Patients

T1D subjects receiving care through the Section of Endocrinology at the General Hospital of Jinan Military Command (Jinan, Shandong, China) were enrolled in a phase 1/phase 2, open-label clinical trial conducted from October 2010 through January 2011. With oversight from a planning committee, the principal investigator designed the trial and received ethical approval for the clinical treatment protocol and consent form from the General Hospital of Jinan Military Command (Jinan, Shandong, China) and ethical approval for the in vitro study protocol and consent form from the University of Illinois at Chicago Institutional Review Board. Written informed consent was obtained from each participant. The trial was conducted with 15 subjects with established T1D (mean duration: 8.5 ± 6.4 years). Patients were qualified for enrollment if they met the 2010 diagnosis standards of the American Diabetes Association and a blood test confirmed the presence of at least one autoantibody to pancreatic islet β cells. Exclusion criteria included clinically significant liver, kidney, or heart disease; pregnancy; immunosuppressive medication; viral diseases; or diseases associated with immunodeficiency.

Stem Cell Educator design

In previous studies we isolated multipotent cord blood stem cells (CB-SCs) from human cord blood [16]. The CB-SCs display embryonic cell markers (for example, transcription factors OCT-4 and Nanog, stage-specific embryonic antigen (SSEA)-3 and SSEA-4) and leukocyte common antigen CD45, but they are negative for blood cell lineage markers [9, 16]. We identified a hydrophobic material from FDA-approved (USP Class VI) Petri dishes that tightly binds CB-SCs without interfering with their immune modulating capability. We designed a chamber for co-culture of lymphocytes and CB-SCs that includes nine discs of the material with adherent CB-SCs sandwiched between a top cover plate and a bottom collecting plate (Figure 1). The device was manufactured in a Class 100 K clean room and gamma-irradiated prior to introducing CB-SCs [16]. In the Stem Cell Educator, lymphocytes separated from a patient's blood are slowly passed through the stacked discs of material with adherent CB-SCs and lymphocytes collected through a hole in the bottom plate are returned to the patient. The materials used to produce the device are approved for in vivo use according to the United States Pharmacopeia (that is, Grade Class VI Plastic).

CB-SC culture

Human cord blood units derived from healthy donors were purchased from the Maternal and Child Health Hospital (Jinan, Shandong, China). All cord blood samples were screened for alanine aminotransferase and pathogenic antigen antibodies (including anti-HCV, anti-HBsAg, anti-HIV, and anti-syphilis Abs), and only pathogen-free cord blood units were used for isolating CB-SCs. Human cord blood-derived stem cells (CB-SC) were generated as previously described with the following modifications [14, 16]. Cord blood mononuclear cells were plated in serum-free culture medium (Lonza, Walkersville, MD) and incubated at 37°C, in 8% CO₂. After 2 to 3 weeks, CB-SCs growing at 80% to 90% confluence were prepared for clinical trial. Endotoxin level was < 0.05 EU/ml.

Treatment and follow-up

Twelve participants received a single treatment with the Stem Cell Educator (Tianhe Stem Cell Biotechnology^®, Jinan, China), and three received a single treatment with the Stem Cell Educator without adherent CB-SCs (that is, sham or process-only control) (Figure 1). A 16-gauge IV needle was placed in the left (or right) median cubital vein, and the patient's blood was passed through a Blood Cell Separator MCS+ (Haemonetics^®, Braintree, MA, US) at 35 mL/min for 6 to 7 hours to isolate lymphocytes in accordance with the manufacturer's recommended protocol. The collected lymphocytes were transferred into the device for exposure to allogeneic CB-SCs (or process control without CB-SCs), and other blood components were returned to the patient. After 2 to 3 hours in the device, lymphocytes were returned to the patient's circulation via a dorsal vein in the hand under gravity flow control (2 to 3 mL/min) with physiological saline. Approximately 10,000 mL of blood was processed during the procedure resulting in approximately two repeated educations for the lymphocyte fraction. Patients were hospitalized for two days to monitor temperature and conduct routine laboratory blood tests for adverse reactions following treatment. Follow-up visits were scheduled 4, 12, 24, and 40 weeks after treatment for clinical assessments and laboratory tests (Additional file 1).

