Article Text

Download PDFPDF

Extended report
Intravenous administration of expanded allogeneic adipose-derived mesenchymal stem cells in refractory rheumatoid arthritis (Cx611): results of a multicentre, dose escalation, randomised, single-blind, placebo-controlled phase Ib/IIa clinical trial
  1. Jose M Álvaro-Gracia1,
  2. Juan A Jover2,
  3. Rosario García-Vicuña1,
  4. Luis Carreño3,
  5. Alberto Alonso4,
  6. Sara Marsal5,
  7. Francisco Blanco6,
  8. Victor M Martínez-Taboada7,8,
  9. Peter Taylor9,
  10. Cristina Martín-Martín10,
  11. Olga DelaRosa11,
  12. Ignacio Tagarro11,
  13. Federico Díaz-González12,13
    1. 1Hospital Universitario de La Princesa, IIS-IP, Madrid, Spain
    2. 2Hospital Universitario Clínico San Carlos de Madrid, Madrid, Spain
    3. 3Hospital General Universitario Gregorio Marañón, Madrid, Spain
    4. 4Hospital de Cruces, Bilbao, Spain
    5. 5Hospital Vall d'Hebron, Barcelona, Spain
    6. 6INIBIC—Complejo Hospitalario Universitario A Coruña, A Coruña, Spain
    7. 7Hospital Universitario Marqués de Valdecilla, Santander, Spain
    8. 8Facultad de Medicina, Universidad de Cantabria, Santander, Spain
    9. 9Kennedy Institute of Rheumatology, University of Oxford, Oxford, UK
    10. 10IRYCIS, Hospital Universitario Ramón y Cajal, Madrid, Spain
    11. 11TiGenix, Madrid, Spain
    12. 12Department of Medicine, Universidad de La Laguna, La Laguna, Spain
    13. 13Complejo Hospitalario Universitario de Canarias, Tenerife, Spain
    1. Correspondence to Dr José M Álvaro-Gracia, Hospital Universitario de La Princesa, Unidad de Terapias Biológicas, Servicio de Reumatología, IIS-IP, Madrid 28006, Spain; jalvarogracia{at}


    Objectives To evaluate the safety and tolerability of the intravenous administration of Cx611, a preparation of allogeneic expanded adipose-derived stem cells (eASCs), in patients with refractory rheumatoid arthritis (RA), as well as to obtain preliminary clinical efficacy data in this population.

    Methods It is a multicentre, dose escalation, randomised, single-blind (double-blind for efficacy), placebo-controlled, phase Ib/IIa clinical trial. Patients with active refractory RA (failure to at least two biologicals) were randomised to receive three intravenous infusions of Cx611: 1 million/kg (cohort A), 2 million/kg (cohort B), 4 million/kg (cohort C) or placebo, on days 1, 8 and 15, and they were followed for therapy assessment for 24 weeks.

    Results Fifty-three patients were treated (20 in cohort A, 20 in cohort B, 6 in cohort C and 7 in placebo group). A total of 141 adverse events (AEs) were reported. Seventeen patients from the group A (85%), 15 from the group B (75%), 6 from the group C (100%) and 4 from the placebo group (57%) experienced at least one AE.

    Eight AEs from 6 patients were grade 3 in intensity (severe), 5 in cohort A (lacunar infarction, diarrhoea, tendon rupture, rheumatoid nodule and arthritis), 2 in cohort B (sciatica and RA) and 1 in the placebo group (asthenia). Only one of the grade 3 AEs was serious (the lacunar infarction). American College of Rheumatology 20 responses for cohorts A, B, C and placebo were 45%, 20%, 33% and 29%, respectively, at month 1, and 25%, 15%, 17% and 0%, respectively, at month 3.

    Conclusions The intravenous infusion of Cx611 was in general well tolerated, without evidence of dose-related toxicity at the dose range and time period studied. In addition, a trend for clinical efficacy was observed. These data, in our opinion, justify further investigation of this innovative therapy in patients with RA.

    Trial registration numbers EudraCT: 2010-021602-37; NCT01663116; Results.

