Article Text

Extended report
Mesenchymal stem cells are conditionally therapeutic in preclinical models of rheumatoid arthritis
  1. Anastasia Papadopoulou1,2,
  2. Minas Yiangou2,
  3. Evangelia Athanasiou1,
  4. Nikolaos Zogas1,2,
  5. Panayotis Kaloyannidis1,
  6. Ioannis Batsis1,
  7. Athanasios Fassas1,
  8. Achilles Anagnostopoulos1,
  9. Evangelia Yannaki1
  1. 1Gene and Cell Therapy Center, Hematology Department–BMT Unit, George Papanicolaou Hospital, Thessaloniki, Greece
  2. 2Department of Genetics, Development and Molecular Biology, School of Biology, Aristotle University, Thessaloniki, Greece
  1. Correspondence to Evangelia Yannaki, George Papanicolaou Hospital, Gene and Cell Therapy Center, Hematology Department–BMT Unit, Thessaloniki 57010, Greece; eyannaki{at}u.washington.edu

Abstract

Objective The role of mesenchymal stem cells (MSC) in experimental arthritis is undoubtedly conflicting. This study explored the effect of bone marrow-derived MSC in previously untested and pathogenetically different models of rheumatoid arthritis (RA).

Methods MSC were tested both in an induced (adjuvant-induced) and a spontaneous (K/BxN) arthritis model. Arthritis was assessed clinically and histologically. The proliferation of splenocytes and fibroblast-like synoviocytes (FLS) in the presence of MSC was measured by radioactivity incorporation. Toll-like receptor (TLR) expression was measured by real-time PCR. T-regulatory cell (Treg) frequency, T-cell apoptosis and cytokine secretion were monitored by flow cytometry.

Results MSC, in vitro, strongly inhibited critical cell populations; splenocytes and FLS. In contrast, MSC proved ineffective in vivo, unless they were administered before disease onset, an effect implying that the inflammatory arthritic milieu potentially abrogates MSC immunomodulatory properties. In order to alleviate inflammation before MSC infusion, the authors administered, at arthritis onset, a short course with a proteasome inhibitor, bortezomib, whereas MSC were infused when established disease was expected. The bortezomib plus MSC group demonstrated a significantly decreased arthritis score over arthritic, MSC-only, bortezomib-only groups, also confirmed by histology and immunohistochemistry. The bortezomib plus MSC combination restored TLR expression and Treg frequency in blood and normalised FLS and splenocyte proliferation, apoptosis and cytokine secretion.

Conclusion MSC lose their immunomodulatory properties when infused in the inflammatory micromilieu of autoimmune arthritis. Conditioning of the recipient with bortezomib alters the disease microenvironment enabling MSC to modulate arthritis. Should milieu limitations also operate in human disease, this approach could serve as a strategy to treat RA by MSC.

View Full Text

Statistics from Altmetric.com

Mesenchymal stem cells (MSC) hold promise for repairing damaged tissues, as they have been shown potentially to differentiate into a broad spectrum of tissues overcoming the barriers of germ layers.1 ,2 They also exert, in a non-major histocompatibility complex (MHC)-restricted fashion, a number of unique immunological properties by inhibiting the proliferation and cytokine production by T, B, natural killer and dendritic cells.3,,5 These features have paved the way for their clinical use in graft versus host disease (GvHD) after allogeneic haematopoietic cell transplantation,6 and have also provided the rationale for their possible use in the treatment of autoimmune diseases.

Although several reports have demonstrated the efficacy of MSC in animal models of autoimmune diseases such as allergic encephalomyelitis,7 diabetes mellitus8 and autoimmune enteropathy,9 their role in treating experimental arthritis has been thus far conflicting. MSC were shown to ameliorate collagen-induced arthritis (CIA),10,,14 and in sharp contrast, to provide no benefit and/or an adverse effect in CIA.15,,19 The positive MSC effect was shown to be MHC-independent and syngeneic or allogeneic MSC provided similar results.11 ,14 Efficacy was, however, demonstrated within a narrow window of MSC application; MSC were effective at an early disease phase,10 ,12,,14 but failed to ameliorate or reduce the incidence of arthritis when infused in advanced disease.11 ,14

