Fibroblasts are programmed epigenetically to determine the unique structure and function of different organs and tissues. These unique features might contribute to organ-specific disease.
Tissue fibroblasts may be recruited from a number of sources and cell types including the bone marrow, blood, and local stromal cells and act as organ-specific innate immune sentinel cells.
Under inflammatory conditions, fibroblasts become key immune system players by recruiting and modulating the behavior and survival of infiltrating immune cells.
Fibroblasts can be programmed epigenetically through exposure to inflammatory and environmental stress such that they inappropriately prolong inflammation, which becomes persistent.
Within the synovium, persistent abnormal behavior of fibroblasts in some diseases like rheumatoid arthritis results in continued damage to vital joint structures such as cartilage and bone, which, if untreated, will result in deformity and functional impairment.
Recent advances in single cell transcriptomic analysis have demonstrated subpopulations of fibroblasts with discrete markers and functions within the synovium, with the potential to develop novel therapeutic approaches.
What Is a Fibroblast?
The architecture of organs and tissues is closely adapted to their function to provide microenvironments in which specialized functions may be carried out efficiently. The nature and character of such microenvironments are primarily defined by the stromal cells that reside within the tissues. The most abundant cell types of the stroma are fibroblasts, which are responsible for the synthesis and remodeling of extra-cellular matrix (ECM) components. In addition, their ability to produce and respond to growth factors and cytokines allows reciprocal interactions with adjacent epithelial and endothelial structures and with infiltrating leukocytes. Fibroblasts also integrate microenvironmental stimuli such as oxygen tension and pH. As a consequence, fibroblasts play a critical role during tissue development and homeostasis and are often described as having a “landscaping” function.
Fibroblast Identity and Microenvironments
Tissue-resident macrophages in the liver (Kupffer cells) and lung (alveolar macrophages) perform very different functions compared with macrophages in the brain (glial cells) or skin (Langerhans cells), yet they are all members of the monocyte/macrophage family. Until recently fibroblasts had been thought of as ubiquitous, generic cells with a common phenotype even within different tissues. However, we now know that fibroblasts from different organs are more like their macrophage counterparts, with unique morphologic features and repertoires of ECM proteins, cytokines, co-stimulatory molecules, and chemokines specialized to the different microenvironments in which they are found. This characterization also extends to their function as “immune sentinel” cells, expressing innate immune system pattern recognition receptors such as Toll-like receptors (TLRs), which trigger a pro-inflammatory response when ligated by bacterial or viral determinants.
Examination of fibroblast transcriptional profiles with use of microarray techniques reveals that fibroblasts retain a strong memory of their anatomic position and function in the body. Early studies demonstrated that fibroblast transcriptomes (i.e., the global picture of transcribed genes measured using microarrays) could be clustered into peripheral (synovial joint or skin fibroblasts) versus lymphoid (tonsil or lymph node) groups according to their organ of origin, with the potential to shift their transcriptional profiles upon exposure to inflammatory mediators such as TNF, IL-4, or interferon (IFN)-γ. More extensive analysis of expression profiles from primary human fibroblasts by one study has shown large-scale differences related to three broad anatomic divisions: anterior-posterior, proximal-distal, and dermal-nondermal. Genes involved in pattern forming, cell signaling, and matrix remodeling were found to predominantly account for these divisions. The gene expression profile of adult fibroblasts may therefore play a significant role in assigning positional identity within an organism.
More recently, it has become clear that these stable changes in gene transcription are brought about through epigenetic activation and silencing of the HOX family of landscaping genes. Such epigenetic patterning, whereby covalent modifications are made to regulatory regions of DNA or to the histones around which the DNA is wrapped to control access of transcriptional complexes, is a prototype for the stable changes that are also seen in fibroblasts. Epigenetic modifications result in stable changes in gene expression that persist over cellular generations in the absence of mutation of the primary DNA sequence and therefore drive the persistence of disease, as is described in more detail in Chapter 26 .
The problem of distinguishing fibroblasts of differing origin or maturity has historically been difficult because of a lack of specific cell surface markers. Whereas cluster of differentiation (CD) markers have revolutionized the isolation and study of leukocyte subsets, relatively few, poor-quality discriminatory markers have been identified that allow the identification of fibroblast subpopulations. Therefore fibroblasts traditionally have been identified by their spindle-shaped structure ( Fig. 14.1 ), elaboration of ECM, and lack of positive markers for endothelium, epithelial, and hemopoietic cells.
