Study design and population
Between April 1998 and January 2017, a total of 248 consecutive patients with LRC Class III β-thalassemia major underwent allogeneic HSCT in the Research Institute for Oncology, Hematology and Cell Therapy (RIOHCT), affiliated by Tehran University of Medical Sciences, and were enrolled in this single-center prospective randomized controlled trial. Out of these patients, 24 were excluded from the study due to comorbid medical conditions. Finally, 224 eligible patients were randomly assigned to receive co-transplantation of bone marrow-derived MSCs with HSCs (n = 83) or to receive HSCs alone (n = 131).
In order to assess liver fibrotic changes after HSCT, in the second phase of the study, 47 patients participated in the MSC plus HSC group and 30 patients in the HSC only group (Fig. 1). In this phase, participants who met any of the following criteria were excluded from the study: (1) patients who declined to participate in the study; (2) any severe underlying or pre-existing medical condition other than thalassemia major and its obvious complications; (3) those who lost a regular follow up visit; (4) patients who did not gain complete hematological recovery after transplantation; (5) patients who had received drugs with hepatic metabolism 6 months before the study; and (6) patients with class I and II thalassemia major.
Hemoglobin electrophoresis and mutation analyses were performed on all patients to confirm the diagnosis of β-thalassemia major. Patients were classified using the LRC before HSCT based on evidence of portal fibrosis in liver biopsy, the presence of hepatomegaly (liver size greater than 2 cm palpable below the right costal margin), and inadequate chelation therapy. Patients with none of the above risk factors were classified as Class I, patients with 1 or 2 of these risk factors were classified as Class II, while patients with all three risk factors were classified as Class III.33,34
The Ethics Committee of Tehran University of Medical Sciences approved the study (Ethical code: IR.TUMS.MEDICINE.REC.1398.093, Research code: 97-02-36-38570), and the study was conducted under the Declaration of Helsinki. Written informed consent was obtained from all participants and their parents.
Evaluation of hepatic fibrosis
All participants underwent laboratory tests, including serum ferritin level and liver function tests, hepatic T2* magnetic resonance imaging (MRI), liver biopsy, and FibroScan (liver elastography) before and after HSCT.
Hepatic T2* MRI
Hepatic T2* MRI was performed in the Department of Radiology using a 1.5 Tesla Magnetom Siemens Symphony scanner (Siemens Medical Solutions, Erlangen, Germany). All the patients were placed in a supine position and entered the magnet cradle, using the head-first configuration. A 10 mm slide thickness through the liver core scanned at 12 different echo times (TE 1.3 to 23 ms) was used for the calculation of hepatic T2 values. The repetition time (TR) was 200 ms, the base resolution matrix was 128 pixels, the field of view was 39.7 cm × 19.7 cm, the flip angle used was 20°, and the sampling bandwidth was 125 kHz. Each image was taken within 11–13 s of breath-holding using a gradient-echo sequence. T2* values were calculated by the software (CMR tools, Imperial College). Results of hepatic T2* were categorized as normal (T2* > 30 ms), mild (6.2 < T2* < 30 ms), moderate (3.1 < T2* < 6.2 ms), severe (2.1 < T2* < 3.1 ms), and very severe (T2* < 2.1 ms).35
FibroScan (Transient elastography)
All patients were assessed by transient elastography (FibroScan®, EchoSens, Paris, France) on the same day as the liver biopsy examination. This method was performed by a skilled gastroenterologist who was blind to clinical and histological data. Measurements were taken in the right lobe of the liver, through intercostal spaces while the patient was lying in dorsal decubitus, with the right arm at maximal abduction. Using ultrasound, a portion of the liver at least 6 cm thickness and free of large vessels was recognized for examination. Only procedures were considered reliable that at least ten validated measurements were performed on each patient, the success rate was at least 60%, and an interquartile range (IQR) of the median stiffness value was lower than 30%.36 The results were expressed in kilopascals (kPa). A cut-off value less than 7.0 kPa was not considered significant for liver fibrosis.37
Histological data
Liver biopsy was performed by an expert radiologist with an 18-gauge (18G) needle (Bard Peripheral Vascular, Biopsy System Max Core®, USA) under ultrasound guidance. Liver biopsy specimens were fixed in formalin and paraffin embedded. Sections of liver tissue were stained with hematoxylin-eosin and Masson trichrome and read by one experienced liver pathologist blinded to results of liver transient elastography and T2* MRI. The METAVIR scoring system was used for assessment of liver fibrosis and necro-inflammatory activity.38 Fibrosis was staged on a 0 to 4 scale as follows: F0: no fibrosis; F1: portal fibrosis without septa; F2: periportal fibrosis and few septa; F3: bridging fibrosis and numerous septa without cirrhosis; F4: cirrhosis. Activity was graded as follows: A0: none; A1: mild; A2: moderate; A3: severe.
Isolation of Mesenchymal Stem Cells
All patients received MSC and HSC from the same donor. For the generation of MSCs, about 40–90 mL of bone marrow aspirate was obtained from HLA-matched siblings or other HLA-matched relative donors before bone marrow transplantation.
