Serum Antibody Reactivity to Recombinant mig and Whole Cell Antigens in Mycobacterium avium Infection

Zbl. Bakt. 284, 348-360 (1996) © Gustav Fischer Verlag, Stuttgart· Jena . New York

Serum Antibody Reactivity to Recombinant mig and Whole Cell Antigens in Mycobacterium avium Infection * GEORG PLUMl, MARTINA BRENDEN!, PAULO SANTOS 2 , EVELYN SCHWARZ!, ULRICH WAHNSCHAFFE3, GOTTFRIED MAUFF\ and GERHARD PULVERERI Institut fiir Medizinische Mikrobiologie und Hygiene der Universitiit zu Kaln, Cologne, Germany 2 Centro de Histocompatibilidade do Centro do Lusotransplante, Coimbra, Portugal 3 Medizinische Klinik und Poliklinik mit Schwerpunkt Gastroenterologie, Universitiitsklinikum Benjamin Franklin, Berlin;Germany 4 Hygienisches Institut Hamburg, Germany 1

Received March 4, 1996 . Accepted March 15, 1996

Summary

Mycobacterium avium is a significant opportunistic pathogen in immunocompromised patients. Moreover, the prevalence of infections in patients without known predisposing conditions has also been increasing in recent years. Patients would greatly benefit from early diagnosis of disseminated infection. Serodiagnostic tests have already been promising in tuberculosis and immunocompetent patients but studies in HIV-infected patients and humoral response to M. avium antigens resulted in conflicting data. We have evaluated the use of the phagocytosis-induced MIG protein of M. avium as a diagnostic antigen. Serum antibody levels of M. avium-infected, HIV-negative patients were significantly elevated for the recombinant MIG (p <::: 0.001) and' also for M. avium whole-cell antigens (p < 0.025) as compared to controls. In contrast, HIV-infected patients with disseminated M. avium infection demonstrated also elevated levels of antibody for the whole-cell antigen (p < 0.00001) but a decreased reactivity for the MIG antigen (p < 0.007). The recombinant antigen proved to have no cross-reactivity with M. tuberculosis antigens as antibody levels were decreased in tuberculosis patients (p < 0.001). Therefore, a simultaneous serological test using recombinant MIG and the whole cell antigens might be helpful in the sometimes problematic diagnosis of M. avium infections in patients without predisposing conditions.

Introduction

Mycobacterium avium is now the most common cause of disseminated bacterial infection in patients with AIDS in the United States (12). The same trend is probably present in other industrialized countries with a low incidence of tuberculosis. Due to

* Dedicated to Prof. Dr. H. Brandis on the occasion of his 80th birthday.

Humoral Response to Mycobacterium avium Antigens

349

progress made in the prophylaxis of Pneumocystis carinii pneumonia and the treatment of other opportunistic infections, more patients are likely to suffer from disseminated M. avium infection (11). Prevalence of M. avium infection depends on the diagnostic methods used for postmortem diagnosis resulting in significantly higher rates than diagnosis based on blood cultures (10, 17,23,26). This suggests that this infection is severely underdiagnosed at a time when patients would greatly benefit from treatment (4). In view of the recent achievements in therapy and prophylaxis (3, 5, 8, 9,14,20,31), the rapid diagnosis of MAC infections is increasingly important. However, velocity in detecting the presence of the organism can not be the only issue. Important for an adequate design of a therapeutic regimen may be the knowledge of the primary focus of infection, the number of organs involved and the bacterial load. Most commonly, blood cultures are used for diagnosis but these have several drawbacks. They provide no quantitative measures for the bacterial load, they have to be performed on multiple samples for satisfactory sensitivity, they need several days for obtaining positive results and they are costly. An economic test giving the clinician rapidly a quantitative or semiquantitative idea of the bacterial load in a given patient would be desirable and help to distinguish between bacterial colonization and invasion of the host. Serodiagnosis has been shown to be a rapid test for the detection of approximately 50% of cases with smear-negative pulmonary and extra pulmonary tuberculosis (29). However, attempts to establish a serological assay for M. avium infection in HIVinfected patients have failed so far (16, 19,27,30). The aim of the present study was to evaluate whether the use of an M. avium-specific antigen induced to express after phagocytosis would provide a helpful marker for the diagnosis of disseminated infections.

