Stromal Cell Progeny of Murine Bone Marrow ... - Semantic Scholar

Stromal Cell Progeny of Murine Bone Marrow Fibroblast Colony-Forming Units Are Clonal Endothelial-Like Cells That Express Collagen IV and Laminin By Steve Perkins and Roger A. Fleischman Studies of human and murine bone marrow explants have demonstrated the existence of stromal cell precursors that give rise to colonies of adherent cells in short-term cultures. Because previous data suggested that these colonies were composed of fibroblasts, the precursor cells were termed fibroblast colony-forming units (CFU-F). However, we have recently shown that the stromal cells which support hematopoiesis in murine long-term bone marrow cultures (LTBC) express collagen IV and laminin, markers associated with an endothelial cell lineage, but are negative for collagen I and 111, markers associated with a fibroblast cell lineage. Because these conflicting results suggest major functional differences between the stromal cells observed in long-term cultures and the short-term assay, we re-examined the lineage of CFU-F-derived stromal cells. Using two-color immunofluorescence, we characterized virtually all of the cells comprising individual "CFU-F" colonies derived from mouse radiation chimeras. Identification of donor (hematopoietic) or host (stromal) origin was based on surface staining for strain-specific H-2 surface antigens, and, for endothelial or fibroblast properties, on cytoplasmic staining for laminin and collagen IV, or collagens Iand 111, respectively. The results demonstrate that a large proportion of the cells in CFU-F colonies are donorderived and fail to stain with any of the antisera specific for

nonhematopoietic cells. In addition, these donor-derived cells exhibit marked phagocytic capacity and stain positively with monoclonal antibodies characteristic of the monocyte-macrophage hematopoietic cell lineage (antiT200, anti-Mac-1, F4/80). However, the remainder of the cells are host-derived cells that stain positively with antisera t o collagen IV and laminin. In contrast, stains for collagen types Iand 111 were negative under conditions that allowed for strong staining of control skin fibroblasts. In separate studies, using mixtures of two genetically distinct bone marrows, the cells expressing collagen IV were further shown to be clonal in origin within individual colonies, directly demonstrating that the CFU-F assay provides a quantitative measure of the numbers of marrow stromal cell precursors. Thus, the current studies establish a remarkable similarity between the hematopoieticmicroenvironment in the short-term CFU-F assay and the long-term culture system: the majority of adherent cells are hematopoietic cells of the monocyte-macrophage lineage, while the remainder are stromal cells whose precise lineage remains uncertain, but whose pattern of collagen expression is more consistent with an endothelial rather than a fibroblast cell origin. 0 1990 by The American Society of Hematology.

T

cally consistent with histiocytes and endothelial cells have been described by a number of investigator^.'.^ The description of these other cell types has therefore created uncertainty regarding the exclusively fibroblast nature of CFU-Fderived cells. Moreover, in another important model of the hematopoietic microenvironment, long-term bone marrow cultures (LTBC), the majority of stromal cells have recently been shown to express cytoplasmic collagen IV and laminin, markers consistent with an endothelial rather than a fibroblast cell l i ~ ~ e a g eBecause . ~ . ~ the stromal cells in LTBC also support long-term hematopoiesis, the discrepancy between the apparent cell lineage of LTBC- and CFU-F-derived stromal cells brings into question the relevance of CFU-F assays and CFU-F numbers for studies of hematopoiesis. In an attempt to clarify the relationship between these two systems, we used a previously described in situ two-color immunofluorescence assay4 to determine unequivocally the origin and lineage of the cells comprising individual CFU-F colonies. As in our previous studies, we used mouse radiation chimeras as a model system to distinguish contaminating hematopoietic cells of donor origin from stromal cells of host origin, identified by their surface immunofluorescence staining for strain-specific histocompatibility antigens. The hostderived stromal cells were further simultaneously characterized by staining for expression of lineage-specific cytoplasmic markers, such as collagens I, 111, IV, and laminin. Finally, in separate experiments, the clonality of the stromal cells was analyzed by immunofluorescence staining of individual colonies derived from mixtures of bone marrow from two strains carrying different histocompatibility antigens.

