A Fibroblast Is Not a Fibroblast Is Not a Fibroblast

Abstract

Fibrosis after injury is a huge public health concern, leading to morbidity, mortality, and expenditure of billions of health care dollars. Recent mouse studies have shown that dermal fibroblasts are heterogeneous. New research using single-cell RNA sequencing to identify major fibroblast populations in humans is paving the way to a better understanding of fibroblast heterogeneity and fibrosis.

Introduction

Inhibiting fibrosis is an important goal for scientists and physicians. When an organ is exposed to a destructive stimulus, fibrosis similar to scarring after cutaneous wound healing occurs. After spinal cord injury or myocardial infarction this results in dense fibrotic tissue, and the resulting morbidity and mortality is significant (Wynn and Ramalingam, 2012). In skin, the fibrotic response to injury is a scar characterized by excess collagen deposition and a lack of dermal appendages. Preventing imperfect repair has been a goal of regenerative medicine for several decades, but the holy grail of scarless wound healing has been elusive (Walmsley et al., 2015). Many treatments that minimize scarring and fibrosis have been proposed, but no pharmacologic therapies have been successful. Nonetheless, the annual market for these treatments is estimated to be $12 billion (Sen et al., 2009). Research by Tabib et al. (2018) and Philippeos et al. (2018) brings us closer to understanding skin fibrosis by identifying fibroblast populations in human skin.

Fibroblast heterogeneity in mice

For many years, fibroblasts have been considered to be a homogeneous cell type responsible for producing extracellular matrix (ECM). However, recent advances in developmental biology involving lineage tracing and transplantation assays have allowed researchers to identify subpopulations of fibroblasts. Work by Driskell et al. (2013) showed that dermal fibroblasts in mice arise from two distinct lineages: one that forms the upper dermis and another that forms the lower dermis. The lower dermis includes the reticular fibroblasts responsible for the bulk of ECM deposition. After wounding, the lower dermal lineage was shown to mediate repair, recruiting participation of the upper dermal lineage only during reepithelialization. (Driskell et al., 2013)

Building on this study, research by Rinkevich et al. (2015) identified a subpopulation of dermal fibroblasts in mice that is marked by the embryonic expression of En1) and is responsible for most ECM deposition in both acute and chronic progressive forms of fibroses. Using a double transgenic mouse model in which En1 lineage-positive fibroblasts (EPFs) were labeled with membrane-bound GFP, the authors showed that EPFs were responsible for collagen deposition during late embryonic development, cutaneous wound healing, radiation-induced fibrosis, and melanoma tumor stroma formation. Furthermore, flow cytometry showed dipeptidyl peptidase-4 (DPP4) to be a cell surface marker that enriches for EPFs. When the small molecule diprotin A, a selective allosteric inhibitor of DPP4 peptidase activity, was applied to healing wounds via hydrogel delivery, scarring was significantly decreased. (Rinkevich et al., 2015)

Taken together, these and other studies clearly show that fibroblasts result from distinct lineages, thought to be regulated by epidermal Wnt signaling (Driskell and Watt, 2015), and are functionally heterogeneous. Further research is necessary to convert this understanding into clinical applications.

Fibroblast populations in humans

Tabib and coworkers (2018) extend these research efforts from mice to humans by using single-cell RNA sequencing to identify fibroblast populations in human skin from six healthy control individuals. Single-cell RNA sequencing was performed on all cells obtained from enzymatically digested skin. This all-inclusive unbiased technique allowed the authors to detect potentially rare cell types (Tabib et al., 2018).

To analyze the data and identify fibroblast subpopulations, a smart local moving algorithm was used. Based on differential gene expression, 19 distinct clusters of cells were identified. Using this technique, the authors distinguished two major fibroblast populations characterized by expression of SFRP2/DPP4 and FMO1/LSP1, as well as five minor fibroblast populations featuring CRABP1, COL11A1, FMO2, PRG4, and C2ORF40. Known mouse dermal fibroblast markers did not stratify human dermal fibroblasts, suggesting evolutionary divergence. DPP4 is expressed by fibroblasts in both humans and mice, underlying its potential importance in fibrosis.

