Macrophage-derived GPNMB accelerates skin healing

Abstract

Healing is a vital response important for the re-establishment of the skin integrity following injury. Delayed or aberrant dermal wound healing leads to morbidity in patients.

The development of therapies to improve dermal healing would be useful. Currently, the design of efficient treatments is stalled by the lack of detailed knowledge about the cellular and molecular mechanisms involved in wound healing. Recently, by using state-of-art technologies, it was revealed that macrophages signal via GPNMB to mesenchymal stem cells, accelerating skin healing. Strikingly, transplantation of macrophages expressing GPNMB improves skin healing in GPNMB-mutant mice. Additionally, topical treatment with recombinant GPNMB restored mesenchymal stem cells recruitment and accelerated wound closure in the diabetic skin. From a drug development perspective, this GPNMB is a new candidate for skin healing.

Skin is the largest organ in the adult mammalian organism. It covers the body’s entire external surface, separating the organism from the environment. The skin is essential for regulating temperature and dehydration, preventing infections, and providing a mechanical barrier against injury, irradiations, chemicals, and pathogens. It continually renews its outermost cell layer during the whole life, presenting remarkable healing capacity (1). Nevertheless, in certain conditions, such as major traumatic injuries or genetic skin diseases, the skin becomes unable to recover its original functionality, resulting in open wounds or scar formation, leading to increased susceptibility to infection and impaired sensitivity, respectively (2). Encouraging skin healing that preserves function will have positive effects on patients’ lives. The development of efficient treatments may benefit greatly from the emerging understanding of the cellular and molecular processes involved in skin healing.

The skin is composed of three main layers: epidermis, dermis and hypodermis (3). These layers consist of diverse immune and non-immune cell populations. The epidermis consists of a stratified squamous epithelium of keratinocytes delimited by a basal membrane, containing melanocytes, Langerhans cells, and Merkel cells (4, 5). The dermis contains resident immune cells, fibroblasts, sweat glands, sebaceous glands, blood and lymphatic vessels, nerve endings, and mesenchymal stem cells (6). The hair follicles, present in the epidermis, participate in sebaceous secretions, hair growth, and serve as a reservoir of epidermal stem cells (7). The hypodermis (subcutaneous tissue) is composed of adipocytes, macrophages, fibroblasts, vasculature, and nerves (8). It anchors the epidermal and dermal layers, and it is essential for epidermal homeostasis (9). Although in the last few years, numerous studies have improved our understanding of the cellular complexity in the skin microenvironment (10), the biological processes underlying wound healing in the skin are not fully understood. After skin damage, complex intra- and inter-cellular modifications are triggered to recover dermal integrity (11). Despite the great progress in our knowledge of dermal wound healing in recent years, the exact cells and the underlying cellular and molecular processes that directly contribute to dermal wound healing remain unknown. The lack of a detailed knowledge about the cellular contributors and the molecular mechanisms mediating skin wound restoration restricts the design of effective therapies that would improve healing.

Even though it is well established that macrophages and mesenchymal stem cells are critical in skin wound healing (12, 13), the relationship between these cell types, and how they interact remain unknown. Now, in a recent article in Journal of Investigative Dermatology, Yu and colleagues reveal that a macrophage derived signal mobilizes endogenous mesenchymal stem cells to the wound, improving healing in the skin (14) (Figure 1). The authors showed that Glycoprotein Nonmetastatic Melanoma Protein B (GPNMB) expression increases in macrophages after skin wounding in C57BL/6 mice. Yu and colleagues investigated the role of GPNMB in wound healing by using state-of-the-art techniques, including GPNMB-mutant and diabetic transgenic mouse models, macrophages transplantation, and in vivo treatment with recombinant GPNMB of a mouse model of cutaneous skin healing. These experiments revealed, by using GPNMB-mutant (DBA/2J) inbred mice in which GPNMB expression is defective, that reduced GPNMB correlates with decreased recruitment of mesenchymal stem cells to the wound (14). By transplantation of macrophages from wild-type mice into the injured skin of GPNMB-mutant mice, Yu and colleagues demonstrated that macrophage-derived GPNMB promotes mobilization of mesenchymal stem cells to the wound. Strikingly, the authors reported that topically applying recombinant GPNMB improves skin healing in wild-type and GPNMB-mutant mice (14). Additionally, Yu and colleagues showed that, in the diabetic skin microenvironment after injury, the GPNMB expression in macrophages is impaired. Importantly, topical treatment with recombinant GPNMB restored mesenchymal stem cells recruitment, and accelerated wound closure in the diabetic skin (14). Thus, this study brings a novel possible target for skin healing, urgently needed in the clinic.

