Cell-to-cell contact is critical for the survival of hematopoietic progenitor cells on osteoblasts

www.elsevier.com/locate/issn/10434666 Cytokine 32 (2005) 155e162

Cell-to-cell contact is critical for the survival of hematopoietic progenitor cells on osteoblasts Younghun Jung a,1, Jianhua Wang a,1, Aaron Havens a, Yanxi Sun a, Jingcheng Wang a, Taocong Jin b, R.S. Taichman a,* a

Department of Periodontics and Oral Medicine, University of Michigan School of Dentistry, 1011 North University Avenue, Ann Arbor, MI 48109-1078, USA b Department of Cariology, Restorative Sciences and Endodontics, University of Michigan School of Dentistry, 1011 North University Avenue, Ann Arbor, MI 48109-1078, USA Received 29 March 2005; received in revised form 9 June 2005; accepted 1 September 2005

Abstract Osteoblasts constitute part of the stromal cell support system in marrow for hematopoiesis, however little is known as to how they interact with hematopoietic stem cells (HSCs). In vitro studies have demonstrated that the survival of HSCs in co-culture with osteoblasts requires intimate cell-to-cell contact. This suggests that the osteoblast-derived factor(s) that supports stem cell activities are produced in very small quantities, are rapidly turned over, may be membrane-anchored and/or require the engagement of cellecell adhesion molecules that are yet to be determined. In the present report we found that the survival of hematopoietic progenitor cells on osteoblasts is dependent upon the engagement of VLA-4 (a4b1) and VLA-5 (a5b1) receptors using function blocking antibodies. Cell-to-cell contact is required to support progenitor activity, but can be replaced if receptoreligand engagement of the VLA-4 and LFA-1 complexes is provided through the use of recombinant ligands (fibronectin, ICAM-1, VCAM-1). Moreover, once these receptors were engaged, conditioned medium derived from HSCs grown on osteoblast ligands supported significantly greater hematopoietic progenitors in vitro than did osteoblast-conditioned or HSC-conditioned medium alone. While the molecules present in the co-cultured medium remain to be identified, the data suggest that hematopoietic cells cooperate with osteoblasts to assemble the various marrow microenvironments by directing the synthesis of osteoblast-derived cytokines to improve HSC survival. Ó 2005 Elsevier Ltd. All rights reserved. Keywords: Osteoblasts; Niche; Stem cell; Endosteal

1. Introduction The bone marrow microenvironment is a multifaceted tissue consisting of hematopoietic stem cells, maturing myeloid, erythroid and lymphoid progenitors. These hematopoietic cells are supported in their development by a variety of stromale parenchymal cells including fibroblasts, endothelial cells, reticular cells and adipocytes. During the growth and

* Corresponding author. Tel.: C1 7347649952; fax: C1 7347635503. E-mail address: [email protected] (R.S. Taichman). 1 Drs. Younghun Jung and Jianhua Wang made equal contributions to the work. 1043-4666/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.cyto.2005.09.001

development of mature blood elements, interactions with stromal cells appear to be critical, as stem cells and stromal cells are closely associated in vivo. A number of culture conditions have been shown to support the limited survival and expansion of stem cells, and here too intimate physical contact between hematopoietic precursors and stromaleparenchymal cells is essential. The mechanism by which contact facilitates survival and growth of stem cells remains unclear. Osteoblasts also constitute part of the stromal cell support system in marrow but little is known about their functional relevance to hematopoietic stem cells. Earliest efforts to appreciate these relationships focused on the protective function that bone might serve for the hematopoietic organ [31]. Although the marrow is encased in bone, this does not prove that

156

Y. Jung et al. / Cytokine 32 (2005) 155e162

osteoblasts have a direct hematopoietic function. Most investigation into the role that osteoblasts play in hematopoetic cascades has focused on how osteoblasts induce the formation of osteoclasts from hematopoietic precursors [15,27] and activate osteoclastic bone resorption [3,6,32]. Despite the intriguing anatomic and developmental findings characterizing hematopoietic cells in close approximation with endosteal osteoblasts, surprisingly little is known [2,10,14]. Using osteoblasts derived from human bone explants [17], we and others have determined that human osteoblasts maintain hematopoietic progenitors and long-term culture initiating cells (LTC-IC) in vitro [20,22,23,25,26]. More recently several groups have demonstrated that events centered at the endosteal surfaces are critical for hematopoietic regulation by cells which share an osteoblastic phenotype [2,36]. Osteoblasts may support hematopoietic stem cells by several mechanisms. Essential to any mechanism is the possibility that osteoblasts produce cytokines that potentially affect hematopoietic stem cell growth and development. In fact we demonstrated that proximity of CD34C cells to osteoblasts can induce the synthesis of several cytokines by osteoblasts [23,26]. However, the necessity of contact between hematopoietic precursors and osteoblasts suggests that either the critical factors are produced in very small quantities that close proximity of the two cell types is required [5], or that these factors are presented to HSCs in a membrane-anchored form [12,21,29]. Alternatively, HSCs and stromal cells are believed to adhere to one another largely through interactions between the very late antigen-4 (VLA-4) on the stem cells and stromal cells expressing the VLA-4 ligand vascular cell adhesion molecule-1 (VCAM-1) [13,18]. While it is clear that the involvement of other receptors/ligands are possible [5,9,28,30,35], signaling events associated with receptor engagement are likely critical factors in stem cell growth and development. Finally, a distinct possibility is that contact by stem cells with osteoblasts influences the support function osteoblast by enhancing their production of either soluble or membrane-associated molecules that support stem cell activities. Previously we have shown that close contact between HSCs and osteoblasts stimulates osteoblastic production of IL-6, MIP-a and HGF [23,26], and that close approximation of the two cell types improves stem cell survival [5]. We now show that cell-to-cell contact is required to support progenitor activity, but can partially be replaced if the VLA-4 and LFA-1 complexes are activated through the use of recombinant ligands. We also demonstrate that the soluble activity that supports progenitor cells is likely to be a short-lived or short ranged diffusible molecule(s).

