Safety of a controlled human infection model of tuberculosis with aerosolised, live-attenuated Mycobacterium bovis BCG versus intradermal BCG in BCG-naive adults in the UK: a dose-escalation, randomised, controlled, phase 1 trial

Abstract

Background: Mycobacterium tuberculosis is the main causative agent of tuberculosis. BCG, the only licensed vaccine, provides inadequate protection against pulmonary tuberculosis. Controlled human infection models are useful tools for vaccine development. We aimed to determine a safe dose of aerosol-inhaled live-attenuated Mycobacterium bovis BCG as a surrogate for M tuberculosis infection, then compare the safety and tolerability of infection models established using aerosol-inhaled and intradermally administered BCG.

Methods: This phase 1 controlled human infection trial was conducted at two clinical research facilities in the UK. Healthy, immunocompetent adults aged 18-50 years, who were both M tuberculosis-naive and BCG-naive and had no history of asthma or other respiratory diseases, were eligible for the trial. Participants were initially enrolled into group 1 (receiving the BCG Danish strain); the trial was subsequently paused because of a worldwide shortage of BCG Danish and, after protocol amendment, was restarted using the BCG Bulgaria strain (group 2). After a dose-escalation study, during which participants were sequentially allocated to receive either 1 × 10³, 1 × 10⁴, 1 × 10⁵, 1 × 10⁶, or 1 × 10⁷ colony-forming units (CFU) of aerosol BCG, the maximum tolerated dose was selected for the randomised controlled trial. Participants in this trial were randomly assigned (9:12), by variable block randomisation and using sequentially numbered sealed envelopes, to receive aerosol BCG (1 × 10⁷ CFU) and intradermal saline or intradermal BCG (1 × 10⁶ CFU) and aerosol saline. Participants were masked to treatment allocation until day 14. The primary outcome was to compare the safety of a controlled human infection model based on aerosol-inhaled BCG versus one based on intradermally administered BCG, and the secondary outcome was to evaluate BCG recovery in the airways of participants who received aerosol BCG or skin biopsies of participants who received intradermal BCG. BCG was detected by culture and by PCR. The trial is registered at ClinicalTrials.gov, NCT02709278, and is complete.

Findings: Participants were assessed for eligibility between April 7, 2016, and Sept 29, 2018. For group 1, 15 participants were screened, of whom 13 were enrolled and ten completed the study; for group 2, 60 were screened and 33 enrolled, all of whom completed the study. Doses up to 1 × 10⁷ CFU aerosol-inhaled BCG were sufficiently well tolerated. No significant difference was observed in the frequency of adverse events between aerosol and intradermal groups (median percentage of solicited adverse events per participant, post-aerosol vs post-intradermal BCG: systemic 7% [IQR 2-11] vs 4% [1-13], p=0·62; respiratory 7% [1-19] vs 4% [1-9], p=0·56). More severe systemic adverse events occurred in the 2 weeks after aerosol BCG (15 [12%] of 122 reported systemic adverse events) than after intradermal BCG (one [1%] of 94; difference 11% [95% CI 5-17]; p=0·0013), but no difference was observed in the severity of respiratory adverse events (two [1%] of 144 vs zero [0%] of 97; 1% [-1 to 3]; p=0·52). All adverse events after aerosol BCG resolved spontaneously. One serious adverse event was reported-a participant in group 2 was admitted to hospital to receive analgesia for a pre-existing ovarian cyst, which was deemed unrelated to BCG infection. On day 14, BCG was cultured from bronchoalveolar lavage samples after aerosol infection and from skin biopsy samples after intradermal infection.

Interpretation: This first-in-human aerosol BCG controlled human infection model was sufficiently well tolerated. Further work will evaluate the utility of this model in assessing vaccine efficacy and identifying potential correlates of protection.

Funding: Bill & Melinda Gates Foundation, Wellcome Trust, National Institute for Health Research Oxford Biomedical Research Centre, Thames Valley Clinical Research Network, and TBVAC2020.

Figure 1. Trial profile

(A) Trial profile for group 1 (BCG Danish). (B) Trial profile for group 2 (BCG Bulgaria). *The first participant in group 1C was sequentially enrolled as part of the dose-escalation study and not randomly assigned; this participant was included in the comparison of safety between aerosol BCG and intradermal BCG. †The target for enrolment into groups 1C and 1D was 12 participants per group but was incomplete owing to a shortage of BCG Danish. ‡Group 2D was enrolled as part of the dose-escalation trial but was included in the comparison between aerosol and intradermal BCG.

