Printed in Great Britain. All rights reserved 0145-2126/94 $7.00 + 0.00
CHARACTERIZATION OF BONE MARROW STROMAL CELLS FROM M U L T I P L E MYELOMA
MARIA GRAZIA GREGORETTI,* DANIELA GOTTARDI,* PAOLO GHIA,* LUCIANA BERGUI,* FRANCA MERICO,* PIER CARLO MARCHISIOt and FEDER1CO CALIGARIS-CAPPIO*
*Sezione Clinica; and +Sezione di Istologia, Dipartimento di Scienze Biomediche e Oncologia Umana. Universitfi di Torino, Torino, Italy
(Received 21 March 1994. Revision accepted 30 April 1994)
Abstract--We have cultured multiple myeloma (MM) bone marrow (BM) stromal cells that are able to sustain the in vitro growth of monoclonal B-cells. Our aim was to evaluate which adhesion molecules are expressed and which extracellular matrix proteins are produced by these cells and whether they differ from the stromal cells that can be grown under the same experimental conditions from the BM of monoclonal gammopathies of undetermined significance (MGUS) and of normal donors.
MM BM stromal cells that support malignant B-cell development have a striking proliferative ability that is absent in MGUS and normal donors of the same age group and are formed by four major different cell populations. Two kinds of HLA-DR+, CD10+ fibroblast-like cells can be recognized through the expression (or the lack) of u-smooth muscle actin isoform; further, macrophages and osteoclasts can be identified. Fibroblast-like cells that express o~-smooth muscle actin isoform, often organized along stress fibers in a periodic fashion, may be considered as myofibroblasts. Fibroblast-like cells react strongly with antibodies to CD54 (ICAM-1), integrin i l l , fi3, f15 and some of associated achains. Integrin fll is diffusely exposed on the surface while f13 is clustered in focal contacts in association with vinculin. A still undetermined subpopulation of fibroblasts is highly positive for ~xvfl5 that is clustered at focal contacts as shown by association with stress fiber termini and by interference reflection microscopy. A major difference between MM and normal donor BM stromal cells involves lower deposition and simpler organization of the extracellular matrix proteins (fibronectin, laminin, collagen type IV) deposited by MM fibroblast-like cells. CD14+ macrophages from MM, MGUS and normal donor BM are CD11a+ (o3_), CD11 b+ (aM), CD11 c+ (o~X), CD54+ (ICAM-1), CD56+ (N-CAM), fll and f12 (CD18) integrin positive. The integrin fll is diffusely expressed on the surface, while f12 is concentrated in podo- somes. MM osteoclasts show a weak diffuse staining with CD54 and CD56 MoAbs; fll integrin has a diffuse surface expression, while f13 integrin is concentrated in the podosomes. Normal donor osteoclasts are CD54- and the staining with CD56 is barely visible.
These findings lead us to suggest that the microenvironment provided by MM BM may be significantly different from that of normal BM indicating its potential role in controlling the local proliferation and differentiation of malignant B-lineage cells.
Correspondence to: Prof. F. Caligaris-Cappio, Cattedra Immunologia Clinica, Via Genova 3, 10126 Torino, Italy (Tel: 39-11-6637230-6964579; Fax: 39-11-6637238).
675
a properly organized meshwork of heterogeneous adherent elements that produce extracellular matrix glycoproteins and provide essential matrix-associated growth signals (reviewed in ref. [1]). The Whitlock and Witte technique has allowed the characterization of BM stromal cells that favor the selective growth of B-lineage cells in the mouse [2]. The surface mol- ecules involved in the adhesive interactions between murine stromal cells and B-cell precursors have thus been defined and the essential B-ceU growth-pro- moting role of a variety of cytokines such as IL-7 identified [3-5].
In humans, BM culture systems that support the growth and differentiation of B-lineage cells have
676 M.G. GREGORETTI et al.
been established [6, 7], but a proper human equiv- alent of the Whitlock and Witte technique is still lacking. The knowledge of human BM stromal cells supporting B-cell lymphopoiesis is therefore frag- mentary. As evidence is accumulating that abnor- malities of the BM microenvironment are involved in the pathophysiology of human B-cell malignancies [8, 9], the study of the relationships between BM microenvironment and malignant B-cell growth may help in understanding the relevant features of BM stromal cells involved in B-lymphopoiesis. To this end, multiple myeloma (MM) is a valuable model of investigation for at least three reasons: (i) in contrast with the distribution of normal plasma cells, MM malignant plasma cells accumulate only within the BM [10] indicating that BM stromal cells provide a unique microenvironment suitable for the growth of malignant plasma cells; (ii) while conventional B-cell surface markers are almost absent at the plasma cell level, a large array of adhesion molecules are expressed on the surface of MM malignant plasma cells [11-16]; and (iii) a heterogeneous population of adherent cells can be grown in vitro from the BM of patients with MM: these 's t romal ' elements, used as a feeder layer for autologous peripheral blood (PB) mononuclear cells, support the growth of mono- clonal B-lineage cells from circulating precursors [17].