Study end points

The primary study end points were: 1) feasibility of the Stem Cell Educator therapy; 2) safety of the therapy through 12 weeks post treatment; and 3) preliminary evaluation of the efficacy of the therapy for improving β cell function through 24 weeks. Pancreatic islet β cell function was assessed by measuring basal and glucose-stimulated C-peptide production over time, as described elsewhere [17, 18]. Metabolic control was monitored throughout the study. The secondary study end point was evidence of the efficacy of the therapy in modulating autoimmunity. Baseline blood samples were collected prior to Stem Cell Educator therapy. Detailed descriptions of the methods are included in the Supplementary Appendix.

Statistics

An intention-to treat approach was used, with 12 of 15 patients undergoing Stem Cell Educator therapy and the remaining 3 patients undergoing sham therapy without CB-SCs in the Educator. All patients were included in the safety analyses. The primary efficacy end point was the change in C-peptide secretion between baseline and follow-up.

Feasibility and safety of Stem Cell Educator therapy

No participants experienced any significant adverse events during the course of treatment. Most patients experienced mild discomfort during venipuncture and some soreness of the arm during apheresis, but discomfort and soreness resolved quickly following the conclusion of the procedure. Twenty-four hours post treatment, no significant difference was noted in white blood cell counts relative to baseline (total white blood cell count: 6.95 × 10⁹/L ± 1.98 versus 6.39 × 10⁹/L ± 1.72, P = 0.38; granulocytes: 3.79 × 10⁹/L ± 1.43 versus 3.66 × 10⁹/L ± 1.05, P = 0.77; lymphocytes: 2.31 × 10⁹/L ± 0.9 versus 2.08 × 10⁹/L ± 0.67, P = 0.40; monocytes: 0.49 × 10⁹/L ± 0.13 versus 0.46 × 10⁹/L ± 0.10, P = 0.48). Participants' body temperatures were not significantly changed during the two-day post-treatment observation (36.44°C ± 0.24 versus 36.5°C ± 0.22, n = 15, P = 0.35). No changes were observed in blood cell count or temperature at the 12-week follow-up.

CB-SCs are tightly adherent [9, 16] and not expected to escape the device. To confirm CB-SCs are completely retained in the Educator and not transferred to the patient, we examined cells leaving the device to check for SSEA-3, a CB-SC-specific marker. Flow cytometry confirmed the absence of SSEA-3 in cells leaving the Educator (Additional file 1: Figure S1). These data indicate that the cells returned to the patients are autologous. Additionally, HLA matching is not required prior to Stem Cell Educator therapy because CB-SCs are not transferred to the patient and because CB-SCs have very low immunogenicity [9, 13, 16]. Thus, Stem Cell Educator therapy is a very safe approach.

Efficacy outcomes in improving beta cell function

Participants in Group A (that is, those with moderate T1D and some residual β cell function) exhibited improved fasting C-peptide levels at 12 and 24 weeks post-treatment (Figure 2A and 2B, Table 2), and participants in Group B (that is, those with severe T1D and no residual pancreatic islet β cell function) exhibited successive improvement in fasting C-peptide levels at each follow-up (Figure 2A and 2C, Table 2). C-peptide response following a 75-g oral glucose tolerance test (OGTT) improved among Group A participants at 4 and 12 weeks (Figure 2B). Notably, participants in Group B exhibited essentially no C-peptide production following glucose challenge at baseline (that is, less than the minimum sensitivity of 0.01 ng/ml at all time points) but demonstrated marked improvement at 12 weeks (Figure 2C, Table 2). Improvement was maintained through the final follow-up (that is, 40 weeks post-treatment; P = 0.026) (Figure 2C). Participants in the Control Group did not exhibit significant change at any follow-up (Figure 2A, Table 2).

