Peripheral Blood Fibrocytes: Differentiation Pathway and Migration to Wound Sites¹

Fibroblasts, depending on their tissue source and stimuli for activation, are a heterogeneous population of cell types exhibiting distinct functions. Fibroblasts found in the wound are considered essential for the healing process. The concept that wound fibroblasts can originate from peripheral blood cells goes back almost 100 years (reviewed in Ref. 1). Since then, numerous studies have reported the differentiation of peripheral mononuclear cells into fibroblast-like cells.

In 1994, a distinct population of blood-borne fibroblast-like cells that rapidly enter sites of tissue injury was described (2). Termed fibrocytes, these cells comprise 0.1–0.5% of nonerythrocytic cells in peripheral blood and display an adherent, spindle-shaped morphology when cultured in vitro. Cultured fibrocytes express the fibroblast products collagen (Col)³ I, Col III, and fibronectin, as well as the leukocyte common Ag (CD45RO), the pan-myeloid Ag (CD13), and the hemopoietic stem cell Ag (CD34). In addition, fibrocytes express MHC class II and costimulatory molecules (CD80 and CD86) and have the capacity to present Ag in vitro and in vivo (3, 4). By their morphology, growth properties, and cell surface markers, fibrocytes appear to be distinct from monocytes/macrophages, dendritic cells, and other known APC types. Cultured fibrocytes do not express typical monocyte/macrophage-specific or B cell markers (such as CD14, CD16, or CD19), nor do they express typical surface proteins of dendritic cells or their precursors (such as CD1a, CD10, CD25, and CD38). In addition, fibrocytes isolated from peripheral blood and cultured ex vivo secrete a unique profile of cytokines, growth factors, and chemokines (5).

Based on their presence in wounds and their secretion of proinflammatory cytokines, chemokines, and extracellular matrix proteins, fibrocytes have been postulated to play a role in wound healing and connective tissue formation. Although initial studies performed in sex-mismatched bone marrow chimeric mice suggested that fibrocytes arose from a relatively radioresistant progenitor population (2), the precise origin of these cells and the wound trafficking signals relevant to their directed migration remain unknown. In this study we identify a differentiation pathway of cultured fibrocytes, characterize the signals for fibrocyte migration to wound sites in vivo, and reveal the potential role of fibrocytes in wound contracture.

BALB/c mice (females, 8–12 wk old) were purchased from The Jackson Laboratory (Bar Harbor, ME). All animal procedures were conducted according to guidelines of the institutional animal care and use committee of North Shore University Hospital under an approved protocol.

FITC-anti-α smooth muscle actin (FITC-anti-αSMA) mAb was purchased from Sigma (St. Louis, MO). Biotinylated rabbit anti-Col I and biotinylated rabbit IgG were purchased from Rockland (Gilbertsville, PA). Anti-mouse CCR3, CCR5, CCR7, or CXCR4 polyclonal Abs and FITC-anti-goat IgG Ab were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). All other Abs were purchased from BD PharMingen (San Diego, CA). TGF-β1 (active), secondary lymphoid chemokine (SLC), and stromal-derived cell factor (SDF) were purchased from R&D Systems (Minneapolis, MN).

Fibrocytes (human and mouse) were purified from peripheral blood and cultured as previously described (2, 5). Briefly, PBMCs were isolated from human Leukopaks (purchased from the Long Island Blood Center, Long Island, NY) by centrifugation over Ficoll/Paque (Pharmacia, Piscataway, NJ) following the manufacturer’s protocol. After 2 days of culture on tissue culture flasks in DMEM (Life Technologies, Gaithersburg, MD) supplemented with 20% FBS (HyClone Laboratories, Logan, UT), penicillin, streptomycin, and l-glutamine, nonadherent cells were removed by gentle aspiration, and media were replaced. After 10–12 days, adherent cells were lifted by incubation in ice-cold 0.05% EDTA (in PBS). The crude fibrocyte preparations (∼70–80% pure based on Col I/CD11b staining) then were depleted by immunomagnetic selection of contaminating T cells (∼13%), B cells (∼3%), and monocytes (∼11%) using pan-T, anti-CD2; pan-B, anti-CD19; and anti-CD14 Dynabeads, respectively (Dynal, Great Neck, NY). The resultant cultured, enriched fibrocyte populations were ≥95% pure based on Col I/CD11b staining, with T cells and monocytes contributing ∼3 and 2%, respectively. Typically, between 0.4 and 5 × 10⁴ fibrocytes were isolated per milliliter of human blood.