However, growing evidence indicates that fibroblasts are not a homogeneous population but exist as subsets of cells, much like tissue macrophages and dendritic cells (DCs). It is likely that connective tissue contains a mixture of distinct fibroblast lineages with mature fibroblasts existing side by side with more immature fibroblasts that are capable of differentiating into other connective tissue cells. Over the past 15 years studies have begun to identify novel markers that demarcate distinct subpopulations of stromal cells during development and have the potential to act as markers for different subpopulations of fibroblasts, each with different roles. Such markers include smooth muscle actin, which marks out a population of secretory, activated cells termed myofibroblasts, and more recently discovered markers such as CD248 and gp38 (podoplanin) ( Table 14.1 ; also see later discussion). More recently, the development of single cell transcriptomic techniques has enabled the discovery of discrete stromal and leukocyte clusters in tissues that allow correlation between protein-based markers and function among cells sorted from disaggregated tissue. This opens the door to new therapies that delete or promote differentiation of specific subpopulations within target tissues. Fibroblasts have been defined in terms of their embryologic origins and lineage relationships and are generally considered to be mesenchymal in origin. However, cell populations that appear to blur the distinction between hemopoietic and nonhemopoietic populations have now been identified. In addition, other unexpected shifts in lineage have been reported, including differentiation from neural stem cells into myeloid and lymphoid hemopoietic lineages. Classification by such lineages is therefore becoming increasingly untenable.
|Marker||Associated Cell Type||Synovial Location||Functional Significance|
|CD55||Fibroblast-like synoviocyte||Lining layer||Receptor/ligand for synovial macrophage CD97|
|VCAM-1||Fibroblast-like synoviocyte||Lining layer||Activated lining layer fibroblasts; adhesion molecule|
|α-SMA||Myofibroblast||Variable, minority subpopulation||Secretory, profibrotic fibroblast|
|CD248/endosialin||Pericyte||Sublining fibroblasts, pericytes||Acute inflammation, cancer and vasculogenesis|
|gp38/podoplanin||Pericyte and lymphoid endothelium||Lining layer fibroblasts, pericytes, lymphoid endothelium||Structural, proangiogenic lymph node role ; promotes motility in cancer|
|5B5/prolyl-4-hydroxylase||Broad fibroblast marker in vivo||Lining and sublining cells||Marks collagen synthetic machinery|
|S-100A4/FSP-1/Mts-1||—||Lining and sublining cells, invasive regions||Cancer, invasiveness roles via motility and impaired apoptosis|
|FAP||Associated with α-SMA + fibroblasts||Lining layer||Role in cancer fibroblasts, protective if ectoenzyme blocked in rheumatoid arthritis|
Origins of Fibroblasts in Tissue
Both inflammation and wound healing are characterized by the formation of new tissue. However, recent findings suggest that the new cells that form the remodeled tissues might not be derived from the proliferation of resident cells in the adjacent noninjured tissue, as was previously assumed. This finding is important because in both rheumatoid arthritis (RA) and fibrotic pathologic conditions, fibroblasts accumulate in excessive numbers despite apparently low proliferative rates. The principle origin for fibroblasts is from primary mesenchymal cells, and upon appropriate stimulation, fibroblasts can proliferate locally to generate new fibroblasts; however, although an increase in fibroblast numbers caused by local proliferation does occur, fibroblasts also may arise from other sources ( Fig. 14.2 ). The first of these sources is local epithelial to mesenchymal transition (EMT). EMT is an essential, physiologically important developmental mechanism for diversifying cells in the formation of complex tissues. However, fibroblasts also appear to be derived by this process in adult tissue after epithelial stress such as inflammation or tissue injury. EMT both disaggregates epithelial cells and reshapes them for movement. The epithelium loses polarity as defined by the loss of adherens junctions, tight junctions, desmosomes, and cytokeratin intermediate filaments. Epithelial cells also rearrange their F-actin stress fibers and express filopodia and lamellipodia. A combination of cytokines and matrix metalloproteinases (MMPs) associated with digestion of the basement membrane is believed to be secreted and is important in the process. The transition of epithelial to mesenchymal cell populations occurs in cancer and in diseases of the lung and kidney, and the process has been implicated in fibrotic disease. Early evidence suggests that a similar process may occur within the RA synovium.
An alternative explanation for the accumulation of stromal cells in people with chronic inflammatory conditions such as RA lies in the possibility of blood-borne precursors. In the mid-1990s it was discovered that vascular precursors (angioblasts) circulate in the blood of healthy people and that they could be recruited to sites of vasculogenesis in a rabbit ischemic hind limb model. This finding demonstrated that circulating mesenchymal precursors exist outside the hemopoietic system. Subsequent work has confirmed the presence of circulating cells of a mesenchymal phenotype in human subjects. These cells bear a remarkable resemblance to the synovial fibroblasts found in the joints of people with RA, which accumulate in large quantities in the joint lining despite little evidence of proliferation. Interestingly, one group showed that an influx of such cells preceded inflammation in a mouse collagen-induced arthritis model, suggesting that a role may exist for blood-borne stromal cell precursors in the initiation of inflammatory diseases. Furthermore, synovial fibroblasts themselves may migrate in the bloodstream, at least between distant sections of human cartilage in severe combined immunodeficiency (SCID) mice, raising an intriguing parallel to cancer and the radical concept of RA as a metastatic disease of the stroma.
Another circulating precursor cell that could account for the accumulation of fibroblasts in disease is the fibrocyte. Fibrocytes constitute 0.1% to 0.5% of nonerythrocytic cells in peripheral blood and rapidly enter sites of tissue injury and contribute to tissue remodeling in models of inflammatory lung disease. Fibrocytes are adherent cells with a spindle-shaped structure. They express class II major histocompatibility complex (MHC) and type I collagen and arise from within the CD14 + (monocyte) fraction of peripheral blood. They are capable of matrix elaboration and differentiate along fibroblast lineages under the influence of cytokines, particularly transforming growth factor (TGF)-β. The mere fact that a cell type apparently arising from within the monocyte lineage may become a “mesenchymal” stromal cell such as a fibroblast implies a further degree of plasticity and blurring of the apparently clear dividing line that was previously thought to exist between hemopoietic and nonhemopoietic lineages.