2. Materials and methods 2.1. Isolation of human OBs Enriched human osteoblast cultures were established using modifications of methods described [20]. Normal human

trabecular bones were obtained from patients undergoing orthopedic surgery in accordance with the University of Michigan’s Investigation Review Board. Bones cleaned of loosely adherent tissue were ground to a uniform particle size (size % 1 mm2) (BioComp Minimill, W. Lorenz, Jacksonville, FL) and incubated in 1 mg/ml bacterial collagenase (Type P, Boehringer Mannheim, Indianapolis, IN). The explants were placed in Dulbecco’s minimal essential medium with low Ca2C and 10% heat inactivated FBS (Invitrogen Corp., Carlsbad, CA). When confluent monolayers were evident, the cultures were maintained in calcium replete Ham’s F12/DMEM. To verify that these cells reflect an osteoblast phenotype, the cultures were screened by RT-PCR for the expression of the osteoblast-specific protein osteocalcin (osteocalcinC) as previously detailed [20,21]. 2.2. Isolation of human CD34CLinÿ HSCs Human bone marrow cells were obtained from healthy adult volunteers by iliac crest puncture and aspiration into preservative-free heparin (protocol approved by the University of Michigan Investigation Review Board). Mononuclear cells were isolated on FicolleHypaque (specific gravity 1.077, Sigma Chemical Corp., Saint Louis, MO), and were allowed to adhere overnight to plastic at 37  C in modified Dexter’s medium (IMDM medium, 10% FBS, 10% equine serum, 1 mM hydrocortisone, penicillin/streptomycin (Invitrogen Corp)). Non-adherent cells were recovered and HSCs isolated by immunomagnetic selection for the CD34CCD36ÿCD38ÿCD45ÿ phenotype with a cocktail of monoclonal antibodies for lineage depletion including CD2, CD3, CD14, CD16, CD19, CD24, CD36, CD38, CD45RA, CD56, CD66b, and glycophorin A (StemCell Technologies, Vancouver, BC). To evaluate cell purity, HSC preparations were stained with a PE-conjugate of the anti-CD34 antibody, HPCA-2 (Becton Dickinson, San Jose, CA) and sorted by FACS (University of Michigan Cancer Center Flow Cytometry Core). 2.3. Adhesion assays These observations were made using a cell-to-cell adhesion assay. CD34C bone marrow cells were labeled with 2.5 mg/ml of the lipophilic dye 2#,7#-bis-(2-carboxyethyl)-5-(and-6)carboxyfluorescein or acetoxymethyl ester BCECF-AM (B-3051; Molecular Bioprobes, Inc., Eugene, OR) in IMDM for 30 min at 37  C, and washed in PBS. Thereafter, the cells were rested for 30 min to reduce non-specific background, and subsequently resuspended in PBS to deliver 104 cells/well in the adhesion assay. Primary OB monolayers were prepared for adhesion assays by washing twice with PBS. Adhesion assays were performed in PBS containing Ca2C/Mg2C where the hematopoietic cells (CD34C cells) were added to a final reaction volume of 100 ml. Antibodies to the particular adhesion molecule or IgG control were added first to the hematopoietic cells. Thereafter the plates were spun at 500 rpm for 5 min at 4  C and allowed to incubate for an additional 10 min at 4  C. The non-adherent cells were removed in three