HE DEVELOPMENT of in vitro models of the hematopoietic microenvironment has facilitated studies of bone marrow stromal cells and their role in the regulation of hematopoiesis. One of the most widely used of these techniques, originally described by Friedenstein et al, detects individual colonies of adherent cells arising from short-term cultures of bone marrow explants.' Based on data that suggested a fibroblast lineage for these cells, the stromal precursor cells were originally termed fibroblast colonyforming units (CFU-F). However, in addition to fibroblasts, cells with marked phagocytic capacity and cells morphologi-

From the Division of Hematology/Oncology. Department of Internal Medicine. University of Texas Southwestern Medical Center at Dallas, TX. Submitted June 26, 1989; accepted October 2, 1989. Supported by a grant from the Nasher Family Cancer Research Program. S.P. is the recipient of a Clinical Oncology Career Development Award from the American Cancer Society. R.A.F. is the recipient of a John A . and George L. Hartford Fellowship. Presented in part at the 29th Annual Meeting of the American Society of Hematology, December 7 . 1987, Washington, DC, and published in abstract form (Blood 70:161a. 1987). Address reprint requests to R.A. Fleischman, IUD. PhD, Division of Hematology/Oncology. University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75235. The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C.section 1734 solely to indicate this fact. 0 1990 by The American Society of Hematology. 0006-4971/90/7503-0019$3.00/0

620

Blood, Vol75, No 3 (February 11, 1990: pp 620-625

6 21

COMPOSITION OF MARROW CFU-F

MATERIALS AND METHODS

Mice. All mice were obtained from Jackson Laboratories (Bar Harbor, ME). Eight- to twelve-week-old C3H females (H-2Kk/k) served as bone marrow donors and 2- to 4-month-old (C57BL/6 x C3H) F1 females (H-2Kb/') served as recipients for construction of radiation chimeras, prepared as described previously: H-2Kb was therefore unique to the recipients. Chimeras were sacrificed 2 to 6 months after engraftment. The H-2Kk/' and H-2Kblbstrains used in the mixing experiments were C3H and C57BL/6, respectively. CFU-F cultures. The method of Friedenstein et al' was used with the following modifications. Marrow suspensionswere prepared from donor mice as previously described: The cells were suspended at a final concentration of 106/mL in Iscove's modified Dulbecco's medium (Sigma Chemical Co, St Louis, MO) supplemented with 20% fetal calf serum (Hazelton Biologies, Lenexa, KS), penicillin 50 U/mL, and streptomycin 50 & n L . Cultures were incubated in a humidified atmosphere with 5% CO, at 37OC in 35-mm suspension dishes (Miles Scientific, Naperville, IL), each containing a sterilized 22-mm square glass coverslip (American Scientific, McGaw Park, IL), and 1 mL of the cell suspension. At 1 week all the medium and nonadherent cells were removed and replaced with fresh medium. The number of colonies obtained was between 10 and 20 per lo6cells plated, a number consistent with published data.' One to 2 days before immunofluorescence staining, 1,000 U/mL of ab-interferon (Lee Biomolecular Research Laboratories, Inc, San Diego, CA) was added to the cultures to enhance the expression of H-2K surface antigens? Surface and cytoplasmic immunofluorescence. Immunofluorescence was performed at 1 to 2 weeks in culture. Procedures and reagents for surface indirect immunofluorescence with the mouse monoclonal antibody (MoAb) specific for H-2Kb have been previously reported.4 This antibody, directed against the host-specific surface antigen, was able to distinguish donor-derived cells from host-derived cells with a high degree of accuracy in these CFU-F cultures, as reported previously for adherent cells in LTBC: Thus, the MoAb directed against the H-2Kb determinant stained more than 99% of the adherent cells derived from host strain mice (H-2Kklb),while none of the cells from donor strain mice (H-2Kk/') were stained (data not shown). Therefore, positive staining with this antibody unequivocally identified cells as host in origin, while negative staining identified donor cells with a high degree of accuracy (more than 99%). The reagents and methods for surface staining directed against T200 (Ly-5): Mac-1: and F4/80: and for cytoplasmic immunofluorescence for collagen IV, laminin, collagens I and 111, and fibronectin have also been previously described in detail: Additional antisera to collagens I and 111 were obtained from Chemicon International, Inc, El Segundo, CA, and sheep anti-human/calf collagen I was kindly provided by Dr Hynda Kleinman, National Institutes of Health, Bethesda, MD. Rat MoAb MECA-IO, raised against antigenic determinants on endothelial cells, was kindly provided by Dr E.C. Butcher (Stanford Medical School, Palo Alto, CA).'' Forty-eight hours before staining with the MECA antibodies, 100 U/mL of y-interferon (provided by Dr M. Ziff, University of Texas Southwestern Medical School, Dallas) was added to each culture. Collagen types I and IV (Collaborative Research, Bedford, MA), were used in blocking studies of the anti-collagen IV anti-serum at concentrations of 1.7 nmol and 0.8 nmol, respectively. Other methods. Procedures for phagocytic assays, cytochemistry, microscopy, and photography have been previously described: For the quantitative analysis, discrete colonies of between 100 and 1,000 cells were selected and all the cells in the colonies counted.