Fibroblasts Versus Mesenchymal Progenitor Cells
The potential role of circulating mesenchymal cell precursors (variously termed mesenchymal stem cells [MSCs], mesenchymal stromal cells, or mesenchymal progenitor cells [MPCs]) as sources of tissue fibroblasts is highlighted by the remarkable capacity of these cells to differentiate into other members of the connective tissue family, including cartilage, bone, adipocyte, and smooth muscle cells. This ability was initially demonstrated in bone marrow stromal cells, RA synovial fibroblasts, and circulating mesenchymal cells. Therefore, a characteristic mesenchymal phenotype could be defined on the basis of the hypothesis that the rheumatoid synovium becomes populated by a large proportion of circulating mesenchymal progenitor cells exported from the bone marrow. However, the property of trilineage differentiation (“pluripotentiality”) is now a property of many adult tissue fibroblasts, although varying somewhat between fibroblasts from different tissues, implying a hitherto unsuspected degree of plasticity in the body’s stromal cell populations. The two previously separate fields of mesenchymal precursor cell biology and largely disease-centered fibroblast biology have therefore rapidly converged. However, the concept of bone marrow stromal precursors remains interesting; in chimeric murine models with bone marrow green fluorescent protein (GFP) expression, arthritic joints contained significantly more GFP + cells than did nonarthritic joints, supporting a bone marrow origin for expanded fibroblast populations.
Physiologic Characteristics and Functions of Fibroblasts
Production of ECM Components
Ensuring ECM homeostasis is one of the primary functions of fibroblasts. To perform this function, fibroblasts must be able to produce and degrade ECM, as well as adhere to and interact with existing matrix components. Fibroblasts produce a number of ECM molecules—both fibrous proteins and polysaccharide gel components such as collagens, fibronectins, vitronectin, and proteoglycans—which are then assembled into a three-dimensional network. This mechanism provides a framework through which other cell types, which use varying strategies to navigate through the ECM, can move. It also provides a substrate for the deposition of haptotactic (tissue rather than fluid-based) gradients of chemokines and stores of growth factors to direct cellular movement and behavior in a regional fashion. The types of ECM molecules produced by individual populations of fibroblasts differ from tissue to tissue, reflecting the diversity of fibroblasts in different organs. For example, dermal fibroblasts produce significant amounts of type VII collagen, which adheres to the epidermal and dermal layers in the skin. Fibroblasts in other organs such as the lung and kidney produce mainly interstitial, fibrillar collagens (particularly types I and III).
In the synovial membrane, fibroblasts also have a barrier function, in that they provide the joint cavity and the adjacent cartilage with lubricating molecules such as hyaluronic acid, along with plasma-derived nutrients. Anatomically, the intimal synovial membrane is an unusual structure in that barrier function is maintained in the absence of a laminin-rich basement membrane, as occurs in epithelial structures. In addition to lacking a basement membrane, cellular contacts between the fibroblast-like synoviocytes also lack tight junctions and desmosomes. However, strong homophilic adhesion between synoviocytes is mediated by the adhesion molecule cadherin-11 (see later discussion), which is largely responsible for fibroblast organization into synovial tissue. In the presence of disease, fibroblasts must migrate to sites of tissue injury or remodeling and interact with ECM molecules through specific surface receptors. Through such receptors, fibroblasts must sense changes in both the structure and the cellular composition of connective tissues. They respond dynamically by adjusting the production of ECM components and cross-linking them into the appropriate matrix.
Attachment to and Interaction With Extra-cellular Matrix
Integrins are key mediators of both cell-to-matrix and cell-to-cell adhesive interactions. They are expressed as transmembrane heterodimers containing one α- and one β-subunit, of which at least 25 αβ combinations are known ( Table 14.2 ). The main adhesion molecules responsible for the attachment of fibroblasts to collagen are α 1 β 1 and α 2 β integrins, whereas other β 1 integrins such as α 4 β 1 and α 5 β 1 mediate attachment of fibroblasts to fibronectin and its spliced variants. In addition, α v integrins are responsible for attachment to vitronectin.
