Dr K. Le Blanc, MD, PhD, Department of Haematology, Huddinge University Hospital, M54, SE-141 86 Stockholm, Sweden. E-mail: firstname.lastname@example.org
We aimed to study the effects of mesenchymal stem cells (MSCs) on alloreactivity and effects of T-cell activation on human peripheral blood lymphocytes (PBLs) in vitro. MSCs were expanded from the bone marrow of healthy subjects. MSCs isolated from second to third passage were positive for CD166, CD105, CD44, CD29, SH-3 and SH-4, but negative for CD34 and CD45. MSCs cultured in osteogenic, adipogenic or chondrogenic media differentiated, respectively, into osteocytes, adipocytes or chondrocytes. MSC added to PBL cultures had various effects, ranging from slight inhibition to stimulation of DNA synthesis. The stimulation index (SI = (PBL + MSC)/PBL) varied between 0.2 and 7.3. The SI was not affected by the MSC dose or by the addition of allogeneic or autologous MSCs to the lymphocytes.
Suppression of proliferative activity was observed in all experiments after the addition of 10,000–40,000 MSCs to mixed lymphocyte cultures (MLCs). Lymphocyte proliferation was 10–90%, compared with a control MLC run in parallel without MSCs. In contrast, the addition of fewer MSCs (10–1000 cells) led to a less consistent suppression or a marked lymphocyte proliferation in several experiments, ranging from 40 to 190% of the maximal lymphocyte proliferation in control MLCs. The ability to inhibit or stimulate T-cell alloresponses appeared to be independent of the major histocompatibility complex, as results were similar using ‘third party’ MSCs or MSCs that were autologous to the responder or stimulating PBLs. The strongest inhibitory effect was seen if MSCs were added at the beginning of the 6 day culture, and the effect declined if MSCs were added on day 3 or 5. Marked inhibitory effects of allogeneic and autologous MSCs (15,000) were also noted after mitogenic lymphocyte stimulation by phytohaemagglutinin (median lymphocyte proliferation of 30% of controls), Concanavalin A (56%) and protein A (65%). Little, if any, inhibition occurred after stimulation with pokeweed mitogen. Low numbers of MSCs (150 cells) were unable to inhibit mitogen-induced T-cell responses.
MSCs have significant immune modulatory effects on MLCs and after mitogenic stimulation of PBL. High numbers of MSCs suppress alloreactive T cells, whereas very low numbers clearly stimulated lymphocyte proliferation in some experiments. The effect of a larger number of MSCs on MLCs seems more dependent on cell dose than histocompatibility and could result from an ‘overload’ of a stimulatory mechanism.
Mesenchymal cells in human bone marrow contribute to haematopoiesis by promoting cell-to-cell interactions, the expression and presentation of growth factors and cytokines and the secretion of extracellular matrix proteins [1–3]. Mesenchymal progenitor cells (mesenchymal stem cells – MSCs) can be isolated from the bone marrow and have been characterized phenotypically as non-haematopoietic cells, as they do not express CD34 or CD45 [4–6]. Human MSCs can be identified through flow cytometry by the monoclonal antibodies SH-2, SH-3 and SH-4 [6, 7]. They show extensive proliferative capacity in vitro without the loss of phenotype and are called stem cells because they retain the ability to differentiate into several mesenchymal lineages, such as bone, cartilage, adipose and muscle, in vitro and in vivo[5, 6, 8]. MSCs constitutively secrete a large number of cytokines and promote the expansion and differentiation of haematopoietic stem cells (HSCs) in vitro. Co-transplantation of human MSCs along with human HSCs enhances human haematopoietic engraftment in nonobese diabetic/severe combined immunodeficiency (NOD/SCID) mice . In humans, autologous, in vitro culture-expanded, MSCs can be infused intravenously without causing toxicity . Co-transplantation of autologous MSCs and HSCs may increase HSC engraftment .