Prompted by the contradictory effects of MSC in CIA, we aimed to explore the effect of bone marrow MSC in two different in pathogenesis and previously untested animal models of rheumatoid arthritis (RA): the adjuvant-induced arthritis (AIA) and the spontaneous K/BxN arthritis model. Both models share clinical and histological features with human RA, including inflammation of the joints, synoviocyte proliferation, pannus formation, synovitis and cartilage and bone erosion.20 It was recently shown by our group that bortezomib, the first marketed proteasome inhibitor and a drastic nuclear factor kappa B inhibitor, ameliorates advanced AIA by reducing inflammation and bone loss.21 We thus tested the hypothesis that a short course with bortezomib, at disease onset and before MSC infusion, could attenuate the highly inflammatory milieu and maximise the in-vivo performance of MSC, providing clinical benefit in experimental arthritis.

Materials and methods

Animals and reagents

In the AIA model, the donor–recipient pairs consisted of 7-week old, Fisher–Fisher rats in the syngeneic and pseudo-autologous transplantation setting and Fisher–Lewis rats in the allogeneic condition. In the K/BxN model, C57Bl/6 mice served as allogeneic-MSC donors to K/BxN mice generated by crossing KRN with NOD.Ea16 transgenic mice.

AIA was induced by intradermal, tail base injection of 0.1 ml complete Freund's adjuvant (FCA) (0.6 mg of heat-inactivated Mycobacterium butyricum; Difco Laboratories, Detroit, Michigan, USA). Inflammation usually starts 11–13 days after induction whereas disease is established on days 20–24. In the K/BxN model, severe spontaneous erosive arthritis (SSEA) arises 30–35 days after birth.22

Bortezomib (Velcade) was kindly provided by Janssen-Cilag , International NV, Beerse, Belgium.

The study was approved by the Institutional Animal Care and Use Committee.

MSC culture

Mouse whole bone marrow cells or rat mononuclear cells obtained by Ficoll gradient (Biochrom, Berlin, Germany), were plated at a concentration of 16.5×104 cells/cm2 in α-MEM or IMDM, Berlin, Germany medium, respectively, containing 20% fetal bovine serum, 100 U/ml penicillin/100 μg/ml streptomycin, 2 mM l-glutamine (Gibco, CA, USA).Cells were incubated at 37oC in a 5% carbon dioxide atmosphere, and after 7 days adherent cells were detached (rat MSC: trypsin–0.25% EDTA; Gibco/murine MSC: TrypLE Select, Gibco) and expanded. MSC at passage 3–5, either as wild-type or as cytokine-primed, were used in the experiments. Cytokine-activated MSC were obtained by pulsing MSC with 50 ng/ml tumour necrosis factor α (TNFα) or/and interferon γ (IFNγ) for 8 h before MSC or MSC supernatant were added to splenocyte or fibroblast-like synoviocyte (FLS) culture, respectively.

Splenocyte and FLS proliferation assay

Concanavalin A (ConA)-stimulated splenocytes, prepared as previously described,21 were cultured in the presence or absence of MSC supernatant or co-cultured with syngeneic or allogeneic MSC at a ratio of 1:0.2, 1:0.1, 1:0.05, for 3 days. Cells were pulsed with 0.4 μCi 3H-thymidine/well for 18 h and collected with a cell harvester (Perkin Elmer, MA, USA).

FLS were seeded at a density of 2×104 cells/well (2×104 cell/well) and cultured in the presence or absence of MSC supernatant. After 3 days, 0.5 μCi3 H-thymidine/well was added and cells were treated as previously described.21

The incorporated radioactivity was measured by a scintillation counter (LS6500; Beckman Coulter, CA, USA).

Chemotaxis assay: invasion assay

Splenocytes

Normal or AIA–ConA-stimulated splenocytes were co-cultured for 72 h with allogeneic MSC (splenocytes:MSC 1:0.2) and observed under a light microscope. They were also cultured for 72 h in the upper chamber of a transwell system (5.0 μm; Costar, Corning Inc, NY, USA) the bottom chamber of which contained MSC or medium; migration towards MSC or medium was quantitated as %CD3+/CD45+ cells by flow cytometry.

