Mesenchymal Stromal Cells Fail to Prevent Acute Graft-versus-Host Disease and Graft Rejection after Dog-Leukocyte-Antigen Haploidentical Bone Marrow Transplantation

Abstract

Mesenchymal stromal cells (MSC) have been shown to have immunosuppressive effects in vitro. To test the hypothesis that these effects can be harnessed to prevent graft-versus-host disease (GVHD) and graft rejection after hematopoietic cell transplantation (HCT), we administered a combination of three different immortalized marrow-derived MSC lines (15–30×10⁶ MSC/kg/day, 2–5 times/week) or third-party primary MSC (1.0×10⁶ MSC/kg/day, 3 times/week) to canine recipients (n=15) of dog-leukocyte antigen-haploidentical marrow grafts prepared with 9.2 Gy total body irradiation. Additional pharmacological immunosuppression was not given after HCT. Prior to their in vivo use, the MSC products were shown to suppress alloantigen-induced T cell proliferation in a dose-dependent, major histocompatibility complex-unrestricted and cell contact-independent fashion in vitro. Among 14 dogs evaluable, 7 (50%) rejected their grafts and 7 engrafted with ensuing rapidly fatal acute GVHD (50%). These observations were not statistically different from outcomes obtained with historical controls (n=11) not given MSC infusions (p=0.69). Hence, survival curves for MSC-treated dogs and controls were virtually superimposable (median survival, 18 vs. 15 days, respectively). Finally, outcomes of dogs given primary MSC (n=3) did not appear to be different from those given clonal MSC (n=12). In conclusion, our data fail to demonstrate MSC-mediated protection against GVHD and allograft rejection in this model.

INTRODUCTION

Mesenchymal stromal cells (MSC) suppress proliferation of alloantigen-activated lymphocytes in vitro in a dose-dependent, cell contact-independent and major histocompatibility complex (MHC)-unrestricted fashion [1–4]. Motivated by the immunosuppressive properties observed in vitro, MSC derived from bone marrow have been evaluated in animal models and in human patients for treatment and prevention of graft-versus-host disease (GVHD) and allograft rejection [4–12]. Thus far most of the evidence documenting the in vivo immunosuppressive effects of MSC is based on case reports, small case series and phase II studies. [13] While results have not been overwhelmingly positive, no acute or long-term adverse events, including ectopic tissue formation, after MSC infusion have been reported [14–18].

One obstacle that may limit the effectiveness of MSC in vivo is the relatively low numbers of MSC that can be generated for clinical applications. In a recent phase II study of MSC for treatment of steroid-resistant, severe, acute GVHD, 55 patients were treated with allogeneic MSC at a median dose of 1.4 (range, 0.4–9) × 10⁶ cells per infusion [17]. Eighty-nine percent of patients in this study received a total of only one or two MSC infusions. Even though survival of patients who responded to MSC in this study seemed higher (52% at 2 years) than previously described for patients with a similar severity of acute GVHD, randomized controlled trials unequivocally proving the usefulness of MSC for GVHD prevention or treatment have yet to be published. One recently completed phase III study of MSC for treatment of steroid-refractory acute GVHD has thus far only been presented in abstract form and showed no benefit of this intervention with respect to the prospectively defined primary endpoint of a durable complete response for ≥ 28 days [19]. A subgroup analysis, however, suggested a possible therapeutic benefit for patients with liver and gut involvement.

The effectiveness of treating or preventing GVHD by infusing MSC is reportedly inconsistent. We hypothesized that these inconsistencies could be attributed to several variables: (i) suboptimal numbers of MSC infused, (ii) inter-donor variations in the quality of the marrow harvest, or (iii) varying ratios of MSC-subpopulations leading to functional differences in bulk cultures. To address this concern, we used immortalized clonal populations of canine MSC to provide a consistent product for infusion. First, we demonstrated that their in vitro immunosuppressive potential was comparable to primary MSC. Next, we tested their ability to prevent GVHD and allograft rejection in the canine dog-leukocyte antigen (DLA)-haploidentical hematopoietic cell transplantation (HCT) model. Results indicated that even though clonal MSC could be produced with high efficiency, and were infused frequently and in high numbers, they failed to either mitigate or prevent GVHD, or decrease the likelihood of rejection.

MATERIALS AND METHODS

Immortalized clonal populations of bone marrow-derived MSC

Clonal MSC lines were generated from marrow derived from one donor dog according to procedures previously described for human MSC lines [20]. In brief, canine marrow-derived mononuclear cells were depleted of hematopoietic lineage cells by immune absorption after incubation with anti-CD45 antibodies. The remaining cells were cultured until confluent and then immortalized with a retrovirus containing the human papilloma virus E6/E7 genes. Infected cells were plated at low concentrations and growing clones were isolated with cloning rings. Twenty cloned dog MSC lines designated DS 1–20 were established. Five DS lines were subsequently analyzed in more detail for their ability to suppress the allogeneic MLC in vitro. The 3 lines (DS 1–3) with the strongest suppressive activity in vitro yet the greatest phenotypical differences between each other were chosen for further in vivo experiments. The differences in mRNA profile and cytokine phenotype between DS 1–3 are shown in Figure 1 and Table 2.

Figure 1. Clustering analysis of genes expressed by DS 1–3.