Mouse PBMC were isolated from BALB/c mouse blood (heparinized) obtained by cardiac puncture following CO₂ asphyxiation. Mouse blood was mixed with PBS (2:1) and layered over Ficoll/Paque (Pharmacia; 15 ml blood over 30 ml Ficoll) and centrifuged according to the manufacturer’s protocol. Mouse fibrocytes were cultured in DMEM supplemented with 10% FBS and 10% mouse serum (Sigma), penicillin, streptomycin, and l-glutamine, as previously described (4). After 10–12 days, the adherent crude fibrocyte preparation (∼75% pure based on Col I/CD11b staining) were lifted using 0.05% EDTA in PBS and depleted by immunomagnetic selection of contaminating T cells, B cells, and monocytes using pan-T (anti-CD90), pan-B (anti-B220) Dynabeads (Dynal), and anti-mouse CD14 attached to Dynabeads, respectively. Following immunodepletion, the cultured, enriched fibrocyte preparations were verified to be ≥95% pure by Col I⁺/CD11b⁺ staining as determined by flow cytometry. Approximately 0.8–4 × 10⁴ fibrocytes/ml mouse blood (∼1–1.2 ml blood/mouse) were purified.

Human adult dermal fibroblasts were purchased from Clonetics (San Diego, CA) and cultured according to the manufacturer’s recommendations. The human intestinal smooth muscle cell line was obtained from American Type Culture Collection (Manassas, VA) and cultivated according to recommended procedures.

Initial studies were aimed toward elucidating the cellular origin of peripheral blood fibrocytes. Therefore, we fractionated whole blood supplied as Leukopaks (see Fig. 1 A) and cultured the various fractions in vitro. Adherent cells were collected from overnight cultures of human PBMCs (total), and CD14⁺ cells were enriched from the PBMC fraction by depletion of T and B cells (CD14⁺). CD14⁻ cells (including all PBMCs except CD14⁺ cells) were purified by depletion of the CD14⁺ cells from the total PBMC preparation. Using the Transwell two-chamber system (0.4 μm; Corning Costar, Cambridge, MA), CD14⁺, CD14⁻, or total cells (3 × 10⁶ cells/ml in DMEM/10% FBS) were cultured in either the upper or lower chambers, as indicated. After 7 days of culture, the cells that grew in the lower well were collected and analyzed for fibrocyte-like differentiation by Col I/CD11b staining and flow cytometry. Similar results were observed with cells prepared from three other donors.

For studies investigating a requirement for T cells in fibrocyte differentiation, the CD14⁺ cell fraction (see above) was purified from purchased Leukopaks and cultured with autologous T cells isolated using T cell enrichment columns (R&D Systems). T cell purity was ≥95%, as assessed by flow cytometry using anti-CD3 Abs (PharMingen). After 7 days of coculture, the resulting population was analyzed for the percentage of fibrocytes by Col I/CD11b staining and flow cytometry. Similar results were observed using fibrocytes isolated from three different donors.

For single Ab staining, cells (10⁵ aliquots) were resuspended in PBS containing 3% BSA and 0.1% sodium azide (FACS buffer) and incubated with the indicated Abs (or labeled isotype control Abs) for 30 min at 4°C. In cases where the primary Abs were not labeled, cells were washed and incubated with revealing Abs diluted in FACS buffer. After washing the cells in FACS buffer, fluorescence data were acquired on a FACSCalibur flow cytometer (BD Biosciences, San Jose, CA) and analyzed using CellQuest software (BD Biosciences). At least 5000 cells were analyzed per condition. To analyze preparations for Col I/CD11b staining, cells were prepared as described above and first incubated in FACS buffer containing biotinylated Col I Ab (or biotinylated rabbit control IgG), then washed and incubated sequentially in FACS buffer containing FITC-strepavidin (PharMingen) and PE-CD11b (PharMingen). Intracellular staining for αSMA was performed as previously described (6, 7). Briefly, cells were fixed and permeabilized using the Perm/Fix kit (PharMingen) according to the manufacturer’s recommendations and incubated with FITC-anti-αSMA mAb (Sigma).