Fibroblast-like synoviocytes

The invasiveness of AIA FLS was evaluated in a transwell system (8.0 μm; Costar) as previously described.21 Briefly, 2×104 FLS were seeded in the inner matrigel-coated (matrigel basement membrane matrix; BD Becton Dickinson, NJ, USA) wells and IMDM/20% fetal bovine serum with or without MSC supernatant was added in the outer wells. After 72 h, the cells that invaded matrigel were fixed with 4% paraformaldehyde, washed, stained with haematoxylin–eosin and observed under a light microscope.

Experimental design

AIA model

The in-vivo effect of MSC was tested by using different cell doses (0.5–30×106), multiplicity of infusions (1-3), MHC compatibility (syngeneic, allogeneic), routes of administration (intravenous, intrasplenic, intraperitoneal, intra-bone marrow), timing of delivery (before disease onset, early disease, advanced disease) and cell manipulation processes (wild-type or TNFα plus IFNγ-primed MSC).

AIA and SSEA model

Animals pretreated with three intraperitoneal doses of bortezomib (AIA rats: 0.25 mg/kg; K/BxN mice:1 mg/kg), received allogeneic or pseudo-autologous (from arthritic AIA donors) MSC (AIA: 20×106/SSEA: 12×106), by two intraperitoneal injections, when advanced disease was expected.

Control groups consisted of PBS, bortezomib-only and MSC-only-treated animals. Animals were killed 5 days after MSC infusion. In the AIA model, a number of PBS and bortezomib-only-treated rats were killed just before MSC infusion, in order to evaluate the effect of bortezomib on disease parameters.

Flow cytometry analysis

For T regulatory cell (Treg) analysis, peripheral blood was stained using CD4-FITC (BD), CD25-PE (BD), Foxp3-PerCP (eBioscience, San Diego, CA, USA) antibodies and the Foxp3 fixation/permeabilisation concentrate and diluent (eBioscience). Apoptosis was measured as annexinV+/7AAD- events (BD), and cytokine secretion in splenocyte and FLS supernatant was quantitated by cytometric bead array (BD). Results were obtained on a FACSCalibur device (BD) and analysed with the CellQuest Pro6 software.

Real-time quantitative reverse transcription PCR

Total RNA was isolated using the QIAmp RNA kit (Qiagen, Düsseldorf, Germany), reverse transcribed with RT2 First Strand kit (Superarray Biosciences, Maryland, USA) and amplified by real-time PCR using the ABI 7500 detection system (Applied Biosystems, CA, USA), Toll-like receptor (TLR) 2, 3, 4, TNFα, interleukin (IL)-6 primers and quantitative PCR master mix (Superarray Biosciences). The comparative CT method was used for relative quantification using GAPDH as the endogenous control.

Polymerase chain reaction

Total DNA was isolated using the QIAmp DNA Kit (Qiagen). PCR primers for sry gene (sex determining region y, 5′-CC CGCGGAGAGAGGCACAAGT-3′ and 5′-TAGG GTCTTCAG TCTCTGCGC-3′) and for beta actin gene (5′-ACCCACACT GTGCCCATCTA-3′ and 5′-GCCACAGGATTCCATACCCA-3′) were designed using Primer3 software. The reaction was held by platinum Taq DNA polymerase (Invitrogen, CA, USA) in a DNAEngine thermocycler (BioRad, CA, USA).

Statistical analysis

Results are expressed as mean±SEM. Differences between experimental conditions were analysed by the analysis of variance general linear model. Analyses were followed by Tukey's post-hoc test for multiple comparisons. A p value less than 0.05 was considered statistically significant.

A description of MSC and FLS characterisation assays, FLS culture, assessment of arthritis, joint histology, immunohistochemistry and fluorescence in-situ hybridisation is found as supplementary material, available online only.