Exhaustive cell counts were not performed on colonies larger than 1,000 cells. However, their staining appeared to be similar to the more common smaller colonies. RESULTS

One to 2 weeks after explantation, the cells comprising individual CFU-F colonies were simultaneously typed for donor or host origin, and expression of cytoplasmic collagens and laminin, as described in Materials and Methods. In addition, the cells were also characterized for phagocytic capacity by incubation with small latex particles for 24 hours before antibody staining. The results of this analysis showed that CFU-F-derived colonies contain two distinct populations of cells: (1) donor-derived cells exhibiting marked phagocytic capacity, and (2) host-derived cells expressing cytoplasmic collagen IV and laminin (Table 1). As indicated in Table 1, more than 99% of the donorderived cell population in CFU-F colonies also stain positively with the macrophage-specific MoAb F4/80, as well as with MoAbs directed against Mac-1 and T200, antigens found exclusively on hematopoietic cells. As expected for hematopoietic cells, the donor-derived cells do not express a variety of cytoplasmic antigens normally associated with nonhematopoietic cells, such as collagens I, 111, IV, laminin, and fibronectin. Even in mice analyzed as early as 8 weeks after transplantation, no collagen-IV-positive cells of donor origin were observed. These results provide strong evidence that the donor-derived cells in CFU-F colonies are hematopoietic cells of the monocyte-macrophage lineage. The host-derived cells in CFU-F colonies, on the other hand, are nonphagocytic cells that express collagen IV and laminin, cell products associated with endothelial cells. An example of the cytoplasmic staining for collagen IV is shown in Fig 1A. In contrast, fibroblasts explanted from mouse skin Percentage of Lineage-Specific Markers Expressed by CFU-F-Derived Cells of Donor and Host Strain Origin

Table 1.

Cell Origin

Lineage Marker

Collagen IV, laminin. Collagen 1, Ill* T200. F4/80, Mac- 1 t MECA- 1O t Phagocytosis

D"

t0.5 e0.5 >99 c0.5 r95

Host

799 c0.5 c0.5 750

c5

The CFU-F colonies (n = 19) were derived from radiation chimeras (n = 5) that had been engafted for 2 to 6 months. Colonies ranged in size from 1 18 to 2,179 and in donor cells from 23% to 99% of the total cells. Data are expressed as the percent donor or host cells positive for a marker. *In these studies, host cells were identified by the expression of the strain-specific H-2Kb antigen. In a few chimeras, the finding of occaslonal phagocytic cells of host origin (less than 5%) was consistent with low levels of residual host hematopoiesis. To facilitate comparison with completely reconstituted chimeras, phagocytes of host origin were excludedfrom the quantitative analysis of collagen and laminin staining. tBecause all of the host-derived cells proved to be collagen-IVpositive in completely reconstitutedchimeras, in these studies the hostand donor-derived cells were identified by the presence or absence of collagen IV, respectively.

