Better Testing for TB
words by Barbara Hurley / photograph by John Emerson
t’s not always what you see in Italian opera, with people coughing blood,” Marila Gennaro explains in a voice softened with the accent of her native Sicily. “TB may be asymptomatic.” In addition, those being treated often feel better in the short term and stop taking medication even though they still have the disease.
“Even uncomplicated cases take at least six months to cure, and the risk of a drug-resistant strain of TB emerging when the course of medication is not completed can create a terrible issue,” she notes. “What was treatable becomes untreatable.”
Gennaro joined the Public Health Research Institute (PHRI), now a part of New Jersey Medical School, in the mid-80s in New York City at a time when that city was experiencing a rapidly rising TB rate and spending millions of dollars to control its spread through programs like directly observed therapy to deal with compliance issues. Support for TB research — both from PHRI and from the National Institutes of Health — encouraged Gennaro to make this the focus of her work.
Gennaro’s research, however, targeted not treatment but diagnosis. She thinks of TB as coming in “two flavors,” disease and infection. She estimates that one-third of the world’s population — 2 billion people — is latently infected with TB. Even though the risk of developing full blown TB is small — 5 to 10 percent — this translates into millions of people. And these people are now scattered all over the globe.
Immigration has made TB an issue not only in those countries where the infection is endemic, but in those places that draw immigrants to a better life. “Stress can activate TB,” she says. “And what is more stressful than immigration itself, uprooting oneself from home to a new place.” Identifying who is infected is critical. So, how to tell?
For more than a century, the tuberculin skin test that relies on an immune system reaction to proteins associated with TB has been the standard. But its shortcomings go beyond the need to return after two days to have the results read. Those who have the test often — researchers in an infectious disease laboratory, for example — may develop a positive result to the test itself that is hardly distinguishable from actual infection. In addition, those who have received the TB vaccine can also have a positive response without being infected, making this standard test virtually useless.
Gennaro likes to solve puzzles, and here was a difficult one. The DNA of the bacterium that causes TB has some areas that are not present in the genome of the strain used for the TB vaccine. What was the significance of these “regions of difference”? Do they make proteins that would give a specific reaction to a test that would identify those truly infected? Gennaro’s research identified those particular proteins. At Colorado State University, testing on a colony of guinea pigs soon demonstrated that Gennaro was on to something big and — she recognized immediately — something that would be profitable. Her first call was to a patent attorney. She laughingly describes that conversation as one-sided, since she really didn’t understand the attorney’s legalese, but she understood the need to protect her intellectual property. Her work laid the foundation for a new way to test for TB infection that not only was more accurate but also could be done quickly ex vivo by a blood test. The discovery is licensed to Oxford Immunotec, a medical diagnostics company based in the UK that develops novel new tests for various diseases. It has made the test available in the U.S., where it is already helping to promote accurate diagnosis of TB. The discovery also earned the scientist a prestigious Research and Development Council of New Jersey Thomas Alva Edison Patent Award in 2009.
What’s next? Gennaro reports that in countries with high rates of TB almost everyone is infected, so testing for infection is typically meaningless. The need now is to discover what differentiates the bacterium that will transform from dormancy to disease. Gennaro hopes that this will lead to a better, quicker diagnosis, since the longer it takes to begin treatment, the more serious — and more contagious — the disease.
She is optimistic. But even if the results aren’t immediately apparent, she believes passionately in science for its own sake.
“Science is listening to the music of nature. It’s beautiful in itself,” she adds. “It doesn’t have to be useful. Even if it’s not applicable to anything specific, it increases knowledge, gives a greater understanding of how things work, of the complexity of even what seems simple.” And if this knowledge can help someone, it is a bonus, a great personal reward.
“My family had great difficulty understanding how I could become a doctor but not practice,” she acknowledges. “My interest wasn’t in the one person who comes to a physician’s office for help, but in the thousands of sick people who possibly can be helped by scientific research.”
Gennaro loves what she does. This mother of two tells her children to find something that they love to do, something that gets them up each morning eager to go wherever they will do their life’s work. “I have trouble understanding people who say, ‘Thank god it’s Friday’.” On her Fridays she is obviously already looking forward to whatever challenge Monday may bring.
New TB Diagnostics
Tuberculosis (TB) is an infectious disease that still kills millions of people every year. It targets young adults in particular, and thus undermines critical societal groups such as young workers and young mothers.