Figure 1. Macrophage-derived GPNMB role in dermal wound healing.

Skin healing is a complex process involving the participation of multiple cell players.

The study of Yu and colleagues now reveals that macrophages signal via GPNMB to mesenchymal stem cells (MSCs) promoting wound healing in the skin (14).

Transplantation of macrophages expressing GPNMB into the injured skin of GPNMB-mutant mice accelerates skin restoration. With the appearance of state-of-art technologies, future studies will reveal in detail the cellular and molecular components in skin healing microenvironment.

Here, we discuss the findings from this study, and evaluate recent advances in our understanding of the skin microenvironment after injury.

PERSPECTIVES/FUTURE DIRECTIONS

The use of typical knockout mice models has proved useful in understanding the role of key molecules in physiological and pathological conditions. Nonetheless, these technologies have their limitations. They produce broad changes in gene function throughout the body, affecting several organs and cellular types. Thus, they are limited in a sense that they do little to identify the specific roles of a gene in specific cell types. Additionally, the animal becomes knockout since its zygote, stage; thus, important changes may happen due to an alteration in a specific gene during development, and may be overseen in adult animals. The main findings from Yu and colleagues’ study are based on the data obtained from DBA/2J mice (14). The inbred mouse strain DBA/2J has several important genes mutated besides Gpnmb, for instance Cdh23 and Fscn2. Additionally, these mice are used as a model of several disorders, including glaucoma, deafness, host response to infections, schizophrenia, and cardiovascular disease (15–21). For this reason, it is possible that other genes modified in those mice play important roles during skin healing, and may influence the role of GPNMB. Moreover, as the molecular functions of genes may depend on a specific cell type in which they are expressed, restricting gene manipulation to specific cells in the skin at a specific time-point may be very useful to understand the functions of a protein. Thus, conditional gene manipulation methodologies offer a powerful tool. It is known that Gpnmb gene is expressed by macrophages in different pathophysiologic conditions (22–25). Note, however, that the expression of Gpnmb is not restricted to macrophages only. GPNMB is produced by various cell types in addition to macrophages, including dendrytic cells, T cells, melanocytes, fibroblasts, keratinocytes, and others (26–30). Consistent with this, in DBA/2J mice in which GPNMB is mutated in all cells, the changes observed during skin healing could be due to GPNMB derived from other dermal cell populations that can contribute to GPNMB production after skin injury. Yu et al. (2017) now proposed that macrophages accelerate wound healing through the expression of GPNMB. However, GPNMB has not been conditionally deleted from dermal macrophages, so there is no direct evidence that macrophages are the only/main functionally important source of these signal. These issues may be addressed by analyzing the effect of genetic ablation of macrophages on skin healing after lesion by using CD11b-DTR or CD169-DTR mice (31, 32). Nevertheless, the effect of deleting macrophages in the skin may affect other aspects of macrophages’ behavior, and need to be taken into consideration when analyzing those mice. Additionally, as in those mice macrophages would be deleted not only form the skin, a way to delete specifically dermal macrophages should be deleted. One possible solution is to administer the diphtheria toxin exclusively dermally to those mice. Also, the generation of GPNMB-floxed mice to be crossed with macrophage-specific Cre drivers, such as CD169-Cre mice (33), will allow us to specifically delete GPNMB in macrophages. However, this model is not yet available commercially, and those mice should be created in the future. In addition to studies using genetic mouse models, transcriptomic and single cell analysis represent fundamental tools that will help us understand the roles of macrophages within the skin healing microenvironment. These studies may be done by isolating via flow cytometry macrophages’ subpopulations from the skin before, and, at different time points, after lesion both from mice and humans. The benefits from these studies include understanding how exactly single macrophages respond to dermal injury.

