Product Development
Bringing COVID-19 vaccines to authorization faster through data sharing
Data sharing and assay harmonization can accelerate the path to COVID-19 vaccines
Developers of leading COVID-19 vaccines won’t be able to meet global demand on their own, but they do have the power to accelerate global coverage by helping each other reach authorization and streamlining development for the next candidates. It will require working together to harmonize preclinical assays and share patient-level data.
Vaccine developers have been gathering safety and immune response data through Phase I and II trials at record speed during the pandemic. The next step -- and the highest hurdle to clear -- will be demonstrating the immune response translates into protection against infection, or at least reduces disease severity.
According to FDA guidance on clinical development of COVID-19 vaccines, companies are required to prove that their vaccines protect participants. That involves running large Phase III trials in regions with a high incidence of infection, which come with high costs, long timelines, and considerable risk given how rapidly the pandemic moves through regions (see “FDA’s Vaccine Guidance Hits Mark”).
“Usually you know where your patients are, but here, the pandemic is moving around. Usually it takes three, four or five months to start up a trial at a certain site, but by that time in the pandemic the virus has come and gone,” said Mathai Mammen, global head of R&D for the Janssen Pharmaceutical Cos. unit of Johnson & Johnson (NYSE:JNJ).
Companies are tasked with predicting where the virus will be prevalent in the coming months, and have to “sign up 30, 50 or even 100,000 patients for a trial in the hopes of catching enough incidence to compare vaccinated people to placebo,” he added.
Vaccines from Moderna Inc. (NASDAQ:MRNA) and partners Pfizer Inc. (NYSE:PFE) and BioNTech AG (NASDAQ:BNTX) have moved into 30,000 patient Phase III trials. A handful of others, including J&J, are following closely behind with similarly sized late-stage trials expected to kick off this year.
As pioneers in a new indication, there are no shortcuts available to these companies. No one has correlated antibody titers, virus-specific T cells or other immune markers with actual protection.
“Usually you know where your patients are, but here, the pandemic is moving around.”
The first companies to reach the finish line could do a service to their followers: The “immuno-bridging” data they produce will allow other vaccine makers to map their antibody titers to protection.
The challenge is that the front runner vaccine developers are all using different assay protocols, and there are no industry-wide assay standards to address variability in immune readouts. Nor is there a precedent for developers to share the patient-level data required to build the correlates.
Industry and government are taking the first steps to address this. A condition of funding from Operation Warp Speed is that recipients contribute samples to be run through standardized assays in labs at NIH’s National Institute of Allergy and Infectious Disease (NIAID).
In parallel, a small, unofficial group of leading vaccine developers has committed to sharing patient-level data. The task now is to bring more companies on board.
Bridging the gap
The idea behind immuno-bridging is to use an immune response as a surrogate endpoint for protection. Immune responses can be collected much faster than protection data, within weeks of vaccination, and they don’t require high transmission rates at trial sites.
According to Hanneke Schuitemaker, VP and head of viral vaccine discovery and translational medicine and viral vaccines disease area stronghold leader at Janssen, immuno-bridging doesn’t eliminate the requirement for Phase III trials because valuable safety data are collected from those studies, but it can help salvage trials when the number of cases is in flux.
“If something fails to demonstrate efficacy because of reduced incidence of COVID-19, which may happen when for instance immediate lockdowns reduce the transmission rate, it might be possible to use immuno-bridging from a vaccine for which efficacy could be demonstrated and for which an immune correlate of protection could be established to get an estimate of vaccine efficacy,” she said.
It also may cut target enrollment by eliminating the requirement to show protection in the first place.
While the current guidance calls for full Phase III protection studies, Schuitemaker told BioCentury there’s precedent for authorizing vaccines on immune correlates.
“Our Ebola vaccine was tested in multiple Phase III studies that established safety data, but the efficacy signal was based on immuno-bridging from animal models because there was no way to determine efficacy in the field. This vaccine was recently approved in Europe, and it’s not the only precedent,” said Schuitemaker.
“The issue is everyone picked a different assay for how they want to measure performance.”
The Phase III COVID trials aim to prevent symptomatic disease rather than detectable virus as the primary endpoint because it isn’t clear whether a vaccine will prevent infection or decrease disease severity. Both are valuable outcomes.
If the results aren’t binary (infection/no infection), immuno-bridging is still an important tool, said Mammen. The immune response threshold may just need to be set at a point where it will predict protection from symptomatic disease rather than detectable infection.