Cultured, enriched peripheral blood-derived mouse fibrocytes (>96% pure) were stained with a membrane-inserting red dye, PKH-26 (Sigma), following the manufacturer’s protocol. Labeling efficiency, assessed by flow cytometry, and viability, assessed by trypan blue exclusion were >85%. PKH-labeled cells (5 × 10⁵) in 100 μl PBS were administered into the tail vein (i.v.) of BALB/c mice (n = 2/group). Immediately following injection of the enriched fibrocytes, a full-thickness round skin wound (5-mm diameter) was made in the dorsal subscapular area of each recipient mouse by excision with skin punch equipment, as previously described (8). Wound sites were removed 4 days later and examined for the presence of fluorescent fibrocyte cells by microscopic analysis of thin frozen sections and by quantitative flow cytometric analysis following proteolytic digestion. For quantitative flow cytometric analysis, excised skin (250 μg biopsy/animal) was chopped into small fragments, then incubated for 1 h at 37°C in RPMI 1640 containing 10% FBS, 2 mg/ml collagenase, and 20 μg/ml DNase I. The resulting single-cell suspension was examined by flow cytometry to determine the number of fluorescent fibrocytes present using calibration beads as previously described (9).

Total RNA was isolated from cultured, enriched fibrocytes (>95% pure) using RNAzol B (Tel-Test, Friendswood, TX). The cDNA was prepared from 1.0 μg RNA using 0.25 ng oligo-(dT)_12–18 and Moloney murine leukemia virus reverse transcriptase following the protocol supplied by the manufacturer protocol (Life Technologies). Two-microliter aliquots of cDNA were amplified by PCR using Supermix (Life Technologies) in a Perkin-Elmer model 9600 thermal cycler using specific primers PCR pairs as previously described: αSMA (10); CCR3 (11); CCR4, CCR5, and CXCR3 (12); CCR6 (13); CCR7 (14); CXCR4 (15); and β-actin (sense primer, 5′-GTGGGGCGCCCCAGGCACCA-3′; antisense primer, 5′-CTCCTTAATGTCACGCACGATTTC-3′). Thermal cycling (25–30 cycles, in 25 μl) was performed as follows: denaturation at 94°C for 0.5 min, annealing at 55°C for 0.5 min, and extension at 72°C for 1 min. PCR products were separated by electrophoresis through 2% agarose gels and viewed under UV light after ethidium bromide staining. To control for potential genomic DNA contamination, PCR were performed without the reverse transcription step, and no DNA amplification products were detected.

Chemotaxis assays were performed using Costar Transwell inserts (8-μm pore size) according to the manufacturer’s protocol. Cultured, enriched fibrocytes (≥95% pure) were resuspended at 1 × 10⁶ cells/ml in DMEM containing 0.1% BSA. Medium alone (negative control) or medium containing SLC or SDF (600 μl) was added to individual wells of a 24-well plate. Transwell devices then were inserted, and the fibrocytes (100 μl) were layered on top of the membrane (n = 3 wells/condition). After 3 h the transmigrated cells were collected and counted by flow cytometry using calibration beads (Coulter, Miami, FL), as previously described (9). Similar results were observed with two additional donors. For checkerboard analysis of SLC-directed chemotaxis of fibrocytes, 100 ng/ml SLC was added to either the top or bottom chamber alone and to both the bottom and top chambers (see Fig. 5 B).