Results

MSC chemoattract T lymphocytes and suppress splenocyte proliferation without inducing T-cell apoptosis

Rat and mouse MSC were phenotypically characterised by flow cytometry and differentiation into osteocytes and adipocytes (see supplementary figure S1, available online only).

Under microscopy, ConA-stimulated normal or AIA splenocytes were observed to migrate actively into close proximity to the co-cultured, spindle-shaped MSC and in a transwell system, CD3+/CD45+ cells migrated towards MSC at a higher frequency compared with control migration (figure 1A,B).

Figure 1

The in-vitro effect of mesenchymal stem cells (MSC) on splenocytes and fibroblast-like synoviocytes (FLS) from animals with arthritis. (A) Spontaneous migration of concanavalin A (ConA)-stimulated-splenocytes from normal (I) or adjuvant-induced arthritis (AIA) (II) rats towards MSC in co-culture (red arrows). Magnification ×200. (B) Quantification of T-cell chemoattraction by MSC in a transwell system, compared with the no MSC condition. Migration is presented as the ratio of the percentage of migration of T lymphocytes in the presence of MSC to the percentage of migration of T lymphocytes in the absence of MSC. (C–D) Proliferation of ConA-stimulated splenocytes from AIA (C) and severe spontaneous erosive arthritis (SSEA) (D) animals in the presence of increasing concentrations of syngeneic or allogeneic MSC. (E) Proliferation of ConA-stimulated AIA splenocytes in the presence of increasing concentrations of syngeneic or allogeneic MSC supernatant. (F) Proliferation of AIA FLS in the presence of increasing concentrations of syngeneic or allogeneic MSC supernatant. Proliferation is presented as the percentage of maximal response. The maximal response (100%) was considered the proliferation of splenocytes or FLS in the absence of MSC or MSC supernatant. (G) FLS invasiveness of AIA rats in the (II) presence or (I) absence of MSC supernatant in a transwell system. Magnification ×100. Experiments were performed independently, at least three times for each condition described (n=6/condition for B–F). **p≤0.001 versus arthritic cells autologous MSC or MSC supernatant; ##p≤0.001 versus normal cells without MSC, by Tukey's test.

ConA-challenged, syngeneic or allogeneic splenocytes from AIA and SSEA animals were strongly inhibited, in a dose- dependent manner, in the presence of increasing concentrations of MSC or MSC supernatant derived from confluent cultures (figure 1C–E). The MSC proliferation in the presence of ConA was negligible compared with splenocytes (see supplementary figure S2, available online only), thus in splenocytes plus MSC co-cultures, the measured proliferation was attributed to splenocytes.

The MSC inhibitory effect was not associated with increased T lymphocyte apoptosis, as the frequency of annexin+/7AAD-/CD3+ lymphocytes in co-cultures was similar to their control counterpart (splenocytes without MSC, see supplementary figure S3, available online only), suggesting MSC-induced T-cell anergy rather than apoptosis.3 ,7

Inhibition of FLS proliferation and attenuation of FLS invasiveness by MSC

Rat FLS were characterised morphologically and by immunocytochemistry (see supplementary figure S4, available online only). FLS is a critical cell population in RA, which invades and degrades cartilage, characterised by a ‘tumour-like’ phenotype. MSC supernatant suppressed the proliferation of AIA FLS in a dose-dependent manner (figure 1F) and reduced the number of invading FLS in an in-vitro assay, compared with control invasion (AIA FLS without MSC supernatant; figure 1G).

Wild-type or cytokine-stimulated MSC fail to provide clinical benefit in early and advanced arthritis

No clinical benefit could be seen in AIA rats when low syngeneic or allogeneic MSC doses (0.5–1×106) were administered intravenously, as single or multiple infusions, at early disease stages (see supplementary table S1, available online only). Similarly, repeated higher doses of syngeneic or allogeneic MSC (cumulative dose 20–30×106) administered by different routes, provided no benefit in early or established disease, unless they were infused before AIA onset (days 1, 5 or 4, 9 post-FCA; supplementary table S1, available online only; figure 2A).

