Lung-resident mesenchymal stromal cells are tissue-specific regulators of lung homeostasis

Abstract

Until recently, data supporting the tissue-resident status of mesenchymal stromal cells (MSC) has been ambiguous since their discovery in the 1950–60s. These progenitor cells were first discovered as bone marrow-derived adult multipotent cells and believed to migrate to sites of injury, opposing the notion that they are residents of all tissue types. In recent years, however, it has been demonstrated that MSC can be found in all tissues and MSC from different tissues represent distinct populations with differential protein expression unique to each tissue type. Importantly, these cells are efficient mediators of tissue repair, regeneration, and prove to be targets for therapeutics, demonstrated by clinical trials (phase 1–4) for MSC-derived therapies for diseases like graft-versus-host-disease, multiple sclerosis, rheumatoid arthritis, and Crohn’s disease. The tissue-resident status of MSC found in the lung is a key feature of their importance in the context of disease and injuries of the respiratory system, since these cells could be instrumental to providing more specific and targeted therapies. Currently, bone marrow-derived MSC have been established in the treatment of disease, including diseases of the lung. However, with lung-resident MSC representing a unique population with a different phenotypic and gene expression pattern than MSC derived from other tissues, their role in remediating lung inflammation and injury could provide enhanced efficacy over bone marrow-derived MSC methods. Through this review, lung-resident MSC will be characterized, using previously published data, by surface markers, gene expression patterns, and compared with bone-marrow MSC to highlight similarities and, importantly, differences in these cell types.

INTRODUCTION

Defining Mesenchymal Stromal Cells

Identifying the characteristics and molecular mechanisms of adult cell lineages with multipotent differentiation potential provides a unique and promising opportunity in tissue regeneration and the treatment of many diseases. The tissue parenchyma is maintained by many supportive cellular residents of the surrounding stroma, and failure of these supportive cells results in disease and tissue dysfunction (34, 35, 119, 146, 148). Stromal residents of particular interest in wound repair are mesenchymal stromal cells (MSC). In the mid-20th century, Friedenstein et al. (54) first identified a cell in bone-marrow extracts, a tissue with high turnover and regenerative capacity, that has multipotent and self-renewing potential. Culturing this population of bone marrow-derived multipotent cells yields a clonogenic and plastic adherent phenotype with a fibroblastic morphology, coined “mesenchymal stem cells” by Caplan (24) in 1991. These colonies derive from a clonogenic cell known as colony-forming unit (CFU)-fibroblast, which additionally outlines a defining assay of this cell type, the colony-forming assay. In the literature, MSC are referred to as mesenchymal stem cells, “mesenchymal stromal cells,” “multipotent stromal cells,” and “mesenchymal progenitor cells” but are most widely identified as “MSC.” It should be noted that the term mesenchymal stem cells typically refers to the in vitro proliferation of these cells types, particularly when the population is CD271^pos/CD140a^neg, meeting stringent criteria for stem cell classification, which are considered mesenchymal stromal cells in vivo, as limited data on the physiologic relevance of their differentiation in vivo is currently available (55, 146). There has been a recent call to rename MSC as “Medicinal Signaling Cells” to more clearly define their functionality in therapeutics, thus demonstrating the quickly evolving nature of this field (25, 42). Like other stem cells, MSC undergo self-renewal while maintaining their multipotency and have the capacity to differentiate into cell types of mesodermal lineage such as chondrocytes, osteocytes, and adipocytes. More recently, in vitro studies have suggested that MSC may also transdifferentiate into cell types of ectodermal and endodermal lineages, such as through mesenchymal to epithelial transition (MET); however, this has yet to be validated in vivo (33, 112). These outlined features of MSC contribute to their role in tissue repair and wound healing and serve as the rationale for the ongoing characterization and potential application of MSC and their products in therapeutics.

Phenotype

The conflict in nomenclature for MSC throughout the literature galvanized the release of a position statement from the International Society for Cellular Therapy (ISCT) in defining the criteria for MSC classification, listed as follows: cells must be enriched for CD73, CD90, and CD105 surface proteins while being negative for the hematopoietic markers (CD11b, CD34, CD45) as well as CD14, CD19, CD79a, and human leukocyte antigen-1 (HLA)-DR, and they must be plastic-adherent and successfully differentiate into adipocytes, chondrocytes, and osteocytes in vitro (46). The lack of myosin heavy chain (MHC) markers contributes to their ability to seemingly “evade” immune detection, an important phenotypic feature of MSC that contributes to their therapeutic potential. In fact, there are promising data on the success of MSC allografts. However, these data are not universally obtained, since some reports of poor engraftment are noted, and current studies on MSC allograft rejection suggest the term “immunoprivileged” more accurately describes this phenomenon (6, 146).

Role in Tissue Regeneration and Maintenance

MSC define the stromal microenvironment and modulate tissue regeneration via the following functions: differentiation into essential cell types; extracellular matrix deposition; and paracrine actions targeting surrounding epithelial, endothelial, and immune cells (86, 146). MSC are capable of homing to injury sites where they integrate and elicit modulatory effects on the surrounding cells and direct efforts toward tissue repair (79). Specifically, MSC sense and respond to stress signals in the tissue, effectively becoming “activated,” and promote cell survival, regeneration, and angiogenesis with the release of an abundance of growth factors while modifying the activation and transcription of nearby cells for cytoprotection, proliferation, and repair (19, 26, 38, 44, 114, 117). MSC have an important role in maintaining tissue integrity by inhibiting reactive oxygen species, a common byproduct of inflammation and injury, thereby preventing fibrosis secondary to oxidative stress-induced tissue damage (105). The secretion of antimicrobial peptides by MSC also contributes to the protection of tissue from damaging microbial interactions, and, for this reason, MSC have been implicated in the treatment of sepsis (88, 93, 107, 140). Notably, MSC also project protrusions to cells in need and directly pass mitochondria to the target cytoplasm. Interestingly, mitochondrial donation by MSC plays an important role in recovering cellular function and in repairing damaged tissue (98, 134).