Harmonizing immunity
The first step in implementing immuno-bridging to predict COVID-19 vaccine efficacy is to introduce a single assay protocol.
Assays measuring virus-specific immunity produce highly variable results, making it impossible to compare levels of immunity measured by different procedures.
While a neutralizing antibody titer measured by one protocol may be sufficient to predict antiviral immunity, the same titers measured by another may not.
Even the same assay run under different conditions can yield different results.
“Harmonization of assays is important because these are not completely objective immune measurements,” said Mammen. “It’s a little condition- and reagent-dependent. If the titer is 100:1, you may get 300:1 and I may get 50:1. It’s consistent within my lab but not across labs.”
The best time to set assay standards would have been before vaccine development began, but that didn’t happen. Companies were focused on the scramble to start development as quickly as possible.
The result is a growing collection of preclinical and clinical immunology findings that aren’t comparable.
“The issue is everyone picked a different assay for how they want to measure performance of their vaccine,” said Schuitemaker. “It’s not easy to switch assays now, and since there were simply no standards in the beginning of the pandemic and at the beginning of vaccine development, it is now not easy to compare the data.”
Another barrier to setting assay standards is from CROs contracted to run the immune response assays, some of which aren’t willing to divulge details of assay protocols for competitive reasons.
The Operation Warp Speed policy will get the ball rolling by standardizing assays used in late-stage testing for a small group of candidates. Companies that receive funding are required to share patient samples from Phase III trials to be run through standardized assays at NIAID labs.
“That will be a way that at least a few of the companies will have harmonized assay data,” said Mammen.
Operation Warp Speed has invested in the development of six vaccine programs to date, most of which also have funding from the program for manufacturing and procurement. A seventh program, mRNA vaccine BNT162 from Pfizer and BioNTech, is included in the program under a procurement deal but is not receiving development funding (see “Warp Speed’s Totals”).
Whether patient-level data from the Operation Warp Speed-backed Phase III trials will be made publicly available has not been disclosed. The program also hasn’t publicly specified which assays or protocols NIH will use.
Making the information public would enable other vaccine developers to align their assays early in development.
Another open question is how long it will take to generate the patient-level data from the assays, which will depend on capacity constraints at NIAID labs.
“Those labs will get quickly overwhelmed because there’s only so much capacity and with the number of samples coming in from multiple companies, I think it will be a little slow perhaps,” Mammen added.
Schuitemaker agreed. “I think there will be five to six companies each doing 30,000 people Phase III trials. The testing of all the samples from these trials needs to be divided over different laboratories to meet capacity.”
Data sharing adoption
The next barrier will be getting vaccine developers to participate in data sharing.
Companies routinely share summary-level data when clinical trials are complete, but it’s the granular patient-level data that’s required to draw immune correlates. Those data are usually proprietary.
According to Mammen, one of the reasons companies aren’t interested in sharing patient-level data is rooted in game theory. The first company to get a positive Phase III readout has made the big investment and done all the work, but the others gain almost as much from the accomplishment. There’s little incentive to win the race if data sharing is in play.
Most developers are creating COVID-19 vaccines for profit, with the exception of Johnson & Johnson and AstraZeneca plc (LSE:AZN; NYSE:AZN), and sharing data could cut into a first-to-market revenue advantage.
Mammen argued that gaining buy-in will require convincing companies that the benefits of sharing data outweigh the risks.
On top of the obvious public health benefit, companies gain data from peers that can help advance their own programs and there’s little value in being first to market during the pandemic when there’s plenty of room for multiple vaccines.
He told BioCentury that J&J is having unofficial discussions around data sharing with AstraZeneca, Moderna, Sanofi (Euronext:SAN; NASDAQ:SNY) and GlaxoSmithKline plc (LSE:GSK; NYSE:GSK).
Broader discussions about data sharing are under way via the COVID R&D Alliance, a group of more than 20 top biopharmas working towards COVID-19 countermeasures (see “Pharmas Align Behind Crowdsourcing Solution”).
In a podcast interview with BioCentury, Anne Heatherington, head of the Data Sciences Institute at Takeda Pharmaceutical Co. Ltd. (Tokyo:4502; NYSE:TAK), told BioCentury a subgroup of the Alliance is putting together plans to share patient level data, an initiative born during the pandemic that she hopes will survive into the post-COVID environment.