Figure 2

The in-vivo effect of mesenchymal stem cells (MSC) on adjuvant-induced arthritis (AIA) and severe spontaneous erosive arthritis (SSEA). (A) The effect of high dose allogeneic MSC administered intraperitoneally at various disease phases. (B) Proliferation of concanavalin A-stimulated splenocytes in the presence of MSC primed with pro-inflammatory cytokines. (C) Proliferation of AIA fibroblast-like synoviocytes (FLS) in the presence of cytokine-primed and wild-type MSC supernatant. Proliferation is presented as the percentage of maximal response, described in Figure 1(F). **p≤0.001 versus cells without MSC or MSC supernatant, by Tukey's test. (D) Evaluation of arthritis in rats treated with cytokine-stimulated allogeneic MSC and untreated AIA rats. (E–F) Experimental design in AIA (E) and SSEA (F). Bortezomib was initiated at disease onset and allogeneic MSC were infused when advanced disease was expected. Untreated, MSC-only and bortezomib-only animals served as controls. The animals were killed 5 days after the last MSC infusion. A small number of bortezomib-only rats was killed on day 21, early after the last drug dose and just before MSC infusion, in order to assess the direct effect of bortezomib. (G) Evaluation of arthritis in MSC-only, bortezomib-only, bortezomib plus MSC-treated or untreated AIA rats. (H) Evaluation of arthritis in MSC-only, bortezomib-only, bortezomib plus MSC-treated or untreated SSEA mice. Values are the mean±SEM arthritis score from disease onset until the day they were killed. (I) Representative photos of a bortezomib plus MSC-treated rat (ii) and an untreated control (i). (J) Evaluation of arthritis in AIA rats treated with (pseudo)-autogeneic MSC, bortezomib-only, bortezomib plus (pseudo)-autologous MSC or untreated AIA rats. Each experiment was performed at least three times (n=4–5 animals/group per experiment). *p<0.05 versus arthritic animals; **p≤0.001 versus arthritic animals, by Tukey's test.

When an in-vivo inflammatory condition was simulated in vitro, IFNγ plus TNFα-primed MSC maintained their immunosuppressive capacities on splenocytes and FLS (figure 2B,C); however, they failed to ameliorate established AIA (figure 2D).

A short bortezomib course allows MSC to modulate autoimmune arthritis in an induced and a spontaneous disease model

The finding that MSC were beneficial in AIA only when injected before disease onset, ie, before the development of inflammation, prompted us to try to circumvent the susceptibility of MSC to the milieu context, by pretreating AIA and SSEA animals with bortezomib, a proteasome inhibitor, which we have shown to exhibit a strong anti-inflammatory capacity in AIA.21 Allogeneic MSC were administered when advanced disease was expected, after a short course with bortezomib initiated at disease onset (figure 2E,F).

When they were killed (AIA day 29; SSEA day 56), the bortezomib plus MSC group demonstrated a significantly decreased arthritis score over AIA or SSEA, MSC-only, bortezomib-only groups (figure 2G–I). Not unexpectedly, there was significant but transient clinical benefit in the bortezomib-only groups early after drug discontinuation (AIA rats day 21; SSEA mice day 51) but these animals eventually developed clinically significant arthritis (figure 2G,H). To simulate the clinical setting more closely, we also administered (pseudo)-autologous MSC derived from diseased AIA rats to arthritic recipients; autologous MSC strongly inhibited splenocyte proliferation in vitro (see supplementary figure S5, available online only) also being effective in mitigating arthritis in vivo, after recipients' conditioning with bortezomib (figure 2J).

Histological improvement of the joints of bortezomib plus MSC-treated animals

A significant histological improvement was observed in the joints of bortezomib plus MSC animals compared with arthritic, bortezomib-only, MSC-only groups, both in the AIA and SSEA model. In particular, bortezomib plus MSC-treated animals preserved a rather normal joint architecture with markedly decreased inflammatory cell infiltration and without or limited pannus formation, bone destruction, chondroplasia and synovial membrane thickening (figure 3). Improved histology was also seen in bortezomib-only animals killed early after drug discontinuation (Figure 3A,B).

