WHEN ANDREW SPEAKER traveled to Europe last May infected with what was believed to be extensively drug-resistant tuberculosis (XDR-TB), he did more than just precipitate an international health scare. His case, complicated by discrepancies among his early TB test results, thrust into the limelight the methods that clinicians and reference labs use to diagnose regular and drug-resistant forms of TB.
Effective TB treatment requires accurate diagnosis, and this remains a challenge. In Speaker's case, three weeks after his initial diagnosis of XDR-TB, the Centers for Disease Control & Prevention (CDC) downgraded his condition to multi-drug-resistant TB (MDR-TB), which is easier to treat. Most TB cases, however, are in parts of the world that lack CDC's infrastructure. A recent World Health Organization (WHO) assessment estimates that labs detect only 5% of MDR-TB cases.
Ideally, TB diagnostics for the developing world should quickly and accurately detect all forms of infection, as well as identify drug resistance to first- and second-line medications. Tests should be available at local clinics where most TB patients seek care. Although no diagnostic in the works meets all these requirements, there are promising tests in the pipeline that could improve the current situation. Breakthroughs in chemical instrumentation and prudent design of diagnostics at the molecular level may contribute to future advancements in TB diagnosis.
The task of worldwide research and development coordination for TB diagnostics falls to a few organizations. WHO issues policy recommendations to guide the development of new TB diagnostics. It also houses the Stop TB Partnership, a global umbrella organization advocating the eradication of TB. WHO also collaborates closely with the Foundation for Innovative New Diagnostics (FIND), an independent nonprofit organization dedicated to bridging academic research, the diagnostics industry, and developing countries.
Current diagnostics aren't equipped to deal with TB's many manifestations. The most contagious patients have a persistent cough from bacteria concentrated in their lungs. Typically, health care workers use microscopes to check the sputum these patients cough up for Mycobacterium tuberculosis (Mtb), the bacteria that cause TB. So-called sputum-smear microscopy is a 100-year-old test that detects only half of active infections and can't distinguish drug-resistant cases of TB from drug-susceptible ones. The detection rate is even lower among patients also infected with HIV. In those cases, TB spreads through the bloodstream into other tissues, leaving too few bacteria in the lungs to detect in a smear.
Sputum-smear microscopy may be the most commonly used tool, but it "is cumbersome, and technicians are overburdened," says Richard O'Brien, head of product evaluation and demonstration at FIND.
Another commonly used diagnostic, the omnipresent chest X-ray, is unreliable and suffers from poor sensitivity. Lack of specificity plagues the U.S.'s most routinely used diagnostic, the tuberculin skin test. The problem is that the proteins injected under patients' skin in this test aren't TB-specific. For this reason, anyone who received a TB vaccine or was infected with related bacteria would likely give a false positive result.
Culturing Mtb from a patient's sputum, the most sensitive method for detecting TB, is painstakingly slow, with results taking two to six weeks to arrive. Once those results are in, it takes even longer to test whether the cultured Mtb is drug-resistant. While patients wait to learn their culture's responses to a battery of drugs, they may be receiving the wrong treatment.
What's more, most people living with TB don't have an active, transmissible infection. According to WHO, 33% of the world's population has what's called latent TB infection. Scientists are hunting for new diagnostics that can reliably detect latent infection and evaluate the risk of developing full-blown TB on a case-by-case basis.
DESPITE THESE challenges, the costs of developing new diagnostics have been relatively small when compared with those of developing new vaccines and new drugs, according to WHO. For five different tests that have hit the market in the past 15 years, costs ranged from $1 million to $10 million each, depending on the format of the diagnostic. Now, Stop TB estimates $516 million will be required to develop and implement the coming generation of tests through 2015. However, the estimated public and private funding for TB diagnostic development during this time frame is roughly $80 million, far short of the needed budget.
In the developing world, it is still challenging to reliably detect TB in general, not necessarily MDR and XDR strains. Some diagnostics being developed focus on this particular problem.
