Umbilical cord blood mononuclear cell therapy induces clinical remission of steroid-dependent or -resistant ulcerative colitis patients

To compare the efficacy and safety of umbilical cord blood mononuclear cells (CBMNC) and azathioprine (AZA) in the treatment of patients with steroid-dependent or -resistant ulcerative colitis. One hundred and six patients diagnosed with steroid-dependent or -resistant ulcerative colitis were studied retrospectively, including 36 patients treated with CBMNC and 70 treated with AZA. To reduce confounding bias due to retrospective nature of this study, the propensity score matching system was applied to equipoise the pretreatment data of two groups. After matching, 35 matched pairs (1:1) were created. The ratios of clinical remission, clinical response and endoscopic mucosal healing, Mayo score, and major complications were compared between two groups at weeks 8, 16, and 36 after treatment. The results demonstrated that the ratios of clinical remission (80% vs. 57%, P < 0.05) and mucosal healing (74% vs. 51%, P < 0.05) were significantly higher in CBMNC-treated patients compared with those in AZA-treated patients at week 8. The erythrocyte sedimentation rate was significantly decreased in CBMNC group compared with that in AZA-treated group (14.5 ± 3.9 mm/h vs. 18.0 ± 5.7 mm/h, P < 0.01) at week 8. In AZA group, 2 patients had neutropenia and 3 patients had elevated alanine aminotransferase levels, whereas no obvious side-effects were observed in CBMNC-treated group. Our results reveal that CBMNC therapy appears to be an effective and safe strategy for patients with steroid-dependent or -resistant ulcerative colitis. Further prospective studies are needed to define the potential roles and mechanisms of CBMNC in the treatment of refractory ulcerative colitis.


INTRODUCTION
Ulcerative colitis (UC) is a common gastroenterological disorder which is characterized by chronic progressive inflammatory diseases in the gastrointestinal tract. Although the exact causes of UC remain unclear, recent studies showed that UC appears to be a disorder of the host immune response to microbiota which manifested by a state of local immune hyperactivity [1,2]. It is also widely accepted that genetic and environmental factors are involved in the pathogenesis of UC [3,4]. Finally, all these factors lead to the infiltration of leukocytes and lymphocytes into the intestinal mucosa which is believed to be critical in the formation of mucosal lesions in UC patients [5].
During the last decade, studies from countries of Europe, North America and Asia have shown that the annual incidence of UC is 24.3/100,000, 19.2/100,000, Research Paper www.oncotarget.com and 6.3/100,000, respectively [6]. To date, the established therapies for UC include 5-aminosalicylates, corticosteroids and immunosuppressants. The adverse effect is major limitation in the use of corticosteroids, especially for the steroid-dependent or -resistant UC patients. Although immunosuppressants such as azathioprine (AZA) or 6-mercaptopurine are dominant drugs most frequently used in inducing and maintaining remission in refractory UC patients, there is still about 30% of UC patients dropped the AZA therapy due to the side-effects or lack of clinical efficacy [7,8]. Recently, biological agents such as infliximab and adalimumab provide a new therapy for inflammatory bowel disease (IBD), but the looming risks including opportunistic infection and malignancy, specifically, lymphoma limit their clinical application [9]. Some refractory UC like steroid-dependent or -resistant UC patients even face a significant long-term risk of colectomy [10]. Thus, new optimal therapies aim at a cure for UC are warranted, which should not only focus on blocking intestinal mucosal inflammation, but also enhancing intestinal mucosal proliferation and coordinately remodeling during the ulceration-healing process [11].
Stem cells (SC) are emerging as a promisingly therapeutic approach for the treatment of active UC owing to its self-renewal, multipotency, immunosuppressive and tissue-repair promotion functions [12,13]. Umbilical cord blood mononuclear cells (CBMNC) comprising a mix of monocytes, immature lymphocytes, endothelial, hematopoietic and mesenchymal stem cells (MSC) [14,15] have been successfully applied in hematopoietic malignancy for almost 2 decades [16][17][18]. Their advantages including economic and simple retrieval, enrichment of hematopoietic progenitors, enhanced proliferation rate [19], and weak cellular immune responses in vivo [20,21] make them a promising cellular treatment for UC. To date, emerging evidence suggests that adult SC contribute to tissue regeneration by enhancing microcirculation in murine IBD model [22]. Currently, clinical trials mainly focus on SC treatment for refractory CD and perianal fistulizing CD, only a few clinical trials focus on UC were reported [23,24]. Based on the above evidences, we hypothesized that the CBMNC therapy to be an effective and safe strategy for patients with refractory UC. Therefore, we conducted this study to investigate the efficacy and safety of CBMNC therapy for the treatment of steroid-dependent or -resistant UC patients.