Figure 3

Joint histology and total histological score in adjuvant-induced arthritis (AIA) and severe spontaneous erosive arthritis (SSEA). (A) AIA rat joints: (I) normal, (II) AIA, (III) mesenchymal stem cell (MSC)-only, (IV) bortezomib-day 29, (V) bortezomib plus MSC, (VI) bortezomib-day 21 joint. The encircled areas depict arthritis lesions and are labelled as: P, pannus; OC, osteochondroplasia. Magnification ×400. (B) Total histological score in the joints of AIA rats. (C) SSEA joints: (I) normal, (II) SSEA, (III) bortezomib-only, (IV) MSC-only, (V) bortezomib plus MSC joint. The encircled areas depict arthritis lesions and are labelled as: P, pannus; O, bone erosion. Magnification ×400. (D) Total histological score in the joints of SSEA mice. The histological score in (B) and (D) assessed inflammatory cell infiltration, synovial hyperplasia, pannus formation, cartilage damage and bone resorption. Bars show the mean±SEM. ##p<0.0009, versus arthritic animals, by Tukey's test.

Immunohistochemistry in the ankle joints of bortezomib plus MSC-treated rats demonstrated limited T-cell infiltrates and reduced synovial vessel density, compared with AIA and bortezomib-only joints, in which intense T-cell infiltration and vascularisation were evident in pannus and subsynovial sites. Decreased angiogenesis, not accompanied, however, by reduced inflammatory cell infiltration or improved histology, was also seen in MSC-only-treated rats (figure 4).

Figure 4

T-cell infiltration and angiogenesis in adjuvant-induced arthritis (AIA) rat joints. (A, I–V, left panel) T-cell infiltration and (B, I–V, left pannel) angiogenesis in AIA rat joints by immunohistochemistry for CD3+ and FVIII+, respectively, in normal (II), AIA (II), mesenchymal stem cell (MSC)-only (III), bortezomib-only (IV) and bortezomib plus MSC joints (V). (A–B, right panels) Quantitation of CD3+ and FVIII+ cells in five random optical fields in joints of normal, AIA, MSC-only, bortezomib-only, bortezomib-only day 21 and bortezomib plus MSC animals. Magnification ×100. Bars show the mean±SEM. #p=0.008 versus AIA rats, AIA rats, by Tukey's test.

Restoration of disease parameters following treatment with bortezomib plus MSC

The clinical and histological amelioration of bortezomib plus MSC-treated AIA and SSEA animals was associated with restoration of splenocyte proliferation and apoptosis to normal levels (figure 5A, B). The apparently ‘unorthodox’ finding of the hyporesponsive AIA splenocytes compared with normal splenocytes (figure 5A), is in agreement with the literature demonstrating the paradoxical hyporesponsiveness of ex-vivo-stimulated RA T cells, compared with normal T cells.23,,25

Figure 5

Evaluation of arthritis parameters in adjuvant-induced arthritis (AIA) and severe spontaneous erosive arthritis (SSEA) animals after treatment. (A) Proliferation and (B) apoptosis of splenocytes from AIA and SSEA animals after bortezomib plus mesenchymal stem cell (MSC) treatment compared with normal and arthritic AIA and SSEA animals. The proliferative response is presented as the percentage of maximal response. The maximal response (100%) was considered the proliferation of normal splenocytes. (C) Toll-like receptor (TLR) expression in blood by real-time PCR. (D) T-regulatory cell (Treg) frequency in blood by flow cytometry. (E) Interleukin (IL)-6, IL-10 and interferon gamma (IFNγ) levels in fibroblast-like synoviocytes (FLS) and splenocyte culture by cytometric bead array in AIA rats after bortezomib plus MSC treatment compared with normal and arthritic AIA rats. (F) Proliferation of bortezomib plus MSC-treated FLS compared with normal and AIA FLS. The proliferative response is presented as the percentage of maximal response, considered the proliferation of normal FLS (100%). (G) IL-6 and tumour necrosis factor α (TNFα) levels in FLS and peripheral blood by real-time PCR in AIA rats after bortezomib treatment and before MSC administration, compared with normal and arthritic AIA rats. Results are expressed as means±SEM (n=3–5 animals/group per experiment). #p<0.05 versus arthritic animals; ##p<0.0009 versus arthritic animals; $p=0.09 trend to significance versus arthritic animals, by Tukey's test.

