Print this page

PAGES: 97-100 DOI: Full paper
Use of the 2,3-Diacyl-trehalose and the Purified Protein Derivative in the Serodiagnosis of Tuberculosis in AIDS

Maria Helena F Saad, Afranio L KritskiI, Eduardo Werneck-BarrosoI, Solange CavalcanteII, Monica Antônia S FerreiraIII, Leila de Souza FonsecaIII +

Laboratório de Hanseníase, Departamento de Medicina Tropical, Instituto Oswaldo Cruz
IServiço de Pneumologia, Hospital Clementino Fraga Filho, UFRJ 
IIHospital Evandro Chagas Av. Brasil 4365, 21045-900 Rio de Janeiro, RJ, Brasil
IIIInstituto de Microbiologia, Bloco I, Centro de Ciências da Saúde, UFRJ, 21941-590 Rio de Janeiro, RJ, Brasil


The effect of the human immunodeficiency virus (HIV) infection on IgG production against purified protein derivative (PPD) and 2,3-diacil-trehalose (SL-IV) was investigated by an enzyme-linked immunosorbent assay (ELISA) test. Comparison between the antigens showed that immunocompetent patients produce preferentially antibodies to SL-IV than to PPD (73.3% versus 63.3%). Combination of these results showed an increase of the sensitivity to 80%, which decreased over the spectrum of immunodepression caused by HIV. In the tuberculous HIV seropositive group the sensitivities of SL-IV and PPD were 36.4% versus 40% and 0% versus 22.2% in the tuberculosis/acquired immunodeficiency syndrome (TB/AIDS) group. Combination of these results gave respectively 54.5% and 20%, showing that serological tests have limited value for diagnosis of tuberculosis in HIV infected patients. High antibody levels were observed in HIV seropositive asymptomatic group, but only two individuals were positive for both antigens. In the follow up, one of them developed tuberculous lymphadenitis, indicating that further work is needed to access the value of serological tests in predicting tuberculosis in HIV infected individuals.

key words:

The overlap between human immunodeficiency virus (HIV) infection and tuberculosis (TB) has been shown by the high HIV seroprevalence among patients with active TB (Raviglione et al. 1992, Kritski et al. 1993). From a global perspective, TB represents one of the most common HIV-related opportunistic infections. HIV infection, by progressively impairing cell-mediated immunity, appears to be the highest risk factor for endogenous reactivation or exogenous reinfection of tuberculosis (Hopewel 1992). The diagnosis of TB in HIV infection may be difficult, specially in the acquired immunodeficiency syndrome (AIDS) phase, because of atypical presentation (for both pulmonary and extrapulmonary forms) and frequent sputum smear-negative cases. The delayed diagnosis of TB among HIV seropositive individuals contributes to high mortality rate and transmission of Mycobacterium tuberculosis infection, specially to contacts like other HIV-infected patients and healthcare workers (CDC 1990, Hopewel 1992). Current strategies for the TB diagnosis in developing countries are seriously inadequate (Raviglione et al. 1992). Therefore, a more rapid, sensitive and specific test is needed that may be used to diagnosis TB in HIV-infected patients. In the last few years the characterization and identification of new specific antigens have restimulated the interest in the serologic diagnosis of TB (Daffé et al. 1989, Laszlo et al. 1992, Papa et al. 1993). Among HIV seronegative TB patients, these serologic tests offer high sensitivity and specificity and may provide a wide clinical application (Daniel 1989, Wilkins & Ivanyi 1990, Cruaud et al. 1990). But, their possible benefits for improved diagnosis of TB in HIV seropositive patients have not been determined (Theuer et al. 1989, Hoeppner 1990, Berlie et al. 1991, Barrera et al. 1992).

The purpose of this paper is to report the results of TB serodiagnosis in patients with different levels of immune depression caused by HIV infection.



From July 1988 to December 1990, IgG assays were performed using an enzyme-linked immunosorbent assay (ELISA) in 224 individuals as follows: 81 healthly persons without laboratory evidence of HIV infection (controls, Group A); 37 asymptomatic HIV seropositive patients (Group B); 19 AIDS patients without clinical or laboratory evidence of TB (Group C); 45 TB patients without clinical or laboratory evidence of HIV infection (Group D); 22 TB patients infected by HIV without AIDS symptoms (TB/HIV) (Group E), and 20 AIDS cases (according to CDC-1987 criteria) with tuberculosis (TB/AIDS) (Group F).