A patch under development at Sequella, a Rockville, Md.-based clinical-stage biopharmaceutical company, could complement the tuberculin skin test and sputum-smear microscopy. The patch transfers a TB-specific protein just underneath the skin. Infected individuals develop an inflammatory response to the transferred protein at the site of the patch after two days. "It looks and feels like a poison ivy reaction," says Carol Nacy, Sequella's founder and chief executive officer. The test, which is expected to be available outside the U.S. beginning in 2008, could be used in tandem with existing diagnostics to speed diagnosis of TB that has spread outside the lungs, which usually requires slow-growing bacterial cultures.
Patients with latent TB infections won't have a positive patch test, but there are other indicators of the disease lurking beneath the skin. Human T cells that have encountered Mtb in the past will release chemical signals, such as the protein interferon-gamma (IF-γ), if exposed again. TB diagnostics in this area focus on measuring the release of IF-γ from the infected person's cells. Commercially available assays from Australia-based Cellestis and U.K.-based Oxford Immunotec involve exposing blood samples to a pair of peptides secreted specifically by both latent and active TB. After an overnight incubation, lab workers introduce antibodies specific to IF-γ as a means of detection. The entire test takes one to two days.
Immune-response-based tests have disadvantages, however. HIV infection weakens the immune response that's necessary for such assays to work. These tests have not been widely adopted in the developing world, mainly because the priority there is detecting active infection, and in its current form, these tests give the same positive readout for active and latent infections. Ongoing research aims to adapt the test to evaluate the risk of a latently infected patient developing active TB.
Speaker's case illustrates the need for reliable ways to distinguish between the MDR- and XDR-TB strains. Many diagnostics exploit mutations encoding drug resistance to label a strain as MDR-TB. Other tests in development don't require information about specific mutations.
MDR-TB is resistant to first-line drugs rifampicin and isoniazid. Rifampicin-resistant TB strains are characterized by mutations in a subunit of RNA polymerase that prevent the drug from binding. More than 95% of rifampicin-resistant strains are MDR-TB strains, says Blanca I. Restrepo, a microbiologist at the University of Texas School of Public Health, in Brownsville. Restrepo is validating diagnostics called molecular beacons to identify rifampicin-resistance mutations in Mtb strains.
MOLECULAR BEACONS, which were adapted for TB diagnosis by infectious disease specialist David Alland of the University of Medicine & Dentistry of New Jersey, in Newark, are short nucleotide probes that can be used to interrogate polymerase chain reaction (PCR)-amplified Mtb DNA from a patient's sputum for the presence of mutations in the RNA polymerase gene. Beacons are sensitive enough to detect single-nucleotide mutations that are associated with rifampicin resistance. Results are obtained in less than one week.
Beacons feature a fluorophore and a quencher on opposite ends. Mutated genes can't bind to the beacon, which adopts a looped conformation with a quenched fluorophore. The normal RNA polymerase gene binds to the beacon, unfolding the loop and distancing the fluorophore from the quencher to give a light-based readout. Beacons require sufficient Mtb DNA in sputum to amplify by PCR, and they won't work on latently infected patients. Costs and cross-contamination of DNA samples are bottlenecks to implementing this tool in developing nations.
Unlike resistance to rifampicin, isoniazid resistance is encoded in more than one region in the Mtb genome. Even though 95% of rifampicin-resistant Mtb is also isoniazid-resistant, it's still important to be able to detect isoniazid resistance on its own. Isoniazid is a prodrug that interferes with cell-wall synthesis in Mtb. A catalase-peroxidase enzyme in Mtb converts isoniazid into its active form. Mutant catalase-peroxidase genes are found in about 50% of isoniazid-resistant strains. A commercially available 2-inch nitrocellulose strip can be used to detect catalase-peroxidase mutants and rifampicin resistance to boot.
Hain Lifescience, a German company specializing in diagnostics, developed the strips to detect infectious diseases and genetic disorders. The firm partnered with FIND to implement large-scale evaluations of a strip-based TB diagnostic in the developing world. Lab workers apply PCR-amplified Mtb DNA from sputum to the narrow strips, which are printed with specific probes. Different line patterns on the strip indicate whether a TB strain is resistant to one drug, both (MDR-TB), or neither. Results arrive two days after the sputum is obtained. "This test may reduce the need for culture by 80 to 90%," O'Brien says.