Y. Jung et al. / Cytokine 32 (2005) 155e162

subsequent washes and the remaining fluorescence was quantified in a 96-well fluorescent plate reader (IDEXX Research Products, ME). 2.4. Progenitor and LTC-IC assays To determine if osteoblast-derived paracrine signals mediate survival of late hematopoietic progenitors, HSCs were incubated for 14 days with conditioned medium derived from 4-day co-cultures. Where indicated, the cultures were separated with 0.4 mM TranswellÒ culture inserts in the presence or absence of human fibronectin (600 ng/ml), soluble ICAM-1 (sICAM-1; 600 ng/ml), or soluble VCAM-1 (sVCAM-1; 400 ng/ml) or BSA as a protein control (600 ng/ ml) which was added daily to the cultures alone or in combination for 2 weeks. Subsequently, survival of recovered hematopoietic progenitor cells (normalized for cell numbers) was evaluated in methylcellulose for colony formation (Stem Cell Technologies, Vancouver, BC) [25]. The CFU frequency was determined at 2 weeks by scoring clonogenic cells as colony forming granulocyte macrophage (CFU-GM), CFU-mixed (CFU-GEMM) or burst forming units-erythroid (BFU-E). As in all investigations w48% of the progenitors were CFUGM, w48% BFU-E and the remaining 2e4% were CFUMix. We have presented our findings as total CFUs. LTC-IC assays were performed as described by Sutherland et al. [19]. Bone marrow stromal cell layers from healthy adult volunteers were passaged several times to eliminate macrophages. Irradiated stromal cell layers were grown in 96-well plates (Costar, Cambridge, MA) in IMDM (GIBCO-BRL, Life Technologies, Grand Island, NY), with each well containing 200 mL of 10% fetal calf serum, 10% horse serum, 10ÿ5 M hydrocortisone, 3.3 mM L-glutamine, 10ÿ4 M b-mercaptoethanol and antibiotics. Serial dilutions of the experimental populations were placed into 20 replicate wells in a total of 200 mL. Wells without HSCs were included to control for residual HSC activity in the stroma. Half the medium was replaced at weekly intervals for 5 weeks. At the conclusion, wells were harvested and recovered cells replated in colony forming methylcellulose assays. 2.5. Microarray studies For our investigations HSCs and osteoblasts from three individuals were co-cultured for 24 h. RNA was isolated from the combination of nine culture conditions and compared to pooled RNA at the initiation of the cultures. Protocols and instrumentation setups as recommended by Affymetrix Co. (Santa Clara, CA) including total RNA sample isolation, biotin-labeled cRNA synthesis, Affymetrix HG-U133A array hybridization with cRNAs, and washing, staining, and scanning of arrays were followed. Data analysis was performed initially with the accompanying software from Affymetrix (Data Mining Ver.3.3) to obtain average difference intensities. Later, average gene intensity values were determined using a software package (DNA-Chip, dChip version 1.1) and analyzed by the statistical software R to calculate expression values

157

using a Robust Multi-array Average (RMA) in consultation with the University of UMCCC Microarray Core Facility [11]. This program fits a model to the data to calculate expression values [1]. After calculating the expression values, the data were analyzed using significance analysis of microarrays (SAM), a program that calculates various statistical tests with adjustments for multiple comparisons using false discovery rate (FDR) [33]. 2.6. Cytokine antibody array Primary human osteoblasts were plated to an initial density of 2.0 ! 104 cmÿ2 in 24-well plates (Becton Dickinson, Franklin, NJ, USA) in the presence of b-glycerol phosphate (10 mM) and L-ascorbate (50 mg/ml) for 7 days. Bone marrow CD34Ccells were maintained in IMDM and equal numbers of CD34C per osteoblasts were co-cultured for 24 h together but separated by TranswellÒ Inserts (0.4 mm in pore size; Corning Costar) in DMEM/F12 (1:1 v/v) containing 1% antibiotic stock, 1% L-glutamine and 10% fetal bovine serum. The expression of 36 cytokines was evaluated using Antibody Arrays (Panomics, Redwood City, CA). Membranes were exposed to blocking buffer for 1 h at 25  C, and incubated with conditioned medium or control medium up to 2 h at 25  C. After washing three times, the arrays were processed with biotinconjugated angiogenesis antibody mix and streptavidine HRP conjugate according to the manufacturer’s protocol. 2.7. Statistical analysis Each experiment was repeated 2e3 times and the data were expressed as mean G SD. Differences between multiple data groups were assessed using one-way ANOVA. Statistical testing was performed using Instat 1.14 (GraphPad, San Diego, CA), with statistical significance set to p ! 0.05.

3. Results Previously we observed that the survival of CD34C cells co-cultured with osteoblasts requires intimate cell-to-cell contact [5]. When these interactions were prevented a vast majority of the hematopoietic cells died over a 2-week culture period. To characterize the adhesion molecule(s) that mediate the initial binding of hematopoietic cells to osteoblasts, labeled CD34C bone marrow cells were plated onto human osteoblasts at 4  C and for 15 min under conditions that do not favor transcription. We observed that antibodies directed against CD11 at 4  C (aL), CD29 (b1), CD31 (gpIIa), CD44, CD54, CD58 (LFA-3), ICAM-1, ICAM-3, LFA-1 (aLb2), Mac-1 (aMb2), VCAM-1, VLA-4 (a4b1), VLA-5 (a5b1), the vitronectin receptor (aVb3), and L-selectin failed to alter the binding of CD34C bone marrow cells to primary osteoblasts ([5] and data not presented). Recruitment of receptor engagement is likely to play a major role in establishing survival of HSCs on osteoblasts. Therefore an initial attempt was to identify the adhesion molecules