Hair follicles are complex anatomical structures that harbor their own resident cells, released to the inter-follicular area after lesion, capable of influencing skin healing (34). As hair-bearing areas heal quicker than areas lacking hair follicles (35), it has been proposed that optimal skin healing depends on cells contained in the hair follicles. The relationship between cells within and outside the hair follicle is currently being unraveled (36). Nonetheless, further work is needed to explore the role of cells resident in the hair follicle in wound healing. Recent studies have shown that resident macrophages are important for the activation of hair follicle stem cells (37, 38). It remains poorly understood what is the role of macrophages resident in the hair follicle in wound healing.

Stem cells have the capacity to self-renew, or differentiate into distinct cell populations (39–51). Mesenchymal stem cells have been identified in a variety of tissues, including the skin, where, by the formation of new cells, they guarantee the tissue repair after injuries. Yu and colleagues suggest that deficiency of GPNMB attenuates the recruitment of mesenchymal stem cells after skin injury (14). Their identification of mesenchymal stem cells is based on the International Society for Cellular Therapy (ISCT) criteria: specific molecular markers and multipotent differentiation potential in vitro (52). Nevertheless, these criteria are already considered minimalist, as they presuppose that practically every non-clonal culture of cells from any tissue could be classified as a stem cell under the right culture conditions (53). Thus, currently, to be identified as a stem cell the cell needs to have self-renewal capacity at a single clone level, and to present endogenous differentiation capacity in vivo. Based on those initial criteria (52), it was suggested that mesenchymal stem cells have a common origin. Due to the broad organ distribution of mesenchymal stem cells, the attention had turned to pericytes as candidates to be the mesenchymal stem cells (54–74). Mesenchymal stem cells/pericytes have been shown to be heterogeneous in the skin (54, 58), and differ in their functions as well (75). Thus, future studies should reveal whether a specific mesenchymal stem cell subpopulation is responsible for wound healing. Also it remains unknown how GPNMB affects mesenchymal stem cells. Is this effect direct or indirect? Are there receptors for GPNMB on mesenchymal stem cells? Can these receptors be activated by other molecules? Are they absent in diseases in which wound healing is ineffective? The molecular mechanisms of the action of GPNMB on the skin wound should be explored further in more detail in future studies, by analyzing gene expression of different cell subtypes, including mesenchymal stem cells, after application of GPNMB.

Macrophages are heterogeneous; two main interconvertible subtypes were termed M1 and M2, based on molecular markers, and their functions (76). In the early stage of wound repair M1 macrophages predominate, while M2 macrophages are more abundant in the later stages (77). M1 macrophages exhibit a pro-inflammatory phenotype, aiding in the removal of damaged tissues, while M2 macrophages produce anti-inflammatory cytokines, and promote wound healing (78). Genetic ablation of macrophages, by using CD11b-DTR mice, during wound repair, impairs skin healing, and is associated with reduction of M2-related cytokines (79), indicating the importance to elucidate the cellular and molecular mechanisms involved in M2 polarization during wound healing. Yu and colleagues suggest that GPNMB activate mesenchymal stem cells infiltrating the wound to promote a shift toward anti-inflammatory M2 macrophages (14). However, these experiments have been done by in vitro cultures, and it cannot be discounted that modification of cell properties by their artificial preparation in vitro may influence their behavior. Thus, whether mesenchymal stem cells act accordingly in physiological conditions in vivo, remains unclear.

Non-healing wounds affect more than 5 million people in the United States (80). Impaired healing can result from several diseases, anaemia, skin atrophy, deformity, neuropathy, microvascular disease, local factors, or the toxic effects of drugs used in treatment. Yu and colleagues demonstrated that recombinant GPNMB improved wound healing in the skin of diabetic mice (14). It will be interesting to explore whether GPNMB roles are restricted to certain disorders or are applicable to the healing of all type of wounds in the skin. Moreover, wound restoration is important in multiple tissues, and the healing process varies among organs. As GPNMB expression is not restricted solely to the skin, is this protein important for healing in other sites as well? Also, as undesirable scar is formed during the healing process (60), the effect of GPNMB on the scar formation should be explored in future works. Scar tissue is composed of extracellular matrix proteins, and connective tissue matrix-producing cells. The molecular processes that drive scar tissue formation in the skin remain poorly understood. Mesenchymal stem cells were proposed to be able to differentiate into extracellular matrix-forming cells (55, 60, 63, 64, 81). As GPNMB may be involved in the recruitment of mesenchymal stem cells, it is possible that in the absence of GPNMB scar formation is affected. Analyses of transgenic mice models with GPNMB deleted in specific cells from the skin microenvironment will address this issue.

