NEW YORK (GenomeWeb) – ImmunID announced this week it has obtained patent protection in China, adding to IP in Europe, Japan, and the US, and completing a global IP strategy protecting the firm's technology base.
Headquartered in Grenoble, France, ImmunID is developing a method called ImmuneTracker for profiling T cell repertoires as a companion test for immune checkpoint inhibitors.
Immunotherapies have shown promise for cancer patients with previously untreatable cancer, but they only work on a subset of patients, and the side effects of supercharging the immune system can be devastating.
One way to determine likely responders to these drugs may be to profile the T cell repertoire, but typical PCR- and sequencing-based methods can be cumbersome, costly, and time-consuming, according to ImmunID CEO Bernhard Sixt.
A co-founder and former CEO of diagnostics firm Agendia, Sixt was appointed CEO of ImmunID in 2013.
The ImmuneTracker technology is based on a method the firm calls multi-N-plex. It is essentially a set of 23 multiplexed qPCR assays used to detect all V and J gene recombinations in T cells, ultimately reporting out levels of 276 analytes.
Specifically, it runs using a 96-well format plate on a thermal cycler such as a Roche LightCycler or Cobas platform, Sixt said. A plate is made up of four rows of 24 wells, so one row of wells is a single test. Since each patient sample is run in duplicate, one plate can produce two test results.
"We are happy to be able to use PCR, because it is a widespread, very robust technology, but [standard] PCR has a huge weakness," he said. Namely, multiplexing in one tube using other technologies to attempt to detect as many analytes as ImmunTracker does will likely result in biases toward targets expressed in higher concentrations, drowning out lower-concentration ones.
The splitting method used by ImmunID results in less bias, and V and J rearrangement detection yields products of differing lengths which can be separated by multifluidic migration, specifically using a lab-on-a-chip from Perkin Elmer.
"We get a very high resolution of the residual repertoire because the competition is less ... it uses proven, robust technology, and it's simple and very quick," Sixt said.
Patient DNA is isolated and concentrated from 5 microliters of blood collected in EDTA, using commercially available methods from Qiagen and EMD Milipore.
In total, the process takes five hours from start to the issuing of a clinical report.
The system also includes a software component to analyze the data from the 23 multiplex qPCR reactions run in duplicate, called Constel'ID.
The firm's IP limits the number of multiplexed qPCR reactions for V and J rearrangements that its competitors can run in parallel to two, Sixt said, because "everything else is patented."
To resolve all 276 combinations instead using a single reaction, next-generation sequencing is most often used. Such was the case in a Science Translational Medicine study reported last year by GenomeWeb, for example, which found high-frequency clonotype maintenance to be an indicator of more benefit from a particular immunotherapy.
But sequencing can be biased by the initial PCR amplification step, which would tend to preferentially amplify more highly expressed clones, Sixt noted, while splitting the reaction to 23 smaller ones overcomes this problem.
ImmunID recently collaborated with researchers at Memorial Sloan Kettering to investigate whether the method was capable of retrospectively sorting responders and non-responders to the anti-CTLA-4 antibody, Yervoy (ipilimumab).
The results, published in June of this year in the Journal for Immunotherapy of Cancer, were "really stunning," Sixt said.
Of 12 samples, the technology called five responders and classed seven as "immune incompetent" non-responders.
"After un-blinding, it was a 'wow' effect, because all of the seven we said should not respond did not, and from the five we said should respond, four of them responded," Sixt said.
He cautioned that the samples were selected for clear response and non-response, and the evaluation had a small sample size. But the 100 percent negative predictive value and 80 percent response rate prediction was "an extremely clear signal."
The researchers also showed that responders had a greater diversity of T cell combinations, and that their overall expression signature was not dominated by a few clones, but rather had more evenness in the representation.
ImmunID also launched a nationwide prospective clinical trial of more than 200 melanoma patients in France, called PREDICT-ID, in January of this year, Sixt said. That trial measures the predictive value of ImmuneTracker immune profiles for response to Yervoy.
"Immunotherapy costs between $100,000 and $250,000 per patient, per year," Sixt said, but it is also "a real revolution in cancer; people for the first time are experiencing that you can effectively treat metastatic patients who are basically already written off according to usual wisdom, and some of them stand up and go home."
The rate of full response, however, is around 4 percent to 20 percent, he noted, and "if you unleash the immune system, you get quite nasty side effects because the immune system is a very powerful beast — it can kill the cancer but it can also make a lot of trouble for the patient."
Thus, success depends on predicting response and side-effects, Sixt said.
The firm has data on other immune checkpoint inhibitors, he said, noting that preliminary results for anti-PD1 and anti-CTLA-4 inhibitors are promising.
Business plans
ImmuneTracker is a CE-marked test kit, so it could theroretically be sold in Europe, Sixt said. The data created by this kit must be uploaded to ImmunID's internet portal to be evaluated by the firm's Constel'ID software, then ImmunID sells the report to the clinician.
The firm is in discussions with the US Food and Drug Administration to determine what needs to be done to have the test cleared in the US.
"We believe it should be a relatively quick process because being ISO 13485 certified and having a CE mark is relatively close to what the FDA would require for technical validity," Sixt said.
The firm will initially seek Class I approval for the kit, to prove technical validity. For 510(k) or PMA, clinical utility data will be needed. "We are in the process of validating our initial findings in a nationwide clinical study in France, and we are about to start a US/European study, and one study in the UK," Sixt said.
ImmunID used to do fee-for-service, and still has one contract ongoing. But it changed strategy three years ago in order to "capture the value of guiding very powerful and expensive treatments toward effective usage."
The firm has also been inspected and is about to acquire a CLIA number, Sixt said.
"That means that the central service laboratory in Europe, which will be one out of three, will soon be able to accept American clinical patient samples."
ImmunID will thus begin commercialization in the US using this CLIA service starting early next year and already has customers signed up.
Companion testing
There are numerous immune checkpoint inhibitors that might theoretically benefit from being guided by ImmunTracker data.
As previously reported by GenomeWeb, AstraZeneca is developing the PD-L1 inhibitor MEDI4736; Roche/Genentech is developing MPDL3280A, a PD-L1 inhibitor; Bristol-Myers Squibb is advancing an anti-PD-L1 drug BMS-936559 and the PD-1 inhibitor Opdivo (nivolumab); and Merck is developing a PD-1 inhibitor Keytruda (pembrolizumab). The FDA recently approved Opdivo in advanced melanoma and advanced squamous NSCLC; and Keytruda in advanced melanoma.
These four firms are studying these immunotherapies in NSCLC with the help of an IHC test from either Ventana or Dako. Perkin Elmer has also begun a collaboration with OncoSec to test its Vectra quantitative pathology imaging system as a means of determining responders to Keytruda.
The second generation PD-1 and PD-L1 drugs seem to have a different mechanism from anti-CTLA-4 antibodies, and Sixt said preliminary data shows ImmuneTracker can stratify patients toward these or first-generation therapies.
Futhermore, combinations of these drugs produce more dramatic results, but also more dramatic side effects, Sixt said.
"If you have a tool to direct people who have high response toward one or the other drug, you can save them the trouble of first taking the wrong one and, secondly, save them unnecessary combination treatment," Sixt said.
"We believe ultimately we can help guide physicians through the maze of all these drugs and combinations."
The firm has signed a deal with a UK immunotherapy company, Scancell, and has others in discussion, Sixt said. "The smaller biotech companies are signing up with us because we can help them pre-select their patients, which makes their trials very effective, quick, and easy," he said.