158

Y. Jung et al. / Cytokine 32 (2005) 155e162

activated following the establishment of osteoblasteCD34C cell co-cultures. For these investigations HSCs and osteoblasts derived from six different individuals were utilized to establish nine separate co-cultures. Each of these cultures was performed separately and the RNA was pooled and analyzed in duplicate. Due to difficulty in removing the HSCs from the OBs without compromising the quality of the recovered RNA, our analyses were performed on RNA obtained from cells prior to, and following the 24 h co-culture on Affymetrix HG-U133A chips and the data analyzed as described in Section 2. Initially our analysis revealed that there are 259 genes that are up regulated at 24 h, and 14 genes that are down regulated (Tables I and II, Supplemental Data). Inspection revealed that 30 of these signals were repeated at least once suggesting that 206 genuine gene candidates resulting from the co-culture were differentially expressed. A significant proportion of the differentially expressed cDNAs represent intracellular signaling ligands (16.5%, n Z 34) and cell surface receptors (13.5%, n Z 28). Molecules associated with the assembly of the extracellular matrix or its degradation comprised 7.2%. Molecules associated with intracellular signaling, novel sequences and intermediate metabolism comprised the majority of the remaining signals. Among the candidates of extracellular signaling molecules, we noted that IL-6, LIF, MIP-1a and SDF-1 were identified as being enhanced. This observation was most gratifying as we had previously reported that these activities are critical components of an HSCeOB microenvironment [23]. Other notable cytokine messages for BMP-2, CCL7, FGF2b, GRO1a, GRO3, IGF1, IL-1b, IL-8, IL-11, LIF, PDGF were observed. Elevations in mRNA for fibronectin, lysine hydroxylase-like proteins, laminin and type I collagen suggest that the presence of hematopoietic cells also induces osteoblastic activities. Among the adhesion molecules that were most significantly increased (2e3 fold induction) were CD54 (intercellular adhesion molecule-1). To delineate the nature of the adhesion molecules which contribute to the survival of hematopoietic progenitor cells focusing on fibronectin and CD54, CD34C bone marrow cells were seeded onto and incubated with confluent human osteoblasts for 2 weeks. During this period, function blocking antibodies directed to a variety of integrins that bind to fibronectin or CD54 were added daily to the cultures. At the conclusion of the 2-week period, the hematopoietic cells were recovered, counted, and plated in methylcellulose (CFU assay). As we have demonstrated before, in the absence of osteoblasts few if any hematopoietic progenitors were recovered after the culture period (Fig. 1). On osteoblasts survival of the hematopoietic cells was decreased in the presence of antibodies to b1 (CD29) or a4 (CD49d) integrins. Similarly, antibody to ICAM-1 (CD54) also decreased the number of recoverable progenitor cells at the end of the culture period. Functional blocking antibodies to vitronectin (CD51), a5 (CD49e), b2 (CD18) or ICAM-2 (CD102) did not alter the numbers of progenitor cells recovered from the co-culture. From these findings we hypothesized that soluble ligands (fibronectin, VCAM-1 or ICAM-1) to VLA-4 (CD29, CD49d or (a4b1)) and ICAM-1 (CD54) receptors may be

No Antibody CD18 (β2) CD 29 (β1)

*

CD49d (α4)

*

CD49e (α5) CD51 (Vitronect) CD54 (ICAM-1)

*

CD102 (ICAM-2) IgG Control No OB

* 0

10

20

30

40

50

60

70

80

CFU / Well Recovered 2 Weeks Fig. 1. Survival of HSCs on osteoblasts is dependent upon the engagement of VLA-4 (a4b1) and ICAM-1 (CD54) receptors (ligand for LFA-1). Human CD34C cells were seeded onto confluent osteoblast monolayers in the presence or absence of anti-integrin function blocking antibodies to CD18 (b2), CD29 (b1), CD49d (a4), CD49e (a5), CD51 (Vitronectin (Vitronect)), CD54 (ICAM-1), CD102 (ICAM-2) at 10 mg/ml each added daily. IgG and no osteoblasts served as controls (No OB). After 2 weeks, the recovered cells were evaluated for CFU in methylcellulose. The data indicate HSC survival depends upon co-contact and engagement of VLA-4 (CD29, CD49d) and ICAM-1 receptors (CD54). * Indicates significant difference from No antibody control ( p ! 0.05).

able to substitute for the need for direct physical contact of HSCs with osteoblasts to ensure HSC survival. To test this possibility, primary human CD34C bone marrow cells were incubated in the presence or absence of primary human osteoblasts, however contact between the cells was disrupted using culture inserts (TranswellÒ BD Costar Corp). Human fibronectin, recombinant human soluble ICAM-1 or VCAM-1 was added daily to the cultures alone or in combination over a 2-week period. The concentrations used were based upon the protein levels that normally appear in the serum [4,7,16,34]. At the conclusion of a 2-week culture period, the hematopoietic cells were recovered and the survival of normalized late progenitors (CFU assay) was determined in a methylcellulose assay. As shown in Fig. 2, receptor engagement with either fibronectin or VCAM-1, and engagement of the LFA-1 receptors with recombinant ICAM-1 increased the survival of the hematopietic progenitors on osteoblasts in the absence of direct cellecell contact. The data suggest that hematopoietic cells may cooperate with osteoblasts to assemble the various marrow microenvironments. As HSCs die in the absence of a source of growth factors, experiments were designed to determine if the conditioned media from HSCeosteoblast co-cultures could maintain HSCs in vitro in the absence of stromal support. For these investigations, CD34C cells were seeded directly onto confluent human osteoblast monolayers. Controls included culturing of the HSCs or the osteoblasts alone. Each day, 50% of the cell-free conditioned medium was collected, frozen, and upon thawing added (daily 50%) to naı¨ve cultures of CD34CLinÿ cells in the presence of fibronectin for a period of 2 weeks. Direct culture on osteoblasts served as positive control. At the end of 2 weeks the cultures were assayed for

