NEW YORK (GenomeWeb) – Sequencing of multi-drug resistant tuberculosis strains collected during a 2009 outbreak in Swaziland has revealed that approximately 30 percent contain a mutation that is undetectable by most molecular tests of drug resistance, including Cepheid's widely adopted GeneXpert MTB/RIF test.
In a letter to the editor published yesterday in the New England Journal of Medicine, researchers associated with Médecins Sans Frontières/Doctors Without Borders examined 125 MDR TB strains collected in Swaziland during the outbreak. They found 38 had a mutation outside the "hotspot" in the rpoB gene sampled by the Xpert MTB/RIF assay, where mutations confer resistance to rifampin, one of the first-line antibiotic treatments.
The mutation, called I491F, was reported in 2011 in studies from Hong Kong and Australia, but was not a dominant strain in those locations. The Swaziland survey also uncovered two more strains that carry I491F plus a second mutation conferring treatment resistance.
The inability to detect the rpoB I491F outbreak strain is disturbing because Xpert MTB/RIF, which is designed to detect Mycobacterium tuberculosis and rifampin resistance, is used throughout Swaziland as the first-line diagnostic test for TB and MDR TB as recommended by the World Health Organization, the authors of the NEJM study wrote.
And a high prevalence of undetectable drug-resistant TB in Swaziland is particularly devastating because Swaziland also has the highest prevalence of HIV in the world, with 26 percent of the adult population infected. More than 80 percent of people with TB have HIV co-infection, and TB is the leading cause of death among people with HIV in that country.
To begin to tackle the problem, Stefan Niemann of the Research Center Borstel in Germany, an author on the study, told GenomeWeb that he and his co-authors used Sanger sequencing to evaluate the 2009 outbreak strains at the behest of MSF.
"In Swaziland, there was an eight-fold increase in the prevalence of multi-drug resistance over the last few years, so [MSF] wanted to know what the reason is for that," Niemann explained.
"We thought initially that the increase in MDR TB might be related to [increased] transmission, and that was confirmed," he added.
However, the group was also "quite surprised" by fact that such a high percentage of the strains would not be detectable by Xpert MTB/RIF. "Finding a strain that has this particular mutation was completely unexpected," said Niemann.
Analysis of strains' full genomes demonstrated a very low diversity, "indicating ongoing recent transmission," Niemann said. These are "random mutations that are then selected by treatment failures," he said.
To combat the problem of drug-resistant TB, the WHO, the Foundation for Innovative New Diagnostics, and the Bill and Melinda Gates Foundation, among others, collaborated extensively to accelerate the development of an easy-to-use, point-of-care TB test that could also assess resistance to some degree.
They invested in the MTB/RIF assay on the Cepheid GeneXpert platform, and have since negotiated concessionary pricing to place these products in a number of high-burden countries. The test launched in 2009 and was endorsed by WHO at the end of 2010.
Since then, uptake of the test has been "a huge success, from our standpoint," David Persing, executive vice president and chief medical and technology officer at Cepheid told GenomeWeb.
"It pushes this technology to all four corners of the globe, and it really is a good demonstration of what can be done in the future for other tests," Persing said.
The firm has placed over 8,000 GeneXpert systems worldwide, about 3,500 of which were purchased via the high-burden developing country (HBDC) program that Cepheid began in 2011, Persing said.
By of the end of 2014, 116 of 145 countries in the program had purchased at least one GeneXpert platform under the special pricing, WHO reports.
MDR TB strains are resistant to the antibiotics isoniazid and rifampin. Extensively drug resistant, or XDR TB, is defined as strains resistant to these two drugs as well as a class of drugs called fluoroquinolones and at least one of the three injectable drugs, kanamycin, capreomycin, and amikacin.
Some XDR TB strains may, however, respond to drugs on a list of other possible treatments, including ethambutol, pyrazinamide, and streptomycin. By the end of 2011, 77 countries worldwide had reported at least one case of XDR TB, including the US, according to WHO.
Unfortunately, strains containing the I491F mutation are "not only resistant to isoniazid and rifampin, they are also resistant to two other first-line drugs, ethambutol and pyrazinamide, so, actually, they are full first-line resistant," Niemann said.
Persing noted that the targets of Xpert MTB/RIF were chosen in collaboration with WHO and other groups. "The goal was to identify mutations that were found among all MDR clones that were in circulation at the time the assay was designed," he said. "And we did that. We basically accounted for all the major clones in circulation — the assay worked, and still works very well to identify those clones."
The original assay was based on sequence analysis of a collection that was obtained from extensive strain collection of MDR TB clones and strains from a variety of sources, including WHO, Persing explained.
However, he added, "you can never trust a mutant; they can always figure out a way to escape and get more fit and become more transmissible, so we're taking this very seriously."
In a follow-up email, Persing also noted that "in light of the selective pressure of test-and-treat programs based on the GeneXpert, it is perhaps a bit surprising that this MDR strain has not already become recognized as the dominant clone in circulation in that part of the world … GeneXpert implementation did not start until after these outbreaks occurred, so if this were already widespread problem, I would have expected to see it by now."
Persing reiterated that Cepheid will need to be responsive to mutations arising in the TB genome, because it is the nature of the molecular diagnostics business.
"We've seen this before," he said. "There are studies in Sweden of diagnostic tests all of a sudden losing their ability to detect a chlamydia strain, followed by documented outbreaks of chlamydia, unbeknownst to anybody because the tests were not detecting these outbreaks," Persing said.
