Helen McShane and colleagues reply to Deborah CohenBMJ 2018; 360 doi: https://doi.org/10.1136/bmj.k236 (Published 26 January 2018) Cite this as: BMJ 2018;360:k236
All rapid responses
I read with interest, but also sadness, the article by Deborah Cohen on the failure of the MVA85A TB vaccine candidate clinical trial. I believe we need to move on.
McShane’s primary detractor, Professor Beverley, has argued that the animal model data obtained using MVA85A was invalid and hence a clinical trial should never have gone forward. I do not claim to know Beverley’s background, but I would respectfully suggest that it might not equal my own 37-years of experience with animal models of TB, something I have devoted my career to. To my sight, MVA85A was never a “rush job”. It was carefully and systemically evaluated over several years in a series of well-designed studies in small animal models. The results of these studies in mice indicated that MVA85A induced a protective T cell response of the TH1 variety, establishing a state of acquired specific resistance to subsequent challenge. This of course is hardly surprising, given that Ag85 is an immunodominant antigen of the bacillus, and a component of several current vaccine candidates. Accordingly, I find it difficult to understand why Prof. Beverley finds this unconvincing, or for that matter why his fellow co-detractor, Prof. Garner, felt it necessary to perform a meta-analysis of the animal data. I am far from being an expert here, but I was under the understanding that the purpose of this test is to take very large data sets such as, say, several clinical trials involving thousands of participants, and try to make sense of otherwise ambiguous results. To apply meta-analysis to cages of five mice is frankly ludicrous.
I think most of us agree that non-human primate models are an important step in moving vaccines forward, and so MVA85A was then tested in macaques at the Porton Down biosafety Level-III facility. What I have learned in my long career is that animal modeling of TB is not rocket science, and that there are multiple variables that can confound even the most straightforward testing protocol. In this regard the NHP models have been particularly problematic, not just for the MVA85A studies but for other promising candidates as well, and one can be faced with scenarios in which the vaccine looks good, but the controls, positive and/or negative, do not cooperate. This is not a criticism, quite the reverse, and the primary centers involved – Pittsburgh, Tulane, and Porton – have all been working very hard to try to standardize these difficult [and still relatively new] models. The recent efforts by Porton to develop a more realistic “low dose aerosol exposure model” are particularly stellar.
Our field desperately needs a new TB vaccine, but progress is glacial. I first tested a candidate in 1985, and established the NIH Testing Program a decade later; it is now the year 2018. And there is still no guarantee a new candidate trial will work; one only has to look at recent events in the HIV and malaria fields for parallel examples. I would also recommend that MVA85A not be regarded as an “ineffective vaccine”; it was tested in the field as a “BCG-booster” and while it did no harm it failed at a statistical level. My program has noted that the clinical strains prevalent in the trial site area appear to be generally of low fitness and strongly inhibited just by BCG by itself, and hence to show a boosting effect would be very difficult. In fact, my laboratory has provided direct evidence for this potential possibility1.
I believe that the animal studies by McShane in Oxford, and Williams and Sharpe at Porton, were performed properly, honestly, and scrupulously carefully. It is my personal opinion that these three women are completely beyond reproach.
University Distinguished Professor
Colorado State University
1. Henao-Tamayo M, Shanley CA, Verma D, Zilavy A, Stapleton MC, Furney SK, Podell B, Orme IM. 2015. The efficacy of the BCG vaccine against newly emerging clinical strains of Mycobacterium tuberculosis. PLoS One, 10; e0136500
Competing interests: No competing interests