Baseline characteristics
106 patients who were diagnosed with steroiddependent or -resistant UC in Chinese PLA General Hospital of Jinan Military Command were randomly spilt into CBMNC-treated group and AZA-treated group according to their initial treatment strategy. To reduce sample selection bias, propensity score matching (PSM) was applied to equipoise the pretreatment data of these two groups. After performing PSM for the entire population, 35
In AZA-treated group, 77% (27/35) of patients showed clinical response, 57% (20/35) achieved clinical remission and 51% (18/35) achieved endoscopic mucosal healing, while no improvement was observed in 8 patients (23%). Moreover, the Mayo score decreased from 6.9 ± 1.6 to 3.2 ± 1.9 after treatment (P < 0.01); ESR decreased from 34.2 ± 14.4 mm/h to 18.0 ± 5.7 mm/h (P < 0.01); CRP levels decreased from 37.3 ± 13.6 mg/l to 10.3 ± 5.3 mg/l after treatment (P < 0.01); and the defecation frequencies decreased from 5.6 ± 1.1 times per day to 2.6 ± 0.8 times per day after treatment (P < 0.01). The ratios of clinical remission and endoscopic mucosal healing were markedly higher in CBMNC-treated group compared with those in AZA-treated group (P < 0.05). In addition, the level of ESR was found to be decreased in CBMNCtreated group compared to that in AZA-treated group (P < 0.01). But there was no significant difference in clinical response ratio, Mayo score, CRP levels and defecation frequency between two groups after 8 week of treatment (all P > 0.05).
After  36, CBMNC group still showed high ratios of clinical remission and endoscopic mucosal healing compared to AZA group (P < 0.05). The ESR level was lower than that of AZA-treated group, but the difference was not significant (P = 0.08).
Although both groups showed improve of clinical remission and endoscopic mucosal healing, the ratios of clinical remission and endoscopic mucosal healing were significantly higher in CBMNC-treated group compared to AZA-treated group at weeks 8, 16 and 36 (all P < 0.05, Figure 1).
During and after CBMNC treatment, all patients tolerated well and no major complications occurred. In AZA group, 2 patients (5.7%) suffered from neutropenia and asked to discontinue AZA therapy after 7 month treatment and received leicosen (Jibeier Pharmaceutical Industries, China, registration certificate number: H32025444) (20 mg, tid) therapy; Signs of hepatotoxicity (alanine aminotransferase increase more than 2 times of ULN) were reported in 3 patients (8.6%) after 6 month of treatment, after Polyene Phosphatidylcholine (Sanofi aventis, China, registration certificate number: H20059010) therapy, the level of alanine aminotransferase (ALT) in all 3 patients returned to normal and 2 patients asked to discontinue AZA therapy.