    • Rheumatoid Arthritis
    • B cells
    • T Cells
    • Treatment
    View Full Text

    Statistics from


    Mesenchymal stem cells (MSCs) are multipotent cells that can be isolated from a variety of adult tissues of mesodermal origin, such as bone marrow, adipose tissue, placenta, umbilical cord or synovium.1–5 These cells possess immunosuppressive properties, as they inhibit the proliferation and function of major immune cell populations, including T cells, B cells and natural killer (NK) cells.6–12 In addition, they modulate the activity of dendritic cells and induce regulatory T cells, both in vivo and in vitro.11 ,13 ,14 This modulator role can be a consequence of both cell contact mechanisms and paracrine effects.15 ,16 Furthermore, MSCs show low immunogenicity, as they display low levels of MHC I (HLA-A, B and C) and the absence of MHC II (HLA-DR, DQ and DP) and co-stimulatory molecules, such as CD40 (tumour necrosis factor receptor; TNFR), CD80 (B7-1) and CD86 (B7-2).17–19 Consequently, MSCs appear to be adequate candidates for allogeneic therapy targeting autoimmune and inflammatory diseases, without the requirement of suppression of host immunity.20

    Despite the success of biological agents in rheumatoid arthritis (RA),21–23 there is still a significant number of patients who do not respond to these drugs, showing the need of new therapies. It has been described that MSCs diminish in vitro the inflammatory response of peripheral blood mononuclear cells (PBMCs) from patients with RA with active disease.24 In particular, expanded adipose-derived stem cells (eASCs) suppress responses of collagen II-reactive T cells from patients with RA, by inhibition of T-cell proliferation, production of inflammatory cytokines, release of the anti-inflammatory/suppressive cytokine IL-10 and generation of antigen-specific regulatory T cells. Moreover, eASCs inhibit the production of inflammatory factors by activated synovial cells involved in cartilage and bone damage.24–26 Consistently, eASCs have demonstrated therapeutic effect in experimental arthritis.27 However, little is known about the effect of eASCs in humans with RA.

    The primary aim of this study was to assess the safety of intravenous (iv) infusions of eASCs in refractory RA. Indeed, MSCs have generally been reported to be well tolerated,28 but theoretical safety concerns on their therapeutic use needed further analysis.29–38 Additionally, we obtained preliminary efficacy information from a study with a controlled design.


    Study design, cell therapies and other treatments

    It is a multicentre, dose escalation, randomised, single-blind (double-blind for efficacy), placebo-controlled, phase Ib/IIa study, with a follow-up period of up to 6 months, conducted between April 2011 and January 2013. To safeguard patients' safety as much as possible, safety assessments were single-blinded (blinded patients and non-blinded investigators), whereas efficacy assessments were double-blinded. The overall study scheme is shown in figure 1, and the dose escalation scheme is detailed in figure 2.

    Figure 1

    Study scheme. After a screening visit 2–4 weeks prior to study start, patients were randomised to one of the four study cohorts. Treatments were administered as three intravenous doses 1 week apart (days 1, 8 and 15). Evaluations were conducted weekly for the first month, and then monthly till the end of follow-up at month 6. From month 3 onwards rescue medication with any DMARD was allowed, so that the relevant time-window for effectiveness assessment was considered till month 3.

    Figure 2

    Dose escalation scheme. Three patients were initially included in cohort A. After the evaluation of acute toxicity 40 days later, the dose escalation to cohort B was started (another three patients) and the expansion of the cohort A began (randomisation 3:1 to cohort A or placebo). Similarly, after the evaluation of acute toxicity in the three first patients in cohort B 40 days later, the dose escalation to cohort C was started (another three patients), and the expansion of the cohort B was initiated (randomisation 3:1 to cohort B or placebo). Finally, after the evaluation of acute toxicity at cohort C in the three treated patients 40 days after, the cohort C was expanded (randomisation 3:1 to cohort B or placebo).