Contributing to their effectiveness in therapeutics is the ability of MSC to differentiate into chondrocytes, adipocytes, and osteocytes. While there is limited data to suggest this differentiation in vivo, the cell types that MSC can differentiate into are particularly relevant to their role in regenerative medicine (27, 113, 118). Chondrocytes produce type II collagen, which serves as a scaffold within the extracellular matrix, an essential step in ameliorating tissue damage. The scaffold promotes a supportive matrix for the seeding of new cells to repopulate the damaged tissue region (80). In contrast, the role of adipocytes in wound healing is characterized by the release of adipokines, such as adiponectin and leptin, which are critical in angiogenesis and restoring the epithelial cell populations within damaged tissue (132). Osteocytes are critical for maintaining the delicately orchestrated homeostasis of bone, which is constantly undergoing remodeling in osteoblastic and osteolytic cycles. Osteocytes extend multiple processes that reach throughout the matrix, expanding their surface area to sense the status of surrounding cells, which underscores the influence of osteocytes in maintaining homeostasis in bone (29). With more thorough evidence for in vitro differentiation of MSC to many cell types, such as insulin-producing cells, epithelium, and endothelium, further in vivo characterization of MSC differentiation potential is needed (36, 59, 103, 112, 129, 135).

Immunomodulation by MSC

MSC regulate tissue homeostasis in their multipotent state via their secretome, through direct cell-to-cell interactions, as well as a combination of the two, ultimately exerting immunoregulatory functions. MSC can assist the innate immune response due to their expression of a variety of Toll-like receptors and have been documented as proinflammatory; however, the majority of their documented functions are anti-inflammatory, likely due to the interest of using these cells for treatments of inflammatory and autoimmune diseases (82, 109, 146, 154). Waterman et al. (154) describes the potential for MSC polarization, such as is documented in macrophages, as MSC1 or MSC2 phenotypes based on the pro- or anti-inflammatory effects of the activated MSC. MSC can elicit direct effects on immune cells such as neutrophils, macrophages, dendritic cells, lymphocytes, mast cells, and natural killer (NK) cells (83, 111, 146, 156). As summarized in Fig. 1, MSC drive anti-inflammation directly through modulating immune cell activities. MSC suppress the chemotaxis, inflammatory cytokine release, and degranulation of mast cells. Macrophage polarization toward an anti-inflammatory M2-like phenotype is mediated by MSC interaction. This ultimately promotes the synthesis and secretion of anti-inflammatory cytokine IL-10 while suppressing TNF-α and IL-6 release. The proliferation of NK cells and T and B lymphocytes is inhibited by MSC, which also drives a shift in the secretion of inflammatory cytokines to anti-inflammatory cytokines. Additionally, MSC can suppress the activation of B lymphocytes, dendritic cells, NK cells, and neutrophils. Taken together, these downstream effects of MSC actions result in a dampening of inflammation and polarization toward the inflammation-resolution state (1, 32, 70, 156).

Fig. 1.

A schematic of the immunomodulatory effects of mesenchymal stromal cells (MSC) on target immune cells (1, 32, 70, 156).

MSC GAIN RESIDENT STATUS IN THE LUNG

MSC Are Found in All Organs

With MSC first isolated from bone marrow (BM-MSC) in the 1960s, their presence in other tissue types was assumed to be the result of intravascular migration from bone marrow to sites of tissue injury. To date, BM-MSC remain the best-characterized MSC in the literature. These populations are numerous and more readily available for study of their differentiation capacity, phenotype, and roles in immunomodulation and tissue homeostasis. In 1973, MSC were additionally derived from adipose tissue; however, their multipotency was not defined until 2001 (116, 161). Finally, in 2006, da Silva Meirelles et al. (41) described murine mesenchymal stromal cells derived from nearly all postnatal tissues; more specifically, MSC were identified in all the murine tissues they collected. This paper demonstrated a potential role of MSC populations in every tissue type and substantiated the hypothesis that MSC might have resident status in all tissues. MSC have since been isolated repeatedly from other tissues, including placenta, dental pulp, amnion, synovial fluid, cartilage, brain, and skin to name a few. While MSC populations represent a smaller percent of cells in these tissue types compared with bone marrow, there are many papers documenting their presence in tissues (3, 8, 37, 41, 66, 72, 82, 95, 122, 128, 137).

MSC Are a Self-Renewing Population in the Lung

In 2005, human bronchial fibroblast cultures were derived from human bronchial biopsies. These cultured cell types resembled MSC phenotype and multipotency, but this was only a step in the direction of discovering lung-resident MSC (127). In 2007, a study of lung allografts found MSC within the bronchoalveolar lavage fluid (BALF) obtained from transplanted lungs. Multipotent MSC populations were discovered in the BALF with 97% of the sex-mismatched MSC populations having the XY chromosomal status of the donor (91). Therefore, MSC derived from lung allografts are clearly either long lived or self-renewing. Using a different in vivo approach, Jun et al. administered genetically green fluorescent protein (GFP)-marked bone marrow grafts to assess the tissue status of MSC in the lung in a bleomycin-induced fibrosis murine model (77). These findings identified a majority GFP-negative MSC population in the lung homogenates 20 wk following the bone marrow graft, as assessed by flow cytometry, providing evidence supporting the hypothesis of a lung-resident MSC population. While these data provided strong evidence in the case of lung-resident MSC, the limitations of these studies are that MSC derived from BALF or homogenates do not adequately nor directly demonstrate lung tissue residence, since they did not identify the precise location of MSC in the lung. In fact, da Silva et al. (40) proposed a need to determine the true identity of these cell types in vivo.

