Cytotherapy, 2015; 17: 956e968

Clinical efficacy of autologous stem cell transplantation for thetreatment of patients with type 2 diabetes mellitus: a meta-analysis

ZHENG-XU WANG, JUN-XIA CAO, DUO LI, XIAO-YAN ZHANG, JIN-LONG LIU,JUN-LI LI, MIN WANG, YISHAN LIU, BEI-LEI XU & HAI-BO WANG

Biotherapy Center, the General Hospital of Beijing Military Command, Beijing, China

AbstractBackground aims. In this study, we investigate whether bone marrow mononuclear cells (BM-MNC) or peripheral bloodmononuclear cells (PB-MNC) have therapeutic efficacy in type 2 diabetes (T2D). Methods. Search terms included stem cell,bone marrow cell, peripheral blood cell, umbilical cord blood and T2D in MEDLINE, the Cochrane Controlled TrialsRegister, EMBASE, the Wanfang Database, the China Science and Technology Periodical Database and China Journal Net.Results. Fifteen trials met our inclusion criteria (n ¼ 497). One group included 266 cases with BM-MNC therapy and theother group contained 231 cases with PB-MNC treatment. Glycosylated hemoglobin was decreased after BM-MNC or PB-MNC therapy compared with that before (12 months: P < 0.001; 6 months: P < 0.001; 3 months: P < 0.05). Fasting plasmaglucose was reduced in BM-MNC therapy group compared with control after 12-month follow-up (P < 0.001) and afterBM-MNC therapy compared with that before (9 months: P < 0.001) but was not obvious in other stages. Meanwhile, theanalysis showed that C-peptide level increased after BM-MNC and PB-MNC therapy compared with the control therapy (12months: P < 0.001) and with that before therapy (6 months: P < 0.05). Insulin requirement reduction was also observed inpatients receiving BM-MNC therapy (3, 6, 9 and 12 months: P < 0.05). Conclusions. To a certain extent, BM-MNC or PB-MNC therapy for T2D demonstrated superiority of glycemic control, increased insulin biosynthesis and elevated insulinsecretion from existing b-cells and might prevent islet cell loss.

Key Words: bone marrow mononuclear cell, meta-analysis, peripheral blood mononuclear cell, stem cell, type 2 diabetes mellitus

Introduction

The global prevalence of diabetes in 2012 was esti-mated to be more than 10% among adults. Accordingto the report of the World Health Organization, thetotal number of patients with diabetes is projected toreach 366million in 2030. Of the diabetic population,95% are of type 2, characterized by two defects,namely progressive and inexorable b-cell dysfunction,which superimposed on insulin secretion and sensi-tivity [1e3]. Diabetes can result in multi-systemchronic complications, particularly micro- andmacro-vascular complications, with high morbidityand mortality rates. Chronic hyperglycemia can alsodamage the eyes, kidneys, nerves, heart and bloodvessels. Once a person is diagnosed with diabetesmellitus, they will generally need to take drugs or in-sulin all their life, which can cause a great deal ofdisruption to their work and life in general [3].

Recently, research has focused on stem cells togenerate functional b cells [4]. In addition to primarypancreatic b cells, studies on regeneration of

Correspondence: Zheng-Xu Wang, MD, PhD, Biotherapy Center, the General HDistrict, Beijing, 100700, China. E-mail: zhxwang18@hotmail.com, zhxuwang@v

(Received 1 September 2014; accepted 23 February 2015)

ISSN 1465-3249 Copyright � 2015, International Society for Cellular Therapy. Phttp://dx.doi.org/10.1016/j.jcyt.2015.02.014

functional insulin producing cells suggested variousalternative cell sources including embryonic stemcells, induced pluripotent stem cells and adult stemcells, for example, bone marrow cells (BMCs) orbone marrow mononuclear cells (BM-MNC), whichmainly contain mesenchymal stromal cells (MSCs)and hemopoietic stem cells, peripheral bloodmononuclear cells (PB-MNC), umbilical cord bloodstem cells (UCB), pancreatic stem cells and hepaticstem cells. In addition, the conversion of the gallbladder, skin fibroblasts, blastocyst-derived hypo-blast stem cellelike cells and induced pluripotentstem cells into insulin-secreting cells has been tested[5e8]. Among them, MSCs were demonstrated toinhibit T-cellemediated immune responses againstnewly formed b cells, which, in turn, are able tosurvive in this altered immunological milieu [9]. As anew therapeutic agent, MSCs in the treatment ofdiabetic cardiomyopathy, diabetic nephropathy, dia-betic polyneuropathy, diabetic retinopathy and dia-betic wounds were applied [10,11]. Prochymal, a

ospital of Beijing Military Command, Nan Men Cang, No. 5 Dongchengip.163.com

ublished by Elsevier Inc. All rights reserved.

Autologous stem cell therapy type 2 diabetes mellitus 957

human MSCebased stem cell therapy has beendesignated by the US Food and Drug Administrationfor the treatment of acute graft-versus-host diseaseand for ischemic diseases, neurologic disorders anddiabetes, among others [12].