"This was probably one of the most effective demonstrations of the interplay between a diagnostic test and infection control at the public health level. [The I491F mutation] is yet another one, and it's going to keep happening, so I'm confident that we can be ready for it," he said.
Niemann and his colleagues, meanwhile, have begun surveying MDR strains in neighboring countries, such as Mozambique, but he confirmed that available performance data from GeneXpert testing would suggest the mutation has not yet spread in MDR strains common in other regions of the world, such as Eastern Europe.
However, he suggested that countries may need to avoid focusing entirely on one technology, and aim to have "the possibility to use phenotypic testing, gene sequencing, or some other assays," in order to resolve discrepant results.
"The main problem with applying GeneXpert now in Swaziland as a first-line assay to detect MDR is that 30 percent of the patients are presumably not detected … and they are not treated properly," Niemann said. "This leads to enhanced transmission and higher death rates, so it might be a really dangerous scenario."
"One message, in addition to [identifying] this mutation, is that the MDR epidemic is largely driven by transmission of a few outbreak clones in the area, presumably because they have not been properly detected, and by long [inappropriate] treatment those strains have a selective advantage, and are spreading more frequently than others," Niemann said.
Persing concurred with this assessment. "When you apply a selective pressure — in this case a diagnostic test to detect and then to treat patients with a condition — you can expect this to happen," he said.
Niemann also co-authored a study in Nature Genetics two years ago which showed that diversity of mutations goes down when there is a dominance of outbreak strains in a certain area, and the fitness cost of the mutations in those outbreak strains is particularly low.
He advocates for improved detection and susceptibility testing for second-line drugs to give the patients appropriate treatment, including a phenotypic assay called thin layer agar rapid drug susceptibility testing.
"We are working on next-generation sequencing assays: by analyzing the full genome we can detect the resistance markers," Niemann said. That technology is not yet available in most HBDC countries, but "there's strong initiative towards these comprehensive molecular assays … because these are able to detect all resistance mechanisms in a given strain, and that would then allow for a more individualized treatment regimen," Niemann noted.
Niemann hopes that with future technological advances, "at least core centers, even in Africa, [will] have this technology and can at least work on cases that are particularly interesting."
Adapting Xpert TB tests
There are now a number of ways Cepheid could respond to these data, Persing explained.
"We have seen this, we know … it is just outside the range of detection of our assay. Can we make accommodations to be able to detect it? I think that's quite likely," he said.
"The responsibility we have is to track [new mutations], maintain collections, analyze those collections — we already have extensive consortium activity in which we obtain strains that are targeted by our diagnostic tests for the exact purpose of genetic characterization of those strains to look for new strains that might be occurring," he noted.
Persing also noted that Cepheid has had to accommodate new strains before in many of its tests, including its assays for methicillin-resistant Staphylococcus aureus, Clostridium difficile, and HIV.
Cepheid currently has two TB products in development — the Xpert TB Ultra and XDR cartridges — which the firm might now adapt.
Both the MTB/RIF and the TB Ultra currently scan the rpoB gene for about 20 mutations, while the XDR cartridge "scans for dozens more mutations within different genes that confer resistance to other drugs," Persing said.
TB Ultra will have a multi-copy TB target to boost sensitivity by about 10- to 15-fold, Persing said, into the sensitivity range of liquid culture.
"We're targeting that test as a culture replacement technology," Persing said. "Can we modify that? It's conceivable. That test is targeted for release in early 2016. Will we have time to do that in time for release? I know the TB community is greatly anticipating that test because it will bring a new level of confidence to TB testing, such that they don't have to rely on liquid culture anymore as a backup technology."
The firm could also choose to update the GeneXpert XDR cartridge in development. "That cartridge detects the other mutations that are likely to be found in this strain … and it does so with a new high-resolution melt capability which allows for a broad range of mutations to be detected in one cartridge," Persing explained.
XDR is in field trials now. "We expect release of that product about a year to a year and a half after the release of TB Ultra," Persing said.
"Either of those products could potentially be modified to accommodate this new mutation, and we're looking at ways to do that right now."
The firm will also continue to monitor epidemiological studies. "We don't know what the fitness cost of this mutation is, whether it is a clone that is not likely to spread … it could begin to spread outside this area and then obviously the implications are larger," Persing said.
Importantly, Persing noted that WHO and FIND contributed to the initial design of the first assay, providing advice and access to strain collections. "We will pay close attention to … their findings on this, and additional studies that might happen as a function of the work of MSF and other organizations."
He further added that is "good advice" for countries to leave room for other techniques in their lab arsenals.
For example, one goal of the TB Ultra test is to identify patients who need to have additional work done for drug resistance analysis. "Right now, liquid culture capabilities are spread across large patient populations because the molecular methods just aren't quite sensitive enough to use as a front-line screen for everybody," Persing said. "If you had a test for TB that was exquisitely sensitive, you could focus your liquid culture, drug resistance testing, and surveillance activities much more effectively."
Customers who now possess GeneXpert platforms will also see an advantage with the TB Ultra test, because faster time to result will free up room to run other Cepheid tests.
The firm's quantitative HIV cartridge for viral load monitoring got CE-IVD approval last year, and a qualitative test is likely to be approved in the next quarter, Persing said.
"They can both run on the same platform, so a system brought in through the HBDC program and placed in a laboratory can run the HIV cartridge," Persing said. The firm has also seen interest in a human papillomavirus test cartridge for cervical cancer screening in these countries, he said.