All TB cases were bacteriologically confirmed by isolation and identification of M. tuberculosis using procedures recommended by reference laboratories. All patients had a physical examination that included an examination of extrapulmonar features possibly related to TB and/or HIV infection. The control group was select from the population attending the Chest Service. HIV serology status was determined in all patients after they had agreed to informed consent. Serum samples were tested for antibody to HIV by ELISA (Organon Teknika, Boxtei, Netherlands). All ELISA positive samples were analyzed by the Western blot immune assay (Dupont, Wilmington, DE, USA). The samples were considered positive when they reacted to both the ELISA and the Western blot test.

The antigens used in the ELISA were SL-IV (2,3-diacyl-trehalose), a specie-specific glycolipid from M. tuberculosis, kindly supplied by Dr Hugo David (Institut Pasteur, Paris, France), and PPD. The ELISA for SL-IV was performed as previously described (Cruaud et al. 1990). Briefly, the antigen was dissolved in hexane (2ml/ml, 100ng/well) and incubated overnight at 37oC. Blocked by the addition of 100ml of 5% bovine serum albumin (BSA) at 37°C for 2 hr, and washed with phosphate-buffered saline (PBS) pH 7.4. Sera were diluted at 1:320 in 1% BSA and 100ml were added to each well. After 1 hr of incubation and washing, 100ml of goat anti human IgG b-galactosidase conjugate from Biosys (Compiegne, France) diluted at 1:2000 in PBS-BSA were added. After 1 hr of incubation at room temperature and washing, 100ml of o-nitrophenyl-b-galactopyranozide (0.8 mg/ml, Sigma) and b-mercaptoethanol (90ml) prepared phosphate buffer (0.1 M K2HPO4, 1mM Mg SO4) were added. After 1 hr of incubation, 50ml of a solution at 32% of sodium carbonate were added and A405 values determined with a Titertek Multiskan apparatus (Flow Lab., USA). The incubation temperature was always 37°C. Polystyrene microtiter plates used were from Dynatech Lab. (USA) with round bottom. All sera were tested in duplicate and in each set of the experiments a reference positive and a negative poll were included. For each serum a control well was assayed in parallel without antigen. A modified procedure described by Fonseca et al. (1992) was applied in ELISA for PPD. Polystyrene microtiter plates with flat bottom were coated with 10mg/ml of PPD RT-23 diluted in carbonate buffer pH 9.6, and incubate overnight at 4°C. Wells were washed with PBS containing 0.1% of Tween 20 (PBST) and the remain sites blocked by addition of PBST containing 1% of BSA (PBST-BSA) and incubate at 37°C for 1 hr. Sera diluted at 1:320 in PBST-BSA and 100ml were added to each well. The following steps were performed as for SL-IV. In this study the cut off for PPD and SL-IV was established as the mean plus two standard deviations of the healthy control population sera results.

The chi-square (corrected for continuity), Fisher's exact, and Student's t-tests were used for statistical analyses. The efficiency of the serological method was evaluated by calculating the sensitivity and specificity.



The distribution of ELISA OD values in the detection of antibodies to PPD and SL-IV is depicted in Fig. The cut off points obtained with PPD and SL-IV antigens in the healthy population (Group A) were similar (0.2 and 0.25, respectively), however false positive results were detected only with PPD. Considering these results the specificity required for diagnostic purposes using SL-IV antigen was 100%, while for PPD was 97.6%.

The analysis of the Group B (HIV+) and C (AIDS patients), both without TB, showed similar specificity with PPD (73% and 75%). However, a statistical significant difference was found with SL-IV (73% and 94.7%, p<0.05) in the same groups (Table). Only two individuals from Group B were seropositive for both antigens and they showed no cutaneous hypersensitivity to PPD (Mantoux test), which is used as a criteria to initiate chemotherapy. The HIV positive individuals without TB (Group B) were followed up for 9 to 21 months. At the end of this period, one of two positive individuals for both antigens developed tuberculous lymphadenitis and the other remained asymptomatic.

Using the SL-IV antigen, the sensitivity of the method among group D (TB), E (TB/HIV) and F (TB/AIDS) was estimated to be, respectively, 73.3% (33/45), 36.4% (8/22) and 0% (0/20). When PPD antigen was evaluated the results among Group D, E and F were, respectively 63.4% (26/41), 40.0% (8/20) and 22.2% (4/18) (Table). The specificity of the SL-IV antigen, in Group D, E and F (using the Group A, B, C as control), was respectively: 100%, 73% and 94.7%. And using PPD antigen, these features were respectively: 97.6%, 73% and 75%.