MSCs express intermediate levels of human leucocyte antigen (HLA) major histocompatibility complex (MHC) class I molecules and very low levels of HLA class II and Fas ligand; they do not express the costimulatory molecule B7-1, B7-2, CD40 or CD40L [6, 13, 14]. They should therefore be recognized by alloreactive T cells. Surprisingly, the infusion or implantation of allogeneic, MHC-mismatched MSCs into baboons is well tolerated in most animals [15, 16]. Some reports suggest that baboon and human MSCs escape recognition by alloreactive T cells in primary and secondary mixed lymphocyte cultures (MLCs). The addition of MSCs to primary MLCs may suppress T-cell proliferation [17–19]. Such suppression occurs regardless of donor source, including ‘third party’ MSCs, i.e. cells obtained from neither a stimulator nor a responder cell source.
The immunosuppressive nature of MSCs is of clinical relevance in allogeneic transplantation because it can theoretically reduce the incidence and severity of graft-versus-host disease (GvHD). Preliminary reports on co-transplantation of ex vivo-expanded MSCs and HSCs from HLA-identical siblings show promising results, as none of the first 15 evaluable patients developed severe acute GvHD .
The aim of the present study was to determine the effect of MSCs on lymphocyte proliferation in MLCs in more detail. We evaluated the role of HLA compatibility by MLCs with the addition of MSCs autologous to stimulator or responder cells and ‘third party’ MSCs. As it is unclear what the optimal cell dose of MSCs is in clinical transplants, we also studied the effects of various cell doses of MSCs on lymphocyte proliferation.
Materials and methods
Harvesting and ex vivo MSC culture To isolate human MSCs, bone marrow aspirates of 10–20 ml were taken from the iliac crest of normal donors ranging in age from 9 to 47 years. The harvest of bone marrow from healthy persons for isolation and expansion of MSCs to study their effect in vitro was approved of by the Ethics Committee at Huddinge University Hospital. MSCs were isolated and cultured using a previously reported method . Briefly, heparinized bone marrow was mixed with a double volume of phosphate-buffered saline (PBS) (Unimedic AB, Matfors, Sweden) and centrifuged at 900 × g for 10 min at room temperature. Washed cells were re-suspended in PBS to a final density of 1 × 108 cells/ml. Five millilitre aliquots were layered over a 1.073 g/ml Percoll solution (Sigma, St. Louis, MO, USA) and centrifuged at 900 × g for 30 min. Mononuclear cells were collected from the interface, re-suspended in PBS and centrifuged at 900 × g for 10 min at 20 °C. The cells were re-suspended in human MSC medium consisting of Dulbecco's modified Eagle's medium-low glucose (DMEM-LG) (Life Technologies, Gaithersburg, MD, USA), supplemented with 10% fetal bovine serum (Sigma) and 1% antibiotic–antimycotic solution (Life Technologies) and plated at 3 × 107 cells/185 cm2 in Nanclon Solo flasks (Fisher Scientific, Pittsburg, PA, USA). The serum lot was selected on the basis of optimal MSC growth with maximal retention of osteogenic, chondrogenic and adipogenic differentiation. Cultures were maintained at 37 °C in a humidified atmosphere containing 5% CO2. When the cultures were near confluence, the cells were detached by a treatment with trypsin and ethylenediaminetetraacetic acid (EDTA) (GibcoBRL, Grand Island, NY, USA) and re-plated (passaged) at a density of 1 × 106 cells/185 cm2 flask.
Phenotype and capacity for the differentiation of MSC After two to four passages, the cells were harvested by treatment with 0.05% trypsin–EDTA. For flow cytometry, the detached cells were washed and re-suspended in PBS. Aliquots (5 × 105 cells) were incubated on ice with conjugated, monoclonal antibodies against CD34 and CD45 (Becton Dickinson, San Jose, CA, USA), CD166, CD44 (Immunotech, Marseilles, France), CD29 (Coulter, Miami, FL, USA) and CD105. All incubations with antibodies were done for 30 min. Cells were washed with PBS, pelleted and re-suspended in PBS containing 1% formalin and 0.1% bovine serum albumin (BSA) (Sigma). Non-specific fluorescence was determined using equal aliquots of the cell preparation incubated with antimouse monoclonal antibodies (Becton Dickinson). Flow cytometry was done by collecting 10,000 events on a fluorescence-activated cell sorter scan equipped with an argon laser, and the data were analysed with cellquest software (Becton Dickinson).