|Integrins||α 1 β 1||VLA-1||Laminin, collagen|
|α 2 β 1||VLA-2||Laminin, collagen|
|α 3 β 1||VLA-3||Laminin, collagen, fibronectin|
|α 4 β 1||VLA-4, CD49d/CD29||VCAM-1, CS1 fibronectin|
|α 5 β 1||VLA-5||Fibronectin|
|α 6 β 1||VLA-6||Laminin|
|α L β 2||LFA-1, CD11a/CD18||ICAM-1, ICAM-2, ICAM-3, JAM-A|
|α M β 2||Mac-1, CR3, CD11b/CD18||ICAM-2, iC3b, fibrinogen, factor X|
|α X β 2||P150, 95, CD11c/CD18||iC3b, fibrinogen|
|α E β 2||E-cadherin|
|α 4 β 7||CD49d||Fibronectin, VCAM-1, MAdCAM-1|
|α v β 3||CD52/CD61, vitronectin receptor||Vitronectin, fibronectin, osteopontin, thrombospondin-1, tenascin|
|Ig superfamily||ICAM-1||CD54||LFA-1, Mac-1|
|VCAM-1||α 4 β 1 , α 4 β 7|
|MAdCAM-1||α 4 β 7 , L-selectin|
In addition to conventional integrin-to-ligand binding, additional accessory molecules allow for the integration of adhesive contacts and local growth factor signaling. Syndecans are a family of four single transmembrane domain proteins that carry three to five heparan sulfate and chondroitin sulfate chains, allowing for interaction with a large variety of ligands including fibroblast growth factors, vascular endothelial growth factor (VEGF), TGF-β, and ECM molecules such as fibronectin. Syndecans are expressed on fibroblasts in a tissue-specific and development-dependent manner. Data from syndecan knockout mice indicate that syndecan-4 is involved in wound healing and that the response of syndecan-4–deficient fibroblasts to fibronectin attachment is significantly altered.
Immunoglobulin Superfamily Receptors
The immunoglobulin (Ig) superfamily is a diverse group of transmembrane glycoproteins defined by the presence of one or more Ig-like repeats of 60 to 100 amino acids with a single disulfide bond. Although it includes numerous adaptive immune system genes (e.g., Igs, T cell receptor, and MHC), adhesion proteins such as intercellular adhesion molecules (ICAMs) 1 to 3, vascular cell adhesion molecule-1 (VCAM-1), and mucosal addressin cell adhesion molecule (MAdCAM) mediate both cell-to-cell interactions and adhesive interactions with integrins (see Table 14.2 ).
Cadherins mediate homotypic, calcium-dependent adhesive interactions with the same cadherin species expressed by neighboring cells. Classical cadherins possess five extra-cellular domains, a single-pass transmembrane domain, and a highly conserved cytoplasmic tail. The cytoplasmic tail interacts with β-catenin, which in turn binds α-catenin, forming a linkage between the cadherin-catenin complex and the actin cytoskeleton. Tightly regulated expression of cadherins is essential to embryogenesis but is also critical for tissue morphogenesis and tissue-specific cell differentiation. Cadherins also modulate cell proliferation and invasion through activation of intra-cellular signal transduction pathways, modulation of MMP production, and association with growth factor receptors.
Adhesion Molecule-Mediated Signaling
Importantly, interaction with adhesion molecules not only regulates adhesion and motility but also directly influences activation status, apoptosis, and pro-inflammatory and anti-inflammatory responses in fibroblasts and other cells. The engagement of cell adhesion molecules such as integrin receptors on the surface of fibroblasts results in the formation of focal adhesion complexes, which activate intra-cellular signaling cascades that regulate cell proliferation and survival, the secretion of certain cytokines and chemokines, and matrix deposition and resorption. In particular, integrin-to-fibronectin engagement induces MMP expression, linking adhesion-to-matrix remodeling ( Fig. 14.3 ). Among the signaling molecules that transmit signals from the integrins to the cell interior, focal adhesion kinase (FAK) plays a central role. FAK, a tyrosine kinase, is recruited into newly established focal contacts and, in turn, recruits other adapter proteins such as p130Cas and Grb2. This process leads to phosphatidylinositide 3-kinase (PI3K) and Src-kinase activation and promotes the initiation of a variety of signaling cascades, culminating in phosphorylation of the extra-cellular regulating kinase (ERK) mitogen-activated protein kinases (MAPKs) and activation of transcription factors. Such pathways can also be activated through FAK-independent signaling events, such as through growth factor receptor ligation. The exact mechanisms by which different signals cooperate to mediate a specific response of fibroblasts and how this translates into distinct diseases are not yet fully defined.
Degradation of Extra-cellular Matrix by Fibroblasts
Remodeling of the ECM requires fibroblasts to express an extensive repertoire of matrix-degrading enzymes with varying specificity. Although these matrix-degrading enzymes are crucial to tissue maintenance and repair, inappropriate overexpression of such enzymes is a key factor in the joint damage, particularly to cartilage, that occurs in inflammatory disease. Such enzymes fall into a number of families, including MMPs, tissue inhibitors of metalloproteinases (TIMPs), cathepsins, and aggrecanases, which are covered in detail in Chapter 8 .
With the exception of MMP-2 and the membrane-type (MT)-MMPs, which are constitutively expressed by fibroblasts, MMP expression is regulated by extra-cellular signals via transcriptional activation in fibroblasts. Three major groups of inducers can be differentiated: pro-inflammatory cytokines, growth factors, and matrix molecules. Among the cytokines, IL-1 is perhaps the most potent inducer of a variety of MMPs, including MMP-1, MMP-3, MMP-8, MMP-13, and MMP-14. Fibroblast growth factor (FGF) and platelet-derived growth factor (PDGF) are also known inducers of MMPs in fibroblasts because they potentiate the effect of IL-1 on MMP expression. All MMP promoter regions except MMP-2 contain activator protein-1 (AP-1) binding sites; however, there is good evidence that all MAPK families (ERK, c-Jun N-terminal kinase [JNK], and p38 pathways; see Fig. 14.3 ), in addition to activators of nuclear factor-κB (NF-κB), signal transducer and activator of transcription (STAT), and ETS transcription factors participate in MMP regulation. Matrix proteins (i.e., collagen and fibronectin), and especially their degradation products, also activate MMP expression in fibroblasts, providing the possibility for site-specific MMP activation in regions of matrix breakdown.