We have used this approach [17] to characterize the BM stromal cells that favor malignant B-cell growth in MM in order to answer the questions: (i) which adhesion molecules are expressed and which extracellular matrix proteins are produced by these cells; and (ii) do they differ from the stromal cells that can be grown under the same experimental con- ditions from the BM of monoclonal gammopathies of undetermined significance (MGUS) and of normal donors. To unravel the heterogeneity of human BM stromal cells we have analyzed the different cell populations grown from MM, M G U S and normal BM at the individual cell level by immunofluo- rescence and a combination of antibodies (Abs). We have found that normal and malignant BM stromal samples markedly differ in cell organization. These differences involve lower deposition and simpler organization of the extracellular matrix deposited by MM stromal cells, as well as the integrin expression as witnessed by the appearance and clustering of the integrin heterodimer a'vfl5 at focal contacts in a subpopulation of fibroblast-like cells derived from MM patients. Thus, the microenvironment provided by MM BM may be significantly different from that of normal BM indicating its potential role in con- trolling the local proliferation and differentiation of malignant B-lineage cells,
Materials and Methods
Cells
BM aspirate was obtained after informed consent from 19 patients with MM, age range 48-72, six patients with MGUS, age range 55-70, and six normal donors. MM patients were four stage I, three stage II and 12 stage III according to Durie & Salmon [18] and were studied either at diagnosis before the initiation of treatment (11 cases) or at relapse (8 cases). MGUS patients had a mean follow-up of 5 years. The normal controls for BM stromal cell cultures were BM aspirates from young subjects (<25 years of age), undergoing BM biopsy as a part of clinical staging for Hodgkin's disease who had no histological evidence of BM localizations. The reason for selecting young subjects was that, in 12 preliminary experiments, stromal cells from normal donors matched for age with MM were cultured concurrently with the myeloma samples using the same serum. All normal specimens died out shortly after the initiation of culture and did not yield any growth. In each MM and MGUS patient PB samples were obtained con- comitantly with the BM aspirate.
Ficoll-Hypaque (FH) separation of BM samples was carefully avoided and nucleated cells were obtained on methylcellulose gradient. Cells were washed twice with phosphate buffered saline (PBS), resuspended in Dulbecco's medium and 5% horse serum (HS), and cul- tured immediately. PB mononuclear cells (MC) were sep- arated on FH and frozen in liquid nitrogen.
Cell cultures
BM cells were layered onto coverslips in 24 well culture plates or in chamber slides (Nunc, Inc. Naperville III) at a concentration of 2 x 10 6 cells/ml and cultured at 37°C in 5% CO, for 7 weeks. Half of the spent medium was replaced by an equal amount of fresh medium every 4 days. After 7 weeks of culture, a confluent layer of adherent cells was apparent and no residual myeloid or lymphoid cells could be observed either by morphology or upon staining with CD19, CD3, anti-y, ~t, o:, ~¢, A Abs. In each case, half of the wells were used to characterize adherent cells and half were used as feeders for autologous PBMC (2 × 106/ ml) which were thawed, washed twice with PBS and seeded over the stromal cell layer. As described in detail [17], the BM-PBMC co-cultures were run for 3 weeks at 37°C in 5% COz, supplementing the medium with 20% fetal calf serum (FCS) and replacing every 4 days half of the spent medium with an equal amount of fresh medium.
Cell phenotype
The morphology of stromal cells growing in culture was studied on May-Grunwald-Giemsa (MGG)-stained cover- slips. To identify fibroblasts, macrophages and osteoclasts in culture we used the cytoskeletal organization of F-actin as a marker, as it has a distinct distribution pattern on coverslip-attached cells [19]. Coverslips were fixed with 3% formaldehyde, stained as described [20] with rhodamine- labeled phalloidin (R-PHD, Sigma Chemicals, St. Louis, MO) (which selectively binds to any isoform of F-actin) and directly observed in the fluorescence microscope. In some experiments coverslips were stained with fluorescein- labeled phalloidin (F-PHD, Sigma Chemicals, St. Louis, MO) and a monoclonal (M) Ab that reacts with a smooth muscle actin isoform (aSM1; a kind gift of Prof. G. Gabbi- ani, University of Geneva, Switzerland) revealed by a rhodamine-labeled secondary Ab.