Stem cell therapy offers a new paradigm in themanagement of T2D after its success in an elegantstudy by Voltarelli et al. [13] in patients with type 1diabetes mellitus (T1DM). Phase I/II clinical trials ofintra-arterial pancreatic infusion of total autologousBM and/or BM-derived stem cells are currentlyunder way, applying for the treatment of T2D atFuzhou General Hospital in China (in combinationwith hyperbaric oxygen therapy; NCT00767260), atPostgraduate Institute of Medical Education andResearch in India (NCT00644241), at Shan-dongUniversity in China (NCT00465478), at the Uni-versity of Illinois at Chicago (NCT01415726) and atthe University of Miami (NCT01786707). A totalnumber of 88 registered clinical trials on T2D inphase I/II can be found on the website forclinicaltrials.gov (http://www.clinicals.gov) with thekey word of “stem cell” and “type 2 diabetes” untilAugust 31, 2014. There have been some trialsreporting that stem cell therapy can control patients’hyperglycemia and improve the function of pancre-atic islets [14].We herein performed a systematicreview and meta-analysis of clinical trials to assessthe efficacy and tolerability of stem cells in thetreatment of patients with T2D. The aim of thisReview was to evaluate the impact of stemcellebased therapy on the clinical response andclinical observation results as well as to assess theefficacy of such a treatment by glycosylated hemo-globin (HbA1c), C-peptide, fasting plasma glucose(FPG) and insulin requirement.

Methods

Search strategy and selection criteria

Trials were identified by electronic search in thePubMed database (1976 onward), Embase (1966onward), the Cochrane Central Registry ofControlled Trials (no date restriction), the WanfangDatabase (no date restriction), the China Scienceand Technology Periodical Database (no date re-striction), China Journal Net (no date restriction),reference lists of published trials and relevant reviewarticles. The search strategy included the medicalsubject headings of “diabetes,” “stem cells,” “type 2diabetes,” “mesenchymal stem cell,” “cell therapy,”“bone marrow mononuclear cell,” “peripheral bloodmononuclear cell” and “beta cell” for the full textsearch. The type of the study design (ie, whether thetrial reported the mode of randomization, allocation

concealment, description of withdrawals per arm andblinding) for all of the trials included in the study andthe respective study duration was demonstrated bythe corresponding end point. No language limitswere applied. The initial search was performed inNovember of 2012, with updates in July 2014.Furthermore, we consulted with experts in this fieldand performed manual searches in reference lists.We also searched http://www.ClinicalTrials.govwebsite for the information of prospective andongoing trials. We excluded abstracts that were neversubsequently published as full papers and studies onanimals.

Data extraction and quality assessment

Data extraction was independently conducted by twoauthors with the use of a standardized approach.Disagreement was adjudicated by a third author afterreferring back to the original publications. Wecollected the trial data including authors’ names,journal, year of publication, sample size per arm,regimen used, median or mean age of patients, sex,history of T2D, HbA1c, C-peptide, FPG, insulinrequirement of patients and information pertainingto study design in our meta-analysis.

Definition of outcome measures

HbA1c is a measure of the degree to which hemo-globin is glycosylated in erythrocytes and is expressedas a percentage of total hemoglobin concentration. Itreflects the exposure of erythrocytes to glucose in anirreversible and time- and concentration-dependentmanner and is also determined by the FPG assess-ment. The secondary object was the C-peptide leveland insulin requirement that denote the function ofpancreatic islets.

Statistical analysis

The analysis was carried out by means of pair-wisecomparison of the stem cell containing arms of theidentified trials with the respective nonestem cellarms and also the comparison before and after thestem cell therapy. Treatment effects are reflected byHbA1c, FPG, C-peptide and insulin requirement.The data of HbA1c, FPG, C-peptide and insulinrequirement in each arm were extracted from eachtrial and combined by use of a method by Manteland Haenszel (Review Manager Version 5.0, NordicCochran Centre). To evaluate whether the results ofthe studies were homogeneous, we used Cochran’s Qtest; it is a c2 test with degrees of freedom equal tothe number of studies minus 1 and tests the nullhypothesis that the difference between the study

958 Z.-X. Wang et al.

estimates of mean difference (MD) is due to chance.We also calculated the quantity I2 that describes thepercentage of variation across studies that is due toheterogeneity rather than chance. I2 values of 25%,50% and 75% were used as evidence of low, mod-erate and high heterogeneity, respectively. The MDwas calculated with a fixed-effect model when nostatistically significant heterogeneity existed; other-wise, a random-effects model was used. Values ofP < 0.05 were considered to be statistically signifi-cant. All reported P values resulted from two-sidedversion tests of the respective tests.

Risk of bias across studies

The sample size of the selected studies is not verylarge. The follow-up time of the included studies wasnot sufficiently long, and the patient informationsuch as following doctor’s advice or regarding dietcontrol and medication was not insufficient.

Figure 1. Flow diagram shows record identification, screening andstudy inclusion process.

Results

Selection of the trials

The electronic search yielded 162 references. Aftertitle and abstract review, 137 publications wereexcluded for different reasons (25 for being reviewarticles, 18 for using animal models, 28 for beingcase reports, 37 for being in vitro experiments and 29for being nursing studies). The full texts of 25 articleswere selected as potentially relevant and retrieved formore detailed assessment. A further 10 studies wereexcluded because there were no detailed patientclinical data or therapy response reports and the celltype for treatment. The selection procedure of theclinical trials is shown in Figure 1. As a result, 15articles reporting clinical trials of stem cell-basedtherapy for T2D were selected for meta-analysis.

Characteristics of stem cellebased therapy

After the selection process, 15 eligible clinical trialswith a total of 497 patients were included in thepresent analysis [15e29]. All of the trials were fullypublished. The clinical data of the trials are listed inTable I.