The capacity of MSCs to differentiate along adipogenic, chondrogenic and osteogenic lineages was assessed as described elsewhere . Briefly, for adipogenic differentiation, cells were seeded at a concentration of 2.1 × 104 MSC/cm2 and induced by adding 1-methyl-3-isobutylxanthine (0.5 mmol/l), dexamethasone (1 µmol/l), insulin (10 µg/ml) and indomethacin (0.2 µmol/l) to the MSC medium. This medium was replaced every 3–4 days for 21 days. Adipogenesis was measured by the accumulation of neutral lipids in fat vacuoles, stained with oil red O. To stimulate chondrogenic differentiation, 7.5 × 105 mesenchymal cells were pelleted in a 15 ml polypropylene tube and cultured in DMEM-high glucose (DMEM-HG) (Life Technologies) with insulin (6.25 µg/ml), transferrin (6.25 µg/ml), linolenic acid (5.33 µg/ml), BSA (1.25 mg/ml), pyruvate (1 mmol/l), ascorbate-2-phosphate (0.17 mmol/l), dexamethasone (0.1 µmol/l), proline (0.35 mmol/l) and selenous acid (6.25 ng/ml). Differentiation was induced by adding tumour growth factor-β3 (TGF-β3) (0.01 µg/ml) (R&D Systems, Minneapolis, MN, USA). This medium was replaced every 3–4 days for 21 days. Pellets were formalin-fixed, embedded in paraffin, examined morphologically and stained for the presence of glucosaminoglycans with Alcian green. For osteogenic differentiation, 3.1 × 103 cells/cm2 were grown in MSC medium supplemented with dexamethasone (0.1 µmol/l), ascorbic acid (0.05 mmol/l) and glycerophosphate (10 mmol/l). This medium was replaced every 3–4 days for 21 days. Specimens were stained for mineral using the von Kossa method, or paraffin-embedded for morphological examination .
MLCs and mitogenic stimulation Purified peripheral blood lymphocytes (PBLs) were prepared by centrifuging heparinized blood on Ficoll–Isopaque (Lymphoprep, Nycomed, Oslo, Norway). Separated cells were cultured in RPMI-1640 medium supplemented with HEPES (25 mmol/l), penicillin, streptomycin, l-glutamine (2 mmol/l) (Gibco BRL, Life Technologies, Paisley, UK) and 10% heat-inactivated, pooled human AB serum. Samples containing 1 × 105 cells were cultured in 0.2 ml of medium on microtitre plates (Nunclon, Copenhagen, Denmark). This method has been described in detail elsewhere . To study the effect that MSCs have on lymphocyte reactivity, irradiated MSCs in concentrations ranging from 0.01 to 40% of the total number of responder cells in the wells (i.e. 10–40,000 MSCs) were added at the beginning of the experiment, if not otherwise stated. MSCs were allogeneic or autologous to responder or stimulatory lymphocytes. HLA typing was not performed. Lymphocytes were stimulated with irradiated lymphocytes from one or a pool of five donors as indicated. All experiments were run in triplicate. The standard error (cpm) was about 10% and always less than 20%. For mitogenic stimulation, the following mitogens were used: phytohaemagglutinin (PHA), final concentration 10 µg/ml (Sigma), Concanavalin A (Con A) (5 µg/ml) (Pharmacia Biotech, Uppsala, Sweden), purified protein A from Staphylococcus aureus (10 µg/ml) (Pharmacia Upjohn, Uppsala, Sweden) and pokeweed mitogen (diluted to 1/200) (PWM, Gibco BRL) [24–26]. Triplicates of 1.5 × 105 cells in 0.2 ml were incubated at 37 °C in humidified 5% CO2 air. 3H thymidine with a specific activity of 5 Ci/mmol/l (Radiochemical Centre, Amersham, UK) in 0.02 ml PBS was added to each culture 24 h before harvest. The cells were harvested automatically on a glass fibre filter using a Tomtec harvesting machine (Harvester 96, Tomtec, Orange, CT, USA). The radioactivity was measured by means of a micro β-liquid scintillation counter (Wallac, Turku, Finland). Cultures with PHA and Con A were harvested on day 4 . MLCs and cultures with protein A and PWM were harvested on day 6 [23, 25]. The effects of MSCs on MLCs or mitogenic responses were calculated as the percentage of the control response (100%): MLC = ((APx + MSC )/APx) × 100, where A is the number of responding lymphocytes, P the pool of lymphocytes from five donors, x irradiated stimulator cells and MSC of mesenchymal stem cells in culture.