Three separate mRNA preparations were analyzed for each DS line. First, genes on the canine mRNA-arrays (Affymetrix) with raw expression levels >100 were selected. Second, a group of 130 genes with functional annotations allowing them to be classified as “extracellular” by GeneOntology were chosen and found to include secreted factors, cell surface receptors and extracellular matrix proteins. The clustering analysis by the Agilent GeneSpring GX7.3.1 program was based on average linkage, using the Pearson Correlation. DS2 and DS3 differ from DS1, and from each other.

Table 2.

Constitutive production of selected cytokines by immortalized dog marrow stromal cell lines (DS 1–3) and a primary marrow stromal culture (long-term cultures; LTC)

	GM-CSF	IL-4	IL-6	IL-8	IL-10	MCP-1
DS-1	ND	47.6	95.7	7,857	11.8	787.7
DS-2	ND	55.6	72.2	6,002	10.4	2,168
DS-3	ND	47.6	116.3	ND	11.5	27.4
LTC	ND	11.0	87.5	ND	10.3	311.6
Culture media	16.8	17.1	ND	166.1	5.0	20.2

Concentrations of cytokines in 5-day conditioned media from immortalized canine marrow stromal cell lines, DS 1–3, and primary marrow stromal cultures (long-term cultures, LTC) were determined using the Lincoplex system (Millipore, Billerica, MA). Concentrations are given in pg/mL. “ND” designates not detectable. Culture media incubated without stromal cells (bottom row) served as a control.

Transcriptome comparison of DS lines

MSC lines DS 1–3 were grown to semi-confluency and RNA extracted [21]. Using canine mRNA-arrays (Affymetrix, Santa Clara, CA), three separate mRNA preparations were processed for each line. Only those genes with absolute expression values greater than 100 and adequate annotation to be classified as “extracellular”, “chemokines/growth factors”, or “receptors” by Gene Ontology (GO SLIMS) were considered for further analysis. The clustering analysis by the Agilent GeneSpring GX7.3.1 program was based on average linkage, using the Pearson Correlation [21].

Primary MSC cultures (long-term stromal cultures; LTC)

Ex vivo culture conditions for primary canine MSC were adapted from methods described by Le Blanc et al. [2] and Gartner et al. [22]. Briefly, buffy coat cells from marrow aspirates were plated in T-75 flasks (Costar, Cambridge, MA) at 1 to 2 × 10⁶/mL. Adherent cells were grown in LTC medium containing Iscove's modified Dulbecco's medium (IMDM), 12.5% horse serum, 12.5% fetal calf serum, L-glutamine (0.4 mg/mL), sodium pyruvate (1 mmol/L), penicillin (100 U/mL), streptomycin sulfate (100 pg/mL), hydrocortisone sodium succinate (10⁻⁶ mol/L), and P-mercaptoethanol (10⁻⁴ mol/L) and fed weekly by demidepletion. Stromal layers were maintained at 37°C in an atmosphere of 5% CO2. After reaching confluency, the adherent layers were trypsinized once and grown to confluency again in a T-225 flask to deplete hematopoietic cells before harvest and infusion. Experiments were performed with 3-week-old LTCs after cells were confluent.

Phenotypical characterization of stromal cell cultures

Based on availability of canine cell-specific monoclonal antibodies, expression levels of informative stromal and cell-lineage-specific markers were assessed by flow cytometry or semi-quantitative polymerase chain-reaction (RT-PCR). For RT-PCR, mRNA was extracted from stromal cell cultures and transcribed into cDNA using µMACS One-Step cDNA kits (Miltenyi Biotec, CA). PCR reactions were performed using 1.1x PCR Master Mix plates (Thermo Scientific, CA) containing Thermoprime Plus DNA Polymerase, 1.5 mM MgCl₂, and 75 mM Tris-HCl. One hundred nanograms of cDNA template and 2.5 µM of primers (Supplemental Table 1) were added to each reaction. cDNA was amplified during 40 cycles at 94°C for 30 seconds, annealing at 58–62°C for 45–60 seconds, and extension at 72°C for 1 minute. Amplified products were separated on 2% agarose gels and visualized by ethidium bromide fluorescence for gene expression. Canine G3PDH was used as a control for input cDNA and was equivalent in all samples. Expression levels of markers determined by flow cytometry (Supplemental Table 1) were compared to those of isotype-matched control antibodies.

Canine mixed lymphocyte cultures (MLC) and mitogen stimulation

10⁵ responder peripheral blood mononuclear cells (PBMC) were cultured with 10⁵ irradiated (24 Gy), allogeneic, DLA-mismatched stimulator PBMC at a concentration of 10⁶ /mL for 6 days according to established methods [23]. In some experiments, 10⁵ PBMC were stimulated with concanavalin A (conA) at a concentration of 20 µg/ml for 72 hours. Defined numbers (3–50 × 10³) of irradiated (24 Gy) clonal MSC or primary marrow stromal cells were added to MLC or conA stimulation cultures at the beginning of culture. Cells were labeled with 1 µCi/well of ³H-thymidine for the final 16 hours of culture, harvested onto glass fiber filters, and counted for isotope incorporation.