Fibroblasts as Innate Immune Sentinels
Classically, macrophages have been studied as sources of inflammatory cytokines and chemokines in response to innate immune stimuli and portrayed as immune sentinel cells accordingly. However, when activated by substances released during tissue injury or the products of invading microorganisms, fibroblasts are capable of elaborating a broad repertoire of inflammatory mediators, which fully justifies their classification as immune sentinel cells. Through expression of TLRs 2, 3, and 4, fibroblasts respond to bacterial products such as lipopolysaccharides (LPSs) by activating the classical NF-κB and AP-1 inflammatory pathways, generating chemokines capable of recruiting inflammatory cells, and generating metalloproteinases capable of degrading matrix. However, TLR expression may be increased by pro-inflammatory cytokines TNF and IL-1β within the local microenvironment and may also be activated by endogenous cellular debris such as necrotic cells in synovial fluid, leading to widespread fibroblast activation in disease. As immune sentinels, fibroblasts are able to bridge the innate and adaptive immune responses through expression of the molecule CD40. This molecule was initially assumed to be restricted in its expression to antigen-presenting cells such as macrophages and DCs. However, it is widely expressed by fibroblasts within discrete tissues. CD40 engagement by its ligand CD40L expressed on a restricted population of immune cells, including activated T lymphocytes, is critical for the further induction of pro-inflammatory cytokines and chemokines during an immune response, as well as for antibody production by CD40-expressing B lymphocytes.
Fibroblasts also need to be able to respond to more generic danger signals. The intra-cellular apparatus for response to danger signals such as high levels of urate has recently been identified as the nucleotide-binding oligomerization domain (NOD)-like receptor family, which is made up of NOD and NALP (i.e., NACHT domain, leucine-rich repeat [LRR] domain, and pyrin domain [PYD]-containing protein) receptors. A high local level of urate released by dying cells triggers formation of the active NALP3 inflammasome complex, which results in release of IL-1.
Expression of high levels of NOD-1, NOD-2, and NALP3 (cryopyrin) is seen in the RA synovium and can be induced in fibroblasts by TLR ligands and/or TNF. Furthermore, synergy between TLR and NOD stimulation, with increased IL-6 production in response to NOD-1, TLR2, and TLR4 ligands, has recently been demonstrated. The cytokine IL-17 also regulates multiple TLRs in RA synovial fibroblasts.
Role of Specialized Fibroblast Subsets Within Tissue Microenvironments
Combining surface markers with consistent function has been the key to decades of development in the field of leukocyte biology. By comparison, stromal cell biologists have had remarkably few such stable markers. However, this situation is now gradually changing, and certain areas of developmental biology have spearheaded identification of putative markers (such as CD248) through approaches such as immunization of animals with human fibroblasts and digesting and identifying stromal cell subpopulations in tractable organ systems. One example is the murine thymic stroma, in which subsets with both geographic and functional consistency have been identified. For instance, one study identified CD45 − , gp38 + stromal cells in the thymus as T-zone fibroblastic reticular cells. This population of cells, which is geographically restricted to the T zone, provides a limited pool of essential homeostatic survival factors, IL-7 and CCL19, for T lymphocytes, serving a key niche function for which adaptive immune cells must compete. Also, gp38 marks populations of fibroblastic reticular cells within the lymph node that modulate trafficking of DCs. Single cell based sequencing approaches to such cells have advanced our understanding of stromal diversity in the lymph node, revealing many stromal subpopulations within disaggregated lymph node tissue, with regional geography related to interactions with immune cells.
A further subpopulation of specialized fibroblast-like cells of mesenchymal origin is the pericyte. These cells ensheath small blood vessels (i.e., arterioles, capillaries, and venules) and are involved in vasculogenesis, matrix stabilization, and immunologic defense. Pericytes have been hypothesized to represent the extralymphoid source of mesenchymal progenitor cells and express markers consistent with mesenchymal stem cells. Their further definition with newer stromal cell markers such as CD248 and CD146 will be able to establish a mesenchymal progenitor cell niche.
