Global description of clinical trial

This trial is a first-in-human, phase 1/2, open-label study on safety, tolerability and efficacy of VC-02 in patients with T1D and hypoglycemia unawareness. VC-02 is a combination product of PEC-01 cells loaded into a delivery device. The study is conducted at nine centers in North America and one center in Belgium (ClinicalTrials.gov identifier: NCT03163511). This paper is an interim report on one cadre of 10 patients in the larger study. The patients were treated at City of Hope; the University of British Columbia; the University of California, Davis; the University of Minnesota; and Vrije Universiteit Brussel. A detailed description of the clinical protocol is provided as a Supplementary Protocol.

Inclusion criteria included men and non-pregnant women, a diagnosis of T1D for a minimum of 5 years, hypoglycemia unawareness (Clarke score) or significant glycemic lability, a stable diabetes treatment regimen, willingness to use a CGM device and being an acceptable candidate for surgical implantation. A Clarke score ≥4 confirmed patients qualifying for study participation on the basis of hypoglycemia unawareness. Exclusion criteria included history of islet cell, kidney and/or pancreas transplant; six or more severe unexplained hypoglycemic events within 6 months of enrollment; uncontrolled or untreated thyroid disease or adrenal insufficiency; diabetic complications, such as severe kidney disease or renal dysfunction, proliferative retinopathy, foot ulcers, amputations and/or severe peripheral neuropathy; or detectable stimulated serum C-peptide during the screening period defined as ≥0.07 nmol l−1.

After establishing a safety profile of the product candidate in cohort 1, product efficacy was evaluated in cohort 2. Within cohort 2, patients were enrolled in cadres (groups of patients ranging from n = 4 to n = 11) for testing a particular device configuration and/or implant strategy to improve product engraftment and cell survival outcomes. All device configurations contained the same materials but differed in the application of the pores. Differences in implant strategy involved pharmacological interventions, number of dose-finding units implanted and/or implant sites. The patients described in refs. 17,18 had been enrolled in earlier study groups of cohort 2. Those described in the present report participated in the most recent one (enrollment between August 2020 and October 2021); they received devices with the same materials as the devices in previous study groups but in twofold to threefold higher number and with a different application of perforations (described in patent application US16/347,790). The protocol specified the primary efficacy endpoint as change from baseline to week 26 in plasma C-peptide after MMTT and the following secondary efficacy endpoints for follow-up to maximally 104 weeks:

Change from baseline of C-peptide response to MMTT and percent of patients achieving levels >0.07 nmol l−1;

Change from baseline in average insulin dose and percent of patients with 50% reduction and percent with insulin independence;

Percent of time with blood glucose values <54 mg dl−1, 54 mg dl−1 to <70 mg dl−1, 70 mg dl−1 to ≤180 mg dl−1 and >180 mg dl−1 (CGM device) and change from baseline in time-in-hypoglycemic range (<70 mg dl−1), time-in-euglycemic range (70–180 mg/dl−1) and time-in-hyperglycemic range (>180 mg dl−1);

Frequency of hypoglycemic events (HEs) and percent of patients free of HE.

Clinical protocol

Patients fulfilled entry criteria including hypoglycemia unawareness, plasma C-peptide <0.07 nmol l−1 during MMTT, HbA1c ≤10% and signed informed consent. The clinical protocol used immune suppression with anti-thymocyte globulin (ATG) for induction, and tacrolimus and mycophenolate mofetil for maintenance as in refs. 17,18. ATG dose was adjusted to lymphocyte counts. Two days after initiation of ATG, devices were placed subcutaneously in the abdominal wall and flanks, under general anesthesia. No adverse events were recorded during and after the interventions.

As required by the study protocol, all CGM data were acquired by a Dexcom G6 CGM sensor. Glycemic control and insulin administration were guided by a diabetes educator who was independent of the study team and unaware of plasma C-peptide levels. Regular hospital visits were scheduled for medical follow-up, including systematic survey for adverse events, collection of laboratory data and adjustment of immune suppression dosing when needed. The CGM Medtronic automodus was stopped at least 4 h before start of MMTT, and the patient was switched to a fixed insulin dose that was maintained during the test. The mixed meal consisted of Boost Hi-Protein (volume, 360 ml) and was ingested within 15 min. No insulin bolus was given before the meal.

An MMTT was carried out 4 weeks before transplant and at 3-month post-transplant intervals. Blood was collected every 30 min for measurement of glycemia and C-peptide, with the purpose of identifying recipients who increased plasma C-peptide between baseline and month 6 (primary efficacy endpoint) and those who achieved a level >0.07 nmol l−1 (secondary efficacy endpoint) during a 12-month follow-up. The 90-min timepoint in the MMTT was selected for this assessment, as conventionally done; concurrent glycemia was >250 mg dl−1 for all patients and, thus, considered as a similar glucose-stimulatory condition for all. Glucose responsiveness was evaluated by comparing 0-min and 90-min C-peptide values. For the endpoint of case 3 (month 12), this glycemia criterion was reached only at minute 120, thus providing the listed data. Follow-up was extended to 360 min for the month 9 and month 12 MMTT of the case with highest C-peptide response at month 6 to follow release of both C-peptide and proinsulin during prolonged glucose stimulation.