To determine whether MSC-conditioned medium (CM) was suppressive in MLC, CM was produced from MSC-lines (5-day culture), concentrated 10-fold and then added to allogeneic MLC in a 1:9-ratio with fresh media. To further assess whether cell contact was required for MSC-mediated suppression of T cell proliferation, irradiated MSC were cocultured with responder and irradiated stimulator peripheral blood mononuclear cells in allogeneic MLC established on 24-well plates. In these experiments, MSC were either cultured in direct contact with responders/stimulators or separated by a 0.3 µm porous membrane on transwell-inserts.

Hematopoietic cell transplantation

Dogs

Litters of random-bred dogs were either raised at the Fred Hutchinson Cancer Research Center (FHCRC) or purchased from commercial Class A vendors licensed by the U.S. Department of Agriculture. The dogs weighed from 7.4 to 14.4 (median, 11.4) kg and were between 7 and 56 (median, 9) months old. All dogs were enrolled in a veterinary preventive medicine program that included routine anthelmintics and a standard immunization series [24]. The study was approved by the Institutional Animal Care and Use Committee at the FHCRC, which has been fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care. Littermate donor and recipient pairs were selected on the basis of complete family studies showing haplo-identity for highly polymorphic major histocompatibility complex (MHC)-associated class I and II microsatellite markers [25] and for DLA-DRB1 alleles determined by direct sequencing [26].

Transplant regimen

On day 0, donor marrow was harvested and infused intravenously into allogeneic recipients after a single dose of 9.2 Gy total body irradiation (TBI) delivered at 7 cGy/min (Varian Clinac 4; Varian Medical Systems, Palo Alto, CA). The median number of nucleated marrow cells infused was 3.8 ×10⁸ (range, 2.1–9.7 ×10⁸) cells/kg (Table 2). Recipients were given no pharmacological immunosuppression for prevention of GVHD after transplant.

MSC preparation and infusion

MSC lines DS 1–3 and primary, long-term marrow culture (LTC)-derived MSC were expanded to at least semi-confluency ex vivo using T250 culture flasks (Corning, Kennebunk, MN). In adaptation to methods described for culture of human MSC [20,27], ten-percent heat-inactivated fetal calf serum was substituted for pooled dog serum. MSC were then trypsinized, harvested, washed twice and suspended in 0.9% NaCl prior to infusion. The first MSC infusion was given within 30 minutes after marrow graft infusion. Three different MSC regimens were used (Table 2): (1) n=7 recipients; DS1/DS2/DS3 in a 1:1:1 ratio; total dose per infusion, 30 × 10⁶/kg recipient weight; given on 3 days/week during the first week, and on 2 days/week starting the second week after HCT. (2) n=5 recipients; DS1/DS2/DS3 in a 1:1:1 ratio; total dose per infusion, 15 × 10⁶/kg recipient weight; given on 5 days/week. (3) n=3 recipients; primary marrow stromal cells pooled from seven donors; total dose per infusion, 1 × 10⁶/kg recipient weight; given on 3 days/week.

Supportive care and study termination

Prophylaxis with oral antibiotics was given from the day of TBI until the end of the study. Broader antibiotic coverage was administered when neutrophil counts declined to below 0.5 × 10³/µL or fever developed. Intravenous fluids were administered when dehydration occurred. Irradiated blood transfusions were given either when platelet counts declined below 5 ×10³/µL, or when petechiae and ecchymoses of skin and mucous membranes were observed. Clinical signs of acute GVHD included diarrhea, skin erythema, and elevations of liver enzymes [28]. Dogs were euthanized if they were in poor clinical condition and complete necropsies were performed, which allowed histopathological distinction between GVHD and regimen-related toxicities.

Statistical considerations

The studies were designed to use the fewest dogs necessary to have adequate power to detect differences deemed clinically meaningful in comparison to our historical results [29]. Based on 11 dogs given haploidentical grafts and no pharmacological immunosuppression after transplant, the median survival was 15 days, and long-term survival (>100 days) did not occur. It was predetermined that a group size of 7 dogs treated with MSC provided 80% power to detect an increase in median survival to 37 days, at the 1-sided 0.05 level of significance. Seven dogs also provided 80% power to detect an increase in the proportion of long-term survivors from 5% to 38%, based on an exact binomial test at the 1-sided 0.05 level of significance.

Engraftment and chimerism

Hematopoietic engraftment was assessed by increases in granulocyte and platelet counts following postirradiation nadirs, marrow histology from autopsy specimens, documentation of donor-type hematopoiesis in peripheral blood and marrow by variable number tandem repeat (VNTR) polymorphisms [30,31] and clinical and histologic evidence of GVHD.

In vivo distribution of ¹¹¹In-labeled DS1 cells

DS1 cells were labeled in vitro with ¹¹¹In (30 Bq/cell) [32] prior to injection. Labeled cells were infused into one recipient (G-943) immediately after 8 Gy total body irradiation (delivered at 7 cGy/min as a single fraction). Single Photon Emission Computed Tomography (SPECT) images were obtained immediately after injection, and after 1, 2, 4 and 9 days.

Detection of DS cells in vivo

MSC lines DS 1–3 were generated from one donor dog (DLA-DRB1 3/6), which allowed their detection in blood and tissue samples by VNTR-based polymorphism analysis [30,31]. For this purpose, DNA was extracted from blood and tissue preparations. To detect viable MSC, blood or tissue obtained from recipients at different time points after MSC infusion were cultured in RPMI medium supplemented with 10% heat-inactivated dog serum, which allowed for out growth of the immortalized cells. After 72 hours of initial culture, non-adherent cells and debris were rinsed off, and adherent cells were expanded for 7–14 days. Adherent layers were then trypsinized prior to DNA-extraction and chimerism analysis.