Lung-Resident MSC Are Perivascularly Located

It was not until 2014 that Rolandsson et al. (126) demonstrated that MSC isolated from bronchial wall and parenchymal tissue biopsies of allografted lungs were found to be perivascular in location and enriched for CD90 and CD105 surface proteins. These data substantiated the identity, phenotype, and perivascular localization of lung-resident MSC (lr-MSC). Additionally, as with previous lung allograft data, these lr-MSC populations were donor matched up to 16 yr posttransplant, contributing further evidence that these populations are long-lived and/or self-renewing residents of lung stroma. In demonstrating a lung tissue-resident population of MSC, directed efforts to characterize this population relative to the well-characterized BM-MSC began.

SPECIFIC CHARACTERISTICS OF LR-MSC

Enriched Surface Proteins

The classification of MSC is defined by specific minimal criteria, the first of which comprises distinct surface protein characterization. Most data attained for characterization of lr-MSC sought to establish a phenotypically similar identity to BM-MSC, the most well-characterized MSC population. Much of the literature, therefore, consists of data defining the enriched surface protein profile consistencies between BM-MSC and lr-MSC, with more recent efforts made at outlining major differences. Meeting the minimal criteria for MSC classification by surface proteins, as outlined by the ISCT, lr-MSC have been described as enriched for CD73, CD90, and CD105 in addition to being negative for hematopoietic markers. The role of these enriched proteins has yet to be fully characterized in the context of lr-MSC-lung interactions. However, it is generally accepted that CD73 has a specific role in MSC-driven suppression of the adaptive immune response. CD73-enriched MSC promote the expression of anti-inflammatory genes, decrease the expression of proinflammatory genes in macrophages, and attenuate the recruitment of proinflammatory macrophages (12). Additionally, CD90 expression has been identified in stem cell populations and certain fibroblast and endothelial cell populations. The immunosuppressive effects of MSC have been attributed to their enriched CD90 expression (21). Specifically, CD90 has a defined role in inflammation and wound healing by promoting secretion of extracellular matrix molecules as well as cytokines and as an adhesion molecule on a variety of cell types involved in the production and release molecular mediators at the site of inflammation (96). CD105, also referred to as endoglin, is important in angiogenesis. Interestingly, the CD105 expression of MSC has been considered representative of only a subpopulation of MSC, with conflicting hypotheses as to the nature of this heterogenicity (5, 45, 76, 97, 159). Both lr-MSC and BM-MSC populations have also been shown to consist of subpopulations that are Gli⁺ and implicated in the development of tissue fibrosis; however, no data have been published comparing the expression levels between the two (14, 22, 130). MSC derived from bone marrow and lung have no significant differences in expression of CD9, CD29, CD44, CD166, stromal cell surface marker-1 (STRO-1), neuron-glial antigen 2 (NG2), aquaporin-5 (AQP5), HLA-1, α-smooth muscle actin (α-SMA), HLA class I, or vimentin (124). However, although both CD146 and CD106 are enriched surface markers on both, they are expressed at significantly higher levels on BM-MSC. Conversely, the expression of fibroblast growth factor (FGF) receptor-2 (FGFR2) is significantly greater on lr-MSC compared with BM-MSC, to which FGF10 binds, facilitating the mobilization of lr-MSC (49, 106, 124, 144). Additionally, lr-MSC and BM-MSC are equally negative for the hematopoietic surface proteins assessed in each of the publications. These surface protein findings are summarized in Fig. 2A.

Fig. 2.

Venn diagram comparisons of lung-resident mesenchymal stromal cells (lr-MSC) and bone marrow-derived mesenchymal stromal cells (BM-MSC) highlighting their reported similarities and differences. Proteins or genes that are exclusively found on one cell population are listed in the appropriate circle. *Statistically significant difference in expression between shared proteins, with the protein listed within the circle of the cell type with higher expression. The bolded terms represent protein criteria for mesenchymal stromal cell classification. A: surface protein differences. B: transcriptome differences within cells. C: secretome differences within conditioned cell culture media. D: proteome differences within cells (10, 41, 55, 78, 91, 124, 125, 133, 144, 148).