Glucose, C-peptide and HbA1c levels were analyzed by ACM Global Laboratories using hexokinase, chemiluminescent and ion exchange high-performance liquid chromatography assays, respectively. For one point (case 3, month 9), the CRO value for C-peptide was undetectable, differing with the local hospital value (0.10 nmol l−1) as well as with the prior and following CRO values in this patient, such that we think that it is erroneous and can be replaced by the locally measured value. Proinsulin was measured only in the MMTT samples of case 1, month 12; values were obtained by digital ELISA technology (SIMOA HD1, Quanterix). As in previous protocols, start of immune suppression caused a transient decrease in blood cell counts and HbA1c level.

Implantation of PEC-Direct

The PEC-Direct combination product was manufactured by ViaCyte as perforated device units containing PEC-01 that had been differentiated from the CyT49 pluripotent stem cell line6,16. The cells were loaded in dose-finding and in sentinel units where they were retained between polytetrafluoroethylene membranes with pores in controlled distribution, number and diameter. The density of pores is identical for both devices, as is thickness; therefore, engraftment, diffusion and perfusion are expected to be similar. The scaling of the dose-finding device is 12× for area containing cells; they are estimated to contain 75 × 106 PEC-01 cells versus 7 × 106 PEC-01 cells in the sentinels. The sentinel devices were introduced in the same implant site, on the opposite end of an incision used to place the dose-finding device. These units were prepared at ViaCyte (ref. 16 and Supplementary Information) and shipped to the implant centers within a container that maintains 15–25 °C for up to 5 d.

Histology

Retrieved sentinels with surrounding tissue were fixed in formaldehyde before embedding in paraffin and sectioning (5 µm). For histology, sections were stained with hematoxylin and eosin. To distinguish male donor cells from female recipient cells (case 1), the RNAscope 2.5 HD Assay RED (cat. no. 322350, Advanced Cell Diagnostics) was used according to the manufacturer’s instructions (15 min, Target Retrieval; 30 min, Protease Plus). The probe used was Homo sapiens KDM5D (cat. no. 558161), a gene located on the Y chromosome. Because the CyT49 cell line is of male origin and the recipient is female, KDM5D-positive (red) signals distinguish donor from recipient cells. RNA in situ hybridization was performed at the VSTA facility of Vrije Universiteit Brussel https://vsta.research.vub.be.

For immunohistochemical analysis, sections underwent heat-induced antigen retrieval (2100-Retriever, Aptum Biologics) in citrate buffer (ScyTek Laboratories) before staining with guinea pig anti-insulin and rabbit anti-glucagon (each 1:1,000, in-house produced), mouse anti-glucagon (1:500, g2654, Sigma-Aldrich), rat anti-somatostatin (1:100, ab30788, Abcam), mouse anti-CK19 (1:20, M0888, Agilent), rabbit anti-wide spectrum CK (1:200, z0622, Agilent), mouse anti-CHRA (1:500, Ma5-13287, Thermo Fisher Scientific) and rabbit anti-CD34 (1:150, ab81289, Abcam). Alexa Fluor-conjugated F(ab′)2 fragments of affinity-purified antibodies were used as secondary antibodies allowing multiple labeling (1:500, Jackson ImmunoResearch). Nuclei were stained by DAPI (Sigma-Aldrich) added to the mounting medium (Agilent). Digital images were acquired on an Axioplan 2 microscope (Carl Zeiss) with an Orca-R2 camera (Hamamatsu Photonics) and SmartCapture 3 software (DSUK Ltd.) or acquired with an Aperio CS2 (Leica Biosystems) and visualized with Pathomation software. Morphometry was conducted on .tiff pictures captured with ImageXpress Pico (Molecular Devices) and semi-automatically analyzed with IPlab software (Becton Dickinson).

The composition of the tissue in the inner chamber was determined as described in ref. 40. In previous work, we determined the number of evenly distributed sections that are representative for this type of implant. This number was stained for DAPI (marker for nuclear mass) and for insulin, glucagon and CHRA and CK (cytoplasmic pancreatic cell markers). For each staining, the positive area in a section was semi-automatically quantified and extrapolated to a total volume for the specified marker according to the Cavalieri method: the value was multiplied by the distance to the next section, which yields a volume for that region, and all region volumes were added to obtain the total volume for the implant. Nuclear mass was also expressed as percent of that at start (2.5 µl) as well as translated to total cellular mass by applying a conversion factor that was previously defined in stem cell preparations stained for DAPI and actin. This total cellular mass consists of the mass of infiltrated recipient cells and that of the pancreatic cell types of donor origin. Insulin-positive and glucagon-positive cells were CHRA positive. The sum of CHRA-positive mass and CK-positive mass was taken as mass of donor origin; its percentage over total cell mass allowed calculation of the recovery of initial cell mass as well as of the proportion of infiltrating recipient cells (Table 2).

The presence and phenotype of lymphocytes were identified by immunohistochemical staining for CD4 (SP35), CD8 (SP57), CD20 (L26) and CD68 (KP1). Slides were scanned with an Aperio GT450 at ×40 magnification with visualization with Pathomation software.

Ethics statement

All patients signed an informed consent, which also stated the existence of other treatment forms to be explained by the study physician before taking a decision. The consent included an agreement on collection and use of study data for research purposes and in connection with scientific and medical publications. All study sites were approved by the local institutional review board associated with their institution. The trial was conducted in compliance with all applicable regulations.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.