RESULTS

Transcriptome comparison of DS lines

In DS lines 1–3, 130 genes with absolute expression values greater than 100 and adequate annotation to be classified as “extracellular”, “chemokines/growth factors”, or “receptors” by Gene Ontology (GO SLIMS) were considered for analysis. The “clustering analysis” of the relative expression of these 130 selected genes shows that the three DS lines differ substantially from each other with respect to their mRNA expression profiles (Figure 1). The complete data set showing transcriptomes for these lines using the Affymetrix microarray platform can be accessed online at: http://webapps.fhcrc.org/labs/graf/grantdata.html.

Phenotypic characterization of DS lines and primary marrow stromal cultures

While primary marrow stromal cultures (long-term marrow cultures; LTC) showed variable expression of hematopoietic lineage markers CD13, CD14 and CD45, these markers were undetectable in DS lines 1–3 (Table 1). In contrast, markers widely associated with cultured human MSC [33,34], such as CD29 (integrin β1-chain), CD73 (ecto-5'-nucleotidase) and CD90 (Thy-1), were consistently expressed by DS 1–3 and by most LTC. Moreover, CD105 (endoglin) was variably expressed by the different LTC tested and consistently expressed by all three DS lines. The three DS lines did not express DLA-class II or CD34.

Table 1.

Phenotypic characterization of immortalized dog marrow stromal cell lines (DS 1–3) and randomly selected primary marrow stromal cultures (long-term cultures; LTC 1–5)

Marker	Detection Method	MSC Type Tested
		DS1	DS2	DS3	LTC1	LTC2	LTC3	LTC4	LTC5
MHC-class II	FC	−	−	−	ND	ND	ND	ND	ND
CD10	RT-PCR	+	+	+	+	+	++	+	+
CD13	RT-PCR	−	−	−	++	++	+++	−	+++
CD14	FC	−	−	−	ND	ND	ND	ND	ND
CD29	RT-PCR	+	+++	+	++	+++	+++	++	++
CD34	FC	−	−	−	ND	ND	ND	ND	ND
CD45	FC/ RT-PCR	−	−	−	+++	++	−	−	+
CD73	RT-PCR	++	+++	+++	++	++	++	+	++
CD90	FC/ RT-PCR	+++	+++	+++	++	++	++	−	−
CD105	RT-PCR	++	+	++	+++	+++	+++	−	−
CD106	RT-PCR	+	−	−	+	−	++	−	−

Expression levels of informative stromal cell- and lineage-specific markers by marrow-derived stromal cell lines (DS 1–3) and five randomly selected primary marrow stromal cultures (LTC 1–5) were assessed by semi-quantitative polymerase chain-reaction (RT-PCR) and flow cytometry (FC) as described in “Materials and Methods”. Expression levels of markers of interest were estimated in relation to that of the housekeeping gene, G3PDH, and isotype-matched control antibodies for RT-PCR and FC, respectively. “ND” designates ‘not done’; “MSC”, ‘mesenchymal stromal cells’; “DS”, dog stromal cell line; “MHC”, ‘major histocompatibility complex’. Expression levels: (+++), strong; (++) intermediate; (+) weak; (−) absent.

Constitutive production of cytokines by DS cells and primary marrow stromal cells

The three DS lines were assayed for cytokine production using the Lincoplex system (Millipore, Billerica, MA). All three lines expressed low, but consistent levels of IL-4, IL-6 and IL-10 and did not secrete detectable levels of GM-CSF, INFγ, CXCL1, IL-2, IL-15, IL-18, or TNFα (Table 2). DS1 and DS2 expressed high levels of IL-8, whereas DS3 produced undetectable levels of IL-8. DS1 and DS2 showed a three-fold difference in MCP-1 production. Hence, in terms of production of this limited panel of cytokines, DS1 and DS2 were more similar, differing considerably from DS3. The 12 cytokines assayed were also represented on the Affymetrix microarrays, and two of them, IL-8 and MCP-1, differed in constitutive expression level in DS3 in contrast to the other two lines. Hence, the transcriptome analysis lent weight to cytokine results and supported the concept of functional heterogeneity among the generated clonal MSC lines. Except for non-detectable levels of IL-8, the cytokine profile of primary marrow stromal cells (LTC) was similar to that of DS 1–3.

Suppression of alloantigen-driven T-cell proliferation by DS cells

DS1, DS2 or DS3 cells were added at the initiation of allogeneic MLC at a 1:2 DS cell/responder cell ratio (Figure 2A). The results indicated that the three DS lines had variable inhibitory effects on T cell proliferation. While DS3 resulted in >80% suppression in MLC, DS1 or DS2 resulted in 50–70% suppression.

Figure 2. Inhibition of T cell proliferation in canine allogeneic mixed lymphocyte cultures (MLC) by different marrow stromal cell lines (DS 1–3) under contact and non-contact conditions, and by stromal cell-conditioned media.