Briefly, bone marrow mononuclear cells (MNCs) were then separated using Ficoll-Paque (Stem Cell Technologies, Vancouver, Canada) (400 g, 25 min, 20°C) and washed twice by phosphate-buffered saline (PBS). Mononuclear cells were seeded at a density of 160 × 103 cells per cm2 and cultured in a complete culture medium consisting of the following: Dulbecco’s modified Eagle’s medium–low glucose (DMEM; Gibco, Life Technologies, USA), supplemented with 10% fetal bovine serum fetal bovine serum (FBS; Gibco, Life Technologies, USA), was irradiated with a dose of 25 kGy., and 1% (vol/vol) penicillin/streptomycin (Gibco, Life Technologies, USA) with final concentrations of 100 units/mL and 100 µg/mL, respectively. Cultures were maintained in a humidified atmosphere containing 5% CO2 at 37°. Cells were allowed to adhere for 48 h; henceforth, non-adherent cells were discarded, and culture medium was replaced twice weekly. After reaching ≥ 70% confluence, MSCs were harvested using trypsin/EDTA (Gibco, Life Technologies, USA) and re-plated at 4000 cells per cm2. After final harvesting (usually at passage 2–3), MSCs were cryopreserved with 10% (v/v) dimethyl sulfoxide (DMSO) with USP grade components (Miltenyi Biotec, Germay) and 5% human albumin (Biotest AG, Germany) until the time of bone marrow transplantation. The collected cells were identified to meet the minimal criteria of MSCs according to ISCT (International Society of Cell Therapy) standard. The quality and viability of these cells was reconfirmed after preparation and before each infusion for patient. MSCs are characterized by the following criteria: fibroblast-like morphology, the ability to self-renew, expression of cell surface markers including CD19, CD73, CD105, CD90, and HLA-DR, and no expression of CD45, CD34, CD14, CD11b, CD79, or CD19. Additionally, MSCs must be capable of in vitro differentiation into osteoblasts, adipocytes, and chondrocytes.
The median MSC dose was 2.0 × 106 cells/kg of recipient's body weight. The total volume of the MSCs infusion was 50 ml and infused 3 hours prior to HSCT.
Stem Cell Source, Donor selection, Conditioning regimen, and GVHD prophylaxis
The sources of progenitors were bone marrow stem cells (BMSCs) and peripheral blood stem cells (PBSCs). In bone marrow cases, donor stem cells were harvested from the posterior iliac crest under general anesthesia. Mobilization of PBSCs was performed by the administration of 5 µg/kg/day granulocyte-colony stimulating factor (G-CSF) 4 consecutive days and 5 µg/kg twice daily for the fifth day and then collected using a continuous-flow leukapheresis.
We administered a myeloablative conditioning regimen for all patients, consisted of intravenous busulfan at 3.5 mg/kg/day from day − 9 for 4 consecutive days and cyclophosphamide at 40 mg/kg/day from day − 5 for 4 consecutive days. For HSCT recipients other than HLA identical sibling (HLA-matched unrelated and cord blood recipients), in addition to the mentioned regimen, rabbit ATG (Thymoglobulin; Sanofi-Aventis, Quebec, Canada) was given intravenously at a dose of 2.5 mg/kg/day from days − 5 to -2.
Regarding GvHD prophylaxis, cyclosporine (1.5 mg/kg/day, IV, on day − 2, and then 3 mg/kg/day on day + 7 in PBSC and day + 11 in BMSC) in combination with a short course of methotrexate (10 mg/m2 on days + 1 and 6 mg/m2 on days + 3,+6, and + 11) was administered for patients. Cyclosporine was converted to oral formulation when patients were able to tolerate oral intake. It continued orally for at least six months after HSCT and discontinued in the absence of GvHD.
Definitions and Outcome Measurements
Neutrophil engraftment was defined as the first date for 3 consecutive days with an absolute neutrophil count of more than 0.5 × 109/L without G-CSF subcutaneous injection and platelet engraftment was defined as the first date for 7 consecutive days with a platelet count of greater than 20,000 unsupported by platelet transfusions. Full chimerism was assumed when > 95% of the recipient’s blood cells were from the donor’s origin. When this index fell below 5%, graft rejection was considered. Any rate between the two above was defined as mixed chimerism. Standard criteria were considered to define and diagnose acute and chronic GVHD.39,40
The primary endpoint was evaluation of hepatic fibrotic changes. The secondary endpoints were overall survival (OS), thalassemia-free survival (TFS), non-relapse mortality (NRM), and graft-versus-host disease (GVHD).
Statistical analysis
Homogeneity between the two groups of patients was evaluated using the chi-square test for qualitative variables and median test and Student’s T-test for continuous variables. A two-sided P-value of 0.05 or lower was considered to be statistically significant. Kaplan–Meier curves were derived to determine OS and TFS, and were compared using the log-rank test. Median follow-up time was established with the reverse Kaplan-Meier method. The assumption of proportionality of hazards was checked using Schoenfeld residuals. Analyses were conducted using STATA version 11.2. The linear mixed model was used for the repeated measurement data from the follow-up visits.
Repeated-measure, mixed-effect linear regression models41 were used to examine change over time in serum ferritin, AST (aspartate aminotransferase), ALT (alanine aminotransferase), FibroScan score, hepatic T2* MRI, liver stage, liver grade, and dry weight iron concentration. Evaluation of the intervention effect as a change over time in each of variables among all patients (model 1) and determination of time-by-mesenchyme interaction in each variable (model 2) was conducted. All models included a variable indexing time (pre-HSCT and post-HSCT), dichotomous indicators of intervention (with mesenchymal vs. without mesenchymal), and baseline fixed effects for participant characteristics (i.e., patient’s age, sex Matching, ABO matching, source of HSCT, acute GVHD, and chronic GVHD). In addition, we used post-hoc estimation to disaggregate two-way interaction and generate model-based tools and standard errors, and we used t-tests to evaluate the simple main effects.