Most of the patients in these studies had a goodperformance status; median patient age was 50.4years. In all of the 15 trials, stem cell therapy wasevaluated in patients with T2D. BM-MNC, pe-ripheral blood stem cell or UCB stem cell therapieswere all included in this analysis. We put the com-parison into two groups, respectively, with PB-MNC including UCB in six studies [21,23,26e29]and with BM-MNC in the selected nine studies

[15e20,22,24,25] according to the cell type of theregimens used per arm containing 266 and231 cases, respectively. Considering the cells usedfor the therapy, in two trials hyperbaric oxygen(HBO) combined with BM mononuclear stem celltherapy was used for T2D treatment [15,16],whereas the other 12 trials used only stem cellsincluding Wharton’s jelly MSC [17e28] and oneused the cord bloodederived multipotent stem cellswith educator therapy [29]. Four studies included acontrol group; two of them were divided into twogroups on the basis of patients’ willingness when theclinical trials were performed, and the control groupreceived the therapy with insulin intensificationand conventional therapy [19,21]. The other twowere randomly assigned to intervention and controlarms with the therapy of conventional therapy andsham procedure [16,20]. The number of stemcells transfused into patients in these studies was>1.0 � 107/course.

The patient information from two groups (stemcell therapy alone and control group and before andafter stem cell treatment) of the trials, such as sex,age, history of T2D and stem cell doses, are listed inTable I. However, other clinical information fromthe trials such as performance status, weight, bodymass index, waist, body fat, triglycerides, high-density lipoprotein cholesterol and low-density lipo-protein cholesterol were not collected because ofinsufficient data on some reports of the treatment.

Table I. Baseline clinical and laboratory parameters of patients.

Authors and year Age (years, mean)No. of patients

(male)

History ofNIDDM(years) Regimens (per arm)

Regimens(cell number) dose PMID

Ramakrishnan 2009 (India) [17] 57.5 � 5.9 10 (8) 14.6 � 7.5 BM-MNC 3.5 � 1.4 � 108 19686048Ricordi 2008 (USA) [15] 55.8 � 2.1 25 (17) 13.2 � 1.6 HBO and BM-MNC Unknown 19364067Wang 2011 (China, Wuhan) [18] 46.6 � 6.8 31 (18) 5.7 � 4.4 HBO and BM-MNC 3.76 � 108/kg 22340214Hu 2012 (China, Shandong) [19] 50.4 � 4.9

50.2 � 8.256 (38)54 (36)

8.6 � 6.57.3 � 6.3

BM-MNC Insulinintensification

2.8 � 1.9 � 109 22814142

Khandelwal 2014 (India) [20] 50.4 � 2.354.3 � 2.3

11 (9)10 (7)

13.4 � 1.518.2 � 1.9

BM-MNCSham procedure

3.2 � 1.4 � 108 23561959

Wu 2014 (China, Fujian) [16] 53.3 � 5.859.2 � 5.4

20 (13)20 (14)

�2 to �15 BM-MSCConventional therapy

4.0 � 1.4 � 109 24290656

Ding 2012 (China, Beijing) [21] 47.48 � 10.46

50.39 � 9.37

96 (71)

48 (32)

9.02 � 6.28

9.42 � 5.65

PB-MNC andconventional therapy

Conventional therapy

8 � 108 http://dx.doi.org/103877/cma.j.issn.1674-0785.2012.17.039

Wang 2010 (China, Sichuan) [22] 61.60 � 12.58 5 (4) 13.60 � 5.73 BM-MNC 1.27 � 0.18 � 108 http://dx.doi.org/10.3969/j.issn.1673-8225.2010.23.039

Xu 2011 (China, Guangdong) [23] 45 � 18 18 (10) 1.91 � 0.96 UCB mononuclear 1.1 � 0.3 � 108 2095-0616(2011)22-22-04Yao 2012 (China, Chongqing) [24] 45.67 � 14.92 6 (3) 6.28 � 6.96 BM-MNC Unknown 1000-5404 (2012)01-0074-04Chen 2008 (China, Shandong) [25] 42.4 � 10.5 18 (12) 7.1 � 5.4 BM-MNC 6.1 � 3.4 � 108

Tong 2013 (China, Chongqing) [26] 40.7 � 4.5 3 (3) 6.6 � 4.6 UCB mononuclear 5.29 � 108 23861460Zhao 2013 (China, Shandong) [29] 50.8 � 9.4 36 (21) 9 � 5.6 PB-MNC Unknown 23837842Liu 2014 (China, Tianjin) [27] 52.9 � 10.5 22 (15) 8.7 � 4.3 UC-MSC 2 � 106/kg 24759263Wang 2014 (China, Beijing) [28] 41 � 10 8 (4) 3.7 � 1.1 UCB stem cell Unknown

Autologous

stemcelltherapy

type2diabetes

mellitus

959

A

B

Figure 2. Comparison of HbA1c between the stem cell therapy and control groups (A); comparison of HbA1c before stem cell therapy andafter stem cell therapy (B). The random-effects meta-analysis model (Mantel-Haenszel method) was used. Each trial is represented by asquare, the center of which gives the MD for that trial. The size of the square is proportional to the information in that trial. The ends of thehorizontal bars denote a 95% CI. The black diamond gives the overall MD for the combined results of all trials.

