Findings That Led to $183 Million Autoimmune Drug Development Deal Questioned

Ellie Kincaid

July 31, 2019

A group of scientists from leading institutions is questioning the findings of a prominent immunology article linked to a strategic drug development partnership potentially worth up to €164 million ($183 million).

The original article, published in the journal Cell in 2016, describes studies of the immune systems of patients with autoimmune polyglandular syndrome type 1 (APS-1), a rare disorder. The disease results from mutations in the autoimmune regulator (AIRE) gene that in healthy people plays a role in training immune cells not to attack the body's own tissue. Although niche at first glance, the findings seemed to indicate that these patients produced antibodies that could have therapeutic potential for more common autoimmune diseases.

In an article published in the journal eLife last month, scientists from institutions that include the Karolinska Institute, the University of California, San Francisco, Stanford University, and the National Institute of Allergy and Infectious Diseases report that they attempted to replicate some of the results from the 2016 Cell article and were unable to do so.

They challenge two of the article's findings: that patients with APS-1 produced a much wider range of antibodies against their own bodies' biological molecules than previously described, and that the patients who did not develop type 1 diabetes had antibodies that neutralized interferon-alpha, whereas patients who lacked those antibodies did develop type 1 diabetes, pointing to a "disease-ameliorating" role for the antibodies.

Those interferon-neutralizing antibodies are the focus of a development partnership between the France-headquartered pharmaceutical company Servier and German biotechnology company ImmunoQure, said ImmunoQure cofounder Adrian Hayday, PhD, a professor and researcher at King's College London and corresponding author of the Cell article.

The press release that announced the partnership in February 2018 said the two companies would both work on the antibody's preclinical development and that Servier would be responsible for all further clinical development, such as exploring its use in the treatment of systemic lupus erythematosus and Sjogren syndrome. Servier agreed to pay ImmunoQure up to €164 million ($183 million), which includes an upfront payment of an undisclosed amount and additional payments at milestones in the development process, plus royalties on net sales if the antibody makes it to market.

Despite the new challenge, a Servier spokesperson said the company's internal data confirm the importance of the interferon-alpha pathway in autoimmune diseases, as well as the specificity and neutralizing activity of the antibody cloned by ImmunoQure.

Challenging a "Landmark" Article

On the point about the number of antibodies against their own body's molecules that APS-1 patients developed, the challenging scientists make the case that the statistical analysis in the Cell article was skewed to identify significantly more antibodies in the serum of case patients than of control patients. They performed analyses on their own dataset from APS-1 patients and control persons. "We could show that using random data, we got basically the same result they did," said lead author Nils Landegren, MD, PhD, of the Karolinska Institute in Stockholm, Sweden. "We have a very strong case their results are false."

Landegren and colleagues also ran experiments to test whether they could get results similar to those described in the Cell article that show that APS-1 patients without type 1 diabetes had antibodies that neutralized interferon-alpha, whereas patients who developed type 1 diabetes did not have those antibodies.

"We did our best using an established method, using positive and negative controls showing the expected result," Landegren said, but they could not see any difference between patients who had type 1 diabetes and those who did not. They found interferon-neutralizing activity in both groups of patients. It will be important for other research groups to try to replicate the experiment in future studies and to see how they fare, Landegren said.

Hayday said his group used the statistical analysis method they did because he was more concerned about ruling out false negative results than false positives when looking for antibodies expressed as "private reactivities" in a few patients, not the cohort as a collective group. Further work validated these private reactivities "absolutely irrefutably," he said. He sees the difference in statistics as a difference in approach: "When you undertake a statistical treatment of data, there is no silver bullet, you have to use what makes sense."

The difference in results regarding interferon-alpha-neutralizing antibody and type 1 diabetes, in Hayday's opinion, occurred because the attempted replication was a "shoddy experiment" and the method used was much less sensitive than the method used in the original experiment.

Other companies are pursuing potential therapies that target interferon-alpha in lupus, Bertrand Delsuc, founder and CEO of the European biotechnology business intelligence firm Biotech Radar, wrote to Medscape Medical News in an email. However, in the most recent results from clinical trials that tested different approaches from London-based pharmaceutical company AstraZeneca and the small French biotech Neovacs, both studies failed their primary endpoints. These recent clinical data are "mixed at best" with regard to interferon-alpha's role in lupus, Delsuc said. The conflicting publications offer "nothing really new" as a conclusion, Delsuc said: "reproducibility of literature experiments is globally poor."

Access to ImmunoQure's website currently requires login credentials, but Hayday said that that is unrelated to the recent publication. The website is being redesigned and will be back up again in about 10 days, he said.

One of the scientists who reviewed Landegren's article and Hayday's response for eLife before publication called the original Cell article from 2016 "a landmark paper of recent years." The comments were published in eLife anonymously, as is its practice. That reviewer found Hayday's and colleagues' formal response in eLife about the interferon-neutralizing antibodies to "convincingly demonstrate the original claim," but warned, "be careful about over-generalizing the findings." Another scientist reviewer disagreed: "This appears to warrant a revision to the original conclusions in Cell that 'those with T1D showed only low or negligible neutralization.' " As for the contested statistical analysis, one reviewer wrote that Landegren's and colleagues' "concern is 100% warranted."

The authors of the new critique originally submitted their manuscript to Cell, but it was rejected without an explanation, Landegren said. Journals are not always enthusiastic about publishing work critical of articles they published previously.

Landegren has disclosed no relevant financial relationships. eLife correspondence coauthor Michael Snyder reports serving as founder and consultant for Personalis, being a member of the scientific advisory board of GenapSys, and consulting for Illumina; and coauthor Olle Kampe reports being a board member of Olink Bioscience. Hayday reports holding equity in ImmunoQure, as do other coauthors of the formal response.

eLife. Published online June 27, 2019. Full text

Follow Medscape on Facebook, Twitter, Instagram, and YouTube and follow Ellie Kincaid on Twitter at @ellie_kincaid


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.