(A) Standard MLCs were set up using peripheral blood mononuclear cells (PBMC) from unrelated, dog-leukocyte antigen-mismatched dog pairs. Different MSC-lines, designated DS1-DS3, were irradiated and added at the initiation of MLC at 2:1 responder/MSC ratios. All three DS lines tested had suppressive effects on T cell proliferation. (B) Five-day DS cell-conditioned medium (CM) was produced, concentrated 10-fold and then added to allogeneic MLC in a 1:9-ratio with fresh media. Left panel, results for CM from individual DS-lines. Right panel, results for CM produced by coculture of the three DS-lines (DS1/2/3-CM). (C) DS1 or DS3 cells were irradiated and cocultured with responder and irradiated stimulator PBMC in allogeneic MLC established on 24-well plates. DS cells were either cultured in direct contact with responders/stimulators or separated by a 0.3 µm porous membrane on transwell-inserts (+ Insert). “Allo Control”, allogeneic MLC without addition of DS cells; “Auto Control”, MLC without addition of DS cells using autologous responder and stimulator cells. Shown is one of 3 representative experiments. Results for each experimental condition represent the mean ± standard error of the mean of 6 replicate wells. (*) indicates a statistically significant decrease in proliferative activity compared to Allo Controls (p<0.05; Mann-Whitney test, two-tailed).

MSC-conditioned media (CM) were then tested for their ability to suppress T cell proliferation in allogeneic MLC. For this purpose, 5-day CM was produced and 10-fold concentrated prior to addition to MLC at the initiation of culture (final MSC-CM concentration, 1x). As shown in Figure 2B, MSC-CM was only minimally effective in suppressing the allogeneic reaction, with DS3-CM having the strongest impact (approximately 25% suppression). CM produced by coculture of three DS-lines (DS1/2/3-CM) was equally ineffective in suppressing the proliferative response as CM derived from culture of individual DS-lines (Figure 2B).

Suppression of T cell proliferation by DS cells is contact-independent

To determine whether the relative lack of suppression of T cell proliferation with MSC-CM was due to the lack of cell-contact, transwell-experiments in which DS1 or DS3 cells were set up in the upper chamber and separated by a porous 0.3 µm membrane from the MLC in the lower chamber (Figure 2C). DS cells separated from responder and stimulator cells were as effective in suppressing the allogeneic MLC as DS cells added directly to responder and stimulator cells.

DS cells suppress alloantigen and mitogen-stimulated T cell proliferation in a dose-dependent fashion

DS1, DS2 and DS3-cells, used singly or in combination (DS1/DS2/DS3 in 1:1:1 ratios; DS-Mix), were added to allogeneic MLC (Figure 3A) or mitogen-stimulated PBMC (Figure 3B) in graded numbers. The inhibition of T cell proliferation by DS-cells was cell dose-dependent, with DS3 cells and the DS-Mix having the strongest suppressive effects. Allogeneic MLC performed with titrated numbers of DS1/DS2/DS3 (DS-Mix) in the presence of the prostaglandin-synthesis inhibitor indomethacin (20 µM) reversed DS-cell-mediated suppression of T cell proliferation by 32–56%, suggesting a partial role of prostaglandin-E in DS-cell-mediated suppression of T cell proliferation.

Figure 3. Canine marrow stromal cell lines, DS1, DS2 and DS3, inhibit alloantigen and mitogen-induced T-cell proliferation in a dose-dependent fashion.

Different numbers (3.12– 50.0 × 10³) of DS1, DS2 and DS3 cells, used singly or in combination (DS1/DS2/DS3 in 1:1:1 ratio; DS-Mix) were irradiated and added to 10⁵ canine peripheral blood mononuclear cells (PBMC) that were stimulated with 10⁵ irradiated allogeneic stimulator PBMC in allogeneic mixed leukocyte culture (MLC) or concavalin A (ConA). Shown are one of 5 representative experiments using individual DS-lines, and one of 2 representative experiments using the DS-Mix. Results for each experimental condition represent the mean ± standard error of the mean of 6 replicate wells. (*) indicates a statistically significant decrease in proliferative activity compared to controls cultured without DS-cells (p<0.05; Mann-Whitney test, two-tailed).

In vivo distribution of ¹¹¹In-labeled MSC

To better understand the in vivo distribution of immortalized marrow stromal cells, DS1 cells were labeled with ¹¹¹In and injected intravenously into one beagle dog. The distribution of the label was then monitored over time using a gamma-camera. Based on the assumption that the detected radioactivity was associated with DS1 cells, we found that DS1 cells accumulated in the lung immediately after infusion (day 0), but then left the lung within 24 hours and preferentially redistributed to liver and spleen; weaker signals were detectable in bone marrow and gut (Figure 4). Residual label in liver and spleen could be detected as long as 9 days after infusion.

Figure 4. In vivo distribution of ¹¹¹In-labeled DS1 cells in a beagle dog (G-943).

DS1 cells were labeled with ¹¹¹In in vitro and then injected into a beagle dog immediately after 8 Gy total body irradiation (delivered at 7 cGy/min as a single fraction). The dose of labeled and subsequently injected DS1 cells was 28 × 10⁶/kg recipient weight. Single Photon Emission Computed Tomography (SPECT) images were obtained immediately after injection, and after 1, 2, 4 and 9 days. Shown are anterior views in cranio-caudal (top to bottom) orientation.