Materials and Methods Antigens

Whole-cell antigens were prepared from M. bovis BCG (Pasteur vaccine strain) and Mycobacterium avium strain # 5-8 (serotype 4, clinical isolate, kindly provided by A. Tsang, National Jewish Center for Immunology and Respiratory Medicine, Denver, CO, USA). The mycobacteria were cultured in Middlebrook 7H9 broth with 10% OADC enrichment (Difco), 0,05% Tween 80 and 0,2% glycerol at 37°C on a platform shaker until mid-log phase. Cells were harvested by centrifugation, pellets rapidly frozen and stored at -70°C until further use. After thawing, cells were lysed by sonication on ice in five intervals of 1 minute using a sonifier (Bransonic 250, Branson, Danbury, CT, USA). Cell walls and insoluble matter were sedimented by centrifugation at 20000 G one hour in a refrigerated centrifuge. Supernatant was filtered through a 0.22 11m filter to ensure sterility of the antigen solution. Construction of an expression plasmid for the production of recombinant MIG protein. The isolation of the mig gene has been reported previously (22). Based on the sequence information obtained, a recombinant plasmid was constructed for chemically induced expression of recombinant protein in E. coli and subsequent affinity purification from crude cell lysates through binding of the protein to a metal chelate adsorbent (13). Briefly, a 696 bp internal DNA fragment of the mig gene was obtained by Kpn I and Bsp HI digestion and was subcloned into an E. coli-expression vector to create a QIAexpress type III construct (Qiagen, Hilden, Germany) which places the 6 X His affinity tag at the C-terminus of the protein. For directed insertion of the mig coding sequence (CDS) into the expression vector, the recipient vector arms were prepared in the following manner: Plasmid pQE51 was digested with restriction enzvmes Kpn I and Bgl I and a 993 bp DNA fragment was isolated.

f

pQE51

/I PTre ')

- -I i

I~ +

I I

II

~ Q!I

qfQ

I Imig '

~ .... J'~~

II

""~ ~....

qfi ~

~ b

:!'

*'

1

~ § I.ll'

pQE17

71#

Fig. 1. Partial restriction of expression plasmid pGPC150 and the corresponding region of the M. avium genome. PTrc IPTG-inducible promotor of vector pQE51 (Diagen, Langen, Germany). The hatched area denotes the carboxyterminal affinity tag of six histidine residues. Vertical bars represent the approximate physical location of cleavage sites by restriction endonucleases.

Plasmid pGPC150

M. avium genome

<1:1"

~ ~

w

:-

~

",

S

a-

'"C

o

'o""

Humoral Response to Mycobacterium avium Antigens

351

Plasmid pQE17 was digested with restriction enzyme Bgl II,S' overhanging ends made blunt with T4 DNA polymerase and further digested with Bgll. A 2428 bp fragment was isolated, combined with the 993 bp pQE51 fragment and ligated to the 696 bp mig fragment, resulting in plasmid pGPC150 which was used to transform E. coli M15(pREP4) (Qiagen, Hilden, Germany) to Amp resistance (Figure 1). The correct insertion in plasmid pGPC150 relative to the translational reading frame beginning at the start codon provided by the vector sequence and to the C-terminal 6 X His tag was confirmed by DNA sequencing. Purification of recombinant MIG-6 x His fusion protein from E. coli lysates

E. coli MIS (pREP4, pGPC 150) cells were grown overnight in L-broth (24) containing ampicillin (100 !lg/mL) plus kanamycin (25 !lg/mL) to maintain the expression plasmid pGPC150 and the repressor plasmid pREP4. 25 mL of the overnight culture were used to inoculate one litre of fresh L-broth with antibiotics and incubated for two hours at 37°C with good aeration. Thereafter, isopropyl-~-D-thiogalactopyranoside (IPTG) with a final concentration of 2 mM was added to induce expression from the tac promoter present in the expression vector and the culture was continued as before for five hours. Bacterial cells were harvested by centrifugation at 4°C, washed in ice-cold sonication buffer, 50 mM sodium phosphate pH 7.8, 300 mM sodium chloride, 1 mM phenylmethylsulfonyl fluoride (PMSF). Cells were stored frozen at -70°C until further purification. Pellets of frozen cells were disrupted by sonication and centrifued to remove cell walls as described above. Ammonium sulfate was added to the crude lysate at a final concentration of 1 M and crude lysates were incubated on ice for 15 minutes. After centrifugation for 10 minutes at 20000 G, the clear supernatant was applied to a hydrophobic interaction chromatography column (HIC, Phenyl Sepharose 6 Fast Flow, Pharmacia) that had been equilibrated with buffer A (1 M (NH4hS04 in 50 mM sodium phosphate buffer pH 7.4). After extensive washing with buffer A, the recombinant antigen was eluted with a linear gradient between buffer A and buffer B (50 mM sodium phosphate buffer pH 7.4, 10%vo\lvo1 ethanol). HIC fractions containing the antigen were passed through a nickel-nitrilotriacetic acid (Ni-NTA) adsorbent column which was then extensively washed with buffer (50 mM sodium phosphate pH 6.0, 300 mM sodium chloride, 10% glycerol). The recombinant protein was eluted by an increasing gradient of imidazole buffer (0-0.5 M). Fractions containing the recombinant protein were dialyzed for 48 hours against phosphate-buffered saline (PBS) and then stored at -20°C for later coating of ELISA plates. ELISA