622

PERKINS AND REISCHMAN

Fig 1. Indirect immunoIluoreswnce mining of CFW cdonies (A and C) and control fibroblaem (B and D) with antisera t o collagen iV [A and B) and collagen I (C and D). The cytoplasmic immunofluorescence evident in (A) indicates the prwmce of cdlagen IV within these stromal cells of CFU-F. In contrast. explanted skin fibroblasts stained under identical conditions are negative for collagen IV (8).On the othw hand. tho CFU-F-derived stromal cells in (CI are negative for expression of intracellular collagen I. while control skin fibroblasts stain strongly for this marker (D). Occasional foci of extracellular collagen Iwere observed in some cultures, as shown in (C). Original magnification x630.

are entirely negative for expression of collagen IV (Fig IB). As an additional control to confirm the specificity of these results, blocking studies demonstrated that preincubation of the antiserum with purified collagen IV eliminated the staining of the host-derived cells in CFU-Fcolonies, while preincubation with collagen I did not. Just Over half of the host-derived cells also strongly express the antigen detected

by the MoAb MECA-IO. which stains endothelial cells in mouse thymus and spleen, as well as the endothelial-like stromal cells in mouse LTBC.' An additional 10% to 15% of the host-derived cells exhibited weak staining for MECA-IO. Despite prior data suggesting a fibroblast lineage for CFU-F-derived stromal cells, studies with a variety of antibody preparations directed against collagens I and 111

COMPOSITION OF MARROW CFU-F

623

624

PERKINS AND FLEISCHMAN

demonstrated only occasional small foci of extracellular staining and were entirely negative for intracellular staining. Figure 1C shows an example of this staining for collagen I. AS a positive control, fibroblasts explanted from mouse skin exhibit strong staining with this anti-collagen I antisera (Fig 1D). To determine if the cells comprising CFU-F colonies are clonal in origin, bone marrow cells from two genetically distinct mouse strains were mixed, and at 7 days of culture the CFU-F colonies were stained for strain-specific H-2 antigens and cytoplasmic collagen IV. As shown in Table 2, an analysis of individual colonies demonstrated that the collagen IV positive cells within an individual colony are derived exclusively from a single mouse strain in the majority of colonies examined. On the other hand, macrophages are consistently derived from both strains of mice. This result strongly suggests the clonal origin of the endothelial-like cell population and the polyclonal origin of the hematopoietic cells. Consistent with this conclusion is the observation that the cells staining for collagen IV tend to grow in large clusters, although more than one cluster can be observed in some colonies. The additional observation of occasional isolated collagen IV positive cells may account for the small numbers of nonclonal stromal cells found in a minority of the colonies.

as well as with anti-Mac-1 and anti-T200. These results are consistent with a monocyte-macrophage lineage for this cell population. Similar cells, characterized by morphology and phagocytic capacity, have been described in CFU-F from other species, although the number of such cells may be substantially smaller in canine and human CFU-F.”-13 The remainder of the cells comprising CFU-F colonies are host-derived cells that label strongly with antibodies to collagen IV and laminin, cell products associated with endothelial cells. On the other hand, control fibroblasts from mouse skin are negative for these markers, and blocking studies with collagen IV confirm the specificity of this stain. In addition, the host-derived cells in CFU-F colonies are negative for cytoplasmic collagen I and collagen 111, cell products associated with the fibroblast cell lineage, while control skin fibroblasts are strongly positive for these markers. These data, which suggest that the stromal cells in CFU-F colonies do not express markers of the fibroblast cell lineage, are inconsistent with some but not all previous Conceivably, differences in the specificity and cross-reactivity of antibody preparations to collagen I and 111 could account for the divergent observations, although a variety of anti-collagen antibodies were examined in the present studies and all yielded similar results. It should also be noted that the composition of human cultures, which may have relatively larger numbers of fibroblasts, may not be comparable with mouse cultures. In addition, we observed that in older cultures (more than 3 weeks) a few cells appeared to express low, but significant, levels of collagen I and 111, suggesting that culture conditions and age may influence the phenotype of the stromal cells in CFU-F colonies. More difficult to reconcile with a fibroblast cell lineage, however, are the positive results obtained here with cytoplasmic stains for collagen IV and laminin. Because the two-color immunofluorescence assay demonstrated that virtually every host-derived nonhematopoietic cell in CFU-F colonies is