Consistent with improved β cell function, the median daily dose of insulin was reduced 38% at 12 weeks post-treatment in Group A (36 ± 13.2 units/day at baseline versus 22 ± 1.8 units/day 12 weeks post-treatment) and 25% in Group B (48 ± 7.4 units/day at baseline versus 36 ± 4.4 units/day 12 weeks post- treatment), but no change was observed in the Control Group. The reduced daily dose of insulin in Group A and B was maintained through the last follow-up for this measure (24 weeks). The median glycated hemoglobin (HbA₁C) in Group A was significantly lowered from 8.73% ± 2.49 at baseline to 7.67% ± 1.03 at 4 weeks post-treatment (P = 0.036) and to 6.82% ± 0.49 at 12 weeks post-treatment (P = 0.019). The median HbA₁C in Group B was reduced 1.68% ± 0.42 at 12 weeks post-treatment, but no change was observed in the Control Group (9.0% ± 2.3 at baseline versus 8.7% ± 1.9 at 12 weeks post-treatment, P = 0.86). Thus, the ex vivo immune education of CB-SC leads to regeneration of islet β cells and improvement of β cell function in long-standing T1D subjects.

Efficacy outcomes in autoimmune control

Next, we explored mechanisms underlying CB-SC-mediated immune modulation. Regulatory T lymphocytes (Tregs) play a crucial role in maintaining homeostasis and self-tolerance by inhibiting the action of autoreactive effector T cells [14, 19, 20], but previous attempts to manipulate Tregs for clinical applications have been problematic [21]. We measured changes in the percentage of CD4⁺CD25⁺Foxp3⁺ Tregs in peripheral blood of participants following Stem Cell Educator treatment. The percentage of Tregs in peripheral blood of participants was significantly increased 4 weeks after Stem Cell Educator therapy (Figure 3A), whereas the percentage of Tregs in peripheral blood of participants receiving sham therapy was unchanged from baseline (Figure 3A). TGF-β1 has also been implicated in Treg-mediated immune suppression [22] as well as in maintenance of self-tolerance in T1D animal models subjected to stem cell-mediated immune modulation [9, 15, 23]. We examined TGF-β1 and IL-10 expression to explore whether these pathways are activated following Stem Cell Educator therapy. Participants in the treatment group exhibited significant increases in plasma level of TGF-β1 at the 4-week follow-up (P = 0.001, Figure 3B) but did not exhibit changes in the plasma level of IL-10 (P = 0.44, Figure 3B). Both TGF-β1 and IL-10 failed to show changes in the control group.

We also examined levels of CD28 [24–28] and inducible costimulator (ICOS) [29, 30], which are essential for the establishment, maintenance and efficacy of Tregs [24–32]. Flow cytometry revealed an increase in CD28 and ICOS in lymphocytes 4 weeks after Stem Cell Educator therapy (Figure 3C), but levels of both molecules were unchanged in participants receiving sham therapy (Figure 3C). We also noted other changes at the 4-week follow-up consistent with improved helper T cell 1 (Th1) and Th2-mediated immune function (Figure 3D). Expression of IL-4 and IL-12 was significantly increased (P = 0.016 and P = 0.0093, respectively) and expression of IL-5 and IL-13 was decreased (P = 0.00039 and P = 0.00206, respectively). The production of pro-inflammatory IL-17A was also decreased 4 weeks after treatment (Figure 3D, P = 0.0043). No changes were observed in levels of these cytokines in participants who received sham therapy (Additional file 1: Figure S2).

Autoimmune regulator (Aire), usually expressed in thymic medullary epithelial cells, plays an important role in immune tolerance by mediating ectopic expression of peripheral self-antigens and mediating the deletion of auto-reactive T cells [33, 34]. We found that CB-SCs express Aire (Figures 4A and 4B). To determine the Aire function in CB-SC, we used three pairs of human Aire-specific small interfering RNAs (siRNA) to knock down Aire expression in CB-SCs. Western blots confirmed knockdown of Aire protein expression (Figure 4C) and a corresponding reduction in the expression of programmed death ligand-1 (PD-L1) that contributes to the immune modulation of CB-SC [13, 35] (Figure 4D). Knockdown of Aire also reduced the percentage of Tregs in the co-cultured lymphocyte population (P = 0.028) (Figure 4E). The data indicate that Aire is involved in immune modulation and induction of immune tolerance following Stem Cell Educator therapy.