Y. Jung et al. / Cytokine 32 (2005) 155e162

No OB OB OB+FN OB+ICAM OB+VCAM OB+COMBO

OB Isolate 1 OB Isolate 2

* ** ** ** **

OB+CONTACT No OB+COMBO No OB OB OB+FN OB+ICAM OB+VCAM OB+COMBO

* ** ** ** ** **

OB+CONTACT 0

100

200

300

CFU/Well (Mean +/- SD) Fig. 2. Soluble receptor ligands improve progenitor survival when separated from osteoblasts. CD34C were incubated in the presence (OB) or absence of osteoblasts (No OB) derived from two individuals in TranswellÒ culture inserts or in direct contact (Contact). Human fibronectin (FN; 600 ng/ml), sICAM-1 (ICAM; 600 ng/ml), sVCAM-1 (denoted VCAM; 400 ng/ml) or BSA as a protein control (600 ng/ml) (OB culture) were added daily to the cultures alone or in combination (COMBO) for 2 weeks. Thereafter, the hematopoietic cells were recovered and the survival of late progenitors (CFU assay) was determined in methylcellulose assay. The data are reported as mean CFU/ well G S.D. * and ** indicate a significant difference from osteoblast alone to levels of p ! 0.05 and p ! 0.001, respectively. The data indicate that engagement of VLA-4 (FN or VCAM-1) or LFA-1 (ICAM-1) receptors by CD34C cells increases CFU survival on osteoblasts in the absence of direct contact.

the presence of progenitor cell populations in methylcellulose. As shown in Fig. 3, the culture of HSCs on osteoblasts maintains a significant proportion of the input progenitor cell population. In the absence of recombinant growth factors or

159

osteoblasts, few if any progentitor cells survived over a 2-week period. Conditioned medium from osteoblasts alone was unable to support progenitor cell survival. Importantly, frozen and then thawed HSCeosteoblast conditioned medium was unable to maintain progenitor cell survival (Fig. 3). These data suggest that HSCs may not establish a paracrine loop that supports HSC activities through osteoblast intermediaries that can be transferred to naı¨ve cultures. We next considered the possibility that subliminal levels or short-lived molecules support the in vitro maintenance and differentiation of early blood cells. For these investigations, parallel cultures were established such that the co-culture conditioned medium was not frozen. Rather conditioned medium (25% v/v) was collected and directly added daily to secondary HSC cultures in the presence of fibronectin. At the end of 2 weeks, the cultures were assayed for the presence of CFUs in methylcellulose. In this case, we found that the conditioned media derived from the co-cultures of HSCs grown on osteoblast maintained significantly greater survival of hematopoietic progenitor cells in vitro in the absence of osteoblastic or stromal cell support (Fig. 4). These data suggest that hematopoietic cells cooperate with osteoblasts to assemble the various marrow microenvironments by directing the synthesis of osteoblast-derived cytokines to improve HSC survival. To follow up these investigations, an initial attempt was made to further identify candidate molecules involved in supporting HSCs on osteoblasts. We turned to antibody arrays for our initial investigations. The expression of 36 cytokines was evaluated using antibody arrays (Panomics, Redwood City, CA) in the collected conditioned medium from co-cultures at 0 and 24 h. As expected, IL-6 and MIP-1a levels were slightly elevated in the co-culture conditions which correlated

1100

600 Exp 2

880

500

Exp 3

***

660

400

CFU / Well

CFU Recovered (n=3)

Exp 1

440 220

** 300

*

HSC/OB CM

OB CM

HSC CM

HSC/OB CO-Cult

HSC

0

INPUT HSC

200

100

0

Input HSC

HSC

HSC/OB HSC/CM OB/CM HSC/OB CM

Condition

Condition Fig. 3. Frozen co-culture conditioned medium does not support progenitor cell survival. Primary HSCeosteoblast (OB) co-cultures were established over a 2-week period where 25% of the conditioned medium (CM) was collected (frozen) and replaced daily. Secondary HSC cultures were established in the presence (HSC/osteoblast CO-Cult) or absence (HSC) of osteoblasts, or with the daily addition of the previously frozen CM in the presence of (30 mg/ml) human fibronectin. At the end of 2 weeks the cultures were assayed for the presence of CFUs in methylcellulose containing IL-3, GM-CSF and Epo.

Fig. 4. Short-lived molecules in HSCeosteoblast co-cultured conditioned medium support progenitor cell survival. HSC, osteoblast or HSCeosteoblast (OB) co-cultures were established over a 2-week period. The conditioned medium (CM) (25%) was collected and directly added daily to secondary HSC cultures in the presence of (30 mg/ml) human fibronectin. At the end of 2 weeks the cultures were assayed for the presence of CFUs in methylcellulose. * Indicates significant difference from osteoblast CM alone, ** HSC CM alone, *** from HSC/osteoblast co-cultures (p ! 0.05). The data indicate that HSCeosteoblast CM significantly improves CFU survival.