Bone marrow stroma in multiple myeloma 677
FIG. 1. BM stromal cells in MM. (A) An osteoclast-like cell revealed by MGG staining. (B) Cytoskeleton organ- ization of F-actin assessed by F-PHD: fibroblast-like cells (large arrow) show F-actin in stress fibers, macrophages (small arrow) in podosomes and osteoclasts (large arrow- head) in podosomes arranged in a ring-like fashion at the periphery of the cell. Three-week-old co-cultures of MM BM stromal cells and autologous PBMC show a large number of lymphoid cells and plasma cells that express the monoclonal Ig light chain (C) and adhere to the stromal
cell layer (D). Scale bar: 10 p.
Immunofluorescence (IF) was used to identify lym- phocytes and monocytes and to investigate the major phenotypic features of stromal cells. The range of Abs used included polyclonal anti-human y, ~t, o~, r , )t Abs directly conjugated with tetraethyl-rhodamine-isothiocyanate (TRITC; Dakopatts, Glostrup, Denmark), anti-HLA-DR MAb (Becton Dickinson, Mountain View, CA), and MAbs of CD10 (anti-Calla, Becton Dickinson), CD14 (MO2, Coulter), CD19 (B4, Coulter, Hialeah, FL), CD23 (Coul- ter) and CD38 (A10; ref. [21]) clusters.
To analyse which extracellular matrix glycoproteins were produced by BM stromal cells, MAbs to fibronectin (MoAb IST9---ref. [22] kindly provided by L. Zardi, IST Genova), laminin (Sigma Chemical Co., St. Louis, MO, L8271) and collagen type IV (Heyl GmBH, Berlin, Germany) were used.
The adhesion molecules expressed by the different cell populations were evaluated by means of MAbs to the following molecular structures: CDl la (o;L; LFA-1, Dak- opatts), CDl lb (aM; Mol, Coulter Clone, Hialeah, FL), CDllc (o~X; Leu-M5, Becton Dickinson), CD18 (/32 inte- grin) (CLB LFA1/1, kindly supplied by Dr P. A. T. Tetteroo, Central Laboratories of The Netherlands Red
Cross, Amsterdam), CD44 (Becton Dickinson), CD54 (CL203.4 clone, kind gift of Dr S. Ferrone, New York Medical College, Valhalla, New York) which detects the ICAM-1 molecule, CD56 (Becton Dickinson), which detects the N-CAM molecule,/31 integrin (A1A5, kind gift of Dr M. Hemler, Dana Farber Cancer Inst., Boston, MA), 133 integrin (anti-IIIa, kind gift of Dr W. Knapp, Institute of Immunology, Wien, Austria), chicken gizzard vinculin (clone VIN-11-5, Sigma) that strongly cross-reacts with mammalian vinculin. Some coverslips were also immunostained for VCAM-1 (BBIG-V4; a gift of Dr E. Dejana, Istituto Mario Negri, Milano, Italy), VLA 4 (MAb HP1/12; ref. [23]),/35 (MAb IA9; a gift of Dr R. Pasqua- lini, Dana Farber Cancer Inst., Boston, MA) and/36 (MAb E7P6; a gift of Dr D. Sheppard, University of California, San Francisco).
MAbs were employed in indirect IF as culture super- natants at 1 : 5 dilution or as purified ascitic fluid proteins (1 mg/ml at 1 : 1000 final dilution) and revealed by rabbit- anti-mouse-Ig-TRITC or -FITC (Dakopatts) as previously described [20]. Isotype-matched Abs were used as controls. Slides were viewed in a Zeiss Axiophot epifluorescence microscope equipped with a planapochromat 63 x/1.4NA
678 M.G. GrEoOrESrrl et al.
oil immersion lens and photographed with Kodak T-Max 400 films.
Results
General properties of BM stromal cells grown from MM, MGUS and normal donors
In all MM samples examined the culture system allowed the easy and steady growth of BM stromal cells. After 7 weeks of culture a confluent layer of fibroblast-like cells was observed which were inter- spersed with macrophages (10-15% of the total cell number) and with variable percentages of osteoclast- like cells (Fig. I(A) and (B)). The number of osteo- clasts present in the culture was roughly proportional to the stage of the disease and the extent of lytic lesions observed in the patient's axial skeleton. Once the confluence was reached, primary cultures could be kept alive and functionally capable of supporting PBMC for more than 3 months. However, secondary cultures were never obtained: this failure may be, at least partly, explained by the extremely tight adhesion that stromal cells have with the vessel wall. After 3 weeks of BM/PBMC co-culture variable pro- portions of monoctonal lymphoid cells and plasma cells were tightly adherent to the stromal cell layer
in all MM samples studied confirming previous obser- vations (Fig. I(C) and (D) and ref. [17]).