Human bone marrow-derived MSCs were successfully culture-expanded from all donors. As reported by others, the cultures underwent an initial lag phase of about 5 days . A morphologically homogeneous population of fibroblast-like cells with more than 90% conf luence was seen at a median of 13 days (Fig. 1A). After the first passage, the cells grew exponentially, requiring weekly passages.
The cultured mesenchymal cells, isolated from second to third passage, comprised a single phenotypic population by flow cytometric and immunohistochemical analysis of surface-expressed antigens. MSCs were uniformly positive for the activated leucocyte cell adhesion molecule CD166, the endoglin receptors CD105, CD44 and the β1-integrin CD29 (Fig. 2). Evaluation by immunofluorescence microscopy showed cells positive for SH-3 and SH-4, with no unlabelled cells. In contrast, there was no detectable contamination of haematopoietic cells, as flow cytometry was negative for markers of the haematopoietic lineage, including the lipopolysaccharide receptor CD14, CD34 and the leucocyte common antigen CD45 (Fig. 2).
In vitro differentiation of MSCs
MSCs treated with osteogenic medium underwent a change in their morphology from spindle-shaped to cuboidal, and formed large nodules by day 18 of induction.
Adipogenic differentiation was shown by the accumulation of lipid vacuoles. In all samples, these vacoules were already seen after the first induction treatment and increased with time in size and number until they coalesced. Such lipid vacuoles stained with oil red O (Fig. 1B), but untreated cultures were negative (data not shown). No apparent correlation was seen with donor age or cell passage.
Chondrogenic differentiation was induced by TGF-β3. Paraffin-embedded micromasses had a chondrogenic morphology when stained with haematoxylin–eosin and Alcian green for proteoglycans (Fig. 1C).
Nodules were embedded in paraffin, sliced and evaluated with haematoxylin–eosin. Specimens were then stained for mineral using the von Kossa method (Fig. 1D).
Effects of MSCs on PBLs
To determine whether MSCs could induce a proliferative response by allogeneic lymphocytes, PBLs were cultured with irradiated MSCs in various doses. As shown in Table 1, the addition of MSCs to PBLs caused some stimulation in a few subjects and no response or mild inhibition in others. This finding was independent of the number of MSCs added to the culture, as the stimulation index (SI = (AAx + MSC )/AAx) showed no increase when higher concentrations of MSCs were added. A = PBL, Ax irradiated PBL from the same donor. The SI ranged from 0.2 to 7.3. Whether or not MSCs were autologous to the responding lymphocytes or allogeneic did not seem to affect the background SI. For instance, when 10,000 allogeneic MSCs were added to PBLs, the median SI was 1.7 (n = 13) versus 2.1 (n = 4) when autologous MSCs were added. The ranges of SI were similar for allogeneic and autologous MSCs (Table 1).
Table 1. Allogeneic or autologous mesenchymal stem cells(MSCs) were added to autologos mixed lymphocyte cultures (AAx)*
To determine whether MSCs affect the proliferation of lymphocytes in response to alloantigens, PBLs were stimulated with PBLs from a pool of five donors. DNA synthesis (cpm) in the absence or presence of various concentrations of MSCs was performed. In the first set of experiments, irradiated ‘third party’ MSCs were added on day 0 in diminishing concentrations. As shown in Fig. 3, the suppression of proliferation in MLCs occurred in a dose-dependent fashion when larger numbers of MSCs were added (10,000–40,000). This was not owing to cell crowding, because Trypan blue staining showed >95% cell viability at the time of harvest.