960 Z.-X. Wang et al.

Autologous stem cell therapy type 2 diabetes mellitus 961

HbA1c

HbA1c is a stable marker of glycemic control, andinformation on HbA1c was available in 12 trials[15e21,24,26e29] containing 243 patients withBM-MNC therapy and 213 patients with PB-MNCtherapy (92 patients as the control group withoutstem cell therapy). Regarding the efficacy of the BM-MNC therapy compared with control therapy, theestimated pooled MD for all of the three trials[16,19,20] showed significant reduction in HbA1cwith the inverse variance model (BM-MNC group:MD, �0.87; 95% confidence interval [CI], �1.55to �0.2; P ¼ 0.01) in T2D after 12-month follow-up(Figure 2A). On the other hand, we compared thelevel of HbA1c in the different stages before and afterBM-MNC or PB-MNC treatment to show the effectof the therapy. The estimated pooled MD for 12trials shows a highly significant decrease in HbA1cafter 12-, 6- and 3-month follow-up (BM-MNCgroup: MD, �1.93; 95% CI, �2.89 to �0.98; P <0.001; MD, �1.56; 95% CI, �2.22 to �0.91; P <0.001; MD, �0.71; 95% CI, �1.26 to �0.17; P ¼0.01; PB-MNC group: MD, �1.43; 95% CI, �1.76to �1.10; P < 0.001; MD, �1.58; 95% CI, �1.99to �1.17; P < 0.001; MD, �2.08; 95% CI, �3.01to �1.16; P < 0.001), respectively (Figure 2B). Inall, the comparison of HbA1c in the two groups is thesame, and the BM-MNC group included moreresponders.

Fasting plasma glucose

FPG is the metabolic parameter that indicates thediagnostic criteria for diabetes mellitus. Informationon FPG was available in 11 trials[15e17,20,21,23e27,29] containing 278 patients(91 patients as the control group without stem celltherapy). The efficacy of the cell therapy in the esti-mated pooled MD for all of the three trials[16,20,21] showed significant reduction in FPG withthe inverse variance model (BM-MNC group:MD, �19.30; 95% CI, �38.01 to �0.59; P ¼ 0.04)in T2D after 12-month follow-up (Figure 3A). Onthe other hand, we compared the level of FPG in thedifferent stages before and after BM-MNC or PB-MNC treatment to show the effect of the therapy.The estimated pooled MD for eight trials shows ahighly significant decrease of FPG after 9-monthfollow-up in the BM-MNC therapy group(MD, �83.51; 95% CI, �109.51 to �57.51; P <0.001) but not for 6-month and 3-month follow-up,respectively (BM-MNC group: MD, �34.0; 95%CI, �75.5 to 7.49; P ¼ 0.11; MD, �24.38; 95%CI, �74.23 to 25.47; P ¼ 0.34; PB-MNC group:MD, �42.44; 95% CI, �128.11 to 43.24; P ¼ 0.33)

(Figure 3B). Thus, we observed after a long follow-up that FPG showed some improvements in theBM-MNC group including 235 patients but no sig-nificant change in the PB-MNC group with 43patients.

C-peptide

C-peptide levels are the established biomarker forendogenous insulin synthesis. Therefore, wecollected information on C-peptide that was availablein seven trials [15e17,20,21,24,26,27]. Theseincluded trials contained 271 patients (78 patients asthe control group without stem cell therapy). Theefficacy of the stem cell therapy in the estimatedpooled MD for the two trials showed a significant C-peptide increase with the inverse variance model(MD, 0.97; 95% CI, 0.73 to 1.2; P < 0.001) in T2Dfor 12 months in the BM-MNC group (Figure 4A).On the other hand, we compared the level of C-peptide in the different stages before and after stemcell treatment to show the effect of the therapy. Theestimated pooled MD was highly significantlyincreased after 6-month follow-up (BM-MNCgroup: MD, 1.13; 95% CI, 0.79 to 1.47; P < 0.001;PB-MNC group: MD, 0.60; 95% CI, 0.01 to 1.18;P ¼ 0.04) but not after 3- and 12-month follow-up(BM-MNC group: MD, 1.81; 95% CI, �0.15 to3.77; P ¼ 0.07; PB-MNC group: MD, 0.23; 95%CI, �0.27 to 0.74; P ¼ 0.37; MD, 0.11; 95%CI, �0.78 to 1.00; P ¼ 0.81) (Figure 4B). Generally,C-peptide improvements were observed in the BM-MNC group with 102 cases and in the PB-MNCgroup with 169 cases.

Insulin requirement

Some prospective and randomized studies demon-strated that the stem cell transplantation results inreduction in exogenous insulin requirement in pa-tients with T2D having oral anti-diabetic drug fail-ure and requiring insulin for glycemic control [25].Information on insulin requirement was available insix trials [15e17,20,22,25] containing 119 patients.Regarding the efficacy of the cell therapy, the esti-mated pooled MD showed no significant reductionin insulin requirement with the inverse variancemodel (BM-MNC group: MD, �0.12; 95%CI, �0.24 to 0.01; P ¼ 0.06) in T2D after 12-monthfollow-up (Figure 5A). We compared the level ofinsulin requirement in the different stages beforeand after BM-MNC treatment. The estimatedpooled MD shows a highly significant decrease ininsulin requirement after 3-, 6-, 9- and 12-monthfollow-up, respectively, in the BM-MNC group(MD, �21.96; 95% CI, �37.77 to �6.14;

A

B

Figure 3. Forest plot for FPG. Comparison of FPG between the stem cell therapy and control groups (A); comparison of FPG before stemcell therapy and after stem cell therapy (B). The random-effects model (Mantel-Haenszel method) was used in this analysis.

962 Z.-X. Wang et al.

P ¼ 0.007; MD, �34.29; 95% CI, �51.38to �17.20; P < 0.001; MD, �26.75; 95%CI, �29.18 to �24.32; P < 0.001; MD, �31.73;95% CI, �34.26 to �29.20; P < 0.001) (Figure 5).Therefore, especially for insulin requirement, thecomparison results showed significant improve-ments in the BM-MNC group with 119 patients,whereas for the PB-MNC group in the selected tri-als, we could not collect the data for meta-analysisbecause six of the selected trials reported therelated different data. One reported the number ofthe patients of the insulin requirement [21], and,

although two of them showed the data with the unitof U/kg/d [26,27], one trial only included two pa-tients’ data. Some reported the value of the insulin-releasing test [23,28,29]; therefore, for the insulinrequirement comparison, the PB-MNC groupincluded six studies [21,23,26e29], but we couldnot collect these data in our meta-analysis.