Immediately following tail vein injection of PKH-labeled enriched fibrocytes (>94% pure; 5 × 10⁵ cells/mouse), BALC/c mice received an i.d. injection of SLC, SDF (0.1 or 1 μg in 50 μl), or PBS alone in the scapular region of the back (shaved). The injected site was excised 4 h later and proteolytically digested to produce a single-cell suspension (as described above). The number of labeled fibrocytes per biopsy sample (250 μg) was estimated by flow cytometry using calibration beads (9). This experiment was repeated twice with similar results.

Cellular collagen gel contraction assays were performed as previously described (16). Overnight adherent PBMC cultures, 10-day-old enriched fibrocytes (≥95% pure) previously cultured in the absence or the presence of TGF-β1 (10 ng/ml for 7 days before experiment), or normal human dermal fibroblasts were lifted using cold EDTA/PBS solution. A collagen solution in DMEM was prepared from rat tail Col I according to the manufacturer’s instructions, and combined with cells at 2 × 10⁵/ml (n = 3/cell type). The collagen/cell mixture (400 μl/well) was dispensed into culture plates and allowed to polymerize at 37°C for 30 min. Immediately after polymerization, 2 ml DMEM containing 10% FBS was added to each well. The gels then were detached from the wells by gently shaking the culture plates at various time points (0, 24, 48, and 72 h), and the longest and shortest diameters of each gel were measured. The mean of the linear measurements (n = 3 for each sample) taken at each time point was used to estimate the contractility of the cells. The data are presented as percent gel contraction. This experiment was repeated twice with similar results using cells obtained from different donors.

To determine the origin of fibrocytes, we analyzed the growth and phenotype of adherent human PBMC cultured on plastic (Fig. 1 A). After standard Ficoll separation, the resulting population was ∼40–50% CD14⁺ cells. Following an overnight adherence step, the adherent cell population (total) was >70% CD14⁺ cells exhibiting no detectable Col I staining, as assessed by flow cytometry (data not shown) (5). We have shown in previous studies that after 2 wk cells in these cultures no longer express CD14, but do express Col I (5). Importantly, we found that a cell population enriched for CD14⁺ cells, (i.e., PBMCs depleted of all T or B cells by magnetic beads) gives rise to very few Col I⁺/CD11b⁺ spindle-shaped fibrocytes after 1 wk of culture (data not shown).

Using Transwell culture chambers, we examined the cellular requirements for fibrocyte differentiation (CD11b/Col I⁺) in vitro from circulating blood cell fractions (Fig. 1 B). When a CD14⁻ cell fraction was cultured in the lower well of a Transwell plate and total PBMCs were cultured in the top chamber for 1 wk, no fibrocytes appeared in the lower chamber. Similarly, no fibrocytes appeared in the lower chamber when CD14⁺ cells alone were cultured in the bottom chamber and CD14⁺ cells or total PBMCs were cultured in the top chamber for 1 wk. By contrast, when total PBMCs were cultured in the bottom well of the Transwell chamber and either CD14⁻ cells or CD14⁺ cells (or medium alone, data not shown) were cultured in the top chamber, numerous spindle-shaped fibrocytes (CD11b⁺/Col I⁺) were observed within 1 wk. These data suggest that fibrocyte outgrowth from cultured PBMCs requires cellular interaction between a population of enriched CD14⁺ cells and another peripheral blood cell type or that fibrocyte precursors are only present in the PBMC fraction.

To examine the requirement of cellular interaction, we then added either purified, autologous T or B cells to CD14⁺ cell cultures in various ratios (CD14⁺:T, 0:1, 1:0, 3:1, 1:1, and 1:3) for 7–10 days and found that cocultures of CD14⁺ cells and T cells give rise to fibrocytes (CD11b⁺/Col I⁺; Fig. 1,C). We observed that a CD14⁺cell:T cell ratio of 3:1 was optimal (Fig. 1 C) for culturing fibrocytes. By contrast, no fibrocytes appeared when T cells were cultured alone or in cocultures of B and CD14⁺ cells or when CD14⁺ cells were cultured with T cell-conditioned medium (data not shown). Because fibrocytes do not express T cell markers (CD2, CD3, CD4, and CD8) or typical T cell cytokines (IL-2, IL-4, and IFN-γ), it is unlikely that T cells give rise to fibrocytes.