Proteome and Secretome

The response of MSC to the microenvironment is mediated by dynamic proteomic and secretomic profiles by which BM-MSC and lr-MSC can be distinguished. These marked differences are reported in the literature of MSC of other tissue origins as well, although many of these reports do not include lr-MSC in their analyses (17). Contributing to these vast proteomic and transcriptomic differences are the microenvironments by which each tissue of origin alters the resident MSC; thus, there is considerable evidence supporting tissue specificity of MSC populations (89, 123). The driving force of these secretome expression profiles is the ability of MSC to sense stress and changes in the tissue microenvironment. In response, MSC secrete soluble factors that have paracrine activity and modulate cytokine production, differentiation, activation, and the secretome of target cells. Given this, it stands to reason that protein expression of BM-MSC and lr-MSC varies in comparative studies that assay the proteome and secretome of each cell type (10, 91, 121, 124, 125). In 2016, Rolandsson Enes et al. (125) determined that BM-MSC secreted significantly less monocyte chemoattractant protein-1 (MCP-1, CCL2), a monocyte chemotactic protein essential for macrophage recruitment, than lr-MSC, while there were no significant differences in immunosuppressive effects on lymphocytes between the two subpopulations. In the 2017 study by Rolandsson Enes et al. (124), BM-MSC cell layers and conditioned medium were compared with those of lr-MSC cultures by data-independent acquisition-mass spectrometral analysis. Importantly, lr-MSC had nearly 1,000 more proteins expressed and 288 significantly enriched proteins within the cell layer (CL) compared with BM-MSC CL. Additionally, lr-MSC had 36 significantly enriched proteins in the conditioned medium (CM) compared with the CM of BM-MSC cultures. In contrast, BM-MSC were significantly enriched for 64 CL proteins and 37 CM proteins when compared with lr-MSC cultures (124). These data created two distinct clusters upon hierarchical clustering analyses, distinguishing BM and lr-MSC subpopulations. To further assess the protein data from publications on BM- and lr-MSC where pathway analysis was not conducted, we have curated two separately compiled lists, one list each for BM-MSC and lr-MSC, containing significantly upregulated proteins that were then entered into the pathway analysis software Reactome (51). Among the proteins entered in the software, those that yielded upregulated pathways reaching statistical significance (as determined by Reactome algorithms) were used to generate Table 1. These data highlight some unique differences among these populations of MSC. Of note, there are more pathways involving Wnt, elasticity, motility, MET signaling, and integrins among the genes upregulated in lr-MSC compared with BM-MSC.

Table 1.

Pathway analyses of proteomic data

Pathways Downregulated in lr-MSC	Ref. No.	Pathways Upregulated in lr-MSC	Ref. No.
YAP1 (TAZ)-stimulated gene expression	121, 125	RUNX3 regulates WNT signaling	121, 125
Interferon-γ signaling	121, 125	Elastic fiber formation	121, 125
MPS IIID–Sanpilippo syndrome D	121, 125	Nuclear signaling by ERBB4	121, 125
Antagonism of activin by follstatin	121, 125	MET promotes cell motility	121, 125
Gene and protein expression by JAK-STAT signaling after IL-12 stimulation	121, 125	MET activates PTPN11	121, 125
Interferon signaling	121, 125	MET activates STAT3	121, 125
IL-12 signaling	121, 125	Degradation of extracellular matrix	121, 125
IL-12 family signaling	121, 125	MET activates PTK2 signaling	121, 125
Glycoprotein hormones	121, 125	MET receptor activation	121, 125
Regulation of active cofactor, UDP-glucuronate	121, 125	MET activates PI3K/AKT signaling	121, 125
Peptide hormone biosynthesis	121, 125	Potassium transport channels	121, 125
Transcriptional regulation by RUNX3	121, 125	MET interacts with TNS proteins	121, 125
Signaling by activin	125	MET activates PTPN11	121, 125, 148
Peptide hormone metabolism	121, 125, 148	MET activates STAT3	121, 125, 148
Regulation of PTEN localization	121, 125, 148	Regulation of IGF transport and uptake by insulin-like growth factor-binding proteins	121, 125, 148
Cell proliferation	124	Posttranslational protein phosphorylation	121, 125, 148
Cell growth	124	Extracellular matrix organization	121, 125, 148
Response to wound healing	124	IL-4 and IL-13 signaling	121, 125, 148
Oxidation-reduction processes	124	Laminin interactions	121, 125, 148
Positive regulation of gene expression	124	Chemokine receptors bind chemokines	121, 125, 148
		Activation of GABAB receptors	121, 125, 148
		Degradation of extracellular matrix	121, 125, 148
		Nonintegrin membrane-ECM interactions	121, 125, 148
		GABA receptor activation	121, 125, 148
		Formation of β-catenin: TCF transactivating complex	121, 125, 148
		Platelet degranulation	121, 125, 148
		Response to elevated platelet cytosolic Ca2+	121, 125, 148
		ECM proteoglycans	121, 125, 148
		Activation of G protein-gated potassium channels	121, 125, 148
		G protein-gated potassium channels	121, 125, 148
		Inhibition of voltage-gated Ca2+ channels via Gβ/γ-subunits	121, 125, 148
		Platelet activation, signaling, and aggregation	121, 125, 148
		Neurodegenerative diseases	121, 125, 148
		Deregulated CDK5 triggers multiple degenerative pathways in Alzheimer’s	121, 125, 148
		Myogenesis	121, 125, 148
		Inwardly rectifying K+ channels	121, 125, 148
		Oxidation-reduction process	124
		Cell proliferation	124
		Response to drug	124
		RNA splicing	124
		Protein folding	124
		Positive regulation of IκB/NF-κB signaling	124
		Metabolic process	124
		Reg mRNA stability	124

Table comprises significantly different pathways between lung-resident mesenchymal stromal cells (lr-MSC) and bone marrow-derived mesenchymal stromal cells (BM-MSC), downregulated in lr-MSC compared with BM-MSC (left column) and upregulated in lr-MSC compared with BM-MSC (right column), determined by proteomic analysis. If enrichment analyses were not provided, we conducted our own analyses of the differentially expressed proteins, with P ≤ 0.05. These analyses were conducted using Reactome (https://reactome.org) using a cutoff of P ≤ 0.05 and are listed in italics above. YAP1, yes-associated protein 1; MPS IIID, mucopolysaccharidosis type IIID; JAK, Janus kinase; STAT, signal transducer and activator of transcription; IL, interleukin; PTEN, phosphatase and tensin homolog; RUNX3, runt-related transcription factor 3; WNT, wingless and Int-1; ERBB4, Erb-B2 receptor tyrosine kinase 4; MET, MET proto-oncogene, receptor tyrosine kinase; PTPN11, protein tyrosine phosphatase nonreceptor type 11; PTK2, protein tyrosine kinase 2; PI3K, phosphatidylinositol 3-kinase; AKT, protein kinase B; TNS, tensin proteins; IGF, insulin-like growth factor; GABAB, γ-aminobutyric acid B receptors; ECM, extracellular matrix; GABAA, γ-aminobutyric acid A receptors; TCF, T cell factor; CDK5, cyclin-dependent kinase 5; IκB, inhibitor of κB; NF-κB, nuclear factor-κB.