Y. Jung et al. / Cytokine 32 (2005) 155e162

160

Previously we reported that sialated-N-linked glycoproteins mediate initial tethering of HSCs to osteoblast [5]. However, after these adhesions are established, secondary adhesive events are established that provide firm cellecell contact to osteoblasts that assure the survival of HSCs (survival receptors). Based upon antibody and recombinant ligand studies, engagement of LFA-1 and VLA-4 receptors on progenitor cells appears to be the likely candidate for transducing survival signals in the presence of soluble factors from osteoblasts. Here we demonstrate that recombinant ligands for LFA-1 (ICAM-1) or VLA-4 (VCAM-1, FN) together with osteoblast-conditioned medium are critical for the maintenance of progenitor cells. In this regard it is also interesting to note that engagement of the receptors alone is not sufficient for the survival of the progenitor cells. Adhesive events next to endosteal surfaces appear crucial for the full functioning of early hematopoietic cells as well as for the pathophysiology of myeloproliferative disorders and tumor metastasis. Surface studies show that multiple cell adhesion molecules are likely to engage in even the earliest interactions between hematopoietic stem cells and osteoblasts. Furthermore, final affinity and avidity would extend beyond the identification of a candidate molecule since many adhesion molecules exist in multiple isoforms and glycosylation states and can be displayed in different conformations and cell surface clusters. We have therefore begun to explore this unique tissue compartment with several reduced models of endosteal hematopoiesis [5,23,24]. In our previous investigations we considered the impact of post-translational modifications on the first tethering between blood cells and osteoblasts [5]. In

with the microarray data (and in our previous reports [23,26]). IL-8 and VEGF levels were slightly elevated at time 0 and 24 h, but no significant differences were noted. None of the other cytokines or soluble receptors, however, showed any alterations in protein levels over osteoblasts cultured alone including soluble ICAM-1, VCAM-1, GM-CSF, EGF, IP-10, MIP-1b, MIP-4 or 5, IL-1a or -b, IL-2, IL-3, IL-4, IL-5, IL-7, IL-10, IL-15 or IL-17 (Fig. 5). This may reflect the utilization of the receptor/ligands by HSCs or that changes in their activities are localized to the cell surface. 4. Discussion Hematopoiesis is the process of development, formation, and regulation of the various cell types and formed elements of blood from pluripotent stem or progenitor cells. Under normal conditions several components are required for hematopoietic processes. These include undifferentiated precursor cells, colony stimulating factors, and cofactors which alter the genes expressed by the stem cells. Non-hematopoietic cells of the bone marrow (e.g. stromal cells) are central participants in hematopoietic cell cascades secreting several growth factors. While diverse cell types are known to contribute to the marrow microenvironment, the contributions of each individual stromal cell type are still uncertain. HSCs and osteoblasts are both located near endosteal surfaces and this probably reflects a functional relationship between these two cellular populations. This interaction is, at least in part, dependent on celle cell adhesion and we have observed that HSCs bind tightly to osteoblasts in vitro [25].

A

B

C

1 2 3 4 5 6 7 8 9 10 11 12 13 14

1 2 3 4 5 6 7 8 9 10 11 12 13 14

1 2 3 4 5 6 7 8 9 10 11 12 13 14

A B C D E F

D

1-2

3-4

5-6

7-8

9-10

11-12

13-14

A

Apol/Fas

Leptin

Rantes

ICAM-1

IL-2

IL-7

pos

B

CTLA

MIP-1α

TGF-ß

VCAM-1

IL-3

IL-8

pos

C

Eotaxin

MIP-1ß

IFN-γ

VEGF

IL-4

IL-10

neg

D

GM-CSF

MIP-4

TNF-α

IL-1α

IL-5

IL-12(p40)

neg

E

EGF

MIP-5

TNFRI

IL-1ß

IL-6

IL-15

pos

F

IP-10

MMP-3

TNFRII

IL-1rα

IL-6R

IL-17

pos

Fig. 5. Detection of cytokines in array with high specificity. (A) Different capture antibodies against different cytokines were immoblilized on membranes and incubated subsequently with the CD34C conditioned medium. Multiple biotin-conjugated anticytokine and HRP-conjugated streptavidin were used to detect specific cytokines bound to the capture antibodies. The signals were visualized by ECL. (B, C) Simultaneous detection of cytokines from the conditioned medium of osteoblast and osteoblast co-cultured medium of CD34C/osteoblast cells at 24 h. The signals were ultimately visualized by chemiluminecense. (D) Template for the human cytokines in the array. Positive (pos) and negative (neg) controls.