The sensitivity obtained with both antigens (PPD and SL-IV) was 80.0% (36/45) in Group D, 54% (12/22) in Group E and 20.0% (4/20) in Group F.



Among HIV seropositive TB patients, specially in the AIDS phase, failure to make the diagnosis can have serious consequences; e.g., mortality rate of 45% versus 19% when an early diagnosis is made (Hopewel 1992). Conversely, blind treatment with anti-TB exposes the patient to the risk of unnecessary toxic effects of the drugs. In our study, the performance of the antigen SL-IV and PPD was different between HIV seropositive and seronegative patients. Using the above criteria, we observed a higher sensitivity with SL-IV than with PPD in the immunocompetent group (73.3% versus63.4%). Although immunocompetent patients produce more antibodies to SL-IV than to PPD, the differences between the two antigens were not significant (p>0.05). Our data confirmed the results obtained by Cruaud et al. (1990) on the evaluation of SL-IV antigen for case finding of TB in non-immunocompromised patients. The high specificity and high sensitivity achieved in the ELISA test based on the SL-IV in these groups were similar to those results observed by others, using other specific antigens (Daniel 1989, Wilkins & Ivanyi 1990).

The sensitivity of the SL-IV antigen (36.4%) in the detection of HIV seropositive TB patients (Group E) and in TB patients with AIDS (0%) (Group F), was similar to that achieved by Berlie et al. (1991), using the same antigen (29.4% and 6.8%, respectively). Nevertheless, our results with Group E and F were lower than that observed by Hoeppner (1990) (78% and 67%, respectively) using TB 72 monoclonal antibody probe which binds to the M. tuberculosis specific 38 kD antigen and by Rigous et al. (1990) (70% and 60%) using antigen 60.

As occurred with SL-IV, the sensitivity of PPD in the detection of TB decreased from immunocompetent tuberculous patients to tuberculous AIDS patients, but some cases with AIDS and TB showed some levels of the antibody (22.2%). Using the same nonspecific antigen, Barrera et al. (1992) and Theuer et al. (1989) observed similar results. Figueiredo and Machado (1992), using a crude antigen of M. tuberculosis, in the same group of patients obtained higher sensitive (37.8%).

For diagnostic purposes the SL-IV antigen is of limited value in HIV seropositive patients. One way in which serologic methods might be improved is by combining the results of two or more tests. We evaluated the sensitivity with both antigens (SL-IV and PPD). The sensitivity in HIV seropositive TB patients (54.5%) and in TB patients with AIDS (20%) was slightly lower than that observed by Berlie et al. (1991), using SL-IV associated to PGL-Tb1 antigens.

False positive results were observed with SL-IV and PPD antigens in similar proportion in HIV seropositive asymptomatic patients, but in AIDS patients without TB, the proportion of false positive results was lower with SL-IV antigen. In asymptomatic HIV positive group, only two individuals were seropositive for both antigens and they showed no cutaneous reactivity for PPD. These results led us to follow-up these patients in order to verify the predictive value of the positive serologic test as an early diagnosis of TB, as proposed by others (van Vooren et al. 1990, Berlie et al. 1991). After a period of follow-up ranging from 9 to 21 months, one developed tuberculous lympha-denitis and the other remained asymptomatic. These observations may indicate that the method might help in deciding to use preventive anti-TB therapy in HIV infected patients.

The present study show that serological tests for the detection of anti SL-IV and anti PPD antibody in HIV seropositive patients are of limited value for TB diagnosis. The sensitivity of the immunoenzymatic test for the diagnosis of TB in patients with AIDS was lower than that observed in HIV seronegative or seropositive TB patients. These features most likely result from the underlying immune defect of HIV seropositive patients, especially in the AIDS phase. On the other hand the positivity obtained with tuberculous patients HIV positive might be due to the polyclonal hyper-gammaglobulinemia of IgG and IgA classes and selective defects in antigen recognitions due to intrinsic lymphocytic alterations in these patients (Lane et al. 1985).

In conclusion, our study is preliminar and further work with other specific antigens needs to be done to determine: (a) the value of serologic tests for TB diagnosis in HIV seropositive patients, and (b) the value of predicting who will develop TB to provide a basis for justifying prophylactic drug regimen.



Barrera L, Kantor I, Ritacco V, Reniero A, Lopez B, Benetucci J, Beltran M, Libonatti O, Padula E, Castagnino J, Gonzalez-Montagner L 1992. Humoral response to Mycobacterium tuberculosis in patients with human immunodeficiency virus infection. Tubercle and Lung Disease 73: 187-191.