For enzyme-linked immunosorbent assays, antigen stock solutions were diluted in coating buffer (0.5 M NaHC01 pH 9.6) to a final protein concentration of 40 !lg/mL for the whole-cell antigens and 10 ng/mL for the recombinant MIG antigen. lOO!lL of diluted antigen suspension were dispensed in to 96-well immuno plates (Polysorp, Nunc, Roskilde, Denmark) and incubated overnight at 4°C. On the following day, the antigen suspension was removed, wells were filled with 200 !lL of blocking buffer and incubated for one hour at 20°C. Buffer was again removed and immuno assays using the whole cell antigens were performed as described by Wayne, L. G., et al. 1992 (28). Briefly, for each test of a human serum, a well coated with antigen and a well without antigen were each blocked for two hours at 20°C by the addition of 2 % normal equine serum in PBS. Plates were washed with 0.02% bovine serum albumin (BSA) in PBS. ELiSAs using the recombinant MIG antigen were blocked for one hour only with 0.5% BSA in coating buffer at 20°C. Before applying the human serum to both types of wells (with or without antigen), these were washed five times with washing buffer (PBS 0.03% Tween 80). Sera were diluted 1: 100 in washing buffer and 100 !lL of the dilution were dispensed into wells coated with the different antigens and the corresponding blank wells that were blocked only. After a two-hour incubation, sample fluid was removed and wells were washed five times with

352

G. Plum et al.

washing buffer. As secondary antibody, peroxidase-conjugated goat antihuman Ig (Sigma, St. Louis, USA) was diluted 1: 1000 in washing buffer and 100 III were applied for one hour. After washing five times as before and a final PBS wash, a freshly prepared substrate solution of 0.34 mg/mL 2,2' -azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) (Sigma) 0.06% H 20 2 vol/vol was added to each well. Absorbance A405 was read after 30 minutes and absorbance readings were corrected for the non-specific background by subtracting the absorbance ~O\ of the antigen-free wells from each corresponding antigen-coated well.

Study populations The tested human sera were divided into the following categories based on the clinical evaluation of the subjects: i) 30 healthy volunteers recruited from the medical staff of the University Hospital Cologne, ii) five HIV-negative patients with active M. avium infection, iii) 67 tuberculosis patients, iv) thirty HIV-infected patients CDC B21B3, v) 33 patients with Kaposi's sarcoma CDC C1IC2/C3, vi) seventeen HIV-infected patients with culturally confirmed disseminated M. avium infection CDC C3. Sera of individuals of categories iv) and v) were taken on entry into a previous study on the association of MHC genes and disease progression in HIV-infected individuals (21). Category vi) were HIV-infected patients hospitalized at the University of Cologne hospital for an AIDS-defining infection, neoplasma or other conditon (1) during an 18-month period. Multiple serum samples were obtained where possible. If a mycobacterial infection was suspected, bacteriological culture was performed from multiple sputum, stool and blood samples. The limited number of category ii) patients stems from the fact that active M. avium infection in HIV-negative individuals is a very rare event in Central Europe. In this category, three elderly patients had a chronic pulmonary infection and two children suffered from cervical lymphadenitis.

Tuberculin skin testing in healthy controls Cell-mediated immunity against mycobacterial antigens was tested in the control population by the Mantoux method. Ten units of Tuberkulin GT in a volume of 0.1 ml were injected intradermally on the volar aspect of the forearm. After 48 to 72 hours, the maximal induration was measured and recorded in millimeters by an experienced reader. Statistical analysis: Statistical significance of observed differences in reactivity between the investigated groups of sera were analyzed by the U-test of Mann and Whitney (18). Results

Reactivity of serum antibody with mycobacterial antigens in the healthy control population. The mean absorbance values for M. avium whole-cell antigen (MAV), M. bovis BCG whole cell antigen (BCG), and recombinant MIG antigen in this group were 0.28 OD (0.24),0.44 OD (0.38), and 0.12 OD (0.08), respectively (standard errors in brackets). Twenty nine of 30 individuals (95% percentile) had absorbance values below 0.81 for MAV, 1.16 for BCG and 0.28 for MIG which excellently corresponded to calculated 95% confidence limits based on the assumption of a standard distribution. The line in Figure 2A-Figure 2C indicates the 95 percentile of the healthy control group. PPD skin testing in the healthy control population. Ten of the thirty volunteers had skin reactions with indurations below 10 mm, 14 had indurations ranging from 10 mm to 14 mm, and 6 persons had reactions equal to or above 15 mm. There was no correlation between induration size and serum reactivity to the tested antigens (data not shown).