Infusions of DS cells or primary marrow stromal cells after DLA-haploidentical marrow transplantation do not prevent GVHD or graft rejection

Dogs given DLA-haploidentical marrow grafts without pharmacological postgrafting immunosuppression (n=11) have a median survival of 15 (range, 13–25) days (summarized in Table 3) [29]; seven recipients (64%) rejected their grafts and four (36%) engrafted with ensuing lethal GVHD. Based on this historical experience, we asked the questions whether postgrafting MSC infusions, by virtue of their putative immunosuppressive effects, may (i) increase rates of engraftment and (ii) mitigate GVHD after engraftment. Thus, MSC were infused after transplant according to the three regimens detailed in ‘Materials and Methods’ (Table 3). Three dogs (dogs no. 13–15) were given primary, LTC-derived MSC that were pooled from seven different marrow donors. Even though the doses of primary MSC were substantially lower than those of DS 1–3 cells administered (3 × 10⁶/kg/week vs. 60–90 × 10⁶/kg/week), they were consistent with MSC doses given in clinical GVHD treatment studies [17,18].

Table 3.

Graft composition and outcome of DLA-haploidentical marrow transplants with and without MSC infusions following 9.2 Gy TBI

Dog #	Donor ID	Recip. ID	Recip. Weight (kg)	DLA-DRB1			Bone Marrow Cell Doses			MSC Treatment			Rejec- tion	Clinical and Histological GVHD			Survival (Days)	Hemato- poietic Chimerism at Necropsy
				Donor	Recip.	MSC	TNC (×108/ kg)	CD34 (×106/ kg)	CD3 (×107/ kg)	MSC Type	Number of Doses/ Week	Single Dose (×106 cells/kg)		Skin	Liver	Gut
1	G-897	H-003	13.2	3/6	3/9	3/6	3	0.9	1.02	DS1/2/3a	2–3	30	No	Yes	Yes		17	Donor
2	G-993	G-988	12.6	6/9	9/17	3/6	2.1	0.42	1.28	DS1/2/3	2–3	30	No	Yes	Yes	Yes	18	Donor
3	G-896	H-006	14.3	9/20	6/20	3/6	3.8	2.28	1.44	DS1/2/3	2–3	30	Yes				18	Host
4	H-030	H-029	13.4	3/22	20/22	3/6	5.5	1.1	2.53	DS 1/2/3	2–3	30	No	Yes	Yes		18	Donor
5	H-098	H-097	9.1	2/15	2/17	3/6	3.6	0.36	1.22	DS 1/2/3	2–3	30	Yes				18	Host
6	H-095	H-096	10.6	2/3	2/15	3/6	3.8	1.52	2.55	DS 1/2/3	2–3	30	Yes				16	Host
7	H-092	H-093	7.4	3/9	9/15	3/6	7.1	4.26	4.47	DS 1/2/3	2–3	30	NE				3	NE
8	H-101	H-100	11.5	9/17	9/15	3/6	6.8	5.44	3.67	DS 1/2/3	5	15	Yes				18	Host
9	H-139	H-130	7.8	15/22	19/22	3/6	5.0	6.5	2.8	DS 1/2/3	5	15	Yes				>280**	Host
10	H-053	H-048	9.2	20/22	3/22	3/6	9.7	5	0.97	DS 1/2/3	5	15	No	Yes			14	Donor
11	G-550	H-098	10.4	2/9	2/15	3/6	8.6	19.8	2.84	DS 1/2/3	5	15	No	Yes			13	Donor
12	G-550	H-099	14.5	2/9	2/15	3/6	2.3	1.16	2.07	DS 1/2/3	5	15	No	Yes	Yes	Yes	18	Donor
13	H-178	H-176	11.4	14/15	6/14	*	3.7	2.22	2.52	Primaryb	3	1	No	Yes	Yes		14	Donor
14	H-180	H-175	9.5	6/9	6/14	*	3.5	2.8	3.74	Primary	3	1	Yes				25	Host
15	H-180	H-179	11.5	6/9	6/14	*	5	5	4.9	Primary	3	1	Yes				24	Host
	Median		11.4				3.8	2.3	2.5								18

Without MSC infusions: Historical data [29]
No. of Dogs studied	Recip. Weight (kg)	Bone Marrow Cell Doses Median (Range)			Rejec tion N (%)	GVHD N (%)	Median Survival Days (Range)
		TNC (×108/ kg)	CD34 (×106/ kg)	CD3 (×107/ kg)
N=11	10.1 (6.5–15)	3.1 (1.0– 4.4)	2.8 (2.0– 4.9)	1.3 (1.2– 2.9)	7 (64)	4 (36)	15 (13–25)

^*The DLA-DRB1 types of the different primary MSC donors were as follows: 1/6, 3/9, 3/15, 6/9, 9/17, 9/27, and 15/22.

^**Dog rejected and, after a long duration of profound pancytopenia, survived with autologous hematopoietic recovery.

^aDS cell lines 1–3 were always infused at a 1:1:1 ratio.

^bPrimary MSC products in all 3 treated dogs were generated from 7 different donors as described in ‘Methods’, pooled, and infused at approximately equal ratios. MSC, designates mesenchymal stromal cells; DS, dog stromal cell line; DLA, dog leukocyte antigen; TBI, total body irradiation; GVHD, graft-versus-host disease; TNC, total nucleated cells; NE, not evaluable.