Fibroblast-like Synoviocytes in the Normal Synovium
The normal synovium provides an excellent prototypic model of fibroblast subsets defined by known markers, some of which are responsive to disease. In healthy people the synovium is a delicate, thin, two-layer structure attaching bone and the joint capsule. One layer, a two- to three-cell-thick lining layer, is formed with roughly equal proportions of CD68 + , phagocytic type A macrophage-like synoviocytes, and type B mesenchymal, fibroblast-like synoviocytes (FLSs). This layer serves a barrier function, and FLS secretes lubricative substances, including hyaluronic acid and lubricin, along with secreting the lining layer matrix. The second layer is the sublining layer, which is composed of less densely packed fibroblasts and macrophages in a loose tissue matrix along with blood vessel networks. FLSs in the lining layer are associated with a number of cellular markers (see Table 14.1 ), including CD55 (decay accelerating factor [DAF]), VCAM-1 (which, outside of T cell–to-integrin interactions is generally only expressed by bone marrow fibroblasts providing support for the B cell developmental niche ), uridine diphosphoglucose dehydrogenase (UDPGD), reflecting the ability to synthesize hyaluronan, and the novel marker gp38. Sublining FLSs are instead marked by the nonspecific cellular marker CD90 (Thy-1), which also recognizes endothelium, and by the recently discovered marker CD248, which marks both pericytes and stromal fibroblasts. Gp38 marks cells in the sublining region, including lymphatic endothelium ( Fig. 14.4 ).
As mentioned earlier, the unique lining layer barrier function is not supported by a basement membrane and conventional tight junctions but instead by homotypic interactions between cadherin-11 molecules. Randomly assorted cells expressing classical cadherins, such as cadherin-11, will sort themselves in a cadherin-specific manner, emphasizing their importance in the generation and maintenance of organ integrity. Cadherin-11 mediates selective association of mesenchymal rather than epithelial tissues, a function that is carried forward after embryogenesis in structures such as the joint, lung, and testis. Cadherin-11 knockout mice exhibit a hypoplastic synovial lining that lacks the normal numbers of synovial lining cells and is deficient in ECM quantity. Adhesion between type A and type B synoviocytes is maintained by ICAM-1:β 2 integrin and VCAM-1:α 4 β 1 integrin interactions.
By virtue of their role in defining the geography of specialized tissues, fibroblasts and other stromal cells exist in living organisms within three-dimensional environments, whereas the vast majority of experiments performed using fibroblasts in the laboratory are still conducted within two-dimensional environments. Furthermore, fibroblasts are frequently grown in nonphysiologic stimuli such as serum, to which fibroblasts would not normally be exposed unless tissue damage were to occur. Behavior is significantly different when cells are cultured in artificial three-dimensional environments. It is therefore all the more remarkable that fibroblasts cultured using conventional two-dimensional techniques retain characteristics such as positional memory and unique cytokine profiles.
Recent work has addressed the issue of three-dimensional synovial models. In so-called micromass cultures, FLS, but not dermal fibroblasts, within laminin-containing spheres reproduced a lining layer structure with production of lubricin, support for co-cultured monocytic cells, and expansion of the membrane upon stimulation with pro-inflammatory stimuli such as TNF. Some cells also remained in a “sublining” zone of low density. FLSs therefore have the ability to self-organize in a tissue organoid, which recapitulates some of the key features of the synovium. This finding is further evidence of the robustness of epigenetic programming, which determines site and organ specialization.
Fibroblasts in Rheumatic Diseases
Role of Fibroblasts in Persistent Inflammation
Inflammatory reactions proceed against the backdrop of specialized stromal microenvironments. The response to tissue damage involves a carefully choreographed series of interactions among diverse cellular, humoral, and connective tissue elements. For an inflammatory lesion to resolve, dead or redundant cells that were recruited and expanded during the active phases of the response must be removed. In addition, resident fibroblasts attempt to repair damaged tissue.
It is becoming increasingly clear that fibroblasts are not only passive players in immune responses but also actively determine the switches that govern progression from acute to chronic inflammation, as well as those governing resolution or the progression to chronic, persistent inflammation. The “switch to resolution” is an important signal that permits tissue repair to take place and enables immune cells to return to draining lymphoid tissues (lymph nodes) for immunologic memory to become established. However, in chronic immune-mediated inflammatory diseases such as RA, fibroblasts contribute to the inappropriate recruitment and retention of leukocytes in a site- or organ-dependent manner, leading to tissue- and site-specific initiation and subsequent relapse of chronic persistent inflammatory disease, effectively a “switch to persistence.”
It is now recognized that fibroblasts themselves may undergo fundamental changes while responding to such environmental stimuli. It is known that during wound healing and under profibrotic conditions, some fibroblast-like cells are transformed into myofibroblasts, which are distinct from tissue fibroblasts in terms of both their phenotype and their behavior. The mechanisms underlying such persistent phenotypic change, which is maintained through cellular generations, are highly likely to involve epigenetic modifications of gene promoters and their closely related histones (see Chapter 26 ). This has been shown recently in both human and murine renal fibrotic disease, where hypermethylation of the promoter region of a ras oncogene inhibitor led to gene silencing, ras pathway activation, and hence persistent fibrogenesis. Such fibrotic transformation of fibroblasts is also characteristic of systemic sclerosis, a generalized fibrotic disorder that affects the skin and various internal organs such as the lungs, heart, and gastrointestinal tract (see Chapter 88 ). The overproduction of ECM components, particularly type I, III, VI, and VII collagen, by skin fibroblasts is a hallmark of this disease and is closely linked to the disease-specific activation of these fibroblasts. This pattern of activation includes not only a distinct profile of ECM overproduction but also altered responses to both inflammatory mediators and immune cells. Although the phenotype of fibroblasts in RA is not fundamentally profibrotic in this sense, the hallmark of these cells, both in vitro and in vivo, is also a persistently imprinted phenotype that is maintained even in the absence of continuous stimulation by inflammatory triggers or leukocytes.