Next, we examined whether TGF-β1, a cytokine important for fibroblast proliferation and extracellular matrix production could promote the differentiation and accumulation of fibrocytes within PBMC cultures. The addition of TGF-β1 (1–10 ng/ml) to PBMC cultures on days 3–10 promoted fibrocyte differentiation in vitro, as shown by the enhanced accumulation of cells with spindle-shaped morphology (Fig. 2, A–C). Treatment of these cultures with TGF-β1 increased the expression of Col I by fibrocytes within these cultures in a dose-dependent manner (Fig. 2,D). The mean fluorescence intensities for Col I expression were 11, 24, and 63 for fibrocytes in cultures treated with 0, 1, and 10 ng/ml TGF-β1, respectively (Fig. 2 D). These Col I⁺ cells also stained positively for CD11b (data not shown). Furthermore, there was a dose-dependent increase in the number of fibrocytes within the cultures that stained positively for Col I in response to TGF-β1, with almost a 40% increase in response to 10 ng/ml TGF-β1 compared with untreated cells. Similar results were observed with fibrocyte preparations from three other donors, each showing a 30–45% increase in Col I expression between 0 and 10 ng/ml TGF-β1.

We next sought to quantify the migration into wound sites of transferred cultured, enriched fibrocytes using a mouse model system. Cultured, enriched mouse fibrocyte preparations (>96% pure) were injected (5 × 10⁵/mouse) into the tail vein of mice. Immediately, full-thickness skin punch biopsy wounds (5 mm in diameter) were made in the dorsal scapular area in some mice. The wound sites (and comparable untreated skin tissue) were excised 4 days later, and biopsy specimens were examined for the presence of labeled fibrocytes. As shown in Fig. 3,A, numerous fluorescent cells were found by microscopic analysis of the wound tissue at 4 days. Labeled fibrocytes appeared to be located near newly formed blood vessels at the edge of the wound. Using another group of mice (n = 3/group), single-cell suspensions were prepared from the excised wound or normal tissue (250 μg/biopsy), and labeled fibrocytes were quantified by flow cytometry. Enumeration of migrated labeled fibrocytes revealed that wounded tissue contained significantly more labeled fibrocytes than a similar area of normal skin taken from the same mouse (Fig. 3 B).

Numerous circulating cells, including, neutrophils, monocytes, and T cells, are known to migrate into cutaneous wound sites. This process is organized in part by specific interactions between chemokines and their receptors. We surveyed cultured enriched fibrocytes for chemokine receptor mRNA expression by RT-PCR and found CCR3, CCR5, CCR7, and CXCR4 mRNA (Fig. 4,A), but not CCR4, CCR6, or CXCR3 mRNA expression. We confirmed CCR3, CCR5, CCR7, and CXCR4 protein expression on the surface of human enriched fibrocytes by flow cytometry (Fig. 4,B). Cultured, enriched fibrocytes isolated from mouse blood also expressed CCR7 and CXCR4, as analyzed by cytofluorometric analysis (Fig. 4 C).

Based on the expression of CCR7, a receptor for SLC, and CXCR4, a receptor for SDF, by enriched fibrocytes we used SLC and SDF in an in vitro chemotaxis assay. As shown in Fig. 5,A, SLC significantly induced the migration of fibrocytes, whereas SDF did not. Checkerboard analyses confirmed the chemotactic (but not chemokinetic) response of cultured, enriched fibrocytes in response to SLC (Fig. 5,B). Based on these observations, we investigated whether SLC could promote the migration of transferred, cultured, enriched fibrocytes following an i.d. injection of the chemokine in vivo. Administered at a dose of 1 μg, SLC dramatically induced the accumulation of prelabeled, ex vivo cultured fibrocytes in the skin area surrounding the i.d. injection site compared with the effect of PBS alone (Fig. 5,C). By contrast, SDF injection did not promote fibrocyte chemotaxis in vivo (Fig. 5 C). Immunostaining of a 2-day wound site revealed SLC chemokine expression by the vascular endothelium (data not shown). These results suggest that fibrocytes migrate into early wound sites in part due to an interaction between vascular endothelium-derived SLC and fibrocyte CCR7.