Berlie HC, Petit JC, David H 1991. Use of the SL-IV and the PGL-Tb1 glycolipid antigens in ELISA for the diagnosis of tuberculosis in AIDS patients. Zbl Bakr 275: 351-357.

CDC 1990. Guidelines for preventing the transmission of tuberculosis in health-care settings, with special focus on HIV-related issues. MMWR 39 (no. RR-17).

Cruaud P, Yamashita JT, Casabona NM, Papa F, David H 1990. Evaluation of a novel 2,3-Diacyl-Trehalose-2 Sulfate (SL-IV) antigen for case finding and diagnosis of leprosy and tuberculosis. Res Microbiol 141: 679-694.

Daffé M, Papa F, Laszlo A, David HL 1989. Glycolipids of recent clinical isolates of Mycobacterium tubercylosis: chemical characterization and immunoreactivity. J Gen Microbiol 135: 2759-2766.

Daniel T 1989. Rapid diagnosis of Tuberculosis Laboratory techniques applicable in developing countries. Rev Infec Dis Vol II (Suppl 2): S 471-S 478.

Figueiredo JFC, Machado AA 1992. Reduced anti-Mycobacterium tuberculosis antibody response in tuberculosis patients with acquired immunodeficiency syndrome. Braz J Med Biol Res 25: 611-618.

Fonseca LS, Kritski AL, Sant'Anna C, Janini MC, Ferreira MAS, Saad MHF 1992. Relationship between immune response and clinical manifestations in patients with tuberculosis. Braz J Med Biol Res 25: 49-52.

Hoeppner VH 1990. The antibody response to M. tuberculosis in AIDS. [Abstract]. Am Rev Respir Dis 143: A 26.

Hopewel PC 1992. Impact of Human Immunodeficiency Virus Infection on the Epidemiology, Clinical Features Management, and Control of Tuberculosis. Clin Infect Dis 15: 540-547.

Kristki AL, Werneck-Barroso E, Vieira MAMS, Carvalho ACC, Carvalho CE, Bravo-de-Souza R, Andrade GN, Galvão-Castro B, Castilho EA, Hearst NJ 1993. HIV infection in 567 active pulmonary tuberculosis patients in Brazil. AIDS 6: 1008-1012.

Lane IIC, Depper J, Greene WC, Whalen G, Waldman TA, Fauci AS 1985. Qualitative analysis of immune functions in patients with the acquired immunodeficiency syndrome. N Engl J Med 313: 79-84.

Laszlo A, Baer HH, Goren MB, Handzel V, Barrera L, Kantor IN 1992. Evaluation of synthetic pseudo cord-factor-like glycolipids for the serodiagnosis of tuberculosis. Res Microbiol 143: 217-223.

Papa F, Cruaud P, Luquin M, Thorel MF, Goh KS, David HL 1993. Isolation and characterization of serologically reactive lipooligosaccharides from Mycobacterium tuberculosisRes Microbiol 144: 91-99.

Raviglione MC, Narain JP, Kochi A 1992. HIV-associated tuberculosis in developing countries: clinical features, diagnosis, and treatment. Bull WHO 70: 515-526.

Rigouts L, Collart JP, Gigase P, Kadende P, Kaboto M, Perriens J, Willame JC, Prignot J, Portaels F 1990. Serological diagnosis of pulmonary tuberculosis and tuberculous adenitis in HIV positive and HIV negative patients from Central Africa (Zaire and Burundi). [Abstract]. Am Rev Respir Dis 143: A 266.

Theuer CP, Chaisson RE, Elias D, Schecter G, Glassroth J, Zeiss CR, Phair JP, Hopewel PC 1989. Detection of circulating antibodies to purified protein derivative in tuberculosis patients with and without human immunodeficiency virus infection. [Abstract]. Am Rev Respir Dis 139: A 395.

van Vooren JP, Farber CM, Motte S, De Bruyin J, Legros F, Yernault JC 1990. Assay of specific antibody response to mycobacterial antigen for the diagnosis of a pleural effusion in a patient with AIDS. Tubercle 69: 303-305.

Wilkins EGL, Ivanyi J 1990. Potential value of serology for diagnosis of extrapulmonary. Lancet 336: 641-644.

+Corresponding author
Received 12 December 1994
Accepted 1 September 1995
Grant support: CNPq and FINEP (Brazil)