TB is most typically a pulmonary disease caused by the air-borne bacterial pathogen Mycobacterium tuberculosis. Infection of alveolar macrophages with M. tuberculosis initiates a series of attack and defense maneuvers by pathogen and host, the balance of which ultimately determines the outcome. It is exceedingly difficult to estimate how frequently innate immune responses fully resolve infection. Little doubt exists, however, that infection that is not cleared by innate immune mechanisms most commonly leads to an asymptomatic state. This is referred to as latent M. tuberculosis infection (LTBI). LTBI is very prevalent worldwide: an estimated one-third of the world’s population (two billion people) is infected with M. tuberculosis. In adults, active disease typically occurs when host immunity is temporarily or permanently impaired and, consequently, immune control of the infection is lost. The most dramatic example of such immune compromise is HIV co-infection, which increases the risk of active disease in LTBI cases from 5-10% in a lifetime to 10% per year.
Since LTBI is by definition clinically silent, and no bacteria can be recovered from respiratory secretions of infected individuals, the only marker of LTBI resides in the host immune response. In the early 19th century, Robert Koch, a pioneer of TB research, quickly realized that an M. tuberculosis cell extract, called tuberculin — that had been originally formulated as a vaccine — could be used as a diagnostic reagent. The principle of the test is that individuals who have already encountered tubercle bacilli mount a delayed-type hypersensitivity response when injected intradermally with tuberculin. The response is seen as an area of skin induration, the dimensions of which are used to determine whether a person is tuberculin-skin-test (TST) positive (i.e., infected) or negative (i.e., non-infected).
Albeit with some improvement, the reagent used for TST has essentially remained the same for more than a century. TST has been administered in billions of doses, but its drawbacks are evident. First, components of tuberculin are also found in environmental (non-pathogenic) mycobacteria, and in the attenuated strain M. bovis BCG that is used as TB vaccine. Given the high prevalence of non-tuberculous mycobacterial infection and the practice of universal BCG vaccination in TB-endemic countries, the specificity of TST is low.
Second, since the test measures delayed immune responses, results are read only two days after test administration. This requirement for two visits poses a strain on patients and health care providers, which often results in “loss” of the patient to the test. Third, since the test involves injection of antigen, repeated administration of the test may render a subject reactive to the antigen, thus leading to false positive results.
When my laboratory engaged in TB research, we recognized that finding new diagnostics was key to renewed TB control efforts. How can diagnosis of LTBI be improved? Differences might exist in terms of mediators of the immune response in natural infection vs vaccination. These would represent biomarkers of either condition. However, being trained in medical microbiology and prokaryotic molecular biology, I thought that the differentiating biomarkers should be sought on the bacterial side. At that time, genetic differences between the virulent strains and the attenuated vaccine strain were being unraveled by subtractive hybridization. These investigations pointed out to relatively large regions of DNA found in virulent M. tuberculosis but not in M. bovis BCG. One of them, called Region of Difference 1 (RD1), was particularly interesting because it was deleted from all substrains of the BCG vaccine.
I reasoned that if M. tuberculosis DNA encodes proteins that are not found in the BCG genome, and these proteins induce immune responses, infection with M. tuberculosis should be distinguishable from vaccination. Naturally infected individuals — but not BCG-vaccinated persons — would respond to these “differentiating” antigens. Working with a post-doc and a high-school student intern, we identified all the open reading frames (ORF) on the RD1 region, amplified and cloned them in expression vectors in E. coli, and purified proteins. Even though not all ORFs are genes, and not all heterologous genes are well expressed in E. coli, we expressed a good number of proteins encoded by the RD1 region. Having generated a battery of novel proteins, we had to test them to determine whether (i) they induced immune responses; and (ii) the immune response was specific for M. tuberculosis infection.
We chose guinea pigs for testing because these animals are highly susceptible to M. tuberculosis infection. Moreover, because of their skin color, guinea pigs are amenable to skin testing, once the fur is shaved off a portion of their back. Guinea pigs were divided in three groups, (i) infected with M. tuberculosis, (ii) immunized with M. bovis BCG, and (iii) controls (neither vaccinated nor infected). After an appropriate time interval, each novel RD1 protein (plus some controls) was planted in a separate area of the animal’s shaved skin. The responses were examined daily by investigators blinded to the identity of each protein injected. We found that among these novel proteins, some failed to induce any immune response while others induced responses in all groups, or no response at all.
However, a few proteins reacted only in the M. tuberculosis-infected but not in the BCG-vaccinated group. Thus “differentiating” antigens that evoke responses only in infected animals exist, as I had hypothesized. Some of these antigens have become part of new tests for LTBI diagnosis, the first since discovery of tuberculin