Potent skin-test cocktail’ may simplify diagnoses

May offer way to tell BCG from TB infection

A researcher working to find a better skin test for TB infection says a "cocktail" of antigens she’s created may offer a solution to some problems that have plagued researchers until now. The cocktail approach was developed by Marila Laura Gennaro, MD, associate member of the Public Health Research Institute, a public health research firm in New York City.

One big problem with tuberculins used in skin tests currently available is that they do not distinguish among infection by Mycobacterium tuberculosis (MTB), M. bovis, BCG, or other nonpathogenic environmental mycobacteria. Someone vaccinated with BCG, therefore, may react to a tuberculin skin test as if infected with MTB.

That’s not to say the purified protein derivative (PPD) used to test for TB infection is a poor reagent, says Gennaro. To the contrary, its sensitivity is exactly what makes it a good one. "Since it’s a mix of proteins, it contains all the epitopes needed to elicit a delayed-type hypersensitivity response. Each T-cell clone which has previously encountered the mycobacteria and been activated will be activated by a specific antigen."

The reason for tuberculins’ propensity to cross-react, making it so difficult to distinguish the source of infection, is that they contain proteins made by other mycobacteria as well as MTB; for that matter, they contain proteins made by many other bacteria in general, she adds.

Yet taken singly, a protein may not elicit any reaction at all. That leads to a stumbling block for many researchers: As they continue to isolate new proteins from TB, they keep finding antigens that work well in animal studies but fail to elicit the same kind of consistent reaction among humans.

There are many reasons for that disparity, Gennaro says. For one thing, the dose lab animals typically receive is proportionately much higher than that given to human test subjects. For another, the animals are tested only months after having been infected, whereas an antigen may be sensitive enough to react in humans years after exposure and infection have occurred.

Proceeding on the principle that more antigens will produce stronger reactions, Gennaro decided to produce a "cocktail" of antigens to see whether it would evoke a stronger response. "We took antigens already identified in the literature as being active in skin tests, or which we had identified ourselves."

She and her colleagues first tested them in animals immunized with BCG (a member of the MTB complex and similar to MTB but not pathogenic). Then they tested them in animals immunized with M. avium, a widespread environmental mycobacterium that is usually non-pathogenic; M. avium served as an example of a non-tuberculous mycobacterium.

Some antigens produced reactions in the first group of animals; others in the second group; and some in both. The next step was to make the antigen cocktails, using brews designed to elicit reactions from animals immunized with first one and then the other mycobacterium. As hoped, the cocktails retained the specificity of the antigens they contained; and as the number of antigens increased, so did the size of the reaction.

Having found what appears to be a reliable a way to distinguish mycobacterial from non-mycobacterial infection, Gennaro is hard at work on the next step: to find a mix of antigens that can distinguish among members of the same myco bacterial complex. "That’s what we’re doing now: working with antigens specific to BCG and MTB," she says. "Technically it’s no harder; it’s just a matter of finding the right antigens."

Gennaro presented her work, which also has been published,1 at International Business Communications’ International Conference on Myco bacterium Tuberculosis: Novel Therapeutic, held Dec. 10-11 in McLean, VA.


1. Konstantin L, Manca C, Collangeli R, et al. Use of Mycobacterium tuberculosis complex-specific antigen cocktails for a skin test specific for tuberculosis. Infection and Immunity 1998; 66:3,606-3,610.