Compared to historical controls transplanted according to similar supportive care standards, postgrafting MSC infusions did not significantly improve survival (15 [range, 13–25] days vs. 18 [range, 3–>280] days), rates of rejection (64% vs. 50%) and rates of GVHD (36% vs. 50%) (p=0.69; Fisher's Exact test) (Table 3). One dog (H-093) died on day 3 after transplant from acute gastrointestinal toxicity related to the conditioning regimen and, therefore, was not evaluable for rejection or GVHD. One dog treated with DS cell infusions (H-130) rejected the marrow graft and survived with autologous hematopoietic reconstitution. Even though numbers were too small for meaningful intergroup comparisons, there was no MSC regimen that appeared effective in preventing the two major endpoints of this clinical study namely, rejection and GVHD. While 3 of 12 recipients given DS cell infusions had received DLA-haploidentical MSC (Table 3), 9 recipients had received fully mismatched MSC. There was no indication that DLA-compatibility between the MSC-product and the recipient had an effect on rates of rejection and GVHD. The Kaplan-Meier survival curves for MSC-treated recipients and historical controls are shown in Figure 5.

Figure 5. Survival of dogs prepared with 9.2 Gy total body irradiation followed by bone marrow transplantation from DLA-haploidentical donors.

No pharmacological immunosuppression was given after transplant. (MSC) marrow stromal cell infusions were given after transplant (n=15). (Control) No MSC infusions were given after transplant (n=11) [29]. One dog given MSC (H-130; see also Table 2) rejected the graft and, after prolonged myelosuppression, survived with autologous hematopoiesis (absolute neutrophil and platelet counts exceeded 0.5 × 10³/µL and 20 × 10³/µL on days 40 and 56 after transplant, respectively). Survival in the two groups was not significantly different (p=0.18; log-rank test).

Detection of MSC in blood and tissues

Of the 12 recipients given combined infusions of DS 1–3 cells, seven recipients were analyzed for in vivo detection of MSC after infusion (Table 4). Immediately after intravenous infusion, DS cells were detected in the blood by PCR in only 1 of 7 recipients evaluated. At 24, 48 and 72 hours after infusion, PCR products specific for DS cells could be detected in 3 of 6, 1 of 4, and 1 of 4 recipients, respectively, that were evaluated. However, after 7–14 days DS signals could be detected in the blood of all seven recipients. Tissue samples obtained from six recipients at necropsy were also subjected to PCR analysis (Table 4). Without prior ex vivo expansion of stromal elements, DS cell-specific signals were not detected in any of the samples tested. However, after ex vivo expansion DS cell-specific signals were seen in bone marrow, lung and spleen in 2 of 6 dogs evaluated. None of the dogs infused with DS cells showed clinical or histopathological evidence of tumor formation even though cells were immortalized and clonal in nature.

Table 4.

Detection of third-party DS cells in blood and tissues

Dog	Rec. ID	Blood or Tissue	MSC Detectiori										Necropsy
		Enriched for MSC	In Blood (Hours after Infusion)				In Tissue at Necropsy
#		Prior to Detection by VNTR	0	24	48	72	BM	Lung	Liver	Spleen	LN	Day after Transplant	BM Cellularity (%)	Chimerism
1	H-003	No	Yes	No			No	No	No	No	No	17	50	Donor
		Yes	Yes	Yes			Yes	Yes	ng	No	No
2	G-988	No	No				No	No	No	No	No	18	50	Donor
		Yes	Yes				No	ng	ng	ng	No
3	H-006	No	No	No	No		No	No	No	No	No	18	5	Host
		Yes	Yes	Yes	Yes		Yes	Yes	ng	Yes	ng
4	H-029	No	No	No		No	No	No	No	No	No	18	80	Donor
		Yes	Yes	Yes		Yes	No	No	No	No	No
5	H-097	No	No	No	No	No	No	No	No	No	No	18	0	Host
		Yes	Yes	No	No	No	ng	ng	ng	ng	ng
6	H-096	No	No	No	No	No	No	No	No	No	No	16	0	Host
		Yes	Yes	No	No	No	No	No	No	No	No
7	H-130	No	No	No	No	No	N/A	N/A	N/A	N/A	N/A	Alive	N/A	Host
		Yes	Yes	No	No	No

Mesenchymal stromal cell (MSC) lines DS1-3 were generated from one donor dog, which allowed their detection in blood and tissue samples by VNTR-based polymorphism analysis. For this purpose, DNA was extracted from unmodified or MSC-enriched white blood cell (WBC) and tissue preparations. To enrich for viable MSC, WBC obtained from recipients at different time points after MSC infusion, or minced tissue samples obtained at necropsy were cultured in RPMI media supplemented with 10% heat-inactivated dog serum on 12 well-plates. After 72 hours of initial culture, non-adherent cells and debris were rinsed off, and adherent cells were expanded for 7–14 days. Adherent layers were then trypsinized prior to DNA-extraction and subjected to chimerism analysis as described in “Materials and Methods’. In cases where DS cells were detectable by chimerism analysis after ex vivo expansion of adherent stromal cells (“Yes”), DS cell-specific signals consistently exceeded 80%.

‘BM’ designates bone marrow; ‘LN’, lymph node; ‘ng’, no growth; ‘N/A’, not applicable.