    The MSCs chosen to run this trial (Cx611) were eASCs fulfilling ISCT criteria for MSCs.39 Cells were obtained from adipose tissue, as this is an abundant and accessible source of adult stem cells.40 Lipoaspirates were digested with collagenase. Erythrocytes were removed by lysis, and the stromal vascular fraction was obtained through filtration. ASCs were expanded through successive expansion passages. At various stages during the process, the eASCs were tested for viability, population doublings, morphology, potency, identity, purity, sterility and genetic stability, among other quality controls. The product for clinical use was released after recovery from the cryopreserved cell banks, formulation in the corresponding vehicle and confirmation of compliance with quality specifications.

    Three cohorts (A, B and C) received active treatment, with doses of 1, 2 and 4 million cells/kg, respectively, administered intravenously at days 1, 8 and 15. Placebo was Ringer's lactate solution.

    Background non-biological disease modifying antiarthritic drugs (DMARD) were kept stable, as well as NSAIDs and glucocorticoids (≤10 mg per day of prednisone or equivalent). Rescue therapy with any DMARD, including biologics, was allowed after the third month. Follow-up visits were conducted at weeks 1, 2 and 3 (visits 1–3), and at months 1, 2, 3, 4, 5 and 6 (visits 4–9, respectively).


    This study was performed according to the applicable regulations, to Good Clinical Practice standards (CPMP/ICH/135/95) and to the amended Declaration of Helsinki (Seoul, October 2008) after approval by the corresponding ethics committees. All patients provided written informed consent. An independent Safety Monitoring Board was created to ensure and safeguard the welfare and safety of the patients.


    Eligible patients were adults with a diagnosis of RA for ≥6 months, treated with at least one non-biological agent and with previous failure (inefficacy according to investigator judgement) to at least two biologics (any of the European League Against Rheumatism (EULAR) approved biologics for RA). The wash-out periods for biologics are detailed in the online supplementary table S1. Patients' EULAR Disease Activity Score (DAS28— erythrocyte sedimentation rate (ESR)) had to be >3.2; they had to have four tender joints to palpation and four swollen joints (based on a 68/66-joint count) and had to be receiving treatment on an outpatient basis.

    Evaluation criteria

    Primary endpoint

    The primary endpoint was to determine the safety and tolerability of the intravenous administration of Cx611 through the identification of adverse events (AEs) and serious adverse events (SAEs).

    It was also intended to identify, if possible, the dose-limiting toxicity (DLT). Clinically relevant AEs (ie, grades 3–5) related to Cx611 administration were considered DLTs. Intensity of AEs was assessed following the Common Terminology Criteria for AEs, V.4.0, of the National Cancer Institute, ranging from 1 to 5.41

    Vital signs, physical examination, haematology, serum chemistry, urinalysis and coagulation parameters were collected at baseline and at each follow-up visit. Additional evaluations included arterial oxygen saturation and 12-lead ECG. EDTA blood was collected from patients at baseline and at day 30; distribution of circulating T and regulatory T (CD3+CD4+CD25+Foxp3+) cells in PBMCs was analysed by flow cytometry. Plasma levels of anti-HLA antibodies were measured with Luminex LABScreen Mixed (One Lambda, Canoga Park, California, USA). Positive samples were confirmed by LABScreen Single Antigen (One Lambda).

    Secondary endpoints

    Secondary exploratory efficacy endpoints included the proportion of American College of Rheumatology (ACR) 20, ACR50 and ACR70 patients,42 EULAR response (DAS28-ESR and DAS28-C-reactive protein (CRP))43 ,44 and the Short-Form 36 Health Survey (SF-36) questionnaire.45

    Statistical analysis

    Given the lack of safety and efficacy data of intravenous administered eASCs or other MSCs in patients with RA at the time of study design, the sample size was based on other MSCs clinical trials for other indications.

    A descriptive analysis, including anthropometric data, variables related to the medical history of patients, efficacy endpoints reported at baseline and baseline laboratory parameters, was conducted. Analysis of the primary safety endpoint was performed on the intent-to-treat (ITT) population. Number and percentages of patients who experienced AEs, SAEs, treatment-related AEs and treatment-related SAEs were described for the overall population and by treatment group, as well as those who reported grade 3–5 AEs. These values might be compared between the groups in the maintenance phase using a χ² test; otherwise, a Fisher's exact test was used. Laboratory parameters, particularly of the parameters included in the selection criteria, were described by visit.