DISCUSSION
Our results here described that CBMNC therapy appears to be a safe and effective approach for patients with steroid-dependent or -resistant UC. As compared to patients who received AZA therapy, patients who received CBMNC therapy were significantly more like to achieve endoscopic mucosal healing and clinical remission at week 8 and these improvements sustained for 36 weeks after administration. These findings indicate CBMNC as a promising therapy for refractory UC.
The human intestinal epithelium is renewed every 3-5 days by SC residing in the base of crypt which generate about 300 cells per crypt per day. The mucosal lesions in UC patients caused by infiltration of leukocytes and lymphocytes also accompanied by destruction of the crypt structure. Therefore, how to replenish enough SC maybe the key for treatment of UC [5]. In addition, SC especially MSC have the capacity to modulate the immune response through the suppression of dendritic cell maturation and their antigen-presenting abilities [25]. It also demonstrated that human adult SC derived from adipose tissue have the ability to suppress acute inflammatory autoimmune responses by increasing interleulkin-10 and inhibiting inflammatory and Th1 responses in mice [26]. To this end, SC therapies are receiving more attention from IBD investigators.
Hematopoietic SC and MSC have been evidenced to be used for various systemic organ regenerative therapies for almost two decades. With regard to hematopoietic stem cell transplant (HSCT) therapy for IBD, it was reported that 27 IBD cases (22 CD and 5 UC) who underwent HSCT for hematemesis associated cancer from 1993 to 2004 (19 allogeneic and 8 autologous) [27][28][29][30][31][32]. Among those patients, 24 of 27 patients have achieved clinical remission (0.5 to 8 years) while 3 patients who suffered from relapses underwent autologous HSCT. Burt et al. reported that in the phase I/ II study of autologous HSCT (non-myeloablative) in 24 patients with severe refractory CD to anti-TNF therapy, the percentage of clinical relapse-free survival was 91%, 63%, 57%, 39%, and 19% at 1, 2, 3, 4, and 5 years, respectively [33,34]. These results indicate that CBMNC may act as a promising therapy strategy for the treatment of UC. However, cancer-specific treatment is used for myeloablative treatment in most of above studies which need specified destroy the bone marrow stem cells, resulting in lethal and irreversible marrow failure if SC are not properly reinfused. Non-myeloablative autologous HSCT in autoimmune disease appear to offer substantial benefit compared to myeloablative therapy, but it is still unclear whether autologous HSCT which expand in IBD patients with autoimmune or inflammatory diseases could be fully functional [35]. Moreover, an important finding in a phase III study designated as the autologous HSCT for CD patients has demonstrated that mobilization may be the only benefit of this therapy [24].
For non-myeloablative studies in CD, preliminary results of phase II clinical trial using allogeneic bone marrow-derived MSC for treatment refractory luminal CD also showed prospective efficacy and safety. 12 out of 14 patients were improved, including 8 patients with remission and 7 patients of the endoscopic improvement, and no acute reactions were reported [23]. Therefore, these results highlight the safety and efficacy of CBMNC treatment. In our study, we used non-myeloablative CBMNC therapy for the treatment of UC. Ex vivo isolated CBMNC were infused with saline intravenously. After 8 weeks of treatment, CRP and ESR levels which are associated with the activity of UC were significantly decreased, indicating that the CBMNC treatment relieved the inflammatory response in active UC patients. The defecation frequency was also significantly decreased in CBMNC group compared with AZA group. Moreover, clinical remission and endoscopic mucosal healing ratio were significantly higher in the CBMNC group compared with the AZA group. The improved mucosal healing can prevent relapse and improve the life quality of patients. Similarly, these improvements were observed at 16 and 36 weeks after treatment.
In regard to adverse reactions, there were no obvious complications in CBMNC group. These benefits may attribute to the non-myeloablative treatment in this study. In AZA group, side-effects were complained by 5 (14.3%) out of 35 patients which included neutropenia (5.7%) and hepatotoxicity (8.6%). The incidence of side-effects in the AZA-treated group were lower compared with the European studies (31%-45%) [36,37], this result indicating low frequency of AZA dose-dependent side-effects in these studies. The maintenance dose of AZA was lower in our study (1.5 mg/kg/day) than that in the European study (2-2.5 mg/kg/day) without sacrificing efficacy [38]. The small sample size as well as short follow-up period in our study might be also responsible for low frequency of adverse reactions. Follow-up to week 36, there were no other adverse effects in AZA treatment group.
There were also several limitations in our study. First, it could not avoid selection bias as a retrospective study. Although propensity score-matched analysis was performed to adjust for potential confounding factors, it still did not eliminate unmeasured variables and initial selection bias. A second consideration is our small sample size might not be reflective of a large population and the follow-up period was too short. Besides, the exact mechanisms and long-term adverse effects of this therapeutic remain unknown. Several factors, such as the source and type of SC, the quality control of prepared SC, the procedure of administration (route, dose, pretreatment) and other factors that may influence the transplantation efficacy and how to control the appropriate differentiation in the preferred location are still poorly understood. Despite all of above limitations, this study has shown the potential advantages of CBMNC for the treatment of refractory UC.
In conclusion, CBMNC administration provides effective protection and safety in patients with steroiddependent or -resistant ulcerative colitis. These patients have higher ratio of clinical remission and mucosal healing at 8 weeks and sustained up to 36 weeks after therapy. Further prospective studies are needed to define the roles of the CBMNC therapy in the treatment of refractory UC.