Fibroblast-like Synoviocytes in Rheumatoid Arthritis
In inflammatory arthritis such as RA, the two compartments of the synovium undergo radical change. The lining layer undergoes dramatic hyperplasia, sometimes reaching 10 to 20 cells in depth, with both type A and type B cell populations expanded and becoming merged with the sublining. At the articular borders of the synovium, the thickened synovial lining layer may become a mass of “pannus” tissue rich in FLS and osteoclasts, which aggressively invade the adjacent articular cartilage and subchondral bone, respectively. The sublining layer also undergoes expansion, with sometimes huge infiltrates of inflammatory cells including macrophages, mast cells, T cells, B cells, and plasma cells in addition to DCs. T and B lineage cells may remain in diffuse infiltrates or may coalesce into aggregates of cells varying from simple perivascular “cuffs” a few cells in diameter to structures resembling B cell follicles in up to 20% of samples. This increased activity is supported by further ECM production and neoangiogenesis, although the inflamed synovium remains in a state of relative hypoxia.
As mentioned previously, cadherin-11 serves a vital role in preserving the integrity of the synovial lining layer, and cadherin-11 knockout mice display a hypoplastic lining layer. However, when cadherin-11 knockout mice are evaluated in the K/BxN serum transfer model, invasiveness is reduced with a 50% reduction in inflammation. Similarly, cultured fibroblasts with mutant cadherin-11 constructs also demonstrate impaired invasiveness into cartilage. Cadherin-11 expression is also much higher in RA than in osteoarthritis (OA) or normal synovium. This unique structural molecule may therefore emerge as a therapeutic target ; because of shared roles in invasive disease, targeting of this molecule in breast cancer is currently in development.
Persistent Activated Fibroblast Phenotype in the Rheumatoid Arthritis Synovium
Increased expression of cadherin-11 is but one facet of the persistent, activated phenotype of rheumatoid FLS, which remains stable even after culturing in vitro for many months. These cells play a direct role in tissue damage through secretion of multiple MMPs and cathepsins, which degrade cartilage and bone tissues in the joint. In vitro functional assays such as the laminin invasion assay produce intriguing results, in which the degree of invasion with a given in vitro cultured fibroblast sample correlates with the degree of radiographic progression seen in the joints of the patient from whose samples the fibroblasts were initially cultured. The most compelling evidence for a persistent phenotype is the attachment to and invasion of fibronectin-rich matrix such as human cartilage in the absence of functioning leukocyte immune cells in the SCID mouse model of arthritis. Here, fibroblasts in a tissue construct with human cartilage are implanted under the kidney capsule or skin of immune-incompetent SCID or Rag −/− mice. Multiple-passage cultured rheumatoid FLSs, but not OA or normal FLSs, invade and destroy the co-implanted human cartilage. This model has been used to explore the in vivo mechanisms governing invasiveness. For example, targeting MMP-1 and cathepsin L using ribozymes inhibits cartilage destruction. The effectiveness of glucocorticoids and the relative efficacy of different formulations of methotrexate in preventing erosions have also been examined.
Unbiased approaches to determining the key regulators of fibroblast invasiveness have made rapid recent progress. Transcriptomic approaches linking gene expression in RA fibroblasts and macrophages have revealed invasiveness pathways within fibroblasts regulated by complementary macrophage inflammatory pathways that are strongly driven by IL-1β stimulation. Key genes include periostin osteoblast-specific factor (POSTN) and twist basic helix–loop–helix transcription factor 1 (TWIST1). The impact of TNF and IL-17 stimulation on the transcriptome of RA fibroblasts has been examined, revealing critical hypoxia regulated genes linked to invasiveness, including MMP-2 and the chemokine receptor CXCR4, which is already implicated in disease persistence.
An alternative unbiased approach to dissecting function involves the parallel study of cellular enzymes mediating tyrosine phosphorylation of key signaling molecules (protein tyrosine phosphatases [PTPs]). Investigation of the PTPome of synovial fibroblasts in RA compared with OA revealed a dual role for SH2 domain–containing phosphatase 2 (SHP2); knockdown of this enzyme reduced both invasiveness and survival of RA FLSs, suggesting a pivotal signaling molecule.
Fibroblasts implanted with cartilage migrate to a contralateral cell-free implant, and subcutaneous, intraperitoneal, and intravenously injected fibroblasts will also migrate to sections of human cartilage, suggesting a tropism to damaged cartilage tissue. This important finding raises the question of which cell populations are grown from the synovium when tissue is digested and adherent cells are cultured in vitro: lining layer cells, sublining cells, or a mixture of both? From a methodologic perspective, answering this question is a challenge. However, we do know from transcriptomic approaches that the phenotype remains more stable in tissue culture than might be expected, with little transcriptional divergence over the first two to four passages and the level of differentially expressed genes between parallel cultures rising to greater than 10% only after passage 7.