    Analysis of the secondary efficacy endpoints was performed on the intention to treat (ITT) and per-protocol (PP) populations. Statistical analysis was conducted with SAS package V.9.2.



    Eighteen investigational sites recruited 67 patients. Fourteen patients were screening failures; therefore, 53 patients continued in the trial and received at least one dose of the study treatment (ITT population). Twenty patients were allocated to cohort A, 20 to cohort B, 6 to cohort C and 7 received placebo.

    Major protocol violations occurred in five patients (intake of prohibited medication in four patients and non-compliance with selection criteria in one patient); therefore, the PP population consisted of 48 patients. Ten patients discontinued the study prematurely, four in cohort A, one in cohort B, two in cohort C and three in placebo group.

    Details on demographic and clinical characteristics are provided in table 1. Baseline demographic and clinical characteristics were typical of refractory RA and generally comparable among treatment groups (with the exception in the number of previous DMARDs, which was higher in cohort C compared with the other groups). The mean (SD) number of previous biologics was 2.98 (1.38). The proportions of patients taking corticosteroids and NSAIDs were balanced in the different cohorts.

    Table 1

    Summary of baseline demographic and clinical characteristics of patients randomised to the four study cohorts (intent-to-treat, ITT population)


    A total of 141 AEs were reported, of which 133 were of mild or moderate intensity (94%). Seventeen patients from the group A (85%), 15 from the group B (75%), 6 from the group C (100%) and 4 from the placebo group (57%) experienced at least one AE.

    The most frequent AEs (≥5%) were fever (9; 17%), respiratory infections (8; 15%), headache (6; 11%), urinary tract infections (6; 11%), nausea (5; 9%), arthralgia (3; 6%), asthenia (3; 6%), malaise (3; 6%) and vomiting (3; 6%). The split of the most frequent AEs per treatment group is shown in table 2.

    Table 2

    Patients with adverse events per system organ class and preferred term (>1 event in any cohort; intent-to-treat population), according to MedDRA coding (Medical Dictionary for Regulatory Activities)

    By body system, the most frequently reported AEs were infections. None was serious. No opportunistic infections were reported. A listing of the infections, antimicrobials and outcomes is included in the online supplementary table S2.

    Eight AEs from six patients were grade 3 in intensity (severe), five in cohort A (lacunar infarction, diarrhoea, tendon rupture, rheumatoid nodule and arthritis), two in cohort B (sciatica and RA) and one in the placebo group (asthenia).

    Three SAEs were reported, all occurring in patients treated with Cx611: one lacunar infarction (severe), one peroneal nerve palsy (moderate intensity) and one case of fever (moderate intensity).

    The lacunar infarction was considered a DLT, according to the pre-established definition. This event encompassed three consecutive SAEs (two events of generalised muscle weakness and one event of left hemihypoesthesia and paretic ataxic gait, finally diagnosed as lacunar infarction). These episodes were transient and patient recovered with minimum sequelae. This was the only patient who discontinued the study due to AEs.

    No abnormal laboratory values were reported and no relevant vital signs abnormalities occurred, other than those related to the reported AEs. No malignancies or deaths were reported.

    Clinical efficacy

    This study was not powered to compare clinical efficacy between cohorts, so that it was evaluated only in an exploratory context. ACR20 responses for cohorts A, B, C and placebo were 45%, 20%, 33% and 29%, respectively, at month 1; 25%, 30%, 17% and 14%, respectively, at month 2; and 25%, 15%, 17% and 0%, respectively, at month 3 (figure 3). Additional efficacy measures are summarised in table 3. ACR50 responses tended to be generally greater in patients treated with Cx611, whereas ACR70 responses were very low. Table 3 also shows the results on EULAR good response, low disease activity (DAS28-ESR <3.2), DAS28-ESR and CRP. Of note, in contrast to Cx611-treated cohorts, no patients in the placebo group showed good EULAR response or low disease activity. DAS28-ESR evolution over time showed an overall decreasing trend in the cells-treated cohorts, more marked in the case of cohort C, whereas the placebo arm showed a fluctuating response, with a mean value at month 3 similar to that registered at baseline. DAS28-CRP was similar to DAS28-ESR (not shown). CRP evolution over time showed a tendency to decrease from baseline in cohorts A and C but not in cohort B or placebo. Efficacy results did not differ in the PP and ITT populations.

