Background A gluten-free diet is the only means to manage coeliac disease, a permanent immune intolerance to gluten. We developed a therapeutic vaccine, Nexvax2, designed to treat coeliac disease. Nexvax2 is an adjuvant-free mix of three peptides that include immunodominant epitopes for gluten-specific CD4-positive T cells. The vaccine is intended to engage and render gluten-specific CD4-positive T cells unresponsive to further antigenic stimulation. We assessed the safety and pharmacodynamics of the vaccine in patients with coeliac disease on a gluten-free diet. Methods We did two randomised, double-blind, placebo-controlled, phase 1 studies at 12 community sites in Australia, New Zealand, and the USA, in HLA-DQ2·5-positive patients aged 18–70 years who had coeliac disease and were on a gluten-free diet. In the screening period for ascending dose cohorts, participants were randomly assigned (1:1) by central randomisation with a simple block method to a double-blind crossover, placebo-controlled oral gluten challenge. Participants with a negative interferon γ release assay to Nexvax2 peptides after the screening oral gluten challenge were discontinued before dosing. For the biopsy cohorts, the screening period included an endoscopy, and participants with duodenal histology who had a Marsh score of greater than 1 were discontinued before dosing. Participants were subsequently randomly assigned to either Nexvax2 or placebo in ascending dose cohorts (2:1) and in biopsy cohorts (1:1) by central randomisation with a simple block method. In the three-dose study, participants received either Nexvax2 60 μg, 90 μg, or 150 μg weekly, or placebo over 15 days; in a fourth biopsy cohort, patients received either Nexvax2 at the maximum tolerated dose (MTD) or placebo. In the 16-dose study, participants received Nexvax2 150 μg or 300 μg or placebo twice weekly over 53 days; in a third biopsy cohort, patients also received either Nexvax2 at the MTD or placebo. In the 4-week post-treatment period, ascending dose cohorts underwent a further double-blind crossover, placebo-controlled oral gluten challenge, which had a fixed sequence, and biopsy cohorts had a gastroscopy with duodenal biopsies and quantitative histology within 2 weeks without oral gluten challenge. Participants, investigators, and study staff were masked to the treatment assignment, except for the study pharmacist. The primary endpoint was the number and percentage of adverse events in the treatment period in an intention-to-treat analysis. Both trials were completed and closed before data analysis. Trials were registered with the Australian New Zealand Clinical Trials Registry, numbers ACTRN12612000355875 and ACTRN12613001331729. Findings Participants were enrolled from Nov 28, 2012, to Aug 14, 2014, in the three-dose study, and from Aug 3, 2012, to Sept 10, 2013, in the 16-dose study. Overall, 62 (57%) of 108 participants were randomly assigned after oral gluten challenge and 20 (71%) of 28 participants were randomly assigned after endoscopy. In the three-dose study, nine participants were randomly allocated to Nexvax2 60 μg and three to placebo (first cohort), nine were allocated to Nexvax2 90 μg and four to placebo (second cohort), eight were allocated to Nexvax2 150 μg and four to placebo (third cohort), and three were allocated to Nexvax2 150 μg and three to placebo (biopsy cohort). In the 16-dose study, eight participants were randomly allocated to Nexvax2 150 μg and four to placebo (first cohort), ten were allocated to Nexvax2 300 μg and three to placebo (second cohort), and seven were allocated to Nexvax2 150 μg and seven to placebo (biopsy cohort). The MTD for Nexvax2 was 150 μg because of transient, acute gastrointestinal adverse events with onset 2–5 h after initial doses of the vaccine, similar to those caused by gluten ingestion. In the ascending dose cohorts in the three-dose study, six (55%) of 11 placebo recipients, five (56%) of nine who received Nexvax2 60 μg, seven (78%) of nine who received Nexvax2 90 μg, and five (63%) of eight who received Nexvax2 150 μg had at least one treatment-emergent adverse event, as did all three (100%) placebo recipients and one (33%) of three Nexvax2 150 μg recipients in the biopsy cohort. In the ascending dose cohorts of the 16-dose study, five (71%) of seven placebo-treated participants, six (75%) of eight who received Nexvax2 150 μg, and all ten (100%) who received Nexvax2 300 μg had at least one treatment-emergent adverse event, as did six (86%) of seven placebo recipients and five (71%) of seven Nexvax2 150 μg recipients in the biopsy cohort. Vomiting, nausea, and headache were the only treatment-emergent adverse events that occurred in at least 5% of participants in either study. Among participants given the MTD, eight gastrointestinal treatment-emergent adverse events occurred in four (50%) of eight participants in the third cohort and none (0%) of three participants in the biopsy cohort in the three-dose study, and five events occurred in five (63%) of eight participants in the first cohort and three events in two (29%) of seven participants in the biopsy cohort of the 16-dose study. Median villous height to crypt depth ratio in distal duodenal biopsies was not significantly different between those who received the vaccine at the MTD on either schedule and those who received placebo. Of the participants who completed the post-treatment oral gluten challenge per protocol, interferon γ release assay to Nexvax2 peptides was negative (responders to treatment) in two (22%) of nine placebo-treated participants in the three-dose study versus two (33%) of six who received Nexvax2 60 μg, five (63%) of eight who received Nexvax2 90 μg, and six (100%) of six who received Nexvax2 150 μg (p=0·007); in the 16-dose study, none (0%) of five placebo-treated participants had a negative assay versus six (75%) of eight who received Nexvax2 150 μg (p=0·021). Interpretation The MTD of Nexvax2 was 150 μg for twice weekly intradermal administration over 8 weeks, which modified immune responsiveness to Nexvax2 peptides without deterioration in duodenal histology. The gastrointestinal symptoms that followed the first intradermal administration of the vaccine resembled those associated with oral gluten challenge. These findings support continued clinical development of this potential therapeutic vaccine for coeliac disease. Funding ImmusanT.