Discussion

The importance of preventing diabetes in high-riskindividuals is highlighted by the substantial and

A

B

Figure 4. Comparison of C-peptide treatment with stem cell and control therapy (A). The fixed-effects meta-analysis model (Mantel-Haenszel method) was used in this analysis. Comparison of C-peptide before stem cell treatment and after stem cell treatment (B). Therandom-effects meta-analysis model (Mantel-Haenszel method) was used in this analysis.

Autologous stem cell therapy type 2 diabetes mellitus 963

world-wide increase in the prevalence of diabetes inrecent years. Diabetes mellitus has no permanentcure to date. There are various stem cell sources andmanipulation techniques that can be used to generatefunctional b-like cells in a relatively safe and efficient

manner [30]. Stem cellebased therapy could be oneof the most promising solutions.

Data collected from clinical trials completed todate support the hypothesis that MSCs and BMCscan perform immunomodulatory functions to

A

B

Figure 5. Forest plot for the insulin requirement assessment. The data come from patient treatment with stem cell therapy and controltreatment (A) and before stem cell treatment and after stem cell treatment (B). The random-effects meta-analysis model (Mantel-Haenszelmethod) was used in this analysis.

964 Z.-X. Wang et al.

suppress the adverse immunological response. Themechanisms for the MSCs and BMCs were mainlyinvolved in the secretion of several factors includingtransforming growth factor-b, hepatocyte growth factor(HGF), nitric oxide and prostaglandin E2 [31e33],indicating that MSCs and BMCs act to modulate thehost immune response and protect other cells frominnate and adaptive immune responses. Secretion ofstem cell release molecules such as basic fibroblastgrowth factor, HGF, angiopoietin-1 and -2, vascularendothelial growth factor and cysteine-rich protein 61

byMSCs andBMCs leads to increased vascular densityand blood flow in ischemic myocardium, resulting inincreased perfusion and function [34]. In addition,cardiac levels of interleukin-1b, tumor necrosis factor-aand stem cell release molecules implicated in angio-genesis were shown to be elevated after MSC trans-plantation.The expressions of stromal-derived factor-1,insulin-like growth factor-1, HGF, and pigmentepithelium-derived factor are known to promote repairand regeneration by facilitating circulating progenitorcell recruitment to damaged tissues [35].

Autologous stem cell therapy type 2 diabetes mellitus 965

As far as we are aware, our study represents thefirst meta-analysis on the clinical efficacy of autolo-gous stem cell transplantation for the treatment ofT2D in the multi-country populations. In our study,HbA1c and daily insulin requirement decreasedsignificantly after stem cell transplantation. However,there was no significant change of FPG or improve-ment of C-peptide at the end of several studies.

Our analysis showed significant HbA1c reductionin T2D treated with stem cells after 12 monthscompared with the control therapy. The level ofHbA1c before and after stem cell treatment showed ahighly significant decrease after 12-, 9-, 6- and 3-month follow-up in both the BM-MNC group andthe PB-MNC group. HbA1c is a good indicator forthe blood glucose levels because of its characteristicsof more objectiveness and less susceptibility to diet,exercise and other interference factors. Thus, it al-ways acts as a complementary value to the pointvalues of blood glucose. Therefore, the resultdemonstrated that stem cell therapy may improveglycemic control. According to the HbA1c goal(<7%) recommended by the American DiabetesAssociation for the treatment of adult diabetes, mostof the studies achieved this goal after 6 months, and,although the 3-month follow-up is significant, mostof the studies had not reached this criteria at thattime. Thus, it is obvious that stem cell therapy is asafe approach that produces lasting improvement inmetabolic control for T2D. In addition, there aresome newly reported clinical trials including the largesamples with 240 patients performed in China [36],but the results did not demonstrate when to analyzethe index of the therapy, so we cannot include them.An open question remains regarding how long thetherapeutic effects will be sustained, and there is nodefinitive period for how long this observation shouldlast; therefore, much longer and more elegant trialsare needed.

Our analysis showed a slight FPG reduction inT2D treated with the stem cells after 12 monthscompared with the control therapy. The level of FPGbefore and after stem cell treatment was also shownto be significantly decreased after 9- and 3-monthfollow-up but not for 6-month follow-up, respec-tively, in the BM-MNC group and the PB-MNCgroup. It was reported that MSCs can promoteregeneration of the injured islet [37]; however, thefailure of BMC transplantation to restore b-cellfunction in patients has been reported in somestudies, such as the Osiris trial [38,39]. Furthermore,a recent published study by D’Addio et al. [40] onautologous hematopoietic stem cell transplantationin T1DM reported the interesting finding that in-dividuals had good and poor responses to the ther-apy, which suggests that the CD34þ cell fraction in

the total cell transplant would play a key role;therefore, in different studies, the cell fraction alsoaffects the therapeutic efficacy. In our meta-analysis,we collected the clinical trials including differentstem cell therapy for T2D for the limitation of thepatient cases, and we then classified them into twogroups on the basis of the cell types. We found thatthe different stem cell types, numbers and cultureconditions may play diverse roles in T2D treatmentand would induce a bias of the analysis that shouldbe clarified further. In some trials, it showed a pro-gressive and consistent reduction in plasma glucosein several included studies [22,25] but not in otherincluded studies [15,17,24], which suggests thatautologous stem cell infusion is the potential therapyfor T2D to some extent. In addition, we should pointout that the unit of measurement for FPG should beunited further, which would affect the analysis result.For example, in our analysis, the data in differentsources of studies have more than a 10-magnitudegap, and we conducted a unit conversion (�18for mmol/L to mg/dL) for unification of units ofmeasurement.