    Table 3

    Efficacy results at time points until month 3 (intent-to-treat population)

    Figure 3

    Percent patients achieving ACR20 responses till month 3 (intent-to-treat population). Patients were treated with Cx611 or placebo, administered intravenously at days 1, 8 and 15. Cohort A: 1 million cells/kg (n=20), cohort B: 2 million cells/kg (n=20), cohort C: 4 million cells/kg (n=6) and placebo (n=7).

    After 3 months, we did not observe persistent clinical benefit in patients treated with Cx611, as treatment effect waned or fluctuated (data not shown). The physical and mental subscales (SF-36) improved in the different groups, without statistical differences among them (data not shown).

    Immune readouts

    Distribution of T-cell populations, including Tregs in circulation, was not significantly modified among cohorts (not shown). In terms of immunogenicity, nine patients (19%) generated eASC-specific anti-HLA-I antibodies without apparent clinical consequences. Indeed, 43% of treated patients presented baseline anti-HLA-I antibodies (presensitised), presumably related to previous pregnancies or transfusions. Presensitised patients showed higher frequency of eASC donor-specific antibodies (30% vs 11%). No dose-related impact was observed (see online supplementary table S3). Anti-HLA-II antibodies were not found.


    This is, to our knowledge, the first randomised, placebo-controlled study with MSC in RA. The most important findings of our work are (1) in patients with RA, the intravenous treatment with Cx611 was not associated with significant toxicity and (2) signs of possible efficacy of this therapy in patients with RA could be envisioned.

    An overall favourable safety profile of Cx611 is suggested by the fact that most of the reported AEs were of mild-to-moderate intensity, consistent with the patient profile and considered unrelated to the study treatments by investigators. No life-threatening events (grade 4) or deaths occurred. There was no apparent relationship between dose and tolerability. In fact, the rate of treatment-related AEs was lower in cohort C (higher dose) than in both cohorts A and B, although this is probably due to the small size of the cohort C. Transient fever was the most frequent treatment-related AE observed in patients treated with Cx611. The mechanisms for fever are not clear but could be considered as some form of infusion reaction.28

    There was one DLT, a lacunar infarction (left hemihypoesthesia and paretic ataxic gait), which occurred in a patient of cohort A 8 days after the second treatment administration. It was deemed as likely related because there were no other apparent causes, even though the pathophysiology of this event is unclear. Therefore, a conservative position was adopted in the causality assessment. This SAE encompassed three consecutive SAEs (see Results), in which muscle weakness, the most consistent manifestation, was probably a form of infusion reaction, because, after discontinuation from study, this patient suffered a similar episode subsequent to the administration of intravenous tocilizumab.

    Infections were the most frequently reported AEs by body system, none of them being severe or serious. It is difficult to know their relationship to Cx611, since patients with RA are at a higher risk for serious infections than the general population, and prednisone and biological agents have been shown to increase this risk.46–48

    No venous thrombotic events or signs of pulmonary thromboembolism were detected. This is reassuring given that thrombotic events have been described in animal models in which very high doses of intravenous MSCs were administered.36 In addition, no acute or delayed hypersensitivity reactions or haematological AEs, with the exception of anaemia (three cases), were observed in our patients. No malignancies were found although the size and duration of the trial limit the relevance of this finding. Additionally, clinical experience does not indicate that tumorigenicity associated with MSC-based therapies represents a significant risk.35

    We detected sensitisation against allogeneic ASCs in some patients, without apparent clinical consequences. This has been previously discussed in allogeneic stem cell treatment and in patients with previous pregnancies and transfusions.38 ,49 The duration and impact of these allo-antibodies need to be further analysed. As this study was not designed nor powered for efficacy evaluation, clinical efficacy outcomes should be interpreted cautiously and within an exploratory context. The very refractory profile of the included population increases the difficulty for detecting efficacy. However, a tendency to better response was observed in patients treated with Cx611 than in placebo-treated patients. No conclusion can be made on dose responses due to the limited number of patients. The clinical benefit achieved in patients with RA treated with intravenous Cx611 tends either to wane or fluctuate after 3 months of cell administration. This suggests that RA therapy with these cells would require repeated administrations.