None of the MGUS patients yielded a stromal cell layer of comparable density: sparse fibroblasts with occasional macrophages were seen, while osteoclasts were very rare. Co-cultures of BM stromal cells and PBMC from patients with MGUS revealed no obvi- ous enrichment in monoclonal B-lymphocytes and plasma cells.
Normal donors age-matched with MM and MGUS patients did not grow. On the contrary, the growth pattern of BM stromal cells from young normal donors was similar to that of MM: the difference was a significantly lower number of osteoclast-like cells (not shown).
Phenotypic properties of BM stromal cells grown from MM, MGUS and normal donors
The cytoskeleton organization of F-actin, which has a distinct distribution pattern in fibroblasts, macrophages and osteoclasts (Fig. I(B); see also refs [17, 19D, was used to scan in double staining experiments the major phenotypic features of coverslip-attached cells. Two different populations of fibroblast-like cells were defined on the bases of the expression of the c~-smooth-muscle actin isoform.
FIG. 2. Long term cultures of MM BM show two different populations of fibroblast-like cells. F-PHD staining (A) reveals the fibroblastic nature of the cells, about half of which express the ol-smooth muscle actin isoform often organized along stress fibers in a periodic fashion (B, arrow). Osteoclast-like cells (C) assemble the ol-smooth muscle isoform in short microfilamentous structures (D, arrow) not associated with podosomes (C, large arrow- head). The picture in A is deliberately out of focus to indicate that cells are often organized in superimposed
layers. Scale bar: 10 #.
Bone marrow stroma in multiple myeloma 679
It was found that about half of the fibroblastoid cells do express the ~v-smooth-muscle actin isoform (Fig. 2(A) and (B)) and may indeed be considered as myofibroblasts [25-27]. Moreover, most macro- phages and osteoclast-like cells synthesize and assemble this actin isoform in short microfilamentous structures of unknown function (Fig. 1(C) and (D)). This pattern of heterogeneous actin expression was identical in BM stromal cells obtained from different sources (MM, MGUS, normal donors).
Irrespective of the source, fibroblastoid cells were also HLA-DR+ and weakly CD10+; macrophages were HLA-DR+, CD14+, weakly CD10+; osteo- clasts were HLA-DR+, weakly CD10+, CD14+. BM stromal cells were uniformly CD38-, while a tiny (around 5%) population of small CD23+ elements could be seen (not shown).
Extracelhdar matrix and adhesion molecules expressed by BM stromal cells
The extracellular matrix molecules secreted and organized by BM stromal cells from different sources (MM, MGUS, normal donors) were evaluated with a combination of specific antibodies. In all samples, fibroblasts were found to lie on a complex matrix of fibronectin, laminin and collagen type IV indicating
that high levels of these extracellular matrix proteins were produced and organized by these cells. The amount of extracellular matrix proteins observed in MM BM samples was significantly lower than that observed in normal samples and also revealed a looser organization.
The cell populations found in our preparations were studied for the expression of integrins and some of their counter-receptors. MM fibroblasts reacted strongly with antibodies to CD54 (ICAM-1), integrin [31, [33, f15 and some of associated a' chains. The subcellular localization of fi chains was different, as integrin [31 was apparently diffusely exposed on the surface while fi3 was clustered in focal contacts in association with vinculin (not shown). In approxi- mately half of the fibroblast-like cells strong reaction was detected with mAb IA9 to fi5 at focal contacts, that is at the endings of stress fibers (Fig. 3(A) and (B)). This association of f15 with focal contacts was identified also by interference reflection microscopy (Fig. 3(C) and (D)). Apparently no fi5 was produced and clustered by osteoclast-like cells (Fig 3(E) and (F)) that, instead, organize aVfi3 within the podo- some structures. So far, we have not yet identified the subpopulation of fibroblast-like cells that express and use cvvfl5 via focal contacts.
FIG. 3, f15 expression by MM BM stromal cells. Fibroblast- like cells, revealed by R-PHD staining (A), express /35 integrin in focal contacts at the end of stress fibers (B, arrows). The association of/35 with focal contacts is shown by interference reflection microscopy (C, D). Osteoclast- like cells (E, arrow) do not express/35 integrin (F, arrow), while the fibroblast-like cells are /35 positive. Scale bar:
10u.