Our results also showed no significant insulinrequirement reduction in T2D with stem cell treat-ment after 12 months compared with the controltherapy in the BM-MNC group. The level of insulinrequirement before and after stem cell treatmentshowed a highly significant decrease after 3-, 6-, 9-and 12-month follow-up in the BM-MNC group,but no data were shown in the PB-MNC group forthe limitation of the trials and the shortage of thedata. In addition, we should denote that in all of ourincluded clinical trials, only two of them mentionedthat after the cell transplant 24.2% of patients ob-tained insulin independence [25] and two patients inall six switched to oral hypoglycemic drugs [24]. Inall, the comparison is consistent and sustained at theend of the follow-up in most studies, and, until now,we collected more trials with BM-MNC therapy thanwith PB-MNC. Therefore, T2D patients who receiveself-donated (autologous) bone marrow stem cellsrequire less insulin but receive the treatment for type2 diabetes mellitus and its complications. The dataavailable thus far from animal and human studies isencouraging; however, there is a need for longerdurations of follow-up.

Our analysis showed a significant C-peptide in-crease in T2D treated with stem cells compared withcontrol therapy after 12 months in the BM-MNCgroup. The level of C-peptide before and after stemcell treatment showed a highly significant increaseafter 6 months but not the 12-month follow-up in theBM-MNC group and the PB-MNC group.Regarding the C-peptide and the insulin requirementthat demonstrates the functions of the b cells,

966 Z.-X. Wang et al.

C-peptide is a by-product of insulin synthesis and isreleased at equal levels; it demonstrated the insulinsensitivity, but T2D patients received external insu-lin injections and other treatments that limit the ac-curacy of C-peptide. Some studies showed the areaunder the curve of peptide [21]; therefore the dataanalysis may induce the bias.

Although mechanisms for the MSCs and BMCsthat improved glycemic control remain to be inves-tigated, increased C-peptide levels suggest anincreased insulin biosynthesis in these studies. Theincreased insulin biogenesis could either be the re-sults of expanded insulin-secreting b-cells or elevatedinsulin secretion from existing b-cells. Currentlystudy does not provide mechanistic endogenousmature b-cells. Furthermore, it appears that b-cellmaintenance may depend on the ability of MSCs andBMCs to prevent islet cell loss resulting fromexocrine malfunction or degeneration, which con-tributes to the pathogenesis of diabetes [18]. It wasshown that BALB/c-MSC mice expressed higherlevels of the negative co-stimulatory molecule PD-L1and that promoted a shift toward Th2-like responsesand soft tissue and visceral tumors in treated NODmice [41]. Furthermore, it has also been demon-strated that after additional passages, the immortal-ized BM-MSCs became transformed into malignantcells, capable of forming fibrosarcomas in vivo[42,43]. Thus, the importance of these observationsremains to be explored in humans; however, becauseof their oncogenic potential, MSCs should be furtherconsidered in the clinical application.

In short, the application of stem cell therapy as acure for T2D appears promising and tolerable, withbona fide hope for a permanent cure with no sideeffects.

Limitations

Although our meta-analysis showed that autologousstem cell implantation is safe and effective for themajority of T2D patients, it also has certain contra-indications. All 15 trials included in the analysis weremainly conducted in China, only three of them wereperformed in India and the United States. The totalsample size is not very large. The follow-up time wasnot sufficiently long. Patient information was limited,and that might affect the results observed. Somestudies mentioned that the therapeutic effects wereinfluenced by the degree to which the patients fol-lowed their doctor’s advice such as regarding dietcontrol and medication.

The present study has several limitations. First,from a clinical trial point of view, the two groups inthe study should have exactly the same backgroundexcept for the intervention factor, but, in some

studies, treatment with stem cells is a surgical oper-ation, and, as a result, patients endure more distress.Patients were divided into two groups on the basis oftheir willingness to participate in these studies; thismay lead to a bias and is the most important studylimitation. Second, subjects in each study wererecruited from patients with T2D without severecomplications; moreover, the involved number wasrelatively small. These results must be replicated in alarger cohort.

The reliability of this systemic review might alsobe influenced by other factors. For example, most ofthe included studies did not report clinical random-allocation concealment; therefore, this meta-analysis may have distribution and implementationbias. Furthermore, clinical studies with stem cells arestill in their infancy; on the basis of the encouragingexperimental and clinical evidence currently avail-able, randomized clinical trials considering the celltype, cell number, cell culture condition and infusionmethod are justifiable and should be done understringent compliance with the CONSORT princi-ples. This will certainly involve a large number ofpatients to demonstrate statistical significance for amodest degree of outcome superiority. Such studiesare urgently needed to provide unequivocal evidenceof the clinical usefulness of stem cell therapy in T2D.

In all, according to our analysis, the datademonstrated superiority in glycemic control,increased insulin biosynthesis, elevated insulinsecretion from existing b-cells and prevent islet cellloss. Hence, it was proven that efficacy lies in thepossibility of application of a promising therapymethod for T2D, but it also needs more maturationstem cell therapy development. Furthermore, toclarify the exact treatment mechanism, furtherstudies that address the precise molecules andpathways involved in the factors, including cellhoming, microenvironment improvement, in-teractions between stem cells and islet progenitorcells and the adjunctive drugs or other types of cellsthat interact with stem cells during treatment, will berequired.