Y. Jung et al. / Cytokine 32 (2005) 155e162

vivo, hematopoietic stem cells and their maturing progeny are in close approximation with bone marrow stromal cells [5]. While the mechanisms that stromal cells use to support hematopoiesis remain unclear, cell adhesion molecules including the cadherins, immunoglobins, integrins and selectins are likely critical determinants in the process. The data reported here represent an extension of these findings and demonstrate that close cell-to-cell contacts between these cells are critical for the survival of hematopoietic cells on osteoblasts. Just as important as these cellecell adhesion interactions is the critical observation that osteoblast-conditioned medium alone does not fully maintain progenitor cells. In the present studies we explored if HSCs establish a paracrine loop that supports HSC activities through osteoblast intermediaries. From these investigations we noted that conditioned medium did not fully support the survival of HSCs in short- (CFU assay) or long-term assay (LTC-IC (data not presented)). Moreover, without adhesion ligand binding the survival of the HSCs was minimal. Previously Gupta et al. [8] demonstrated that stromal derived macromolecule(s) in combination with subliminal amounts of cytokines support the in vitro maintenance and differentiation of primitive human LTC-IC. For our investigations parallel cultures were established such that the conditioned medium was not frozen. In this case, we did in fact find that the conditioned media derived from the co-cultures of HSCs grown on osteoblasts maintain the survival of hematopoietic progenitor cells. At present we do not know the identity of these molecules but several candidates appear likely including those messages identified by the microarray data including IL-6, LIF, MIP-1a, SDF-1, BMP-2, CCL7, FGF2b, GRO1a, GRO3, IGF1, IL1b, IL-8, IL-11 and/or PDGF-D. Recently several groups have demonstrated that events centered at the endosteal surfaces are critical for hematopoietic regulation particularly by cells which share an osteoblastic phenotype [2,36]. The studies extend these and our previous investigations and demonstrate that both soluble signals secreted by HSC establish a paracrine loop with osteoblasts that supports early stem cell survival and suggest that HSCs modulate the formation of a hematopoietic microenvironment directly. Moreover they demonstrate that the requirement of a cell-to-cell contact to support hematopoietic activity can be overcome if the receptoreligand engagement of the VLA-4 and LFA-1 complexes is accomplished with recombinant ligands. We also demonstrate that close cell-to cell interaction changes the receptor levels on osteoblasts of key cell adhesion molecules. These findings help to define the specific mechanisms and consequences of HSCeosteoblast interactions in normal and possibly abnormal hematopoietic states. Further investigation is required for our understanding of the interactive biology of blood and bone cells in health and disease. Acknowledgements These investigations were supported in part by the Tissue Procurement Core of the University of Michigan Comprehensive Cancer Center, Grant #CA46952, and The National Institutes of Health (DE13701).

161

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.cyto.2005.09.001.

References [1] Antonellis K, Beazer-Barclay Y, Bolstad BM, Cope L, I Hobbs B, Irizarry R, et al. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 2003;4(2):249e69. [2] Calvi LM, Adams GB, Weibrecht KW, Weber JM, Olson DP, Knight MC, et al. Osteoblastic cells regulate the haematopoietic stem cell niche. Nature 2003;425:841e6. [3] Chambers T. The cellular basis of bone resorption. Clin Orthop 1980; 151:283e93. [4] Christiansen I, Sundstrom C, Kalkner KM, Bring J, Totterman TH. Serum levels of soluble vascular cell adhesion molecule-1 (sVCAM-1) are elevated in advanced stages of non-Hodgkin’s lymphomas [corrected and republished article originally printed in Eur J Haematol 1998 Nov; 61(5):311e8]. Eur J Haematol 1999;62:202e9. [5] Crean SM, Meneski JP, Hullinger TG, Reilly MJ, DeBoever EH, Taichman RS. N-linked sialyated sugar receptors support haematopoietic celleosteoblast adhesions. Br J Haematol 2004;124:534e46. [6] Croker PR, Morris L, Gordon S. Novel cell surface adhesion receptors involved in interactions between stromal macrophages and haematopoietic cells. J Cell Sci Suppl 1988;9:185e206. [7] Gruss HJ, Dolken G, Brach MA, Mertelsmann R, Herrmann F. Serum levels of circulating ICAM-1 are increased in Hodgkin’s disease. Leukemia 1993;7:1245e9. [8] Gupta P, Blazar BR, Gupta K, Verfaillie CM. Human CD34(C) bone marrow cells regulate stromal production of interleukin-6 and granulocyte colony-stimulating factor and increase the colony-stimulating activity of stroma. Blood 1998;91:3724e33. [9] Hardy CL, Megason GC. Specificity of hematopoietic stem cell homing. Hematol Oncol 1996;14:17e27. [10] Hermans MN, Hartsuiker H, Opstelten D. An in situ study of B-lymphocytopoiesis in rat bone marrow. Topographical arrangement of terminal deoxynucleotidyl transferase-positive cells and pre-B cells. J Immunol 1989;142:67e73. [11] Irizarry R, Bolstad BM, Collin F, Cope L, Hobbs B, Speed T. Summaries of Affymetrix GeneChip probe level data. Nucleic Acids Res 2002; 31:e15. [12] Kurt-Jones EA, Beller DI, Mizel SB, Unanue ER. Identification of a membrane bound interleukin-1. Proc Natl Acad Sci USA 1985;82: 1204e8. [13] Liesveld JL, Abboud CN, Duerst RE, Ryan DH, Brennan JK, Lichtman MA. Characterization of human marrow stromal cells e role in progenitor-cell binding and granulopoiesis. Blood 1989;73:1794e800. [14] Lord BI. The architecture of bone marrow cell populations. Int J Cell Cloning 1990;8:317e31. [15] Marks SCJ, Mackay CA, Jackson ME, Larson EK, Cielinski MJ, Stanley ER, et al. The skeletal effects of colony-stimulating factor-1 in toothless (osteopetrotic) rats: persistent metaphyseal sclerosis and the failure to restore subepiphyseal osteoclasts. Bone 1993;14:675e80. [16] Morisaki N, Saito I, Tamura K, Tashiro J, Masuda M, Kanzaki T, et al. New indices of ischemic heart disease and aging: studies on the serum levels of soluble intercellular adhesion molecule-1 (ICAM-1) and soluble vascular cell adhesion molecule-1 (VCAM-1) in patients with hypercholesterolemia and ischemic heart disease [see comments]. Atherosclerosis 1997;131:43e8. [17] Robey PG, Termine JD. Human bone cells in vitro. Calcif Tissue Int 1985;37:453e60. [18] Shirozu M, Nakano T, Inazawa J, Tashiro K, Tada H, Shinohara T, et al. Structure and chromosomal localization of the human stromal cellderived factor 1 (SDF1) gene. Genomics 1995;28:495e500.