Based on their presence within the wound and their expression of Col I and III, we have postulated that fibrocytes mediated wound healing and fibrosis. Gabbiani and coworkers have previously described a population of wound fibroblasts that differentiate into myofibroblasts in the presence of TGF-β (Ref. 17 , reviewed in Ref. 18). These cells are characterized by expression of αSMA, the activity of contracting collagen gels in vitro, and their proposed role in wound closure, inflammation, and fibrosis (reviewed in Ref. 19). Recognizing that TGF-β1 enhances Col I expression by cultured fibrocytes (Fig. 2,D) and that fibrocytes are present in wound tissue for days (20), we next examined whether cultured, enriched fibrocytes express αSMA and exhibit a contractile force. As shown in Fig. 6,A, unstimulated, cultured, enriched fibrocytes were found to express αSMA mRNA, but freshly isolated PBMCs did not. Unstimulated cultured, enriched fibrocytes also express αSMA protein, and the addition of TGF-β1 (10 ng/ml) increased αSMA levels by about 4-fold (Fig. 6,B). Next, we examined the contractile activity of cultured, enriched fibrocytes. We found that untreated cultured, enriched fibrocytes significantly contracted the collagen gels in vitro by ∼20%, whereas PBMCs did not (Fig. 6,C). Pretreatment of fibrocytes with TGF-β1 (10 ng/ml) for 7 days before the assay further increased their contractile activity (Fig. 6 C). This increase in gel contraction by TGF-β1-treated fibrocyte cultures correlated to the enhanced expression of αSMA by fibrocytes in response to TGF-β1.

Previous studies have shown that fibrocytes, a distinct mesenchymal cell type that arises in ex vivo cultures of peripheral blood, exhibit both monocyte- and fibroblast-like characteristics (reviewed in Ref. 21). Fibrocytes initially were identified by their rapid and specific recruitment from the blood to s.c. implanted wound chambers in mice (2). Human fibrocytes were then shown to emerge from cultures of the PBMC fraction of whole blood after 1–2 wk (2). Cultured fibrocytes have been shown to mediate fibrosis (5), Ag presentation and immunity (3, 4), and angiogenesis (C. N. Metz, unpublished observations). In the present study we have examined the differentiation pathway of peripheral blood fibrocytes and explored the role of fibrocytes in wound repair.

Fibrocytes differentiated from an adherent population of CD14⁺-enriched peripheral blood cells when cultured in DMEM and FBS (with no additional growth factors). This differentiation process appears to require T cell interaction. Further studies will be necessary to identify the molecules involved in functionally significant interactions between T cells and fibrocytes that are required for fibrocyte maturation. Similarly, several studies have demonstrated the differentiation of CD1a⁺ dendritic cells from GM-CSF- and IL-4-treated CD14⁺ peripheral blood monocytes (22, 23, 24, 25). The T cells requirement observed for fibrocyte differentiation is reminiscent of the maturation of dendritic cells, also known for their ability to process and present Ags (26). Recent studies have shown that the final step in the maturation of dendritic cells occurs during their association with CD4⁺ T cells; this contact diminishes their responsiveness to IFN-γ by down-regulation of their IFN-γ receptors (27).