Humoral Response to Mycobacterium avium Antigens

353

~=======H=I=v=-n=e=g=at=iv=e========~1 ~1========H=IV==in=re=c=ti=o=n========~ M. tuberculosis infection

M. avium infection

Healthy controls

CDC 82/83

2000 I

CDC C1-3

CDCC3

Kaposi sarcoma

M. avium infection

I-

I

1600 ;+

1200

"

.• +. .;

'r-

------

..... I!.+

"

.... , ,

I

i

:,

!

•••

:

!. • • •

L~~''_.

'~~

1,--"':....--'

Fig.2A. 10-3 arbitrary units of A 405 absorbance values from serum antibody levels against MAV antigen. Each dot represents the absorbance value of an individual patient. The line indicates the 95 percentile of the healthy control group.

~======H=I=v=-n=e=ga=ti=ve======~1 I M. tuberculosis mfection

M. avium infection

Healthy controls

r

" I l~_.,

_

..: I

, " --

f "

,"

400

~

I 0

,

•

L

:

!

,"

I +1

, ....~!- . '"

"

____

I,

' '

,I ~-

.

·1 I

~------c

1---------

I

, ,

'"

BOO

CDC C3

M. avium infection

Kaposi sarcoma

I

1600

1200

CDC C1-3 CDC 82/83

----

2000

HIV infection

,I i

,

, ,

I

t': ', ' l:;'~' 'iJ V i, ' 'I

,

I

,

,

"

I

.:.... ..

, ,

" ,

,, " ,

Fig. 2B. 10-3 arbitrary units of A 40 \ absorbance values for serum antibody levels against BeG antigen. Each dot represents the absorbance value of an individual patient. The line indicates the 95 percentile of the healthy control group. 23

Zbl. Bakt. 284/2-.;

354

G. Plum et al.

~=======H=lv=-n=e=g=at=iv=e======~1 M. tuberculosis infection ,'::_c::-cc: __ ===-~===::;-;==~ Healthy controls

2000

M. avium infection

I

HIV infection

I CDC 82/83

CDC C1-3

CDCC3

Kaposi sarcoma

M. avium infection

'

1600

1200

800

..

400

o

l' .:........' T:' • .. .

L_~'

__ _

l,~1J

. . ...

.. . . . . .' .. . .+... ... . '. ........... . -

#

• . . . . . . . . +••

Fig. 2e. 10-1 arbitrary units of A40S absorbance values for serum antibody levels against MIG antigen. Each dot represents the absorbance value of an individual patient. The line indicates the 95 percentile of the healthy control group.

Reactivity of serum antibody to mycobacterial antigens in patients with active M. avium infection without a known predisposing condition. The mean absorbance values for this group were MAV A40s 1.21 (0.88), BCG A405 0.96 (0.76), and MIG ~05 0.58 (0.53). Three of the five patients had higher absorbance values for MAV and BCG than the control 95 percentile but only the values for MAV antigen and MIG antigen, which was elevated in four of the five patients, reached statistically significant values (Table 1). Reactivity of serum antibody to mycobacterial antigens in patients with active M. tuberculosis infection. To study the effect of conserved antigens on the serum reactivity, 67 serum samples from tuberculosis patients were tested for the same antigens. As expected, a significantly higher prevalence of the whole-cell antigens was observed. With a mean absorbance value of A405 0.83 (0.43) and 36 out of 67 serum samples above the 95 percentile, the reactivity towards MAV was even more increased than was the reactivity towards BCG with a mean absorbance value of A405 0.71 (0.50) and 15 out of 67 serum samples above the 95 percentile. These sera had significantly decreased antibody levels against MIG (Table 1). All serum samples were well below the median value (A 4I1S 0.23) of the control group. Reactivity of serum antibody to mycobacterial antigens in patients with HIV infection. In this population, with M. avium infection being the most frequent single cause of opportunistic bacterial infection, three groups of sera were formed on the basis of the clinical and laboratory data. The individual values for each antigen and each group are given in Figures 2A, 2B and 2e. CDC B2/B3 patients showed no significant differences in their reactivity as compared to the controls. AIDS patients with Kaposi's sar-