Transcriptome

There are increasing reports of distinct gene expression patterns that distinguish MSC of different tissue origin as unique subpopulations (55, 121, 124, 125). The gene expression of these differing MSC subtypes is modulated by their interactions with their respective microenvironments and defined role within their tissue of origin. To visualize identified similarities and differences between the transcriptomes of BM-MSC and lr-MSC, please refer to Fig. 2D and Table 2. Interestingly, the transcriptome and proteome of lr-MSC are of closer resemblance to epithelial cell types than BM-MSC, with pathways upregulated in lr-MSC following the pattern from Table 1. There are more pathways involved in cell adhesion and motility, as well as integrin signaling. From the 2016 paper by Rolandsson Enes et al. (125), among 89 differentially expressed genes in lr-MSC compared with BM-MSC, genes related to epithelial-to-mesenchymal transition [SNAI2 (snail family transcriptional repressor 2), HGF (hepatocyte growth factor), TGF-β2 and 3, TCF4 (transcription factor 4), and CDH2 (cadherin 2)] had significantly higher expression in lr-MSC. Additionally, this paper reported significantly increased expression of the genes commonly expressed in lung-resident cells [FOXF1 (forkhead box 1) and HOXB5 (homeobox B5)] along with the Wnt-signaling protein secreted frizzled-related protein 1 (SFRP1) in lr-MSC compared with BM-MSC (125). The importance of Wnt signaling in lr-MSC subpopulations has been attributed to the regulation of epithelial and myofibroblast differentiation in lr-MSC and the consequential suppression of pulmonary fibrosis (22, 23, 30, 31, 101, 131, 150). The myofibroblast differentiation of lr-MSC is induced by TNF-α and TGF-β1. Kruppel-like factor 4 (Klf4), inhibitor of growth family member 5 (Ing5), and nuclear factor κ-light-chain enhancer of activated B cells (NF-κB) are downstream targets of TNF-α and TGF-β1 with a potential role in attenuating this differentiation (69, 150). These data illuminate the potential role of lr-MSC in the lung due to the need for frequent epithelial regeneration to maintain homeostasis and barrier function of the pulmonary system while evading further functional damage through tissue fibrosis. Additionally, lr-MSC exhibit a greater expression of genes related to cell migration, cell adhesion, and extracellular matrix organization, features that integrate seamlessly with the physiology of the lung in particular, where maintaining the structure of the tissue is critical for proper function, since many airway structures are often a few cell layers thick and must remain elastic (144). These data suggest significant differences among these two MSC populations that shed light on their tissue-specific roles and provide evidence to the impact of tissue specificity of the phenotype, genotype, and function of these cell types.

Table 2.

Enrichment analyses of transcriptomic data

Pathways Upregulated in BM-MSC	Ref. No.	Pathways Upregulated in lr-MSC	Ref. No.
Response to lipopolysaccharide	144	Cell adhesion	144
Response to drug	144	Integrin-mediated signaling pathways	144
Cell-to-cell adhesion	144	Positive regulation of epithelial cell proliferation	144
Response to hypoxia	144	Extracellular matrix organization	144
Aging	144	Cell migration and morphogenesis	55, 144
Embryonic development	55, 121, 125, 144, 148	Negative regulation of ossification	144
Signal transduction	144	Embryonic organ system morphogenesis	55, 121, 125, 144, 148
Response to activity	144	Signal transduction: small GTPase mediated	55, 144
Negative regulation of apoptotic process	144	Cellular response to growth factor stimulus	144
Response to organic cyclic compound	144	Collagen biosynthetic process	144
Positive regulation of angiogenesis, blood vessel development	55, 144	Positive regulation of transcription from RNA	55, 144
Response to γ-irradiation	144	Activation of MAPK activity	144
Inflammatory response	55, 121, 125, 144, 148	Heart development	144
Negative regulation of cell development	55, 121, 125, 144, 148	Positive regulation of cell proliferation	144
Proximal distal pattern formation	55, 121, 125, 144, 148	Negative regulation of mitotic division	55, 121, 125, 144, 148
Skeletal system development	55, 121, 125, 144, 148	Meiotic nuclear division	55, 121, 125, 144, 148
Negative regulation of multicellular organismal process	55, 121, 125, 144, 148	Chromosome segregation	55, 121, 125, 144, 148
		Cell division	55, 121, 125, 144, 148
		Digestive tract development	55, 121, 125, 144, 148
		Mitotic cell cycle process	55, 121, 125, 144, 148

Table comprises significantly different gene ontologies (GO) between lung-resident mesenchymal stromal cells (lr-MSC) and bone marrow-derived mesenchymal stromal cells (BM-MSC), downregulated in lr-MSC compared with BM-MSC (left column) and upregulated in lr-MSC compared with BM-MSC (right column), determined by enrichment analyses on genes. If analyses were not already provided, we conducted our own analyses of the differentially expressed genes, with P ≤ 0.05 from expression results and a false discovery rate P ≤ 0.05 using Panther (http://pantherdb.org). Analyses conducted by us are italicized and bolded where there is overlap between our analyses and those published previously (55, 121, 125, 144, 148).