Acknowledgments

This research work was supported by the NationalNatural Science Foundation of China (grant nos.31171427 and 30971651 to Z.-X.W.); BeijingMunicipal Science & Technology Project; Clinicalcharacteristics and Application Research of Capital(grant no. Z121107001012136 to Z.-X.W.); theNational Natural Science Foundation of China(grant no. 30700974 to J.-X.C.) and the PostdoctoralFoundation of China (grant no. 20060400775 toJ.-X.C.).

Autologous stem cell therapy type 2 diabetes mellitus 967

Disclosure of interests: The authors have nocommercial, proprietary, or financial interest in theproducts or companies described in this article.

References

[1] TODAY Study Group, Zeitler P, Hirst K, Pyle L, Linder B,Copeland K, et al. A clinical trial to maintain glycemiccontrol in youth with type 2 diabetes. N Engl J Med 2012;24:2247e56.

[2] Yang W, Lu J, Weng J, Jia W, Ji L, Xiao J, et al. ChinaNational Diabetes and Metabolic Disorders Study Group.Prevalence of diabetes among men and women in China.N Engl J Med 2010;12:1090e101.

[3] Olokoba AB, Obateru OA, Olokoba LB. Type 2 diabetesmellitus: a review of current trends. Oman Med J 2012;4:269e73.

[4] Hess D, Li L, Martin M, Sakano S, Hill D, Strutt B, et al.Bone marrow-derived stem cells initiate pancreatic regener-ation. Nat Biotechnol 2003;21:763e70.

[5] Hickey RD, Galivo F, Schug J, Brehm MA, Haft A, Wang Y,et al. Generation of islet-like cells from mouse gall bladder bydirect ex vivo reprogramming. Stem Cell Res 2013;1:503e15.

[6] Pennarossa G, Maffei S, Campagnol M, Tarantini L,Gandolfi F, Brevini TA. Brief demethylation step allows theconversion of adult human skin fibroblasts into insulin-secreting cells. Proc Natl Acad Sci U S A 2013;22:8948e53.

[7] Kumar A, Lo Nigro A, Gysemans C, Cai Q, Esguerra C,Nelson-Holte M, et al. Reversal of hyperglycemia by insulin-secreting rat bone marrow- and blastocyst-derived hypoblaststem cell-like cells. PLoS One 2013;5:e63491.

[8] Hosoya M. Preparation of pancreatic b-cells from human iPScells with small molecules. Islets 2012;3:249e52.

[9] Urbán VS, Kiss J, Kovács J, Gócza E, Vas V, Monostori E,et al. Mesenchymal stem cells cooperate with bone marrowcells in therapy of diabetes. Stem Cells 2008;1:244e53.

[10] Shin L, Peterson DA. Impaired therapeutic capacity ofautologous stem cells in a model of type 2 diabetes. StemCells Transl Med 2012;2:125e35.

[11] Burt RK, Loh Y, Pearce W, Beohar N, Barr WG, Craig R,et al. Clinical applications of blood-derived and marrow-derived stem cells for nonmalignant diseases. JAMA 2008;8:925e36.

[12] Prasad VK, Lucas KG, Kleiner GI, Talano JA, Jacobsohn D,Broadwater G, et al. Efficacy and safety of ex vivo culturedadult human mesenchymal stem cells (Prochymal) in pedi-atric patients with severe refractory acute graft-versus-hostdisease in a compassionate use study. Biol Blood MarrowTransplant 2011;4:534e41.

[13] Couri CE, Oliveira MC, Stracieri AB, Moraes DA,Pieroni F, Barros GM, et al. C-peptide levels and insulinindependence following autologous nonmyeloablative he-matopoietic stem cell transplantation in newly diagnosedtype 1 diabetes mellitus. JAMA 2009;15:1573e9.

[14] Zhao Y, Jiang Z, Guo C. New hope for type 2 diabetics:targeting insulin resistance through the immune modulationof stem cells. Autoimmun Rev 2011;2:137e42.

[15] Estrada EJ, Valacchi F, Nicora E, Brieva S, Esteve C,Echevarria L, et al. Combined treatment of intrapancreaticautologous bone marrow stem cells and hyperbaric oxygen intype 2 diabetes mellitus. Cell Transplant 2008;12:1295e304.

[16] Wu Z, Cai J, Chen J, Huang L, Wu W, Luo F, et al.Autologous bone marrow mononuclear cell infusion andhyperbaric oxygen therapy in type 2 diabetes mellitus: anopen-label, randomized controlled clinical trial. Cytotherapy2014;16:258e65.

[17] Bhansali A, Upreti V, Khandelwal N, Marwaha N, Gupta V,Sachdeva N, et al. Efficacy of autologous bone marrow-derived stem cell transplantation in patients with type 2diabetes mellitus. Stem Cells Dev 2009;10:1407e16.

[18] Wang L, Zhao S, Mao H, Zhou L, Wang ZJ, Wang HX.Autologous bone marrow stem cell transplantation for thetreatment of type 2 diabetes mellitus. Chin Med J (Engl)2011;22:3622e8.

[19] Hu J, Li C, Wang L, Zhang X, Zhang M, Gao H, et al. Longterm effects of the implantation of autologous bone marrowmononuclear cells for type 2 diabetes mellitus. Endocr J2012;11:1031e9.

[20] Bhansali A, Ashokumar P, Walia R, Bhansali S, Gupta V,Jain A, et al. Efficacy and Safety of Autologous Bone MarrowDerived Stem Cell Transplantation in patients with Type 2Diabetes mellitus: A randomized placebo-controlled study.Cell Transplant 2014;23:1075e85.