278 242 225

441 380 326

119 75 117

MAV MEAN STDEV MEDIAN P (group vs. conti

BCG MEAN STDEV MEDIAN p (group vs. conti

MIG MEAN STDEV MEDIAN P (group vs. conti

Controls

576 527 370 < 0.001

963 760 1323 ns

1209 879 1721 < 0.025

MAV+ HIV-

18 24 9 < 0.001

709 498 675 <0.002

826 426 866 < 0.00001

TB

158 107 172 ns

326 304 249 ns

288 286 187 ns

CDC ClKS

188 145 167 <0.05

232 312 147 <0.01

378 269 355 ns

----~----"".

.

CDC BS/B3

82 84 64 <0.007

375 291 234 ns

916 448 940 < 0.00001

CDCC3 MAV+

Table 1. 10-3 arbitrary units of A 405 absorbance values for serum antibody levels against MAV, BCG and MIG antigens. MEAN, arithmetic mean; STDEV, standard deviation; MEDIAN, median; controls, healthy controls; MAV+ HIV-, HIV-negative patients with active M. avium infection; TB, HIV-negative patients with active M. tuberculosis infection; CDC B2/B3, HIV-infected patients without AIDS-defining clinical condition as defined by CDC 1993 (1). CDC ClKS, AIDS patients with Kaposi's sarcoma as leading diagnosis at the time of serum sampling; CDC C3 MAV+, AIDS patients with confirmed, dissemiated M. avium infection

v, v,

w

'"

:::s

~.

6'"

;t

it

I:> <:!

I·

~

o

r

8'"

'"

:::s

"0

'"o

~

e.

3 o ....

::r: ::

356

G. Plum et al. 10.3 au

----I MAV BeG

1---- MIG_

1500

1000 '------.

-0--

i ----

------------------.---~-. ----------I

500

o .-o

--------.---------_ ..-

+ 100

200

days

Fig. 3. 10-1 arbitrary units of A405 absorbance values for serum antibody levels against MAV, BCG and MIG antigens of one HIV-infected CDC C3 patients with disseminated M. avium infection. First positive bacteriological culture was obtained on day o.

coma and no clinical signs of M. avium infection at the time of serum sampling had a significantly decreased reactivity to BCG, with a mean A405 of 0.23 (0.31). The reactivity to the MAV whole-cell antigen was not significantly altered. In contrast to these data, the reactivity to the recombinant MIG antigen was significantly increased in seven of the 33 patients showing values above the 95 percentile. During the 18-month survey period on CDC C3-patients (category vi), 17 individuals developed disseminated M. avium infections confirmed by multiple culture. Fifty-seven serum samples were collected, resulting in observation periods for individual patients ranging from one week to 388 days. For comparison with the other groups and for performing statistical tests those sera were selected which had been taken closest to the date of the first M. avium culture in the respective patient. The reactivity of the group to MAV was elevated, with a mean absorbance value of A405 0.92 (0.45) and sera from 11 out of 17 patients showed values above the 95 percentile. However, MIG reactivity was significantly decreased as compared to controls or HIV-infected CDC B2/B3-patients and CDC ClIC3-Kaposi's sarcoma patients. Serum antibody reactivity of the M. aviuminfected patients showed some variation with the time if multiple consecutive samples were obtained but a general tendency towards lower activity was obvious during the progression of the HIV disease. One example of the longitudinal survey is given in Figure 3.

Discussion Sera of the control population showed a limited reactivity to M. avium whole cellantigen (MAV) that had been prepared by the method of Wayne et al. (28). This limited reactivity was paralleled by a low level of reactivity to the recombinant MIG. Per-