ROLE AND POTENTIAL OF LR-MSC IN THE REGULATION OF TISSUE HOMEOSTASIS

Role of lr-MSC in Lung Biology

The pulmonary epithelium, in providing barrier function to the lungs that are continually exposed to pathogens and damaging inhalants, requires continual renewal to maintain pulmonary epithelial integrity. Additionally, the lung epithelium has a unique role as a primary defense in immunity evidenced by its frequent exposure to pathogens and the infrequent rate of pulmonary infections (11). Although there are major voids in the literature outlining the precise mechanisms of action of lr-MSC within the lung, many studies of BM-MSC provide a foundation for understanding how lr-MSC might effectively contribute to lung homeostasis. Overall, MSC mediate anti-inflammatory effects by driving down inflammatory cytokine production and release, preventing the activation and proliferation of proinflammatory immune cells, promoting microvascular remodeling, and differentiating into cell types that contribute to tissue repair and regeneration (34, 55, 57, 60, 65, 74, 105, 158). MSC ameliorate oxidative damage through secreted products and additionally alter the gene expression of surrounding cells through the release of miRNAs (contained in exosomes) that serve to dampen inflammation that is exacerbated by more widespread tissue damage (39, 47, 52, 67, 104). Extracellular vesicles (EV) have been specifically identified as secretory products of MSC, and the characterization of these MSC-derived EV has identified cargo, such as miRNAs and functional proteins, often of an anti-inflammatory nature, which highlights their potential use in cell-free therapeutics (142, 143). It has been reported that EV derived from adipose-MSC and BM-MSC promote an M2-like macrophage polarization in vitro, and both MSC-derived EV or conditioned media attenuate inflammatory lung diseases (73, 84, 99, 100, 120). Ultimately, the function of lr-MSC in the lung can only be inferred through these data until further studies are performed.

In a study by Abreu et al. (1), human BM-MSC were exposed to human BALF from patients with acute respiratory distress syndrome (ARDS) or other lung diseases, such as acutely exacerbated cystic fibrosis. The response of BM-MSC was dependent on the disease from which the BALF was derived, with other lung diseases promoting anti-inflammation in BM-MSC cultures more effectively than BALF from ARDS. This work has alluded to the specific interactions between lr-MSC and the pulmonary microenvironment, evidenced by the phenotypic alterations of MSC resulting from lung inflammation (1). When comparing the immunomodulatory effects of lr-MSC with BM-MSC, Ricciardi et al. (121) determined lr-MSC to exert similar effects on NK cells and B cells to those of BM-MSC cocultures. Furthermore, this study identified an important distinction between the two populations, since lr-MSC more remarkably differentiated to epithelial cells in retinoic acid treatment while also expressing epithelial genes at a higher basal amount (121).We have previously detailed changes in lr-MSC resulting from exposure to organic dust extract, a model of agriculture occupational dust exposure, among which include decreased proliferation and migration of dust-exposed human lr-MSC and increased release of both pro- and anti-inflammatory cytokines: TNF-α, IL-6, IL-8, and AREG, FGF10 (109). Additionally, lr-MSC have been shown to encourage prorepair fates of immune cells and increase T-helper 2 cytokines IL-4 and IL-13 while secreting receptors that neutralize proinflammatory cytokines and inhibiting further inflammatory signaling and cytokine release, including TNF-α and IFN-γ (81, 158). Human lr-MSC isolated from allografts were used for in vitro characterization of their interactions with T cells. These isolated lr-MSC were unable to activate allogenic T cells; however, they were able to inhibit CD4 and CD8 T cell proliferation in response to allogenic and mitogenic stimuli through the release of prostanoids (74).

Role of lr-MSC in Lung Pathology

It is important to note that MSC derived from BALF constitute the majority of lung-MSC studies in diseases of the lung, and ultimately disease may result from either too much or too little action by MSC. For example, bronchoalveolar lavage (BAL)-derived donor-matched MSC from lung transplant patients are implicated in patients with chronic graft rejection, namely, bronchiolitis obliterans syndrome (BOS) (148). In this pathology, fibrotic obliteration of small airways is linked with significant increases in BAL-derived MSC with clear variations of gene expression and phenotype compared with BAL-derived MSC from those without BOS. lr-MSC derived from the BAL of BOS patients exhibit increased collagen deposition and increased α-SMA expression associated with a profibrotic phenotype, and FOXF1 expression directly correlates with lr-MSC in lung tissue and increased numbers of lr-MSC in BALF (148). While Walker et al. (148) hypothesize these cells are mobilized to promote tissue in repair, it is unclear if these cells may also contribute to disease pathogenesis in other cases. Vella et al. (147) determined that expression changes in BAL-derived MSC from BOS patients include increased mRNAs associated with histone deacetylases and methyltransferases, as well as genes involved in protein binding, catalytic activity, transferase activity, and macromolecular complex binding to name a few. Although these data suggest a profibrotic phenotype of MSC that might contribute to the fibrosis that is characteristic of BOS, it still stands that BAL-derived MSC are identified only as a predictor of BOS and less clearly as a direct contributor to the pathology. Another paper identified multipotent perivascular Gli1⁺ cells, expressing MSC markers, in bone marrow, muscle, heart, liver, kidneys, and lung that proliferate and differentiate into myofibroblasts upon tissue injury (87). The identification of Gli⁺ MSC as drivers of tissue fibrosis has been documented elsewhere (23, 130). These findings allude to the potential role of tissue-specific MSC in organ fibrosis and more specifically how the presence of these cell types in the lung might contribute to injury-induced pulmonary fibrosis.