In contrast, with lower concentrations of MSCs, the inhibition was less consistent. In several experiments, the addition of 10–100 MSCs to the MLCs clearly increased lymphocyte activation. Figure 4 shows two typical experiments. Larger numbers of MSCs (10,000–40,000) inhibited MLCs, but fewer MSCs had an inhibitory effect in some experiments (Fig. 4A) and a stimulatory effect in others (Fig. 4B).
Figure 5 shows the kinetics of MLC inhibition. One-way MLCs were established and MSCs were added to the lymphocyte cultures on days 0, 3 and 5. The cells were harvested on day 6. MSC inhibition of lymphocyte proliferation was most marked when MSCs were added at the beginning of the culture, but the addition of MSCs on days 3 and 5 still caused significant suppression of proliferation. The addition of 100 MSCs on day 0 increased the MLC response, an effect that was not seen if MSCs were added on day 3 or 5.
Effects of autologous MSCs in MLCs
The same trend, with inhibition of MLCs at higher MSC concentrations and reduced inhibition at lower MSC concentrations, was seen when MSCs and stimulating lymphocytes were derived from the same donors (Fig. 6A) and when MSCs were autologous to the responding PBLs (Fig. 6B). Again, the low numbers of MSCs (10–1000) added to MLCs had a stimulating effect in some experiments.
Effects of MSCs on mitogenic stimulation
Allogeneic or autologous MSCs inhibited mitogenic responses (Fig. 7). In the presence of 15,000 allogeneic MSCs, PHA response was a median of 30% (range 24–60, n = 4), compared with 100% in the controls not cultured with MSCs. The corresponding responses for MSCs with Con A was 56 (26–99%) and for Staphylococcus aureus protein A (SpA) 65 (27–87%). Lymphocyte response to PWM in the presence of 15,000 allogeneic MSCs was inhibited only in one of four experiments. In two experiments using 15,000 autologous MSCs, PWM responses were 40 and 53% of the controls (Fig. 7B). Lower numbers of MSCs (150) result in lesser inhibition of mitogenic responses.
We expanded MSCs from the bone marrow of healthy subjects. These show the typical features of MSCs with spindle-shaped cells which, after a lag phase of 14 days, grew and became confluent layers of adherent cells. After isolation from second to third passages, cell cultures were uniformly positive for CD166, CD105, CD44, CD29, SH-3 and SH-4 with no detectable contamination of haematopoietic cells (Fig. 2). As previously described, MSCs cultured in osteogenic, adipogenic or chondrogenic media could differentiate, respectively, into osteocytes, adipocytes and chondrocytes (Fig. 1B–D) .
In our study, irradiated MSCs added to lymphocyte cultures induced a proliferative response in some experiments, but not in others (Table 1). This lymphocyte proliferation occurred regardless of the fact that MSCs were autologous or allogeneic to the PBLs. This suggests that the effect of MSCs on lymphocytes was independent of MHC. Moreover, it was independent of the number of MSCs added, ranging from 40,000 to as low as 10 MSCs. MSCs produce interleukin-7 (IL-7) which induce lymphocyte proliferation . However, it is unlikely that as little as 10 MSCs produce sufficient IL-7 to induce proliferation. Other mechanisms will therefore have to be investigated.
When 10,000–40,000 MSCs were added to one-way MLCs, a dose-dependent, inhibitory effect was seen. Suppression was independent of MHC, as it occurred when added MSCs were autologous to stimulatory or responder PBLs or derived from a ‘third party’ donor. These data are consistent with findings in baboon MSCs and with preliminary reports on human cells, where MSCs suppressed MLC responses [14, 17–19]. It is unlikely that the reduction in lymphocyte DNA synthesis is caused by the exhaustion of the culture media or cell crowding in the wells. The RPMI medium in MLCs including MSCs was normal in colour as were control cultures. Trypan Blue staining showed >95% cell viability of both MSCs and lymphocytes at the time of harvest.