Humoral Response to Mycobacterium avium Antigens

357

sons with above-average reactivity to MAV generally also had a higher reactivity to BCG (correlation coefficient 0.74). The well-known dissociation between cellular and humoral response to mycobacterial antigens is reflected by the discordant reactivity levels in the PPD skin reaction and the serum antibody reactivity. If current CDC criteria (2) were applied, 17% of the controls had a positive PPD-test result which corresponded well to the median annual conversion rate for medical staff (9). For establishing the validity of the serological method, we first evaluated sera from patients with confirmed mycobacterial infection but without known immunocompromising conditions. Sera from tuberculosis patients or patients with active M. avium infection without HIV infection were found to have significant serum antibody reactivity to the whole-cell antigens. However, when using these antigens, the capacity to differentiate between the responsible pathogens was poor. The high prevalence of antibodies against M. avium antigens in tuberculosis patients and vice versa is due to the high number of cross-reacting epitopes between species of the genus Mycobacterium. The aim of this study was to evaluate an M. avium-specific purified antigen for a rapid test that would be useful for the diagnosis of this infection. The mig gene is a macrophage-induced gene (22) present in the M. avium serovars most frequently isolated from AIDS patients. Evidence gained from experiments in a macrophage infection model has shown that expression of MIG protein is specifically induced in the intraphagosomal compartment. This could make this antigen an excellent marker to differentiate between humoral responses to colonization or to the invasion of the host. This hope was supported also by findings of Morris et al. (19) that recombinant antigens might be useful as indicators of MAC diseases in HIV infection. Another important feature of the mig gene is its non-presence in the genome of M. tuberculosis complex strains. Therefore, MIG should have a higher species specificity than whole-cell antigens. We expected these characteristics to confer upon MIG a high potential as a specific serological tool in this infection. Construction of a recombinant plasmid for the expression in E. coli provided a powerful means for the production of this antigen. Placing the affinity tag at the C-terminus of the fusion allows the selection of fulllength proteins during the purification procedure by selective binding to a metalchelate adsorbent. A uniform protein was isolated from shorter contaminants which are caused by premature termination of translation due to the different codon usage of mycobacteria and E. coli. However, the application of this purified recombinant protein as an antigen in an immunoassay to determine specific serum antibodies to M. avium was disappointing. In patients who are at the highest risk of developing disseminated disease, no increase in specific antibodies to MIG was found. Patients with HIV-infection and classified as CDC B2/B3 did neither show significantly elevated antibody levels against whole-cell antigens nor the recombinant antigen. Only HIVinfected patients classified as CDC CVC3 with Kaposi's sarcoma as AIDS-defining condition showed significantly elevated levels of serum antibodies against MIG. A possible reason for this difference might stem from the fact that Kaposi's sarcoma patients are rather heterogeneous in respect of their immunosuppression, e. g. the CD4 cell number. On the other hand, Lee ct al. (16) did not see any association between helper T cell number and anti-M. avium glycopeptidolipid reactivity in HIV-infected patients. It has been hypothesized that a period of colonization precedes invasion by M. avium in HIV-infected patients (6). As these bacteria are ubiquitous in the environment (25), contact with the immune system is expected to occur constantly. Our serological data concerning the prevalence of antibodies against MIG and MAV in the healthy controls also indicates that these antigens are part of our natural environment.

358

G. Plum et al.

Interestingly, antibody reactivity against MIG was almost absent in tuberculosis patients. This might indicate that the stronger stimulus by M. tuberculosis eradicates the immune system's response to the M. avium-specific antigen. M. tuberculosis antigens may elicit such a strong cellular response directed towards many cross-reacting mycobacterial antigens that the activated macrophages eliminate all M. avium cells before they can express the MIG antigen. In our study, AIDS patients with confirmed disseminated M. avium infection had elevated levels of antibody against whole-cell antigen MAV but significantly decreased reactivity against MIG. We can only speculate which factors might be responsible for this discrepancy. From the longitudinal survey we know that the specific serum reactivity tend to decrease in these patients. The ability to produce these antibodies may have deteriorated to the point that the MIG protein is no longer active in stimulating a humoral response. However, as high reactivities to whole-cell antigens are maintained, different kinetics of the humoral response against MIG and MAV would have to be postulated. Alternatively, MIG protein may be aT-dependent B cell antigen and T-specific help may no longer be available at this stage. The latter is also a common finding for histopathological examination in tissues of these patients infected with M. avium (15) where granulomas are poorly defined and histiocytes are filled with mycobacteria. Yet another possibility is that a defect of the macrophages at this terminal stage of the HIV infection is responsible. MIG protein is only induced in functional phagosomes acidifying to pH 6.2 or lower (unpublished observation). The phagocytes of the severely immunocompromised patients (CDC C3) developing disseminated M. avium disease may be less capable of acidifying their phagocytic vacuoles and consequently are less exposed to MIG protein in their lymphatic tissue. In conclusion a combined immunoassay of whole cell antigens and MIG may be helpful to differentiate between M. tuberculosis and M. avium infections in patients without HIV infection. In AIDS patients, immunoassays with MIG antigen and/or M. avium whole-cell antigens failed to demonstrate reliable data for the diagnosis of M. avium infections. Acknowledgements. This work was supported by a grant from Bundesministerium fur Wissenschaft, Bildung, Forschung und Technologie (Germany), BMBF-Schwerpunkt Mykobakterielle Infektionen, awarded to C. Plum. Grant number 01 Kl 9313. The continuous support of our work by M. Schrappe, C. Fatkenheuer, and P. Hartmann, Klinik I fur Innere Medizin der Universitat zu Kaln, is gratefully acknowledged.