162

Y. Jung et al. / Cytokine 32 (2005) 155e162

[19] Sutherland HJ, Lansdorp PM, Henkelman DH, Eaves AV, Eaves CF. Functional characterization of individual human hematopoietic stem cells cultured at limiting dilution on supportive marrow stromal layers. Proc Natl Acad Sci USA 1990;87:3584e8. [20] Taichman RS, Emerson SG. Human osteoblasts support hematopoiesis through the production of granulocyte colony-stimulating factor. J Exp Med 1994;179:1677e82. [21] Taichman RS, Emerson SG. Human osteosarcoma cell lines MG-63 and SaOS-2 produce G-CSF and GM-CSF: identification and partial characterization of cell-associated isoforms. Exp Hematol 1996;24: 509e17. [22] Taichman RS, Reilly MJ, Emerson SG. The hematopoietic microenvironment: osteoblasts and the hematopoietic microenvironment. Hematology 2000;4:421e6. [23] Taichman RS, Reilly MJ, Matthews LS. Human osteoblast-like cells and osteosarcoma cell lines synthesize MIP-1a in response to IL1b and TNF-a stimulation in vitro. Br J Haematol 2000;108: 275e83. [24] Taichman RS, Reilly MJ, Emerson SG. Human osteoblasts support human hematopoietic progenitor cells in vitro bone marrow cultures. Blood 1996;87:518e24. [25] Taichman RS, Reilly MJ, Emerson SG. Human osteoblasts support human progenitor cells in in vitro bone marrow cultures. Blood 1996;87:518e24. [26] Taichman RS, Reilly MJ, Verma RS, Emerson SG. Augmented production of interleukin-6 by normal human osteoblasts in response to CD34C hematopoietic bone marrow cells in vitro. Blood 1997;89: 1165e72. [27] Takahashi N, Udagawa N, Akatsu T, Tanaka H, Shionome M, Suda T. Role of colony-stimulating factors in osteoclast development. J Bone Miner Res 1991;6:977e85.

[28] Teixido J, Hemler ME, Greenberger J, Anklesaria P. Role of beta1 and beta2 integrins in the adhesion of human CD34 hi stem cells to bone marrow stroma. J Clin Invest 1992;90:358e67. [29] Toksoz D, Zsebo KM, Smith KA, Hu S, Suggs SV, Martin FN, et al. Support of human hematopoiesis in long-term bone marrow cultures by murine stromal cells selectively expressing the membrane-bound and secreted forms of the human homolog of the steel gene product, stem cell factor. Proc Natl Acad Sci USA 1992;89:7350e4. [30] Torok-Storb B, Chabannon C. Stem cellestromal cell interactions. Curr Top Microbiol Immunol 1992;177:123e31. [31] Trentin JJ. Hematopoietic micorenvironments: historical perspectives, status, projections. In: Tavassoli M, editor. Handbook of the hemopoietic microenvironment. Clifton, NJ: Humana Press; 1989. [32] Turner ML, Sweetenham JW. Haemopoietic progenitor homing and mobilization. Br J Haematol 1996;94:592e6. [33] Tusher VG, Tibshirani R, Chu G. Significance analysis of microarrays applied to the ionizing radiation response [erratum appears in Proc Natl Acad Sci USA 2001 Aug 28;98(18):10515]. Proc Natl Acad Sci USA 2001;98:5116e21. [34] Velikova G, Banks RE, Gearing A, Hemingway I, Forbes MA, Preston SR, et al. Circulating soluble adhesion molecules E-cadherin, E-selectin, intercellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1) in patients with gastric cancer. Br J Cancer 1997;76:1398e404. [35] Vermeulen M, Le Pesteur F, Gagnerault MC, Mary JY, Sainteny F, Lepault F. Role of adhesion molecules in the homing and mobilization of murine hematopoietic stem and progenitor cells. Blood 1998;92: 894e900. [36] Zhang JW, Niu C, Ye L, Huang HY, He X, Tong WG, et al. Identification of the haematopoietic stem cell niche and control of the niche size. Nature 2003;425:836e41.