DISCUSSION

To determine the lineage, origin, and transplantability of the cells derived from CFU-F precursors, we analyzed individual colonies cultured from chimeric mouse bone marrow by simultaneously staining for strain-specific H-2 antigens and lineage-specific cytoplasmic markers. The results of this analysis demonstrate that a major fraction of the cells in these colonies are donor-derived, and thus are transplantable cells of hematopoietic origin. Moreover, these donor-derived cells are predominantly phagocytic cells that stain uniformly with the macrophage-specificMoAb F4/80,

Table 2. Strain Origin of Collagen-IV-Positive and -Negative Cells in CFU-F Colonies, Derived From Mixtures of Marrow From T w o Mouse Strains Hematopoietic Cells Collagen IV (-)

Total Cells/

Colony

No.

319 522 792 306 946 147 435 67 0 540 127 26 1 566 198 137 538 495

138 452 733 247 838 122 342 512 494 107 230 556 160 97 404 473

Stromal Cells

Collagen IV ( % Strain E

34 53 57 62 62 66 65 59 31 25 53 56 89 5 73 56

No.

181 70 59 59 108 25 93 158 46 22 31 10 38 40 134 22

+) % Strain B

0 0

0 0 0 0 3 4 100 100 100 100 97 92 90 82

Only colonies with less than 1,000 cells were analyzed. Each count is from a different colony. The percent cells of strain B origin was determined by surface immunofluorescencestaining for the H-2Kbantigen. Results represent pooled data from two different experiments.

625

COMPOSITION OF MARROW CFU-F

collagen-IV-positive, cells with the collagen-IV-negative phenotype expected for fibroblasts, if present, must be a very small population of cells. In addition, the majority of these cells also stain with the MECA-10 MoAb, which labels endothelial cells in mouse hematopoietic organs and stromal cells in mouse long-term bone marrow cultures. However, as compared with LTBC, the intensity of staining with MECA-10 was uniformly reduced under CFU-F culture conditions. This observed decrease in the expression of the antigen recognized by the MECA-10 MoAb may account for the significant proportion of MECA-10 negative stromal cells in CFU-F, rather than indicating any fundamental difference between the stromal cells in CFU-F and LTBC. The mixing studies described in this report further demonstrate that within individual colonies and under in vitro conditions the stromal cell components of CFU-F are clonal in origin. This result confirms the previous assumption, based largely on the linear relationship between numbers of cells plated and observed colonie~,'~ that the CFU-F assay provides a quantitative measure of stromal precursors. However, in contrast to the clonality of the stromal cells, the hematopietic cells within individual CFU-F are clearly polyclonal in origin. Presumably this finding results from in vitro aggregation of macrophage precursors and/or from preferential macrophage proliferation in or migration to sites of stromal cell colonies. Using a similar approach to that described here, we recently characterized the adherent cells in long-term bone marrow c ~ l t u r e s .Indeed, ~ the relationship between the adherent cells in LTBC and the cells comprising CFU-F

colonies has previously been unclear. For example, the hematopoiesis observed in LTBC has been shown to be dependent on the presence of adherent stromal cells. On the other hand, hematopoiesis is not sustained in short-term cultures, although heterotopic transplantation of CFU-Fderived cells has been reported to support hematopoiesis in vivo.I4 In this regard, the current studies demonstrate a remarkable similarity between the cells comprising the hematopoietic microenvironment of the short-term CFU-F assay and the long-term culture system. Thus, in both cases the majority of adherent cells are hematopoietic cells of the monocyte-macrophage lineage, while the nonhematopoietic cells express collagen IV and l a m i ~ ~ i nFurthermore, .~ as previously shown for the stromal cells in LTBC; the stromal cells in CFU-F are exclusively host in origin in radiation chimeras, and thus are not transplantable or derived from the hematopoietic stem cell, as previously sugge~ted.'~'''Finally, since the stromal cells within each colony are clonal, the CFU-F assay is shown to be a valid quantitative measure of the stromal cell precursors whose progeny are the predominant stromal cell population in both CFU-F and LTBC. The question whether highly purified preparations of these stromal cells can support hematopoiesis in vitro is now under investigation. Such experiments should help to define the role of the bone marrow microenvironment in the maintenance and regulation of hematopoiesis. ACKNOWLEDGMENT

We thank Drs R.G. Smith, E.P. Frenkel, and P.L. Witte for their helpful suggestions.