Identifying the role of lr-MSC in lung homeostasis within the context of bleomycin-induced pulmonary fibrosis has been of interest, since this disease has devastating impacts on individuals treated with bleomycin for their cancer. Hou et al. determined that increased TNF-α as well as Wnt/β-catenin and NF-κB signaling, which are characteristics of bleomycin-induced pulmonary fibrosis, contribute to myofibroblast differentiation of lr-MSC in vivo (69). In this mouse model of bleomycin-induced pulmonary fibrosis, the inhibition of NF-κB prevented TNF-α activation and ultimately reduced the development of fibrosis. This outcome was measured by collagen deposition in tissue sections as well as collagen protein expression. Additionally, the replacement of lr-MSC subsequently resolved fibrosis in this model (77). Furthermore, the importance of lr-MSC in tissue repair has also been documented in an in vivo model of LPS-induced acute lung injury (ALI). Using this model, Wang et al. (153) determined that lr-MSC impacted T cell fate by regulating the balance between regulatory T cells and Th17 cells. In this model of LPS-induced ALI, administration of MSC resolved the increase in vascular permeability, ameliorated inflammation, and decreased collagen deposition within the interstitium. Finally, in a review by Möbius and Rüdiger (108), the loss of lr-MSC in neonates is highlighted as a contributory mechanism to the development of bronchopulmonary dysplasia (BPD), and studies using administration of exogenous MSC seem to attenuate this disease. This review hypothesizes that many factors associated with the immature lung alter lr-MSC function, which in turn impairs pulmonary homeostasis and contributes to BPD pathogenesis. Additionally, this review discusses the rescue effect of MSC administration in driving an adequate therapeutic response and recovery in BPD. Altogether, it still remains to be seen if the epigenetic and phenotypic changes in lr-MSC identified in lung injury or disease contribute to disease etiology, pathogenesis, or sequalae and ultimately how these roles and mechanisms may vary relative to a particular disease.

BAL-derived MSC have been distinguished from lr-MSC in vivo, and this distinction is of physiologic importance (126, 133). This evidence is critical to our understanding of MSC subpopulations, since these are equally considered lr-MSC but have markedly different phenotypes and transcriptomes and since BAL-derived-MSC differentiate into a profibrotic phenotype compared with tissue-resident lr-MSC (133). This work by Sinclair et al. (133) demonstrates that BAL-derived lr-MSC express more surface markers typical of fibroblasts, such as α-smooth muscle actin and collagen 1A1, than the tissue-derived lr-MSC. Ultimately, these data suggest that lr-MSC are recruited into alveolar spaces in lung disease states and additionally differentiate to a profibrotic phenotype. It is this profibrotic dysfunction of lr-MSC recruited to the airways that is hypothesized to play a key role in pulmonary diseases like BOS, BPD, and interstitial lung disease (ILD), since the scale tips away from the prorepair phenotype of tissue-resident lr-MSC. Furthermore, the presence of lr-MSC in the airways and their respective proliferative capacity is a measure of disease status used in the clinical determination of the prognosis of BOS. Therefore, teasing apart the differences in tissue-derived and BAL-derived MSC populations is an essential step in understanding their role in pulmonary homeostasis, and further study is warranted.

Unique Therapeutic Potential of lr-MSC

MSC cell therapies are currently implicated in many clinical trials such as those for graft versus host disease, chronic obstructive pulmonary disease, acute lung injury, cardiac ischemic injury and regeneration, radiation damage, traumatic brain injury, idiopathic pulmonary fibrosis, tuberculosis, cancer, and other diseases related to chronic inflammation and autoimmune disorders (2, 7, 58, 61, 85, 90, 102, 149). Of particular importance is identifying the therapeutic potential of these endogenous mediators of repair and renewal for the treatment of disease specific to their host tissue and potentially in diseases of differing tissue types (162). The majority of these current and previous clinical applications use BM-MSC, regardless of the organ implicated in disease (141). These therapies include direct injection of MSC into circulation, in auto- and allograft-based treatments, as well as cell-free therapies using MSC-derived secretory products. A major pitfall of intravenously administered MSC is their inability to migrate across pulmonary capillaries to the lung stroma (48, 53). Furthermore, these injected MSC have short half-lives. These shortcomings have generated greater interest in using cell-free therapies via the MSC-derived secretome. Among these secreted products are proteins, cytokines, and EV with varying cargo that mediate repair and modulate inflammation within tissue targets. EV are particularly stable packages containing receptors, immunoglobulins, and microRNAs. They have been found in milk and serum while effectively withstanding homogenization and heat inactivation, respectively, and can impact the inflammatory response of the consumer (15, 16, 94, 110, 115). With these cell-free-based therapies, there is no risk of allograft rejection, and the products can be administered intravenously or aerosolized. While worth mentioning, these MSC and MSC-derived cell-free therapies are beyond the scope of this review, which aims to highlight the importance of lung-resident MSC specifically but have been extensively reviewed elsewhere (7, 9, 56, 58, 102, 138, 145, 158).

Zulueta et al. (162) published one of the few studies examining the specific impacts of therapies derived from lr-MSC; here, IB3–1 cells (an in vitro model of cystic fibrosis) were treated with EV derived from human lr-MSC, leading to anti-inflammatory and antioxidant effects. These effects included transcriptional and protein-level decreases in proinflammatory cytokines, such as IL-1β, IL-6, and IL-8 with EV treatment when the IB3–1 cells were stimulated with TNF-α. The mechanism of action in this regulation was determined to be through peroxisome proliferator-activated receptor-γ (PPARγ), leading to downregulation of NF-κB signaling and upregulation of heme oxygenase-1 (HO-1) (162). Further foundational work on characterizing these tissue-resident populations of MSC, including lr-MSC, will contribute a greater understanding toward the development of specialized and targeted approaches to the treatment of disease. How lr-MSC mediate their effects on specific cell types is poorly understood, and therefore more robust data on the interactions between tissue-resident MSC and their host tissue are needed to identify mechanisms by which resident populations can be stimulated or activated without the need for cell-based therapies, and ultimately dictated by their potential roles in the pathogenesis of disease.