680 M.G. GREOORETrl et al.
CD14+ macrophages from MM BM were C D l l a + (crL), C D l l b + (aM), C D l l c + (crX), CD54+ (ICAM-1), CD56+ (N-CAM), /31 and /32 (CD18) integrin positive. The integrin/31 was diffuse on the surface, while /32 was concentrated in podosomes. All macrophages were/33 and/35 negative.
MM osteoclasts showed a weak diffuse staining with CD54 and CD56 MoAbs. /31 integrin had a diffuse surface expression; /33 integrin was con- centrated in the podosomes.
The rare stromal cells which could be grown from the BM of MGUS had the same features. Likewise, fibroblasts and macrophages growing in cultures from normal donor BM had the same general pattern of adhesion molecules, even if the intensity of expression was lower. Normal donor osteoclasts were CD54- and revealed a staining with CD56 weak or barely visible. Irrespective of the source, stromal cells were uniformly negative for VCAM-1, VLA-4 and/36 Abs.
Discussion
The rationale for the present study is that stromal cells play a crucial role in the growth of plasma cell tumors. A stromal feeder layer is required for the survival and expansion of murine plasmacytoma [24]. Monoclonal B-lineage cells develop when peripheral blood mononuclear cells from MM patients are seeded in vitro onto a feeder layer of autologous BM stromal cells [17]. The clinical observation that MM is a disease localized uniquely within the BM micro- environment [10] suggests that BM stromal cells have some distinctive features that favor a close relation- ship in vivo with the malignant B-cell clone. In this paper we have identified the extracellular matrix proteins and the adhesion molecules that are expressed by human BM stromal cells from MM patients; we have then compared them with the stro- mal cells from the BM of MGUS patients and normal donors that have been grown under the same exper- imental conditions.
The growth capability of BM stromal cells from MM differs from that of stromal cells from MGUS. It is similar to the growth pattern of young normal donors, while the growth potential of MGUS BM stromal cells is definitely lower. Since MGUS sub- jects belong to the same age group as MM patients and since BM stromal cells from old normal donors do not grow, it may be concluded that, in MM, BM stromal cells have a growth ability that is unusual for that age group.
The stromal cells that can be grown from the BM of MM patients and are able to sustain the growth of monoclonal B-cells are formed by four major
different cell populations. Two kinds of fibroblast- like cells can be recognized through the expression (or the lack) of c~-smooth muscle actin isoform; further, macrophages and osteoclasts can be ident- ified. Fibroblast-like cells that express or-smooth muscle actin isoform may be considered as myofi- broblasts [25-27]. The biology of myofibroblasts is still elusive [27]: their major trait is their contractile activity and possibly also their motility. It is plausible to suggest that their role in the BM might be to help the egress of mature cells from the intersinusoidal spaces into the circulation. Interestingly, oc-smooth muscle actin isoform positive 'myoid' cells [26] are observed in human trephine biopsies and their num- ber is increased during development, in the presence of an inflammatory process or in malignancies [25, 26]. Likewise, the murine likely equivalent of human myoid cells (barrier cells, ref. [28]) are increased when hemopoiesis is stimulated.
A major finding of this investigation is that the complex avfi5 is expressed by a subset of BM fibro- blasts, even if we still do not have any evidence that they may be related to myofibroblasts. Interestingly, avfl5, that does not associate with focal contacts in tumor epithelial cells that do not express f13 [29], is indeed clustered in focal contacts in this particular cell population of mesenchymal origin (Fig. 3).
BM stromal cells from MM have a looser extra- cellular matrix environment and are better equipped with adhesion molecules like integrins, CD54, CD56 as compared with BM stromal cells from normal donors. It is of interest that osteoclasts express a whole range of adhesion molecules. Several examples indicate that the control of cell differentiation and gene expression is mediated through adhesive inter- actions with the extracellular matrix [1] and there is evidence emerging that growth factors are func- tionally active when bound to extracellular com- ponents like proteoglycans and laminin [30]. Further, a strict interplay exists between cytokines and cell adhesion molecules: cytokines may regulate cell adhesion which, in return, may modify the cellular response to cytokines (reviewed in refs [30] and [31]). On these bases it may be asked which relation- ships BM stromal cells and, more specifically, osteo- clasts may have with the growing plasma cell clone. The following points are relevant to define the plasma cell/microenvironment interactions in MM. First, MM plasma cells express a wide range of adhesion structures, including the proteoglycan syndecan [16], a receptor for hyaluronan-mediated motility [15], CD44 [14], CD56 [11, 12] and frequently also C D l l / CD18+ [13], that may find their natural ligands on the surface of BM stromal cells. Second, MM plasma cells also release cytokines, including IL-lfl [32] and
Bone marrow stroma in multiple myeloma 681
M-CSF [33] that are able to activate BM stromal cells to produce IL-6 [34]. The adhesion of MM cell lines to BM stromal cells alone appears to stimulate the secretion of IL-6 [35]. MM BM stromal cells actively secrete IL-6 as well as IL-8 [36, 37], while IL-6 pro- duction by M G U S stromal cells is definitely lower [37]. It is still controversial whether normal BM stromal cells secrete IL-6 in the absence of stimu- lation [34-39] or whether the production is induced only after activation with inflammatory mediators [34] and what role the culture conditions may have in this process. IL-6 has a well-defined plasma cell growth promot ing activity [40] and IL-8 is able to upregulate the expression of adhesion molecules [41]. Thirdly, a number of cytokines released within the BM microenvironment by plasma ceils or by acti- vated stromal cells or by activated T-cells, including IL-1/3, TNF/3, M-CSF, IL-3 and IL-6 also behave as osteoclast activating factors (OAF: refs [42]-[48]).