Bortezomib plus MSC treatment of AIA rats restored the upregulated blood TLR2, 3 and 4 expression21 (figure 5C), induced blood Tregs (figure 5D) and normalised the levels of IL-10, IFNγ and IL-6 in the splenocytes and FLS supernatant (figure 5E). FLS proliferation and invasion capacity was also restored in bortezomib plus MSC-treated rats (figures 5F and supplementary figure S6, available online only).

The bortezomib-induced alterations in the inflammatory AIA microenvironment just before MSC infusion (day 21), were reflected by reduced TNFα levels in blood, a trend towards decreased IL-6 expression by FLS (figure 5G) and restored splenocyte proliferation compared with control animals on day 21 (see supplementary figure S7, available online only).

MSC fate after infusion

Despite the significant clinical benefit provided by MSC after bortezomib pretreatment, when they were killed (day 29 post FCA), neither the sry gene nor the Y chromosome was detectable by PCR or fluorescence in-situ hybridisation, respectively, in the spleen, FLS, lung, liver, kidney, heart, bone marrow, blood and joints of AIA female recipients infused with male MSC (figure 6).

Figure 6

Mesenchymal stem cell (MSC) tracking after infusion. When they were killed, 5 days after MSC infusion, the sry gene or the Y chromosome were tracked in several tissues of female rats infused with male MSC after bortezomib pretreatment (A–C). (A) Representative PCR results for the sry (top) and beta-actin (bottom) gene in various tissues of a bortezomib plus MSC-treated animal. Lane 1: marker; lanes 2–3: water and female DNA (negative controls); lanes 4–11: blood, bone marrow, spleen, lung, kidney, liver, heart and fibroblast-like synoviocytes (FLS) of a bortezomib plus MSC-treated rat; lane 12: male DNA (positive control). (B–C) Representative fluorescence in-situ hybridisation for the Y chromosome in rat joints from (B) a bortezomib plus MSC-treated animal and (C) a male normal rat (positive control).

Discussion

RA is a chronic, systemic and potentially devastating autoimmune disease for which dramatic developments have been accomplished over the past decades, highlighted by the introduction of biological agents in the armamentarium of disease treatment.26 However, unmet medical needs still remain because a considerable portion of patients fail to respond or develop secondary failure, drug resistance or toxicities.27 ,28

In the current study, we evaluated cell therapy with MSC in experimental arthritis. Existing literature on this topic is exclusively based on CIA models. Despite homogeneity in the animal model, the different MSC sources (bone marrow, adipose tissue, umbilical cord, MSC cell lines), doses (0.5–5×106, single or multiple infusions), routes of administration (intravenous, intraperitoneal, intra-articular), timing of delivery (before disease onset, early or advance disease), MHC match (syngeneic, allogeneic, xenogeneic) used in the literature, may have contributed to the inconsistent results reported to date;29 a positive—either preventive or therapeutic—MSC effect has been described,10,,14 and in sharp contrast, failure of MSC to improve CIA,15,,17 unless cytokine transduced.18 ,19

In this study, we sought to investigate the role of MSC in two different in pathogenesis and previously untested models of RA: the rat AIA and the mouse SSEA model. In AIA, a strong immunosuppressive effect of MSC was demonstrated in vitro affecting two critical cell populations: splenocytes and FLS. Despite their in vitro effect, bone marrow MSC were ineffective in vivo, both in early and in established arthritis, although different doses, routes of administration and MHC match were tested. Importantly, disease amelioration occurred only when MSC were administered before AIA onset, an effect that obviously cannot be clinically translated; however, it indirectly suggested that MSC may lose their functional phenotype when inflammation develops. We initially hypothesised that the adjuvants used in the induced RA models (AIA, CIA), which contain TLR ligands, may directly prime MSC towards a pro-inflammatory, instead of an anti-inflammatory, phenotype and alter their properties in vivo.30 Therefore, we then tested allogeneic MSC in the spontaneous K/BxN arthritis model in which no exogenous stimulus is required to induce disease.22 The in-vitro immunosuppressive effect of MSC was not accompanied, also in this model, by in-vivo efficacy.

