Search

Objective: To better understand the differentiation of stromal cells of the hematopoietic microenvironment, we set out to characterize stromal cells from the different developmental sites of hematopoiesis in the mouse (30 bone marrow, 7 spleen, 3 embryonic and 15 fetal liver, 6 yolk sac, and 6 aorta-gonad-mesonephros lines) for expression of 22 cytoskeletal, membrane, and extracellular matrix proteins., Materials and Methods: Western blotting, immunofluorescence, and flow cytometry were used. Statistical methods included principal components analysis and analysis of variance., Results: Stromal cells from 11 dpc mouse embryos express mesenchymal and vascular smooth muscle cell (VSMC) markers. Principal components analysis on the 70 stromal cell lines isolated from different anatomic sites and developmental stages allows classification of stromal lines along a mesenchymal to VSMC differentiation pathway. Stromal cells do not express endothelial and hematopoietic differentiation membrane antigens, but they do express integrin alpha(5), alpha(6), and beta(1) subunits, vascular cell adhesion molecule-1, CD44, stem cell antigen-1, Thy-1, CD34, and endoglin. The intensity of expression of certain markers differs between lines according to the anatomic site of origin., Conclusions: This study indicates that stromal cells, whatever their anatomic site of origin, follow a VSMC differentiation pathway, suggesting a blood-forming tissue-specific differentiation of mesenchymal stem cells. Differential quantitative expression of distinct sets of markers appears to be correlated with the anatomic sites of origin of the stromal cells.

Previous studies by our group showed that stromal cells from human long-term marrow cultures were mesenchymal cells following a vascular smooth muscle pathway. The present study using 58 immortalized stromal lines from different hematopoietic sites was conducted to verify whether this hypothesis also held true for murine stroma. Principal components analysis performed using cytoskeletal and extracellular matrix proteins allowed the segregation of five factors explaining more than 70% of the variance. Factor I, including osteopontin and vimentin, and factor II, laminins and fibronectins, were representative of the mesenchyme. The remaining three factors were representative of vascular smooth muscle: factor III, including alphaSM actin, SM alpha actinin, SM22alpha, EDa+ fibronectin, and thrombospondin-1; factor IV, metavinculin and h-caldesmon; and factor V, smooth muscle myosin SM1 and desmin. All lines expressed factors I and II; 53 lines expressed factor III, 35 lines expressed factor IV; and 11 lines expressed factor V. A second principal components analysis including membrane antigens indicated the cosegregration of vascular cell adhesion molecule-1 with osteopontin and that of Ly6A/E with vimentin, whereas CD34 and Thy-1 appeared to be independent factors. The heterogeneity of vascular smooth muscle markers expression suggests that harmonious maintenance of hematopoiesis depends on the cooperation between different stromal cell clones.

Nontransformed stromal colony-derived cell lines (CDCLs) consist of a pure stromal cell population that differentiates following a vascular smooth-muscle cell repertoire. Here we study the maintenance of hematopoiesis by this cell population. We show that CDCLs allow the generation for several weeks of stroma-adherent colonies (comprising a cobblestone area) from CD34+, CD34+/CD38+, and CD34+/CD38- cells. Stroma-adherent colony-forming cells (CFCs) from CD34+/CD38- cells reach a maximum at week 4 and limiting dilution analysis gives a frequency of 1 per 10 cells seeded; in contrast to this, CFCs from CD34+/CD38+ cells are optimal by week 2 and the frequency is then only 1 per 120 cells seeded. Stroma-adherent colonies comprise hematopoietic cells from all lineages except the T lymphocytic, with a majority of granulomonocytes. CDCLs also allow the amplification of granulomonocytic colony-forming units (CFU-GMs), since cumulative outputs of CFU-GMs by week 6 are 190 and 8 times that observed at culture inception for the CD34+/CD38- and CD34+/CD38+ cell populations, respectively. Our results suggest that stromal cells from CDCLs allow the maintenance of primitive hematopoietic precursors and induce their proliferation and differentiation. This study underscores the potential role of one of the microenvironmental cell populations, that of myoid cells, in the regulation of hematopoietic precursor behavior.