We have no explanation of why MM BM stromal cells are activated in vitro in contrast with MGUS and age-matched controls. The possibility that mono- clonai tumor cells may persist after 7 weeks of culture and sustain the activated BM microenvironment through the production of key cytokines cannot be formally ruled out as we have not used polymerase chain reaction techniques to identify residual tumor cells. Nevertheless, this possibility is highly unlikely, as cultures were scanned in IF with anti-light chain Abs and always proved negative. Therefore, even if residual tumor cells still persist after several weeks of culture, their number is exceedingly low. We speculate that MM BM stromal cells, when cultured, have already been pr imed in v i vo and are thus able to produce the amount of activating cytokines that lead to stromal cell cross-talk and ensuing growth.
In conclusion, our findings lead to a possible scen- ario based upon the mutual interactions between plasma cells and BM stromal cells that are kept in close physical contact by adhesion molecules and are reciprocally influenced by cytokines. These inter- actions may not only promote malignant cell growth and osteoclast activation but also confine plasma cells within the BM.
Acknowledgements--This work was supported by grants from Associazione Italiana Ricerca sul Cancro and Pro- getto Finalizzato ACRO (CNR). M.G.G. is a recipient of a fellowship from F.I .R.C.D.G. is a recipient of a fellowship from A.I.R.C. The secretarial assistance of Ms Giuliana Tessa is gratefully acknowledged.
References
1. Juliano R. L. & Haskill S. (1993) Signal transduction from the extracellular matrix. J. Cell. Biol. 120, 577.
2. Whitlock C. A. & Witte O. N. (1982) Long term cultures of B-lymphocytes and their precursors from murine bone marrow. Proc. natn. Acad. Sci. U.S.A. 79, 3608.
3. Kinkade P., Lee G., Pietrangeli C. E., Hayashi S. H. & Gimble J. (1989) Cells and molecules that regulate B lymphopoiesis in bone marrow. A. Rev. Immun. 7, 111.
4. Namen A. E., Lupton S., Hjerrilk K., Wagnall J., Mochizuki D. Y., Schmierer A., Mosley B., March C., Urdal J., Gillis S., Cosman D. & Goodwin R. G. (1988) Stimulation of B-cell progenitors by clonal murine interleukin 7. Nature 333, 571.
5. Melchers F., Karasuyama H., Haasner D., Bauer S., Kudo A., Zakaguchi N., Jameson B. & Rolink A. (1993) The surrogate light chain in B-cell development. Immunology Today 61, 60.
6. Wolf M. L., Buckley J. A., Goldfarb A., Law C. L. & LeBien T. W. (1991) Development of a bone marrow culture for maintenance and growth of normal human B cell precursors. J. lmmun. 147, 3324.
7. Kaisho T., Oritani K., Ishikawa J., Tanabe M., Muraoka O.. Ochi T. & Hirano T. (1992) Human bone marrow stromal cell lines from myeloma and rheumatoid arthritis that can support murine pre-B cell growth. J. bnmun. 149, 4088.
8. Manabe A., Coustan-Smith E., Behm F. G., Raimondi S. C. & Campana D. (1992) Bone marrow-derived stromal cells prevent apoptotic cell death in B-lineage acute lymphoblastic leukemia. Blood 79, 2370.
9. Caligaris-Cappio F., Gregoretti M. G., Ghia P. & Bergui L. (1992) In vitro growth of human multiple myeloma: implications for biology and therapy. Hemat. Oncol. clin. N. Am. 6, 257.
10. Barlogie B., Epstein J., Selvanayagam P. & Alexanian R. (1989) Plasma cell myeloma: new biological insights and advances in therapy. Blood 73, 865.
