncoding RNAs and circular RNAs in skin wound repair and provide a path forward for the development of novel treatments where wound healing is impaired.