References 1. Centers for Disease Control and Prevention. Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. Morb. Mortal. Wkly. Rep. 41; RR-17 (1993) 1-19 2. Centers for Disease Control and Prevention. Guidelines for Preventing the Transmission of Mycobacterium tuberculosis in Health Care-Facilities, 1994; PPD Skin Testing and Anergy Testing. Morb. Mortal. Wkly. Rep. 43; RR-13 (1994) 59-63 3. Chaisson, R. E., C. A. Benson, M. P. Dube, L. B. Heifets, J. A. Korvick, S. Elkin, T. Smith, J. C. Craft, and F. R. Sattler: Clarithromycin therapy for bacteremic Mycobacterium avium complex disease. A randomized, double-blind, dose-ranging study in patients with AIDS. AIDS Clinical Trials Group Protocol 157 Study Team. Ann. Intern. Med. 121 (1994) 905-911

Humoral Response to Mycobacterium avium Antigens

359

4. Chin, D. P., A. L. Reingold, C. R. Horsburgh, Jr., D. M. Yajko, W. K. Hadley, E. P. Elkin, E. N. Stone, E. M. Simon-EM, P. C. Gonzalez, and S. M. Ostroff et al.: Predicting Mycobacterium avium complex bacteremia in patients infected with human immunodeficiency virus: a prospectively validated model. Clin. Infect. Dis. 19 (1994) 668674 5. Chium,J., J. Nussbaum, S. Bozzette, J. G. Tilles, L. S. Young, J. Leedom, P. N. Heseltine, and J. A. McCutchan: Treatment of disseminated Mycobacterium avium complex infection in AIDS with amikacin, ethambutol, rifampin, and ciprofloxacin. California Collaborative Treatment Group. Ann. Intern. Med. 113 (1990) 358-361 6. Collins, F. M.: Mycobacterium avium-complex infections and development of the acquired immunodeficiency syndrome: casual opportunist or causal cofactor? Int. J. Lepr. 54(1986)458-474 7. Dautzenberg, B., C. Truffot, A. Mignon, W. Rozenbaum, C. Katlama, C. Perronne, R. Parrot, andJ. Grosset: Rifabutin in combination with clofazimine, isoniazid and ethambutol in the treatment of AIDS patients with infections due to opportunist mycobacteria. Groupe d'Etude et de Traitement des Infections a Mycobacteries Resistantes. Tubercle. 72 (1991) 168-175 8. Dautzenberg, 8., C. Truffot, S. Legris, M. C. Meyohas, H. C. Berlie, A. Mercat, S. Chevret, and J. Grosset: Activity of clarithromycin against Mycobacterium avium infection in patients with the acquired immune deficiency syndrome. A controlled clinical trial. Am. Rev. Respir. Dis. 144 (1991) 564-569 9. Fagan, M.j. and G. A. Poland: Tuberculin skin testing in Medical students: A survey of U.S. Medical schools. Ann. Intern. Med. 120 (1994) 930-931 10. Hawkins, C. c., J. W. Gold. E. Whim bey, T. E. Kiehn, P. Brannon, R. Cammarata, A. E. Brown, and D. Armstrong: Mycobacterium avium complex infections in patients with the acquired immunodeficiency syndrome. Ann. Intern. Med. 105 (1986) 184-188 11. Hoover, D. R., A. j. Saah, H. Bacellar, j. Phair, R. Detels, R. Anderson, and R. A. Kaslow: Clinical manifestations of AIDS in the era of pneumocystis prophylaxis. N. Engl. J. Med. 329 (1993) 1922-1926 12. Horsburgh, C. R. Jr.: Epidemiology of mycobacterial diseases in AIDS. Res. Microbiol. 143 (1992) 372-377 13. Hochuli, E.: Purification of recombinant proteins with metal chelate absorbent. In: Genetic engineering, principle and methods U. K. Set/ow, ed.) Vol. 12., 87-98, Plenum Press, New York (1990) 14. Kemper, C. A., T. C. Meng, J. Nussbaum, J. Chiu, D. F. Feigal, A. E. Bartok, J. M. Leedom, j. G. Tilles, S. C. Deresinski, and j. A. McCutchan: Treatment of Mycobacterium avium complex bacteremia in AIDS with a four-drug oral regimen. Rifampin, ethambutol, clofazimine, and ciprofloxacin. The California Collaborative Treatment Group. Ann. Intern. Med. 116 (1992) 466-472 15. Klatt, E. c., D. F. Jensen, and P. R. Meyer: Pathology of Mycobacterium aviumintracellulare infection in acquired immunodeficiency syndrome. Hum. Pathol. 18 (1987) 709-714 16. Lee, B. Y., D. Chatterjee, C. M. Bozic, P.J. Brennan, D. L. Cohn, J. D. Bales, S. M. Harrison, L. A. Andron, and I. M. Orme: Prevalence of serum antibody to the type-specific glycopeptidolipid antigens of Mycobacterium avium in human immunodeficiency viruspositive and -negative individuals. J. Clin. Microbiol. 29 (1991) 1026-1029 17. Macher, A. M., j. A. Kovacs, V. Gill, G. D. Roberts, J. Ames, C. H. Park, S. Straus, H. C. Lane, j. E. Parrillo, and A. S. Fauci et al.: Bacteremia due to Mycobacterium aviumintracellulare in the acquired immunodeficiency syndrome. Ann. Intern. Med. 99 (1983) 782-785 18. Mann, H. B. and D. R. Whitney: Ann. Math. Statist. 18 (1947) 50 19. Morris, S. L., L. Bermudez, and S. D. Chaparas: Mycobacterium avium complex disease in patients with AIDS: seroreactivity to native and recombinant mycobacterial antigens. J. Clin. Microbiol. 29 (1991) 2715-2719