If MSCs were added at the beginning of the culture, the inhibitory effect was more marked than when they were added on day 3 or 5. A slight inhibitory effect of 10,000 MSCs was seen when MSCs were added as late as on day 5 of culture. This suggests that MSCs also modulate lymphocyte responses after antigenic stimulation has occurred.
In several experiments, low percentages of MSCs could also induce profound inhibition. The addition of a few MSCs (10–1000 cells) has not been evaluated previously with human or animal MSCs. In our hands, MSCs in low concentrations had less inhibitory and sometimes a stimulating effect on the lymphocyte alloproliferative response. This finding may suggest that surface structures on MSCs act as antigens or mitogens. We speculate that antigens on MSCs may have stimulatory effects in low concentrations acting synergistically with HLA-DR antigens in MLCs or after mitogenic stimulation . At higher concentrations, these antigens have an inhibitory effect owing to paralysis of the responding lymphocyte clones. That these antigens are non-specific and independent of MHC is suggested by the finding that such inhibitory and stimulatory effects in MLCs were also seen using MSCs autologous to stimulator or responder lymphocytes.
Higher concentrations of MSCs not only suppressed alloreactive T-cells, but greatly reduced lymphocyte proliferation caused by potent T-cell mitogens PHA, Con A and SpA. Inhibition by MSCs occurred when both autologous and allogeneic PBLs were stimulated. The data suggest that MSCs exert their effects on T cells, as they are activated in MLCs and by PHA and Con A . Protein A stimulates human T and B lymphocytes . PWM induces a T-cell-dependent B-cell stimulation . It remains to be seen whether MSCs affect B-cell stimulation. Although MSCs are unique cells, as they can be induced by proinflammatory cytokines to express HLA class II, the ability to suppress T-cell alloreactivity and mitogen-induced activation may not be a characteristic unique to these cells.
Several studies suggest that stromal cells in the bone marrow have immune-modulatory effects that enhance HSC engraftment, induce donor-specific tolerance and have a positive effect on autoimmune disorders in experimental animal models [29–32]. Human stromal cells also enhance long-term haematopoietic engraftment after in utero transplantation in sheep .
Although MSCs express MHC molecules, it has been suggested that they are not inherently immunogenic [17–19]. No immunologic response was seen in the host when allogeneic ACI rat MSCs were used to fill a femoral gap in Fischer rat hosts . Murine osteoblast infusions across MHC barriers promote the engraftment of allogeneic HSCs .
The finding that MSCs induced inhibition of MLCs and T cells regardless of MHC has clinical implications. For instance, after allogeneic stem cell transplantation, expansion and infusion of autologous or allogeneic MSCs may be used to enhance engraftment and to modulate the immune response. After high dose chemo-irradiation therapy, stromal function is reduced [36, 37]. Regeneration is limited and complete only in patients less than 20 years of age. Co-transplantation of HSCs and MSCs may improve the reconstitution of the stroma and subsequently also haematopoiesis. As MSCs in large numbers can be obtained from many patients, autologous MSCs may be preferred for practical and ethical reasons. However, in older patients, in those with malignant diseases as well as haematological or bone disorders, sufficient MSCs may not be obtained. In such cases, allogeneic MSCs from healthy volunteer donors may be used.
To conclude, MSCs when added to MLCs and mitogen-stimulated lymphocytes mainly induced inhibition in high concentrations, whereas an enhancement of lymphocyte proliferation was sometimes seen using low concentrations of MSCs. This effect seems independent of MHC. Further studies are warranted to determine the mechanism of action of MSCs on lymphocyte proliferation.
We thank Inger Hammarberg for help in the preparation of the manuscript. This study was supported by grants from the Swedish Cancer Society, the Children's Cancer Foundation the Swedish Medical Research Council, the Tobias Foundation, the Stockholm Cancer Society and the Tore Nilsson Foundation.