Figure 2. Adverse events in group 1 (BCG Danish)

Occurrences of respiratory (top) and systemic (bottom) adverse events during the 2 weeks after infection with BCG Danish and before bronchoscopy. Shown as a percentage of total solicited occurrences, for group 1C (n=4; 1 × 10⁵ CFU aerosol BCG) and group 1D (n=3; 1 × 10⁵ CFU intradermal BCG), calculated as the number of participants × 16 timepoints (group 1C adjusted to 15·75 timepoints as one participant had their bronchoscopy on day 13). Adverse events were collected every 12 h for 2 days after infection and then daily.

Figure 3. BCG quantification in group 1 (BCG Danish)

(A) BCG recovered from the vaccine vials, plated onto Middlebrook 7H11 agar. The range of CFU as stated by the manufacturer is represented by the dotted lines (2–8 × 10⁶ CFU). The dots represent individual vials. (B) CFU of BCG per dose, determined from the diluted or concentrated vials after participant infection, plated onto Middlebrook 7H11 agar. Dots represent each reconstituted vial (prepared dose) (participants infected on the same day could have been infected with BCG from the same vial). The middle horizontal solid lines represent the median, and the outer lines show the upper and lower limits of the IQR and range. CFU=colony-forming units.

Figure 4. Adverse events in group 2 (BCG Bulgaria)

(A) Occurrences of respiratory (left) and systemic (right) adverse events during the 2 weeks after infection with BCG Bulgaria and before bronchoscopy. Shown as a percentage of total solicited occurrences, for the aerosol group (n=12, including n=3 from group 2D and n=9 from group 2E [both 1 × 10⁷ CFU]) and the intradermal group (n=12 from group 2F [1 × 10⁶ CFU]), calculated as 12 participants × 16 timepoints. Adverse events were solicited every 12 h for 2 days after infection and then daily. (B) Total (left), respiratory (middle), and systemic (right) solicited adverse events per participant (as a percentage of the total number of solicited adverse events [15 adverse events × 16 collected timepoints], respiratory adverse events [seven adverse events × 16 collected timepoints], and systemic adverse events [eight adverse events × 16 collected timepoints], respectively) in the 2 weeks after infection with aerosol (n=12) or intradermal (n=12) BCG Bulgaria. Adverse events were collected every 12 h for 2 days following infection and then daily. No significant difference between aerosol and intradermal groups was seen by Mann–Whitney U tests. Each dot represents one participant. The middle horizonal lines represent the median, the box boundaries show the IQR, and the outer lines depict the range. CFU=colony-forming units.

Figure 5. BCG quantification in group 2 (BCG Bulgaria)

(A) BCG recovered from the vaccine vials, plated onto Middlebrook 7H11 agar. The range of CFU as stated by the manufacturer is represented by the dotted lines (1·5–6·0 × 10⁶ CFU). Dots represent each reconstituted vial. (B) BCG Bulgaria per vial over an 18-month period. The dotted line represents the lower limit of the range of the stated CFU of BCG Bulgaria per vial (1·5 × 10⁶ CFU). Dots represent each reconstituted vial. Dark blue dots indicate batch 175-2, light blue dots indicate batch 204-1. (C) BCG CFU per dose measured from the diluted or concentrated vials after participant infection, plated onto Middlebrook 7H11 agar. Dots represent each reconstituted vial (prepared dose) (participants infected on the same day could have been infected with BCG from the same prepared dose). (D) Recovery of BCG after nebulisation. After vial reconstitution as per manufacturer instructions, BCG CFU were enumerated before addition to the nebuliser (loaded dose; circles) and after nebulisation (delivered dose; squares) by plating onto Middlebrook 7H11 agar. Two vials were reconstituted and nebulised in a duplicate experiment, in which 46·7% and 47·5% (mean 47·1%) of the BCG was recovered from the original loaded dose; each line represents a separate vial. (E) MGIT culture data from bronchoalveolar lavage samples, presented as time to positivity (h) in participants in the aerosol BCG Bulgaria groups. For dose-escalation groups 2A–C (1 × 10⁴–1 × 10⁶ CFU), the whole bronchoalveolar lavage sample was cultured; for the aerosol comparator groups 2D/E (1 × 10⁷ CFU), bronchoalveolar lavage cells were first removed for immunology and only the fluid was cultured for mycobacterial growth detection. (F) Quantification of BCG in skin biopsy samples taken 14 days after infection in the intradermal group 2F, measured directly by plating the biopsy homogenate onto solid agar. Median CFU recovered 3970 (IQR 1500–7650). Dots represent individual participants. In panels A, C, E and F, the middle horizontal solid lines represent the median; in groups 2A–C, the outer lines show the upper and lower limits of the IQR and range, whereas in groups 2D/E, the box boundaries show the IQR and the outer lines depict the range. CFU=colony-forming units. MGIT=mycobacterial growth indicator tube.