MATERIALS AND METHODS
This study was conducted as a retrospective analysis and was approved by the institutional review boards of Chinese PLA General Hospital of Jinan Military Command and The Second Affiliated Hospital of Shandong University of Traditional Chinese Medicine. The requirement to obtain informed consent was waived because of the retrospective nature of this study, but a written informed consent was obtained from each subject before CBMNC therapy after full explanation of the purpose and nature of the procedure used.
All of the patients were followed and diagnosed through colonoscopy and pathological examination.

Preparation of umbilical cord blood mononuclear cells and administration
Fresh human umbilical cord blood (80-120 ml) was obtained from 36 informed healthy donors. All of samples were tested for hepatitis B, hepatitis C, and HIV, syphilis and cytomegalovirus, and serum ALT. In briefly, umbilical cord blood was first diluted with normal saline in the ratio of 2:1, then transferred 30 ml of the diluted blood into a new tube containing 15 ml Ficoll and centrifuged for 20 minutes at 750 g. Mononuclear cells were isolated and washed twice with normal saline. After lysis of erythrocytes, the final cell density was adjusted into 1 -2.5 × 10 8 /ml. The cell viability (≥95%) was determined by trypan blue and sterility test. The proportion of CD34 + cells were among 1.5% -2.3% as determined by flow cytometry.
Each patient received two consecutive CBMNC treatments at 1 week interval. Approximately 1 × 10 8 CBMNC were mixed with 100 ml normal saline and administered intravenously with electrocardiograph monitoring in 30 minutes.

Post-treatment follow-up
Clinical remission, clinical response, endoscopic mucosal healing were determined according to the changes of Mayo score and laboratory tests were performed at weeks 0, 8, 16, and 36.
The modified Mayo score was used to evaluate clinical efficacy of these therapies [39]. Clinical response was defined as a decrease of at least 3 points and 30% from baseline (Mayo score), with an accompanying decrease of at least 1 point of subscore for rectal bleeding or absolute subscore for rectal bleeding of 0 or 1. Clinical remission was determined as a total Mayo score ≤ 2 points with no individual subscore > 1 point. An absolute subscore for endoscopy of 0 or 1 was defined for mucosal healing [40].
The laboratory tests include ESR and CRP, whole blood cell analysis, liver and renal function, levels of sodium, chlorine, potassium, serum ALT, creatinine, blood urea nitrogen, and total bilirubin were performed at baseline (pre-treatment) and weeks 8, 16, and 36 after initial treatment. Colonoscopy was performed before treatment and at weeks 8, 16, and 36 during the followup in all of patients to observe lesions of the intestinal mucosa including erythema, vascular texture, tissue fragility, erosion, and hemorrhage. Moreover, any adverse events including neutropenia, hepatotoxicity or headaches were recorded during the follow up periods.

Statistical analysis
The primary end point of this study was clinical remission ratio of these two groups at weeks 8, 16 and 36 after treatment. The secondary end points consisted of clinical response ratio, endoscopic mucosal healing ratio, Mayo score, ESR and CRP levels, and defecation frequency. To decrease confounding bias caused by nonrandomized retrospective assignment, PSM was applied to balance the pretreatment data of two groups. In brief, PSM for all of the patients were evaluated by multiple logistic-regression using variables of sex, age, disease course, disease extent, defecation frequency, ESR, CRP and Mayo score. By using the Nearest-neighbor method, a 1:1 matching study group was created. In the PSM population, qualitative variables were analyzed using Fisher's exact test or Chi-square test. Normal distribution quantitative data were analyzed by means of one-way analysis of variance. Skew distributional data were compared using Mann-Whitney U test. SPSS 22.0 (SPSS, Inc., Chicago, IL, USA) and SPSS Sample Power (version3.0; IBM, USA) were used for statistical analyses. The significance level was defined as P value < 0.05.

Author contributions
Concept and design: X.J., Z.L. and L.C. Collection and assemble of data: L.C., Y.G. and Z.Z. Data analysis and interpretation: L.C., Y.G. W.Y and M.S. Manuscript written: X.J., Z.L. and L.C. Reviewed the manuscript: All authors.

CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.