CONCLUDING REMARKS

Limitations and Current Needs in the Field

An important limitation in this field is the similarity between MSC populations and fibroblasts. There is debate among scientists as to whether MSC and fibroblasts are distinct cell lineages, or merely separate stages along a continuum of differentiation, especially since both of these cell types may contribute to fibrotic diseases via their differentiation to myofibroblasts (13, 50, 75, 92, 151). Fibroblasts are also characterized as plastic-adherent and have even demonstrated a similar capacity to differentiate into chondrocytes, osteocytes, and adipocytes, although these data are varied (4, 18, 28, 139). Some have even proposed the use of fibroblasts as an alternative to MSC as they also have immunoregulatory properties (64, 71). In fact, there are a multitude of papers comparing MSC to fibroblasts on the basis of proliferation potential, differentiation capacity, immunomodulation, and gene expression and DNA-regulatory patterns (43, 62–64, 68, 136). Even cell surface markers are largely nonspecific, although some papers have identified markers with the potential to distinguish the two (20, 43, 62). An interesting consideration that is discussed in some of these studies and reviews that may be contributing to the likeness of results is that MSC and fibroblast populations studied in vitro may be heterogeneous and actually contain a mix of both, even in commercially available lines, since they are not easily separated (139). Other studies have shown that fibroblasts are targets of MSC, demonstrating an interplay between the two. Conditioned media from and cocultures with MSC have been shown to suppress the proliferation and myofibroblast differentiation of fibroblasts (155, 157, 160). Everything considered, while necessary and informative, in vitro work must always be taken with a grain of salt as not all culture conditions are physiologically relevant or reproducible in vivo, and culture conditions can even make a cancer cell appear phenotypically normal (152). Thus, better characterization of these populations in vivo is incredibly important.

To date, the data on MSC populations are heavily reliant on in vitro and ex vivo studies. While in vitro work helps characterize cells and expand general knowledge of the role of that cell, in vitro studies lack the complexity of in vivo physiology. Ultimately, the characteristics of MSC populations are more complex than we can determine from in vitro and ex vivo studies alone, and, moving forward, considerable effort must be invested into examining these nuances. Furthermore, in vitro and in vivo characterization of lr-MSC, as well as MSC of other tissue origins, is required to broaden the understanding of MSC-directed disease pathologies and ultimately to improve the efficacy of MSC-based therapeutics. In particular, in vivo work can better illuminate the interplay of MSC within their unique tissue microenvironment to more accurately characterize these unique populations and potentially identify more targeted therapies.

Consideration of Tissue Specificity in Defining MSC Populations

As of now, most studies contributing to preclinical and clinical trials are understood within the context of BM-MSC, which leaves much to be discovered within more tissue-specific approaches. There are only a handful of papers that aim to characterize lr-MSC more specifically and even fewer examining the therapeutic potential of this subpopulation. While BM-MSC and lr-MSC share similarities, which at the very least allow for their classification as MSC, there are striking differences between these cell populations. Many of the similarities determined arise from efforts to apply what is known of BM-MSC to other MSC populations, but this approach is parochial and contributes to a limited understanding of how the tissue microenvironment alters MSC biology. As demonstrated by phenotypic markers, proteomics, transcriptomics, and secretomic data, there is a distinct relationship between these MSC and their respective tissue of origin. Therefore, identifying these functional differences is paramount in addressing how these MSC subtypes impact tissue dysfunction or promote tissue repair and homeostasis.

Ultimately, characterizing the similarities of these cell types was critical in identifying the existence of tissue-resident MSC populations; however, it is of considerable importance at this time to put greater effort into identifying inherent differences to advance our understanding of how MSC function in their home tissue and how these functions can be exploited in the treatment of tissue injury and disease. Although these cell types may meet the criteria for MSC classification, these data suggest that lr-MSC and BM-MSC are not equivalent and therefore may present with variation in therapeutic potential. Ultimately, more consideration of how the tissue microenvironment drives MSC fate and biological function in vivo is needed to appropriately characterize MSC so they may be appropriately used in both the understanding of disease pathogenesis and ultimately in the treatment of disease.

Methods

The publications comprising this review were identified on PubMed using the search terms “lung resident mesenchymal stem cells,” “lung resident mesenchymal stromal cells,” “lung mesenchymal cells,” “mesenchymal stromal cells”, and “mesenchymal progenitor cells.” Gene and protein expression data were compiled from papers that specifically stated the use of normal, nondiseased bone marrow and lung tissue. Additionally, in those studies in which primary cells were derived, we extracted data from papers that used at least two additional methods of confirming MSC populations by the ISCT definition (e.g., CFU assay, surface protein expression, flow cytometry). For the Pathway and Enrichment analyses, gene and protein lists were sorted by significantly (P < 0.05) upregulated and downregulated, only where data comparing BM-MSC and lr-MSC were published together to limit potential bias in pooling data from many different sources. A significance and false discovery rate (FDR) cutoff of P ≤ 0.05 was used when conducting our analyses.

GRANTS

This work was supported by National Institute of Environmental Health Sciences Grant R00ES025819.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

S.N.S. prepared figures; S.N.S. drafted manuscript; S.N.S. and T.M.N. edited and revised manuscript; S.N.S. and T.M.N. approved final version of manuscript.