A first-in-human, open-label Phase 1b and a randomised, double-blind Phase 2a clinical trial in recent-onset type 1 diabetes with AG019 as monotherapy and in combination with teplizumab

Aims/hypothesis We hypothesised that islet beta cell antigen presentation in the gut along with a tolerising cytokine would lead to antigen-specific tolerance in type 1 diabetes. We evaluated this in a parallel open-label Phase 1b study using oral AG019, food-grade Lactococcus lactis bacteria genetically modified to express human proinsulin and human IL-10, as a monotherapy and in a parallel, randomised, double-blind Phase 2a study using AG019 in combination with teplizumab. Methods Adults (18–42 years) and adolescents (12–17 years) with type 1 diabetes diagnosed within 150 days were enrolled, with documented evidence of at least one autoantibody and a stimulated peak C-peptide level >0.2 nmol/l. Participants were allocated to interventions using interactive response technology. We treated 42 people aged 12–42 years with recent-onset type 1 diabetes, 24 with Phase 1b monotherapy (open-label) and 18 with Phase 2a combination therapy. In the Phase 2a study, after treatment of the first two open-label participants, all people involved were blinded to group assignment, except for the Data Safety Monitoring Board members and the unblinded statistician. The primary endpoint was safety and tolerability based on the incidence of treatment-emergent adverse events, collected up to 6 months post treatment initiation. The secondary endpoints were pharmacokinetics, based on AG019 detection in blood and faeces, and pharmacodynamic activity. Metabolic and immune endpoints included stimulated C-peptide levels during a mixed meal tolerance test, HbA1c levels, insulin use, and antigen-specific CD4+ and CD8+ T cell responses using an activation-induced marker assay and pooled tetramers, respectively. Results Data from 24 Phase 1b participants and 18 Phase 2a participants were analysed. No serious adverse events were reported and none of the participants discontinued AG019 due to treatment-emergent adverse events. No systemic exposure to AG019 bacteria, proinsulin or human IL-10 was demonstrated. In AG019 monotherapy-treated adults, metabolic variables were stabilised up to 6 months (C-peptide, insulin use) or 12 months (HbA1c) post treatment initiation. In participants treated with AG019/teplizumab combination therapy, all measured metabolic variables stabilised or improved up to 12 months and CD8+ T cells with a partially exhausted phenotype were significantly increased at 6 months. Circulating preproinsulin-specific CD4+ and CD8+ T cells were detected before and after treatment, with a reduction in the frequency of preproinsulin-specific CD8+ T cells after treatment with monotherapy or combination therapy. Conclusions/interpretation Oral delivery of AG019 was well tolerated and safe as monotherapy and in combination with teplizumab. AG019 was not shown to interfere with the safety profile of teplizumab and may have additional biological effects, including changes in preproinsulin-specific T cells. These preliminary data support continuing studies with this agent alone and in combination with teplizumab or other systemic immunotherapies in type 1 diabetes. Trial registration ClinicalTrials.gov NCT03751007, EudraCT 2017-002871-24 Funding This study was funded by Precigen ActoBio Graphical Abstract Supplementary Information The online version contains supplementary material available at 10.1007/s00125-023-06014-2.


Interim analyses
Following interim analyses were performed: • After the last Day 180 follow-up visit of the last participant enrolled into the Phase 1b part of the study, a preliminary analysis of the data collected up to 6 months was performed for these participants.
• Once all participants enrolled in the adult combination therapy cohort, as well as the 2 participants enrolled into the open-label part of the adolescent combination therapy cohort, had completed the Day 180 follow-up visit, unblinding was done and an interim analysis was performed.
• After the last Day 180 follow-up visit of the last patient participant into the adolescent combination therapy cohort, all participants in this cohort were unblinded and an interim analysis was performed on all, which included an analysis of the primary endpoint of the study.
All decisions on data for blinded participants were taken before breaking the blind for the third interim analysis.Since this is a study of descriptive nature, with no statistical tests, the influence of the interim analyses was considered negligible.

Randomization and blinding
All participants were registered in and randomization was performed through the Interactive Response Technology.Blinding of participants and all study personnel at the site was achieved using identical packaging for both active and placebo treatments.

Lab assessments
The lab assessments have been performed by Eurofins BioPharma Services using validated methods.Glucose was measured using a Roche Glucose kit based on hexokinase enzymatic reaction (spectrophotometry); intra CV ranged from 0.8-1.7% and inter CV ranged from 2.9-4.7%.
HbA1c was measured on a BioRad Variant II Classic system using a dedicated BioRad HbA1c kit.
The methodology was based on HPLC Ion Exchange (chromatography) and CV ranged from 0.4-3.1% (intra-CV) and from 1.4-4.8%(inter-CV).C-Peptide was measured on Siemens Centaur XP system using dedicated Siemens C-Peptide kit.The method was based on immunologic reaction followed by chemiluminescence detection.The intra CV ranged from 3.4-3.9%and inter CV was 5.7%.