The same way that GPNMB’s regulation of macrophages’ function is important for skin wound healing, GPNMB could also regulate other inflammatory processes. For instance, GPNMB has been shown to function as a feedback regulator of inflammation after induction of macrophages by lipopolysaccharide and IFN-γ (24). Future research will examine whether the function of GPNMB is dysregulated in inflammatory diseases.

The skin microenvironment is composed of a mixture of cells that cooperate to perform the necessary roles for the skin functioning of that organ. The interplay between different cellular components of this microenvironment will define the tissue outcomes in distinct pathophysiological circumstances (82, 83). Macrophages and mesenchymal stem cells reside in the specialized skin microenvironment composed of several other cell types, which also may cross-talk with macrophages and mesenchymal stem cells, regulating behavior and inducing their activated state (84–92). Yu and colleagues reveal the cross-talk that happens between macrophages and mesenchymal stem cell during wound healing in the skin (14). However, it remains to be explored how other dermal cells interfere in this communication. Depletion of other cells, as well as specific factor in those cells, in the skin may address their role in the cross-talk between macrophages and mesenchymal stem cells. The skin cellular microenvironment is also shaped by soluble molecules, structural proteins, and the extracellular matrix, providing physical support, regulating adhesion and signaling. Defining and understanding the cellular and molecular mechanisms that influence macrophage and mesenchymal stem cell functions is crucial to determine the roles of these cells in skin healing. Further insights into the cellular and molecular processes involved in wound healing will have important implications for our understanding of skin homeostasis and disease.

Wound healing animal models aim to recreate features of human healing after lesion as closely as possible. As skin wound healing in the mouse may differ from the one in humans, researchers confront challenges in connecting the gap between pre-clinical and clinical works (93). Taking into account the peculiarity of each specie is essential to correctly interpret the data. Despite mouse and human skins having the same layers of cells, they differ in thickness and number of cells. Murine skin is thin (<25μm) and loose, while human skin is thicker (>100μm) and adherent to the underlying tissues (94, 95). As skin thickness affects the biomechanics of healing, this should be considered when evaluating murine studies. Moreover, as mouse epidermis is composed of 3 cell layers, while the human contains 10 (96), percutaneous absorption in humans should be less efficient (97). As Yu and colleagues showed that topical treatment of recombinant GPNMB in murine injured skin accelerates wound closure (14), we should be careful when translating these results into humans, since percutaneous absorption of topical GPNMB may be not as efficient.

Human studies are necessary to validate promising discoveries from animal models. The use of patients have several limitations, due to logistical problems, patient variability with regards to the extent and duration of the lesion, and ethical issues. To overcome these problems, several alternatives have been employed: hypertrophic scar model, in which human skin is grafted in a nude mouse (98), limited by the compromised immune component of wound healing; in vitro monolayer cell cultures (99), limited as the skin is formed by more than just fibroblasts; co-cultures of distinct dermal cell types using trans-well systems (100), limited by the absence of the 3D macroscopic skin structure. As any technique, each method has disadvantages and advantages. The combined use of different tools will reduce the limitations of each model. As the scar is formed by a complex cascade of cellular interactions, unfortunately most of culture methods lack physiological relevance (101). Additionally, mechanical loading should be also considered in an ideal model. The development of methods to culture ex vivo human skin biopsies embedded in specially engineered 3D surfaces is promising (102). The findings from these study could be tested in a 3D ex vivo human skin model by the use of specific antibodies blocking GPNMB, as well as by the application of recombinant GPNMB in this model.

In conclusion, our understanding of the cross-talk between different constituents of skin microenvironment after injury still remains limited, and the complexity of these interactions during wound healing should be elucidated in future works. A great challenge for the future will be to translate the research from experimental models into humans. Improving the availability of human tissue samples will be essential to reach this aim.