REFERENCES 1. Friedenstein AJ, Chailakhjan RK, Lalykina KS: The develop-

ment of fibroblast colonies in monolayer cultures of guinea-pig bone marrow and spleen cells. Cell Tissue Kinet 3:393,1970 2. Piersma AH, Brockbank GM, Ploemacher RE, Van Vliet E, Brakel-van Peer KMJ, Visser PJ: Characterization of fibroblastic stromal cells from murine bone marrow. Exp Hematol 13:237,1985 3. Xu CX, Hendry JH, Testa NG, Allen TD: Stromal colonies from mouse marrow: Characterization of cell types, optimization of plating efficiency and its effect on radiosensitivity. J Cell Sci 61:453, 1983 4. Perkins S,Fleischman RA: The hematopoietic microenvironment: Origin, lineage, and transplantability of the stromal cells in long-term bone marrow cultures from chimeric mice. J Clin Invest 81:1072,1988 5. Simmons PJ, Przepiorka D, Thomas ED, Torok-Storb B: Host origin of marrow stromal cells following allogeneic bone marrow transplantation. Nature 328:429,1987 6. Pober JS, Jimbrone MA, Cotran RS, Reiss CS, Burakoff SJ, Fiers W, Ault KA: la expression by vascular endothelium is inducible by activated T cells and by human gamma interferon. J ExpMed 157:1339,1983 7. Ledbetter JA, Herzenberg LA: Xenogeneic monoclonal antibodies to mouse lymphoid differentiation antigens. Immunol Rev 4753, I979 8. Springer T, Galfre G, Secher DS, Milstein C: Mac-1: A macrophage differentiation antigen identified by monoclonal antibody. Eur J Immunol9:301,1979 9. Austyn JM, Gordon S: F4/80, a monoclonal antibody directed

specifically against the mouse macrophage. Eur J Immunol 11:805, 1981 10. Duijvenstijn AM, Schreiber AB, Butcher EC: Interferongamma regulates an antigen specific for endothelial cells involved in lymphocyte traffic. Proc Natl Acad Sci USA. 83:9114,1986 11. Wilson FD, Tavassoli M, Greenberg BR, Hinds D, Klein AK: Morphologic studies on adherent cells in bone marrow cultures from humans, dogs and mice. Stem Cells 1:15,1981 12. Klinnert V, Northdurft W, Fliedner TM: CFU-F from dog marrow: A colony assay and its significance. Blut 50:81, 1985 13. Castro-Malaspina H, Gay RE, Resnick G, Kapoor N, Meyers P, Chiarieri D, McKenzie S, Broxmeyer HE, Moore MAS: Characterization of human bone marrow fibroblast colony-forming cells (CFU-F) and their progeny. Blood 56:289,1980 14. Friedenstein AJ, Chailakhjan RK, Latsinik NV, Panasyuls AF, Keilis-Borok IV: Stromal cells responsible for transferring the microenvironment of the hemopoietic tissues. Cloning in vitro and retransplantation in vivo. Transplantation 17:331, 1974 15. Quesenberry PJ, McNiece IK, Robinson BE, Woodward TA, Baber G, McGrath HE, Isakson PC: Stromal cell regulation of lymphoid and myeloid differentiation. Blood Cells 13:137,1987 16. Piersma AH, Ploemacher RE, Brockbank KGM: Transplantation of bone marrow fibroblastoid stromal cells via the intravenous route. Br J Haematol54:285, 1983 17. Keating A, Singer JW, Killen PD, Striker GE, Salo AC, Sanders J, Thomas ED, Thorning D, Fialkow PJ: Donor origin of the in vitro hematopoietic microenvironment after marrow transplantation in man. Nature 298:280,1982