Analysis of antigen specific CD4 + T cells in the circulation
PBMCs were plated (5 x 10 6 cells/well) in the presence of anti-CD40 blocking antibody and MHC Class II optimized peptides from preproinsulin (PPI; for a full list of peptides, see ESM Table 1).
As a positive control for the assay, PBMCs were stimulated with a pool of 176 peptides from viral and bacterial antigens relevant to humans (CEFX, IDT Peptides).DMSO was used as a negative control.Cells were stimulated with peptides or DMSO for 18h at 37°C to allow for CD154 and CD137 upregulation on activated cells.The following day, stimulated cells were harvested, stained with an CD154-PE antibody and anti-PE coupled-magnetic beads were added.CD154 + cells were enriched over a magnetic column, surfaced stained with antibodies for identification of CD4 + conventional T cells (Tconv) and type 1 (inducible) regulatory T cells (Tr1 cells), and were fixed and permeabilized for intracellular cytokine staining.The column flow-through (containing CD154-cells) was then stained with a CD137-PE-Cy7 antibody, anti-PE magnetic beads were added, and cells were enriched over a second magnetic column.The eluted cells were surface stained with antibodies to identify CD4 + Tregs and were fixed and permeabilized for cytokine and transcription factor staining (for a full list of antibodies, see ESM Table 2).Cells were analysed on a CYTEK Aurora spectral cytometer, data unmixed, and analysed using FlowJo v10.7.1.
Memory antigen-specific Treg and Tconv cells were gated as CD45RA -CD45RO + .To calculate the absolute number of antigen-reactive cells per million CD4 T cells, an aliquot (1/50) of cells was taken before the harvested cells were enriched for CD154 and CD137 expression.The frequency was calculated as: (#Ag stimulated enriched cells x 10 6 )/(#CD4 T cells in pre-sample x dilution factor).
Microbial antigen and DMSO control data were determined in parallel to the PPI-specific CD4 + T-cell analysis.The microbial antigen control was used as a positive control for the assay and as a non-islet T cell response in relation to PPI T cell frequency and phenotype.The DMSO negative control was used to set the flow cytometry gate for T cells responding to peptides, i.e., responses above the DMSO background.

Analysis of antigen specific CD8 + T cells in the circulation
Thawed cryopreserved PBMC samples (2 x 10 6 cells) from each time point for a single subject were individually stained with unique metal isotope labelled CD45 antibodies.All time points for a single subject were then combined, washed twice with PBS and stained for viability using cisplatin, followed by quenching and washing with protein-containing media.Cells were then pretreated with dasatinib and washed prior to staining with 50 μL solution containing 1 μL of each Class I Tetramer (Tmr) pool in running buffer for 15 minutes at 37°C.Without washing, 50 μL of a frozen surface antibody cocktail was added, and the sample was incubated for an additional 30 minutes at 4°C.Samples were washed, fixed, and stained with a frozen intracellular antibody cocktail.Subsequently, samples were washed, fixed, and stored at 4°C overnight or for up to 1 week prior to acquisition.On the day of acquisition, cells were washed and resuspended in cold ultrapure water containing 1/5th EQ Four Element Calibration Beads by volume and acquired on a Helios CyTOF mass cytometer with a target cell acquisition of 1,000,000 live events at a rate of 500 events/second.Files were converted to .FCS and then randomized and normalized for EQ bead intensity using the CyTOF Software.FlowJo software v10.7.1 was used to export .FCS files and to gate CD8 + T cell and Tmr + populations.
In a separate analysis, CD8 + T cells (both total and PPI-specific) were manually gated in FlowJo for EOMES and TIGIT staining, identifying a population of partially exhausted EOMES + TIGIT + CD8 + T-cells.
Changes in the frequency of (antigen-specific) CD8 + T cells are reported as the log of the ratio of the frequency at a time point (month 3 or 6) to the frequency at month 0 (fold-change, FC).For this analysis, adult and adolescent patients were grouped and all available patients from the PD-PP population were included.One patient (an adult from the combination therapy group) had no HLA compatible with the PPI Tmr, and the frequency of Tmr + cells was below LOQ at all time points.Thus, this patient was not reported.Two patients from the combination therapy group were assayed at 2 months; these data were incorporated into the 3-month time point results.

ESM Table 1. Peptides used for antigen specific T cell detection.
NA: Not applicable, CMV: cytomegalovirus, EBV: Epstein-Barr virus, Tmr: tetramer, AIM: Activation-Induced Marker Tmr specificities are labelled with two metals, but not a third Tmr metal and pooled for staining and analyses.For CyTOF: Unlabeled purified antibodies were conjugated to metal isotopes using Maxpar X8 and MCP9 Antibody Labeling Kits (Standard BioTools) as per manufacturer's instructions.All samples from the same subject were barcoded and combined for staining.Tmr gating was applied to all samples in the same subject, guided by subjects that were negative for TMr HLA's (Tmr+<0.03%,LOQ).Double labelled cells were counted, single and triple labelled cells were excluded.

C
CyTOF Boolean Tmr gating