The strips don't yet work on HIV-positive, latently infected, or other patients without sufficient Mtb in their sputum. However, the test's format is highly expandable as mutation information for resistance to other drugs emerges. Hain Lifescience expects a 2008 release of a strip that will test for resistance to a second-line fluoroquinolone TB drug, and strips are being developed for testing resistance to kanamycin, amikacin, and capreomycin, O'Brien says. Resistance to any fluoroquinolone and at least one of the other three defines XDR-TB, according to WHO.
Some diagnostics for TB can test for drug resistance even if information about specific mutations isn't handy, as long as samples of the relevant antibiotics are available. One still in development is the microscopic observation for detection and susceptibility (MODS) assay.
Improving microscope-based diagnosis is an appealing option in the developing world because retraining clinic workers would take minimal effort. To prepare samples for MODS, workers culture Mtb samples from sputum in liquid media, in which Mtb grows more rapidly than in the traditional agar-based media. They then use a microscope to detect the characteristic tangles and coils in the liquid cultures of growing Mtb. Using media spiked with antibiotics kills off drug-susceptible TB strains, allowing detection of resistant bacteria. Results take five to seven days. Use of affordable fluorescence-based microscopes, which FIND estimates will become available in 2009, could make this assay even more sensitive.
Another test under development by FIND and Biotec Laboratories, a U.K. diagnostics company specializing in TB, is aimed at detecting rifampicin-resistant Mtb strains. The test is based on bacteriophages, which are viruses that infect bacteria. The bacteriophages used in the test selectively infect Mtb and a few other closely related species of bacteria. Instead of waiting weeks for cultures to grow, lab workers add the bacteriophage to a processed sputum solution that's been pretreated with rifampicin. They then kill any bacteriophages that haven't safely entered and infected Mtb. Eventually, the infected Mtb bursts open, spreading new bacteriophages throughout the culture. A fast-growing relative of Mtb is then added to the culture dish. If the patient has rifampicin-resistant TB, surviving bacteriophages will kill off the fast-growing Mtb relative, leaving telltale bacteria-free zones in the culture dish.
"From a health care perspective, rapidly identifying patients who are multi-drug-resistant and infectious is critical," says Andre Trollip, a product manager at Biotec. This technology, he says, is in theory adaptable for testing susceptibility to many different drugs, potentially simplifying detection of MDR- and XDR-TB.
NO DIAGNOSTIC currently available or in development meets all of the guidelines set out by WHO and other organizations. However, research in chemical instrumentation and reagent design might someday improve cost and throughput of TB diagnosis. For example, Dirk-Peter Herten, a physical chemist at the University of Heidelberg, in Germany, teamed up with microbiologists to design an alternative molecular beacon for detecting TB (Anal. Chem. 2005, 77, 7195). Most beacons incorporate a fluorophore and a quencher on either end of a short stretch of nucleotides. Labeling on both ends is expensive and difficult. What's more, if the coupling reaction used to install the quencher is ineffective, the beacons suffer from high background signal.
"We wanted to show that we could simplify the detection schemes and show that the sensitivity of the assay could be better," Herten says. His team designed a beacon that uses a specific nucleotide sequence to quench the fluorophore. The new beacon, which is not yet in clinical development, may lower background readings in some kinds of assays, Herten says.
In the future, mass spectrometry could be used to rapidly detect TB biomarkers and distinguish different strains. The instrument is easy to automate and adapt to high-throughput screening, thereby significantly speeding diagnosis. However, difficulties of cost, portability, and the training necessary to analyze complex readouts have yet to be resolved, according to Plamen Demirev, a physicist at Johns Hopkins Applied Physics Laboratory who uses mass spectrometry to detect pathogens. No single method is likely to be a universal diagnostic for TB, Demirev says. "It's a complex problem, and looking at it from many interdisciplinary angles will certainly help."