DISCUSSION

By administering third-party, bone marrow-derived MSC in large numbers and at high frequency after DLA-haploidentical bone marrow transplantation, we hoped to prevent or at least mitigate acute GVHD. In addition, by giving bone marrow that was not supplemented with a T cell-rich buffy-coat and by not using pharmacological immunosuppression after transplant, this well-established preclinical model allowed us to use graft-rejection as an additional read-out for MSC-mediated immunosuppression in vivo.

The majority of recipients were given a combination of three different marrow stromal cell lines (derived from one donor dog) that shared phenotypical features associated with human MSC [33,34] and the ability to suppress allogeneic T cell proliferation in vitro. In fact, their T cell suppressive potential in vitro was similar to or exceeded that observed with primary LTC-derived marrow stromal cells [12]. Analogous to what has been reported by others with primary MSC in vitro, T cell suppression by DS cells was both unrestricted by antigens encoded by the MHC and contact-independent. Even though DS cells suppressed the T cell proliferative response to alloantigen in a contact-independent manner, DS cell conditioned medium was only minimally suppressive. This finding suggested that soluble factors released during the allogeneic immune response may induce MSC to release a soluble mediator that interferes with T cell function. Several soluble factors, including transforming growth factor-β, hepatocyte growth factor and prostaglandin E2, among others, have been proposed to mediate suppressive effects of human MSC [35–37] and canine MSC [12] in vitro. Our findings also suggest that prostaglandins are involved in DS-mediated immunosuppression in vitro.

Although all three DS lines used in our study suppressed alloantigen and mitogen-induced T cell proliferation in vitro, they differed from each other with respect to their immune-phenotype, mRNA expression profile and repertoire of secreted cytokines. The use of MSC products composed of three “phenotypically” different MSC lines isolated from the same LTC, was done to provide an unlimited amount of a consistent cell product for experimental use that would mimic the heterogeneous composition of primary MSC.

Our results in this established canine model showed that, compared to historical controls (n=11) given 9.2 Gy TBI and DLA-haploidentical marrow grafts without MSC infusions after transplant, recipients of MSC (n=15) experienced similar rates of marrow graft rejection. Moreover, all seven MSC recipients which had donor marrow engraftment (50%) developed vicious and rapidly fatal acute GVHD, identical to the experience with historical controls. As a result, survival curves for MSC-treated dogs and controls were virtually superimposable. Even though a meaningful subgroup analysis was limited by small numbers, there was no MSC regimen that appeared effective in preventing the two major endpoints of this preclinical study, rejection and GVHD. Primary MSC were also ineffective in preventing or treating GVHD and rejection. This finding was consistent with a recently published study by Lee et al. [12] showing that donor-derived primary MSC did not prevent marrow graft rejection after DLA-identical HCT. There was also no indication in our study that the degree of DLA-compatibility between MSC products and recipients affected rates of rejection and GVHD. Finally, it is important to emphasize that, since therapeutic benefits associated with MSC infusions were not found, a significant bias related to the use of historical controls appears unlikely.

Additional analysis showed that ¹¹¹In-labelled DS1 cells accumulated in the lungs immediately after intravenous infusion and, within 24 hours, redistributed to liver, spleen and bone marrow. Hence, the in vivo distribution of DS1 cells was similar to that described for primary marrow-derived MSC studied in other animal models [38,39]. In addition, the in vivo distribution of DS1 cells in this irradiated dog was not different from that in two non-irradiated dogs studied previously (not shown). A limitation of this analysis, however, was the fact that the distribution of the ¹¹¹In-label may not have been indicative of viable cells since detection of ¹¹¹In-signals in parenchymal organs could be due to retention of free ¹¹¹In-label by macrophages. We therefore sought to detect viable DS cells in various organs at the time of necropsy. For this purpose, stromal cells were expanded from organ samples prior to subjecting them to DS cell-specific PCR analyses. Even though expansion of stromal cells from organ samples was not uniformly successful, DS cells could be expanded from the lungs, spleen and marrow of two dogs. These limited data suggested that, after high-dose irradiation of recipients, allogeneic DLA-mismatched DS cells could migrate to and survive in parenchymal organs for several days after infusion. Since detection of DS cells was virtually impossible in organ samples and PBMC without ex vivo expansion of stromal cells, the relative frequency of circulating and tissue-based DS cells after infusion appeared to be low.

In conclusion, our results obtained in the clinically relevant canine model of allogeneic HCT indicate that marrow-derived clonal and primary MSC, even though strongly immunosuppressive in vitro, have no identifiable immunosuppressive activity in vivo. MSC were administered frequently and in doses typically undeliverable in clinical transplantation; however, the MSC regimens used in this study failed to prevent graft rejection and GVHD, the two major immunological barriers of allogeneic transplantation. It is possible that the DS lines used in our study were not representative of primary MSC used in and associated with successful treatment of GVHD in clinical studies [17]. However, outcomes with primary MSC in our study were not different from those with DS cells. Due to the rigorous nature of the model used, it is also possible that the addition of MSC to a regimen of pharmacological immunosuppression might have been more successful. At the very least, this data emphasizes a disconnect between in vitro and in vivo data. Given that the medical literature tends to favor positive study results, our study should serve as a counterpoint emphasizing that the in vivo immunosuppressive effects of MSC have not been unequivocally defined.