Interestingly, the addition of TGF-β1, a multifunctional cytokine that plays a central role in tissue repair and fibrosis, to crude fibrocyte-evolving cultures facilitated fibrocyte differentiation. The role of exogenous TGF-β in fibroblast proliferation and collagen production is well documented (reviewed in Ref. 28). TGF-β significantly up-regulates collagen expression by dermal fibroblasts in vitro (29), by myofibroblasts (30), as well as by proliferative scar xenografts in vivo (31). Many laboratories have confirmed that TGF-β plays a role in the natural wound healing process and that TGF-β is expressed in rodent wound chambers during the early to mid phases (days 4–7) of wound healing (32). Furthermore, in vivo gene transfer with TGF-β1 cDNA into the skin of rats significantly enhanced the rate of wound repair (33). Consistent with these prior observations, we postulate that circulating fibrocyte precursor cells interact with activated T cells, which permits their early differentiation (toward the fibrocyte phenotype), and they then migrate to the wound site (Fig. 7). Within the wound site, these early differentiated fibrocytes might further interact with recruited T cells and fully differentiate and mature following exposure to TGF-β. These fully differentiated, mature fibrocytes express increased levels of αSMA and produce collagen and other extracellular matrix proteins that promote wound healing and contracture.

Fibroblasts have been shown to exhibit increased collagen expression and other matrix components in certain fibrotic disease states (reviewed in Ref. 34). Investigators have previously implicated TGF-β overexpression in fibrosis of the skin (35) and lungs (35, 36). In addition, TGF-β overexpression has been associated with enhanced myofibroblast activity in animal models of pulmonary fibrosis (37). Our findings that TGF-β1 enhanced proliferation, collagen production, and αSMA expression by cultured fibrocytes potentially implicates this circulating cell type in TGF-β-dependent fibrotic responses in vivo.

A role for fibrocytes in wound healing and connective scar tissue formation has been postulated based on their accumulation in wound sites (2). However, the molecular signals that mediate the trafficking of fibrocytes to the wound has not yet been investigated. We examined chemokine receptor expression (mRNA and protein) by cultured enriched fibrocytes and revealed the presence of CCR3, CCR5, CCR7, and CXCR4 and the absence of CCR4, CCR6, and CXCR3. Further studies showed directed chemotaxis of cultured, enriched fibrocytes in response to the ligand of CCR7, SLC (also known as 6Ckine, Exodus-2, and TCA-4), in vitro and in vivo. SLC, a C-C chemokine family member, has been shown to be involved in the organization of lymphoid tissue during development by attracting T cells and mature dendritic cells (38). SLC expression has been observed in sites of inflammation (39). We observed SLC expression by the vascular endothelium within the wound sites. Based on these observations it would be interesting to examine the role of fibrocytes in wound responses using mutant mice lacking SLC expression (40, 41, 42).

The functional role of fibrocytes in wound healing has not been investigated previously. TGF-β has been shown to be the most important cytokine for the trans-differentiation of fibroblasts to contractile wound myofibroblasts that exhibit increased αSMA staining, elevated collagen secretion (reviewed in Ref. 19), and increased stress fibers (17) in response to TGF-β. Myofibroblasts are transiently found in early to mid phase wound tissue and have been proposed to exert a critical contractile force that is required to close wounds. Neither the origin of myofibroblasts nor any trafficking signals necessary for myofibroblast migration to injured tissue are well understood. Myofibroblasts have been postulated to derive from progenitor stem cells, resident tissue fibroblasts, or tissue smooth muscle cells. However, a plausible alternative is that myofibroblasts differentiate from a circulating, rather than a resident, precursor cell type.

In this paper we show that blood-borne, ex vivo cultured, precursor fibrocyte cells have the capacity to differentiate into αSMA⁺, TGF-β1-responsive fibrocyte cells that exhibit characteristics similar to those of wound-healing myofibroblasts. Differentiated fibrocytes and myofibroblasts share many common features: transient presence within the wound, production of numerous proinflammatory cytokines and growth factors, secretion of collagen and other extracellular matrix proteins, and enhanced collagen production in response to TGF-β1. Furthermore, we observed that cultured fibrocytes, like myofibroblasts, express αSMA protein that is enhanced by TGF-β1 treatment and exert a contractile force that would aid in reducing the amount of denuded surface area of wounded tissue. The question remains of whether fibrocytes and myofibroblasts are distinct populations. However, it is reasonable to suggest that fibrocytes derived from a circulating precursor population play an important role during the resolution and repair phase of wound healing.

We thank K. Manogue and J. Chesney for critically reading this manuscript.