[21] Ding M, Wang Y, Wang B, Wang H, Dong S, Cui L, et al.Value of targeting intervention technology in autologousperipheral blood stem cells therapy of type 2 diabetes melli-tus. Chin J Clinicians 2012;17:5126e30.

[22] Wang P, Wu Z, Tian J, Qi L. Autologous bone marrow celltransplantation via spleen artery in 5 patients with type 2diabetes Mellitus. J Clinical Rehabilitative Tissue Engineer-ing Research 2010;23:4355e8.

[23] Xu G, Yang Y, Yin Z, Li P, Xiao Y, Zhao H, et al. Theclinical observation of umbilical cord blood stem celltransplantation on the treatment of type 2 diabetes. ChinaMedicine and Pharmacy 2011;11:22e5.

[24] Yao J, Duan L, Tong Q, Xu Z. Efficacy and safety ofautologous bone marrow stem cells transplantation for type 2diabetes mellitus. J Third Mil Med Univ 2012;1:74e7.

[25] Chen L, Dong JJ, Hou WK, Hou XG, Sun L, Dong M, et al.Sustained effect of autologous bone marrow mononuclearcell transplantation in-patients with diabetes: 12 Monthfollow-up. 67th Annual Meeting of the American-Diabetes-Association. 2008 JUN 06-10 San Francisco. Diabetes 2008;57(Suppl):A544.

[26] Tong Q, Duan L, Xu Z, Wang H, Wang X, Li Z, et al.Improved Insulin Secretion Following Intrapancreatic UCBTransplantation in Patients With T2D. J Clin EndocrinolMetab 2013;9:E1501e4.

[27] Liu X, Zheng P, Wang X, Dai G, Cheng H, Zhang Z, et al.A preliminary evaluation of efficacy and safety of Whar-ton’s jelly mesenchymal stem cell transplantation in pa-tients with type 2 diabetes mellitus. Stem Cell Res Ther2014;5:57.

[28] Wang F, Wu W. Exploring the clinical efficacy of the um-bilical cord blood stem cell transplantation on the treatmentof type 2 diabetes. Journal of Qiqihar University of Medicine2014;35:1452e3.

[29] Zhao Y, Jiang Z, Zhao T, Ye M, Hu C, Zhou H, et al.Targeting insulin resistance in type 2 diabetes via immunemodulation of cord blood-derived multipotent stem cells(CB-SCs) in stem cell educator therapy: phase I/II clinicaltrial. BMC Med 2013;11:160.

[30] Shen J, Cheng Y, Han Q, Mu Y, Han W. Generating insulin-producing cells for diabetic therapy: existing strategies andnew development. Ageing Res Rev 2013;2:469e78.

[31] Baraniak PR, McDevitt TC. Stem cell paracrine actions andtissue regeneration. Regen Med 2010;1:121e43.

[32] Jurewicz M, Yang S, Augello A, Godwin JG, Moore RF,Azzi J, et al. Congenic mesenchymal stem cell therapy re-verses hyperglycemia in experimental type 1 diabetes. Dia-betes 2010;59:3139e47.

[33] Francese R, Fiorina P. Immunological and regenerative prop-ertiesof cordbloodstemcells.Clin Immunol2010;136:309e22.

968 Z.-X. Wang et al.

[34] Schenke-Layland K, Strem BM, Jordan MC,Deemedio MT, Hedrick MH, Roos KP, et al. Adiposetissue-derived cells improve cardiac function followingmyocardial infarction. J Surg Res 2009;2:217e23.

[35] Maguire G, Friedman P. The Systems Biology of Stem CellReleased Molecules-Based Therapeutics. 2013; Article ID784541, 12 pages. http://dx.doi.org/10.1155/2013/784541.

[36] Zhou X, Yu W, Chen L, Zhan Y, Wu J, Chen J. Autologousbone marrow mesenchymal stem cell transplantation for thetreatment of type 2 diabetes. Contemporary Medicine 2013;17:20e1.

[37] Meng F, Liu X, Li H, Pang X, Li Gang. Regulation of bloodglucose by administration of allogenic bone marrow derivedmesenchymal stem cells into diabetic rats through modu-lating regeneration of both a and b islet cells. J Diabetes Res2013;2:7.

[38] Fiorina P, Voltarelli J, Zavazava N. Immunological applica-tions of stem cells in type 1 diabetes. Endocr Rev 2011;32:725e54.

[39] Osiris Therapeutics provides update on groundbreaking stemcell trial for type 1 diabetes. Osiris Therapeutics, Inc. website. http://investor.osiris.com/releasedetail.cfm?ReleaseID_636520; [accessed 10.10.13].

[40] D’Addio F, Valderrama Vasquez A, Ben Nasr M, Franek E,Zhu D, Li L, et al. Autologous nonmyeloablative hemato-poietic stem cell transplantation in new-onset type 1 dia-betes: a multicenter analysis. Diabetes 2014;63:3041e6.

[41] Fiorina P, Jurewicz M, Augello A, Vergani A, Dada S, LaRosa S, et al. Immunomodulatory function of bone marrow-derived mesenchymal stem cells in experimental autoim-mune type 1 diabetes. J Immunol 2009;183:993e1004.

[42] Miura M, Miura Y, Padilla-Nash HM, Molinolo AA, Fu B,Patel V, et al. Accumulated chromosomal instability in mu-rine bone marrow mesenchymal stem cells leads to malignanttransformation. Stem Cells 2006;24:1095e103.

[43] Tolar J, Nauta AJ, Osborn MJ, Panoskaltsis Mortari A,McElmurry RT, Bell S, et al. Sarcoma derived from culturedmesenchymal stem cells. Stem Cells 2007;25:371e9.