This article has been updated from a previous version to correct a math error in the last paragraph.
The US Food and Drug Administration’s approval of Gentris’ GentriSure, a line of six genomic human reference controls for cytochrome P450 2D6 genetic testing, signals that the agency feels the quality of reference controls is just as important as validated genetic tests, the company said.
“As an industry we focus more on the test components and ignore the controls, in this particular area,” Gentris CEO Michael Murphy told Pharmacogenomics Reporter this week. “Up until now it has been acceptable to use left-over patient material or residual patient samples as an alternative. This is likely to change in the near future.”
According to Murphy, the FDA believes that reference controls need to be as valid as the genetic test itself.
Gentris’ controls are intended for use with array-based tests like the AmpliChip, as well as non-array 2D6 tests like those sold by Tm Bioscience and Third Wave Technologies. The AmpliChip, which Affymetrix makes for Roche, is the only FDA-approved test that detects 29 variant alleles of CYP2D6 and CYP2C19.
“I think that FDA’s thinking on this is that since there is an approved pharmacogenomic test in the marketplace … the test and the control [should] have equal importance,” Murphy said. “Then we would want to have reference controls that are used for an approved test go through the same regulatory process.”
Additionally, the “FDA has been very public about trying to regulate better the use of laboratory-developed tests and reagents to ensure high quality and consistency,” Murphy said.
This paradigm was apparent last year when the agency issued an unscheduled draft guidance seeking to encourage molecular diagnostic companies to file for pre- and post-market review tests that use an algorithm to generate results from multiple data points, rather than follow traditional homebrew rules.
The FDA composed the draft following meetings with Genomic Health, whose Oncotype Dx uses a gene-expression profile to calculate the likelihood of breast cancer recurrence. With the guidance the FDA appears to be trying to prevent other companies from taking the same regulatory route as Genomic Health, which began selling its CLIA-approved lab test without seeking regulatory approval [see PGx Reporter 09-13-06].
Although Gentris is not aware of any competitors specifically in the area of reference controls for CYP2D6 diagnostic testing, the company expects other companies will take Gentris’ lead in pursuing FDA clearance.
“We certainly do expect other diagnostic manufacturers to get into this area once they start to appreciate the two things … that are necessary under CLIA regulations for testing patients for diagnostic purposes: the test itself and [the] external controls that validate the laboratory, the platform and the technician,” Murphy said. “Since you need both, you can say both have equal importance.”
Gentris’ ‘Gold Standard’
The FDA’s clearance of GentriSure is a milestone for Gentris, marking the company’s first FDA-approved product. The company validated the efficacy and validity of GentriSure in clinical trials at three external, independent sites. Additionally, the controls were tested on several CYP2D6 test platforms including the AmpliChip, as well as laboratory-developed assays.
Gentris believes its human reference controls are better than competing products because “they are identical to what you would see in patient material.” In other words, Gentris’ reference controls “haven’t been manipulated in any way as far as the genome itself,” Murphy said. “This is important because the gold standard that we want to achieve in the clinical diagnostics market is to have controls that are as close to patient samples as possible.”
Currently, reference controls used by companies are derived from less reliable sources, such as leftover patient samples, cell repositories, and synthetics. These methods present ethical dilemmas, and offer limited accuracy and consistency, Gentris said.
“The guidance in this area suggests that to use patients’ left-over samples without their consent you have to anonymize the sample by removing patient identifiers,” Murphy said. “In theory, the patient should have consented for that same test.”
However, “this is not happening,” Murphy said. What is happening in the industry in some cases is that the samples from a patient are being anonymized, but the material is not being used in the test for which the patient consented.
“We want to make sure that if we’re using a patient’s material that we did so with their approval, which is by having them agree to an informed consent process,” Murphy said.
With Gentris’ product the customer can be sure that the materials are from single donors, the company touted.
“So when you order your 2D6*4 control, you’re getting the same exact product every time, and that allows you to make sure that you’re not changing a reference control by moving from one residual sample to another,” Murphy said.
Challenge for Competitors
Gentris said it spent more than five years and “significant resources” in developing GentriSure and then getting it cleared by FDA.
“I think that FDA’s thinking on this is that since there is an approved pharmacogenomic test in the marketplace, … the test and the control [should] have equal importance. Then, we would want to have reference controls that are used for an approved test go through the same regulatory process.”
“We are a very small venture-backed company. Let’s just say it’s not a very inexpensive proposition,” Murphy said. “It’s very expensive to get any kind of product through the FDA approval process.”
Having gone through the process, Murphy noted that any potential competitors hoping to enter this space will need to overcome scientific hurdles and make a significant monetary investment.
“The reason it’s difficult [to enter this space] is, depending on the frequency of a particular variant you might have to screen up to 10,000 patients before you find that particular variant occurring even one time,” Murphy explained.
He gave the example of CYP2D6*7, a significant mutation due to its impact on a patient’s enzyme activity, that occurs in about only one in 1,000 people, primarily in the Hispanic population.
“It took us over five years to find a patient who would have that mutation and could go through our fairly rigorous informed consent process,” Murphy said.
For anyone wishing to get into this space, the complexity of the science also presents many opportunities for error.
There are two other genes that are very closely related: CYP2D7 and CYP2D8, so-called “pseudogenes,” which share 95-percent homology with the CYP2D6 gene.
“So each of these genes has 10,000 base pairs, and 9,500 of those bases are identical. Depending on the platform that you’re using, [this] could result in a false positive test,” Murphy said. “These ‘pseudogenes’ at one point had a function … but through elevation that function has been lost. So, they … can cause problems for testing the gene that we’re really interested in, which is 2D6.”