11. Van Camp B., Durie B. G. M., Spier C., DeWaele M., Riet I. van, Vela E., Frutiger Y., Richter L. & Grogan T. M. (1990) Plasma cells in multiple myeloma express a natural killer cell-associated antigen: CD56 (NKH-1; Leul9). Blood 76, 377.
12. Drach J., Gattringer C. & Huber H. (1991) Expression of the neural cell adhesion molecule (CD56) by human myeloma cells. Clin. Exp. Immun. 83, 418.
13. Ahsmann E. J. M., Lokhorst H. M., Dekker A. W. & Bloem A. C. (1992) Lymphocyte function-associated antigen-1 expression on plasma cells correlates with tumor growth in multiple myeloma. Blood 79, 2068.
14. Uchiyama H., Barut B. A., Chauhan D., Cannistra S. A. & Anderson K. C. (1992) Characterization of adhesion molecules on human myeloma cell lines. Blood 80, 2306.
15. Turley E. A., Belch A. J., Poppema S. & Pilarski L. M. (1993) Expression and function of a receptor for hyaluronan-mediated motility on normal and malig- nant B-lymphocytes. Blood 81,446.
16. Ridley R. C., Huiquing X., Hata H., Woodliff J., Epstein J. & Sanderson R. D. (1993) Expression of syndecan regulates human myeloma plasma cell adhesion to type I collagen. Blood 81, 767.
17. Caligaris-Cappio F., Bergui L., Gregoretti M. G., Gai- dano G., Gaboli M., Schena M., Zambonin-Zallone A. & Marchisio P. C. (1991) Role of bone marrow stromal cells in the growth of human multiple myeloma. Blood 77, 2688.
682 M.G. GREGORETTI et al.
18. Durie B. G. M. & Salmon S. E. (1975) A clinical staging system for multiple myeloma. Cancer 36, 842,
19. Wulf E., Deboben A., Bautz F. A., Faultisch H. & Wieland T. (1979) Fluorescent phallotoxin, a tool for visualization of cellular actin. Proc. natn. Acad. Sci. U.S.A. 76, 4498.
20. Caligaris-Cappio F., Bergui L., Tesio L., Corbascio G., Tousco F. & Marchisio P. C. (1986) Cytoskeleton organization is aberrantly rearranged in the cells of B- chronic lymphocytic leukemia and hairy cell leukemia. Blood 67, 233.
21. Malavasi F., Caligaris-Cappio F., Milanese C., Del- labona P., Richiardi P. & Carbonara A. O. (1984) Characterization of a routine monoclonal antibody specific for human early lymphohemopoietic cells. Hum. Immun. 9, 9.
22. Carnemolla B., Borsi L., Zardi L., Owens R. J. & Baralle F. E. (1987) Localization of the cellular fibro- nectin-specific epitope recognized by the monoclonal antibody IST-9 using fusion proteins expressed in E. coli. FEBS Lett. 215, 269.
23. Pulido R., Elices M. J., Campanero M. R., Osborn L., Schiffer S,, Garcia-Pardo A., Lobb R., Hemler M. & Sanchez-Madrid F, (1991) Functional evidence for three distinct and independently inhibitable adhesion activities mediated by the human integrin VLA-4. J. biol. Chem. 266, 10,241.
24. Degrassi A., Hilbert D. M., Rudikoff S., Anderson A. O., Potter M. & Coon H. G. (1993) In vitro culture of primary plasmacytomas requires stromal cell feeder layers. Proc. natn. Acad. Sci. U.S.A. 90, 2060.
25. Schmitt-Graff A., Skalli O. & Gabbiani G. (1989) or- Smooth muscle actin is expressed in a subset of bone marrow stromal cells in normal and pathological con- ditions. Virchows Archiv. B cell Path. 57, 291.
26. Galmiche M. C., Koteliansky V. E., Briere J., Herv6 P. & Charbord P. (1993) Stromal cells from human long-term marrow cultures are mesenchymal cells that differentiate following a vascular smooth muscle dif- ferentiation pathway. Blood 82, 66.
27. Gabbiani G. (1992) The biology of the myofibroblast. Kidney Int. 41,530.
28. Weiss L. & Geduldig U. (1991) Barrier cells: stromal regulation of hematopoiesis and blood cell release in normal and stressed routine bone marrow. Blood 78, 975.
29. Leavesley D. I., Ferguson G. D., Wayner E. A. & Cheresh D. A. (1992) Requirement of the integrin beta3 subunit for carcinoma cell spreading or migration on vitronectin and fibrinogen. J. Cell. Biol. 117, 1101.