These models demonstrate the remarkably stable and disease-specific phenotype of cultured RA synovial fibroblasts, which includes high basal and stimulated expression of signature cytokines such as IL-6 and chemokines (discussed later). RA synovial fibroblasts also express characteristic adhesion and immune-modulating molecules such as VCAM-1, galectin-3, and a specific repertoire of TLRs, which initiate innate immune cellular responses. A satisfactory molecular explanation for the stable phenotype of RA synovial fibroblasts has until recently evaded the field. However, epigenetic changes including DNA methylation; histone modifications such as acetylation, methylation, and citrullination; and altered micro RNA (miRNA) expression have now been suggested to underlie the observed persistent changes in fibroblast gene transcription and post-transcriptional repression (see Chapter 26 ). Further characteristic aspects of the RA FLS phenotypes and their biology are discussed extensively in Chapter 75 .
Interactions of Fibroblasts With Leukocytes
Recruitment of Inflammatory Infiltrates Into the Joint
Stromal elements such as synovial fibroblasts are subject to a pro-inflammatory cytokine network within the inflamed synovium. Direct-contact interactions with other infiltrating cells such as T lymphocytes lead to high levels of expression of many inflammatory chemokines (see Fig. 14.3 ). Neutrophil-attracting chemokines are expressed at high levels by stimulated fibroblasts and include CXCL8 (IL-8), CXCL5 (ENA-78), and CXCL1 (GRO-α). Monocytes and T cells are recruited by a range of chemokines found at high levels in the synovium; CXCL10 (IP-10) and CXCL9 (Mig) are highly expressed in synovial tissue and fluid. CXCL16 is also highly expressed in the RA synovium and acts as a potent chemoattractant for T cells. CCL2 (MCP-1) is found in synovial fluid and is known to be produced by synovial fibroblasts; it is considered to be a pivotal chemokine for the recruitment of monocytes. CCL3 (MIP-1α), CCL4 (MIP-1β), and CCL5 (RANTES) are chemotactic for monocytes and lymphocytes and are known products of synovial fibroblasts. CCL20 (MIP-3α) is also overexpressed in the synovium and has a similar chemoattractant profile via its specific receptor, CCR6. CX3CL1 (fractalkine) is also widely expressed in the rheumatoid synovium. A number of chemokine receptors differs between peripheral blood and synovial leukocytes, suggesting that they are enriched in the synovium either though their selective recruitment by endothelial-expressed chemokines or after upregulation by the microenvironment after their recruitment.
Fibroblast Support for Leukocyte Survival
Stromal cell support for the survival of leukocyte populations fulfills a physiologic role in certain organs within the body. The selective recruitment and support of hemopoietic subsets is an essential physiologic function of stromal cells in specific microenvironments. For instance, immature B lymphocytes are completely dependent on factors such as IL-6 produced by bone marrow stromal cells. Although the bone marrow niche plays a critical role in the early development of all hemopoietic leukocyte populations, it also acts as an active reservoir for terminally differentiated leukocyte subpopulations, including CD4 and CD8 T cells and neutrophils. The bone marrow stromal microenvironment therefore maintains not only the selective survival, differentiation, and proliferation of all lineages of immature hemopoietic cells but also, in some cases, the survival of their mature counterparts. The stromal microenvironment plays a crucial role in the maintenance of such survival niches, which are not generic but are highly specific to certain organs and tissues, resulting in site-specific differences in the ability of different stromal cells to support the differential accumulation of leukocyte subsets.
In the case of an inflammatory response, successful resolution requires the removal of the vast majority of immune cells that were recruited and expanded during the active phase of the inflammation. A number of studies have shown that during the resolution phase of viral infections, the initial increase in T cell numbers in peripheral blood that is seen within the first few days is followed by a wave of apoptosis occurring in the activated T cells. This situation is mirrored within tissues, where apoptosis induced by the molecule Fas occurs at the peak of the inflammatory response and may be responsible for limiting the extent of the immune response. In contrast, the resolution phase appears to be principally triggered by cytokine-deprivation–induced apoptosis, during which leukocytes compete for a shrinking pool of survival factors provided by the microenvironment, leading to programmed death of those cells, which are surplus to requirements.
In RA the resolution phase of inflammation becomes disordered. Recent studies have shown that a failure of synovial T cells to undergo apoptosis contributes to the persistence of the inflammatory infiltrate. The T cell survival pathway shares all the essential hallmarks of a stromal cell, cytokine-mediated mechanism (high B cell lymphoma [Bcl]-X L , low Bcl-2, and lack of cell proliferation). Type I IFNs (IFN-α and -β), which are produced by synovial fibroblasts and macrophages, have been identified as one of the principal factors responsible for prolonged T cell survival in the rheumatoid joint ( Fig. 14.5 ). Interestingly, although type I IFN is beneficial in multiple sclerosis (a disease in which tissue scarring and low levels of T cell infiltrates are observed), these results suggest that type I IFN is not likely to be a successful therapy for people with RA, a prediction that has been borne out in clinical trials. It is likely that this mechanism of stromal cell–induced leukocyte survival occurs in many chronic inflammatory conditions in which T cells accumulate.