The fibrogenic potential of the SFRP2/DPP4 population of fibroblasts was not directly explored in this study, but the high expression of COL1A1 in this population suggests that these cells play a role in matrix deposition. Additional correlations were performed with markers of proliferation and sex, as well as gene ontology enrichment analysis. The SFRP2/DPP4 fibroblast population was characterized by (i) negative regulation of signaling pathways, (ii) regulation and sequestering of BMP, and (iii)protein localization to ECM. Recent reports have suggested a role for BMP as a negative regulator of cutaneous scarring (Plikus et al., 2017). Finally, the authors corroborated their single-cell RNA sequencing data using immunofluorescence and identified morphological differences between the two major fibroblast populations (Tabib et al., 2018).

Using a different and complementary approach, Philippeos et al. (2018) also explore human fibroblast heterogeneity in this issue. The authors combine spatial transcriptional profiling of mouse and human dermis with scRNAseq of primary human dermal fibroblasts to identify at least four distinct fibroblast subpopulations in adult human skin: (1) lin-CD90 +CD39+CD26-,(2)lin-CD90+CD36+,(3) lin-CD90+CD39+CD26+, and (4) linCD90+ CD39-RGS5-. (Philippeos et al., 2018)

To determine whether spatial differences in gene expression are conserved between mouse and human, human skin samples were enzymatically-treated to remove the epidermis, and papillary and reticular dermis were separated using a dissecting microscope. RNA sequencing and hierarchical clustering were performed on separate dermal layers revealing distinct gene expression profiles. In both mouse and humans, papillary dermal lineages were labeled by CD39.

Based on this analysis, human papillary (lin-CD90+CD39+) and lower reticular/hypodermal (lin-CD90+CD36 +) fibro-blasts were sorted and cultured. After a single passage, there was a loss of cell surface markers. However, each fibroblast subpopulation had distinct morphology and functionality, in terms of Wnt signaling, IFN responsiveness, and ability to repopulate human de-epidermized dermis (DED).

Finally, by performing scRNA-seq on 184 lin-CD90þ/- cells derived from the abdominal skin of a single donor and automated clustering, five groups of cells were characterized, suggesting at least four fibroblast subpopulations. These data corroborate findings by Tabib et al. (2018) and also suggest the importance of CD26 (DPP4).

DPP4 and fibrosis

As previously discussed, DPP4 is a marker for pro-fibrotic fibroblasts in mice, enriching for EPFs (Rinkevich et al., 2015). Studies of hypertrophic scars and keloids suggest the importance of DPP4, and reports of anti-fibrotic effects of DPP4 inhibitors have begun to surface in the literature (Hu and Longaker, 2016). An early study by Thielitz et al. (2008) suggests that inhibition of DPP4 may be used to treat fibrotic skin disorders, such as keloids. Now studies by Tabib et al. (2018) and Philippeos et al. (2018) demonstrate DPP4 as a marker used to differentiate certain populations of human fibroblasts. Although the data thus far are intriguing, further research is necessary to fully understand the role of DPP4 in dermal fibroblasts.

Future directions

Recent literature highlights the difficulty of understanding fibrosis. A report by Plikus et al. showed that adipocytes regenerate from myofibroblasts during wound repair in mice (Plikus et al., 2017). Previously, reprogramming of one cell lineage into another for regeneration had not been observed in mammals. Although the Plikus et al. data suggest further complexity underlying cutaneous scarring, the data also provide another means to reduce fibrosis.

A better understanding of fibroblast heterogeneity in humans and delineation of the role of each fibroblast subpopulation can have a tremendous impact on fibrosis involving multiple organ systems. In addition to the common local and systemic fibrotic pathways seen in multiple human diseases, the role of fibroblasts in cancer tumor stroma has also been recently investigated. A greater understanding of this process lends promise to cancer therapies as well.

Conclusion

Taken together, these studies shed new light into fibroblast heterogeneity. The ultimate goal remains to identify, isolate, and deplete the fibroblast subpopulation that represents the cellular culprit for fibrosis in humans. Expansion of other fibroblast subpopulations may provide therapeutic applications as well. Such work promises to decrease fibroses across organs systems, facilitating treatment of myriad diseases and reduced mortality as well.

Clinical Implications.

• Recent studies identified fibroblast heterogeneity.

• New research provides a deeper understanding of fibroblast heterogeneity in humans.

• Further investigation promises novel therapies to inhibit fibrosis