The discrepancy of MSC function between in-vitro and in-vivo conditions, in different disease models, implicates that a common endogenous stimulus, missing from the in-vitro conditions, may alter the in-vivo performance of MSC and that the immune privilege ascribed to MSC may not be unconditional and autonomous but rather susceptible to the influence of the microenvironment.

Published and also our in-vitro data demonstrate that MSC in the presence of proinflammatory cytokines retain their immunomodulatory function or that Th1-associated cytokines ‘licence’ MSC to induce immunosuppression,32,,34 thus suggesting the efficacy of MSC in inflamed tissues. However, some in-vivo data have show a lack of MSC efficacy in highly inflammatory conditions. Djouad et al15 first suggested that the absence of a clinical effect with an MSC line in CIA was probably associated with the high level of inflammation, abrogating MSC function. The same group when using primary MSC showed efficacy in CIA, although within a very narrow window of MSC application before disease onset.14 This is in agreement with our findings, in which MSC were effective when infused before disease onset but failed, either as wild-type or as TNFα plus IFNγ-primed MSC, to ameliorate established arthritis. The cytokine storm, observed early in acute GvHD, could also interpret the MSC failure to reduce mortality in murine acute GvHD models.35 ,36

The in vivo failure of MSC to induce immunomodulation in highly inflammatory conditions may reflect the dynamic nature and complex network of specific microenvironments, which cannot be reliably reproduced in vitro. Strategies to alter the inflammatory milieu before MSC infusion may be of critical importance in order to allow MSC to exhibit their potential fully in vivo.

We have recently proposed that bortezomib, the first marketed proteasome inhibitor to treat multiple myeloma, can effectively suppress inflammation and reverse bone loss in AIA.21 Despite the striking preclinical effect of bortezomib and its emerging clinical potential in RA, treatment of a chronic disease like RA will require long-term drug administration and consequently chronic exposure to its side effects, resulting in several cases of patient intolerance.

We sought to manipulate the autoimmune arthritis milieu through the ‘universal’ anti-inflammatory effect of bortezomib, resulting from multiple cytokine inhibition. Indeed, bortezomib created a ‘permissive’ microenvironment for MSC to modulate AIA and SSEA effectively when full-blown disease was expected. In bortezomib plus MSC-treated animals, amelioration of arthritis was correlated with decreased T-cell joint infiltration and neovascularisation, blood Treg induction as well as normalised blood TLR expression, splenocyte and FLS proliferation and cytokine expression pattern. In support of our findings, the pharmacological suppression of TNFα and IFNγ reversed the (proinflammatory) cytokine-induced osteogenic deficiency of bone marrow MSC.37 Importantly, the therapeutic effect of MSC after bortezomib conditioning was also achieved in a pseudo-autologous AIA context, suggesting the potential clinical efficacy of bortezomib plus autologous MSC in human RA.

At the time of killing, MSC could not be found in any tissue tested, thus suggesting that MSC restore tissue integrity not by direct tissue repair but through systemic immunomodulation and probably by a ‘hit and go’ mechanism.

Overall, our data suggest that MSC are susceptible to micromilieu stimuli in autoimmune arthritis and that recipients' conditioning with a short bortezomib treatment generates a permissive milieu enabling MSC to provide significant clinical benefit. Should similar microenvironmental limitations ablate proper MSC ‘licensing’ in human RA, the bortezomib plus MSC combination may prove a novel treatment approach for unresponsive RA.

Acknowledgments

The authors would like to thank D Mathis and C Benoist (Harvard Medical School, Boston, MA, USA) and the Institute de Genetique et de Biologie Moleculaire et Cellulaire, Illkirch- Graffenstaden, France, for kindly providing us with the KRN mice.

References

View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

    • Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
    • Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
    • Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
    • Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
    • Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
    • Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
    • Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
    • Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • Funding Institutional funding was provided by G. Papanicolaou Hospital.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.