This study was performed to evaluate the role of human microenvironmental cells in the metabolism of AcSDKP, a physiological inhibitor of hematopoietic stem cells. Using long-term marrow cultures (LTMCs), whose medium already contained a baseline value of AcSDKP, we found after 2 weeks a net output in the culture supernatant indicating that release by cells from the adherent layer was superior to consumption of the peptide. Since human microenvironmental cells consist of macrophages and vascular smooth-muscle-like stromal cells we generated pure populations of macrophages (by culturing cord blood cells in the presence of granulomonocytic colony-stimulating factor) and of stromal cells (generated by stromal colonies). We found in supernatants of macrophage cultures a significantly (p < 0.01) increased level of AcSDKP (compared with value in medium) while in supernatants of stromal cell cultures the level was decreased. Cell content of angiotensin-converting enzyme (ACE) in stromal cells was higher than in macrophages, which suggests a degradation of AcSDKP by stromal cells because of their higher amount of ACE. Finally, we analyzed the content of AcSDKP in adherent layers of LTMCs (with or without extracellular matrix [ECM] components), macrophages, and stromal cells. We found levels of AcSDKP of 1.5 pMol per 106 cells in extracts from macrophages or from stromal cells. On the contrary, extracts from primary layers of LTMCs contained 3 times more AcSDKP; however, after treatment of primary layers by collagenase, AcSDKP level fell to 1 pMol per 10(6) cells. Immunofluorescence using an anti-AcSDKP monoclonal antibody showed an extracellular network in certain areas of LTMCs. This study shows that 1) macrophages synthesize and release in the supernatant AcSDKP, 2) stromal cells probably degrade the peptide via ACE, and 3) components of the ECM from LTMCs serve as a reservoir for the peptide. These results are reminiscent of what has been described for growth factors, produced by microenvironmental cells, and stored in the ECM in close vicinity to hematopoietic precursors.

We report a new method for generating nontransformed human stromal cell lines with a replicative potential of 20 to 25 doublings yielding 10(6) to 3 x 10(7) cells after 4 to 6 weeks. Cells from week-3 to -6 adherent layers of human long-term bone marrow cultures (LTBMC) were plated in methylcellulose in the presence of 20 U/mL interleukin-1 beta (IL-1 beta) and 200 U/mL tumor necrosis factor-alpha (TNF-alpha). After 2 to 3 weeks, we obtained 180 +/- 14 colonies per 10(5) cells seeded. These well-delineated colonies with a dense central core consisted of up to several hundred tightly packed, identical, large refractile cells. Colonies were determined to be clones by sequential examination of the cultures and the linear relationship between the number of colonies counted and cells seeded. Colony-derived cell lines (CDCL) were developed by seeding individual colonies in long-term culture medium (LTCM) supplemented with 20 ng/mL basic fibroblast growth factor (bFGF). The selection of colonies yielding lines with high proliferative capacity was due to the presence of IL-1 beta and TNF-alpha in the semisolid medium. The most effective concentrations for clonal selection were 200 U/mL TNF-alpha and 20 U/mL IL-1 beta. The growth of CDCL in liquid culture depended on the presence of bFGF, with the most effective concentration at 20 ng/mL. CDCL were able to maintain the output of colony-forming units granulocyte/macrophage and burst-forming unit-erythrocyte (CFU-GM and BFU-E) for several weeks from cocultured CD34+ marrow cells. The weekly CFU-GM and BFU-E output from weeks 2-5 was at least the same as observed when using passaged adherent layers. CDCL represent a progenitor cell population for stromal cells that may prove a suitable model for the study of the relationship between marrow stromal cells and hematopoiesis.

In a previous report, we described a method to generate cell lines derived from stromal colonies (colony-derived cell lines [CDCL]). In a first step, colonies were obtained from plating cells from adherent layers of human long-term bone marrow cultures (LTBMC) in methyl-cellulose in the presence of interleukin-1 beta (IL-1 beta) (20 U/mL) and tumor necrosis factor-alpha (TNF-alpha) (200 U/mL). In a second step, cell lines were derived from individual colonies cultured in liquid medium with 20 ng/mL basic fibroblast growth factor (bFGF). In this report, we describe the phenotype of cells from more than 100 cell lines. CDCL did not contain cells of hematopoietic origin, which indicates that the culture system did not allow the growth of hematopoietic precursors. CDCL did not contain endothelial-like cells, similarly to primary adherent layers. CDCL comprised stromal cells, as defined by membrane antigens recognized by monoclonal antibodies 6-19, Stro-1 and 1B10. These cells belonged to the family of connective tissue-forming cells since they synthesized interstitial collagens I, III, and V; nonplasmatic EDa+ and EDb- fibronectin; tenascin; and chondroitin-sulfate. Study of the time course of CDCL showed that the lines differed from one another according to their proliferative capacity: 32% of CDCL grew quickly, yielding about 10,000 cells in 10 days; 36% of CDCL grew slowly, yielding 1000 cells in 10 days; and the remaining 32% had intermediate proliferative capacity. For CDCL with a high proliferative capacity, a distinctive differentiation pattern could be described. At culture, inception cells from the lines were vimentin and laminin. Over time, several cytoskeletal and extracellular matrix (ECM) proteins indicative of vascular smooth muscle differentiation were expressed, including: the alpha-SM-actin isoform; the actin-binding proteins, smooth muscle myosin-heavy chain (SMMHC), SM1; h-caldesmon; calponin; gelsolin; and the ECM proteins, collagen IV and elastin. In full-grown lines, cells were similar to immature, intimal, vascular smooth muscle cells as found beneath the endothelium in adult aortas. Because of the coupling between proliferation and differentiation, the differentiation pattern seems to be under genetic control. However, since the coupling was not stringent during the whole lifespan of the lines, it is possible that cytokines are also involved, ensuring autocrine regulation of CDCL development.