30. Thiery J. P. & Boyer B. (1992) The junction between cytokines and cell adhesion. Curr. Opin. Cell. Biol. 4, 782.
31. Hynes R. O. (1992) Integrins: versatility, modulation, and signaling in cell adhesion. Cell 69, 11.
32. Cozzolino F., Torcia M., Aldinucci D,, Rubartelli A., Miliani A., Shaw A. R., Lansdorp P. M. & Di Gug- lielmo R. (1989) Production of interleukin-1 by bone marrow myeloma cells. Blood 74, 380.
33. Nakamura M., Merchav S., Carter A., Ernst T. J., Demetri G. D., Furukawa Y., Anderson K., Freedman A. S. & Griffin J. D. (1989) Expression of a novel 3.5 kb macrophage colony-stimulating factor transcript in human myeloma cells. J. Immun. 143, 3543.
34. Nemunaitis J., Andrews F., Mochizuki D., Lilly M. B.
& Singer J. W. (1989) Human marrow stromal cells: response to interleukin-6 (IL-6) and control of IL-6 expression. Blood 74, 1929.
35. Uchiyama H., Barut B. A., Mohrbacher A. F., Chau- han D. & Anderson K. C. (1993) Adhesion of human myeloma-derived cell lines to bone marrow stromal cells stimulates interleukin-6 secretion. Blood82, 3712.
36. Klein B., Zhang X. G., Jourdan M., Content J., Hous- siau F,, Aarden L., Piechaczyk M. & Bataille R. (1989) Paracrine rather than autocrine regulation of myeloma- cell growth and differentiation by interleukin-6. Blood 73, 517.
37. Merico F., Bergui L., Gregoretti M. G., Ghia P., Aimo G., Lindley I. J. D. & Caligaris-Cappio F. (1993) Cytokines involved in the progression of multiple myel- oma. Clin. Exp. lmmun. 92, 27.
38. Kittler E. L. W., McGrath H., Temeles D., Crittenden R. W., Kister V. K. & Quesenberry P. J. (1992) Biologic significance of constitutive and subliminal growth factor production by bone marrow stroma. Blood 79, 3168.
39. Guba S. C., Sartor C. I., Gottschalk L. R., Ye-Hu J., Mulligan T. & Emerson S. G. (1992) Bone marrow stromal fibroblasts secrete interleukin-6 and gra- nulocyte-macrophage colony-stimulating factor in the absence of inflammatory stimulation: demonstration by serum-free bioassay, enzyme-linked immunosor- bent assay, and reverse transcriptase polymerase chain reaction. Blood 80, 1190.
40. Kishimoto T. (1989) The biology of interleukin-6. Blood 71, 1.
41. Detmers P. A., Lo S. K., Olsen-Egbert E., Walz A., Baggiolini M. & Cohn Z. A. (1990) Neutrophil- activating protein 1/interleukin 8 stimulates the binding activity of the leukocyte adhesion receptor CDllb/ CD18 on human neutrophils. J. exp. Med. 171, 1155.
42. Mundy G. R., Raisz L. G., Cooper R. A., Schechter G. P. & Salmon S. E. (1974) Evidence for the secretion of an osteoclast stimulating factor in myeloma. New Engl. J. Med. 291, 1041.
43. Ross Garrett I., Durie B. G. M., Nedwin G. E., Gillespie A., Bringman T., Sabatini M., Bertolini D. R. & Mundy G. R. (1987) Production of lymphotoxin, a bone-resorbing cytokine, by cultured human myel- oma cells. New Engl. J. Med. 317, 516.
44. Bertolini D. R,, Nedwin G. E., Bringman T. S., Smith D. D. & Mundy G. R. (1986) Stimulation of bone resorption and inhibition of bone formation in vitro by human tumor necrosis factor. Nature 319, 516.
45. Stashenko P., Dewhirst F. E., Rooney M. L,, Des- jardins L. A. & Heeley J. D. (1987) Interleukin-lb is a potent inhibitor of bone formation in vitro. Bone Min. Res. 2, 559.
46. Kodama H., Nose M., Niida S. & Yamasaki A. (1991) Essential role of macrophage colony-stimulating factor in the osteoclast differentiation supported by stromal cells. J. exp. Med. 173, 1291.
47. Kurihara N., Bertolini D., Suda T., Akiyama Y. & Roodman G. D. (1990) IL-6 stimulates osteoclast-like multinucleated cell formation in long term human mar- row cultures by inducing IL-1 release. J. lmmun. 144, 4226.
48. Barton B. E. & Mayer R. (1989) IL-3 induces dif- ferentiation of bone marrow precursor cells to osteo- clast-like cells. J. Immun. 143, 3211.