    Apart from anecdotal experiences from isolated cases,50–52 only one study has been published so far on MSC for RA treatment.53 This corresponds to an ongoing cohort that included 136 patients with heterogeneous RA with inadequate response to various medications, who received DMARDs plus intravenously umbilical cord MSCs (UC-MSCs). A non-randomised control group comprised 36 patients that started treatment with DMARDs plus medium 2 years later was used as comparator. UC-MSC treatment seemed to induce a significant clinical benefit that was maintained for 3–6 months and it was well tolerated. Therefore, our findings of a lack of short-term serious toxicity associated with the therapeutic administration of Cx611 to patients with RA, and preliminary evidence of clinical efficacy, are in line with the results of this study.

    One area of debate can be the choice of allogeneic cells instead of autologous. The main reason for it is that the allogeneic product, being stored in a cell bank, can be readily made available to the patients when required. Using autologous cells would complicate substantially the therapy, since liposuction (sometimes difficult) is required, and cell expansion would need a much longer time to obtain enough cells. The limitations of this study include the single-blind design for safety (considered adequate in the best interest of patients) and the limited number of patients in some cohorts. In contrast, the randomised placebo-controlled design and the double-blind assessment of efficacy performed for the first time to our knowledge with intravenous eASC are its main strengths. Our results suggest that the intravenous administration of Cx611 is in general well tolerated and without dose-related toxicity at the dose range and time period studied. In addition, signs for a potential therapeutic effect of these cells were observed in a highly challenging population of patients with refractory RA. This opens the possibility of further research to investigate the duration of the therapeutic effects, the optimal dosing strategy, and the most suitable patient profile for this treatment.


    The authors thank Marie Paule Richard, Mary Carmen Diez, Wilfried Dalemans and Eleuterio Lombardo for the critical reading of the manuscript.


    View Abstract


    • Handling editor Tore K Kvien

    • Collaborators Additional list of investigators Javier Ballina, Hospital Central Asturias, Oviedo. Ricardo Blanco Alonso, Hospital Marqués de Valdecilla, Santander. Sagrario Bustabad, Complejo Hospitalario Universitario de Canarias, Tenerife. Eugenio Chamizo, Hospital General de Mérida. Antonio Fernández-Nebro, Hospital Carlos Haya de Málaga. José Luis Marenco, Hospital Nuestra Sra. Valme, Sevilla. Emilio Martín-Mola, Hospital La Paz, Madrid. Federico Navarro, Hospital Virgen Macarena, Sevilla. Arturo Rodríguez, Hospital Santa Creu i Sant Pau, Barcelona. José Andrés Román-Ivorra, Hospital La Fe, Valencia. Raimon Sanmartí, Hospital Clinic i Provincial, Barcelona. Jesús Tornero, Hospital de Guadalajara.

    • Contributors JMA-G, VMM-T, PT, FD-G and ODR were involved in study design, data analysis and in drafting and reviewing the manuscript. IT was involved in data analysis, and in drafting and reviewing the manuscript. JAJ, RG-V, LC, AA, SM and FB were involved in patient recruitment, treatment and follow-up, and in reviewing the manuscript. CMM was involved in anti-HLA antibodies detection and manuscript review.

    • Funding This work was funded by TiGenix and the European Union's Seventh Programme for research, technological development and demonstration under grant agreement no 279174 to TiGenix.

    • Competing interests JMA-G, VMM-T, PT and FD-G received advisory fees from Tigenix. ODR and IT are employees of Tigenix.

    • Patient consent Obtained.

    • Ethics approval Ethics committee at Hospital de la Princesa, Madrid, and others.

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

    • Data sharing statement Data are available upon request to the corresponding author.

    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.