It has been shown that cord blood collected at birth can be used to successfully engraft a human lymphocyte antigen (HLA)-matched sibling suffering from a malignant disease. It has been further suggested that this source of cells may be used in unrelated but HLA-compatible patients. These wide indications would imply the establishment of cord blood banks comprising 10(5) or more samples. In this report we show that it is possible to fractionate and freeze cord blood samples without major loss in granulomonocytic or erythroblastic progenitors (CFU-GM and BFU-E). Density separation should be carried out using Percoll of density 1.080. Separated samples should be frozen and thawed in the presence of DNase I. This procedure should allow the storage of approximately 10 mL samples in cryotubes containing a number of CFU-GM and BFU-E sufficient to engraft a patient weighing less than 30 kg. These data provide a rationale for establishing cord blood banks.

We have investigated the distribution of membrane molecules on CD34+ hematopoietic cells isolated from human bone marrow (BM) and cord blood (CB). A distinct CD10+ population was present in BM, but it was not detected in CB. Most CD34+ CD10+ cells in BM were B-cell precursors (BCP), because they expressed CD19. However, CD40 and CD37 were found on the majority of CD34+ cells from either BM or CB, demonstrating that these antigens are not restricted to B-lineage CD34+ cells. CD40 and CD37 were lost during culture of CD34+ cells in the presence of interleukin 3 (IL-3), indicating transient expression early in myeloid development. CD13 antigen was detected on virtually all CD34+ cells from BM and CB. Accordingly, CD13 was present on CD34+ CD10+ cells, demonstrating that this structure is not restricted to myeloid CD34+ cells. In contrast, myeloid CD33 antigen was not detected on CD34+ CD10+ cells. Expression levels of CD13 and of CD33 were heterogeneous in BM, reflecting diversity within the resident CD34+ population. CD25 and CD71 were found on a proportion of CD34+ cells from either BM or CB and maintained during culture in IL-3, consistent with a distribution on activated cells. Finally, a variety of adhesion receptors were present on CD34+ cells. These included the alpha 4 beta 1 (VLA-4), alpha 5 beta 1 (VLA-5), and alpha L beta 2 (LFA-1) integrins, as well as ICAM-1, LFA-3, H-CAM, and LAM-1. Expression of adhesion receptors was remarkably similar in BM and CB, and it followed an all-or-nothing pattern that failed to delineate CD34+ subsets. Taken together, our data show that although CD34+ cells from BM constitute a more heterogeneous population, resident and circulating CD34+ cells largely display the same cell-surface molecules.

The cytoskeleton of stromal cells from the adherent layer of human Dexter-type cultures has been studied. It was found that the stress fibers contained actin specific for smooth muscle, mainly the alpha SM actin isoform. The intermediate filaments consisted of vimentin, and there were no desmin filaments. This pattern was similar to that of cultured vascular smooth muscle cells. The detectability of the alpha SM actin isoform is coincident with the appearance of stromal cells in long-term marrow cultures and may provide a useful marker for stromal cells. The potential in vivo cellular counterpart for stromal cells generated in the Dexter-type culture system is discussed.

Human granulocyte colony formation has been observed in serum-free methylcellulose cultures with Iscove medium, delipidated bovine serum albumin, iron-saturated transferrin, alpha-thioglycerol, oleylpalmitoyl lecithin, cholesterol, and linoleic acid using serum-free human placental-conditioned medium (SF-HPCM) as the source of colony stimulating factor (CSF). Dose-response curves for SF-HPCM indicated a lower sensitivity to colony-stimulating activity in serum-free cultures than in serum-dependent cultures. Gel filtration of SF-HPCM revealed that CSF fractions with molecular weights in the range of 30 kD are inefficient in serum-free cultures, while fractions with molecular weights in the range of 40 kD stimulate granulocyte colony formation in both types of cultures. These results demonstrate that serum constituents modulate the effects of one of the stimulating factors for granulocyte colony formation, and that serum-free culture conditions are essential for establishing the growth factor requirements of the granulocyte lineage.

This report describes an IgG1 mouse monoclonal antibody derived after immunization of mice with washed stromal cells from human, long-term bone marrow cultures. The antigen recognized by the antibody (BMS-1) is a carbohydrate-containing prosthetic group that is common to and specific for multiple horse serum proteins. These proteins are avidly ingested by stromal cells and concentrated in endocytic vesicles. Cultured smooth muscle cells took up the horse proteins in a similar manner to marrow stromal cells while cultured marrow fibroblasts, endothelial cells, and hepatoma cells did not. These data indicate that marrow stromal cells specifically accumulate horse serum proteins which might partially explain the horse serum requirement for long-term marrow culture maintenance. The data also suggest further similarities between marrow stromal and smooth muscle cells and additional differences between marrow fibroblasts and marrow stromal cells.