360

G. Plum et al.

20. Nightingale, S. D., D. w: Cameron, F. M. Gordin, P. M. Sullam, D. L. Cohn, R. E. Chaisson, L.]. Eron, P. D. Sparti, B. Bihari, and D. L. Kaufman et a!.: Two controlled trials of rifabutin prophylaxis against Mycobacterium avium complex infection in AIDS. N. Eng!. J. Med. 329 (1993) 828-833 21. Plum, G., E. Siebel, C. Bendick, H. Rasokat, M. Braun, H. Krickeberg, and G. Mauff: Major histocompatibility complex class I to III allotypes in patients with AIDS-related complexlWalter-Reed 5, disseminated Kaposi's sarcoma and in normal controls. Vox Sang. 59; Sl (1990) 15-20 22. Plum, G. and]. E. Clark-Curtiss: Induction of Mycobacterium avium gene expression following phagocytosis by human macrophages. Infect. Immun. 62 (1994) 476-483 23. Reichert, C. M., T.]. O'Leary, D. L. Levens, C. R. Simrell, and A. M. Macher: Autopsy pathology in the acquired immune deficiency syndrome. Am. J. Patho!' 112 (1983) 357382 24. Sambrook,j., E. F. Fritsch, and T. Maniatis: Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York (1989) 25. VonReyn, C. F., R. D. Waddell, T. Eaton, R. D. Arbeit, j. N. Maslow, T. W. Barber, R.]. Brindle, C. F. Gilks, j. Lumio, j. Uihdevirta, A. Ranki, D. Dawson, and]. O. Falkinham: Isolation of Mycobacterium avium complex from water in the United States, Finland, Zaire, and Kenya. J. Clim. Microbiol. 31 (1993) 3227-3230 26. Wallace, ]. M. and.J. B. Hannah: Mycobacterium avium complex infection in patients with the acquired immunodeficiency syndrome. A clinicopathologic study. Chest 93 (1988) 926-932 27. Wayne, L. G., L S. Young, and M. Bertram: Absence of mycobacterial antibody in patients with acquired immune deficiency syndrome. Eur. J. Clin. Microbiol. 5 (1986) 363-365 28. Wayne, L. G., D. Hollander, B. Anderson, H. Sramek, C. M. Vadheim, and]. I. Rotter: Immunoglobulin A (IgA) and IgG serum antibodies to mycobacterial antigens in Crohn's disease patients and their relatives. J. Clin. Microbio!. 30 (1992) 2013-2018 29. Wilkins, E. G. Land]. Ivanyi: Potential value of serology for diagnosis of extrapulmonary tuberculosis. Lancet 336 (1991) 641-643 30. Winter, S. M., E. M. Bernard,]. W. Gold, and D. Armstrong: Humoral response to disseminated infection by Mycobacterium avium-Mycobacterium intracellulare in acquired immunodeficiency syndrome and hairy cell leukemia. J. Infect. Dis. 151 (1985) 523-527 31. Young, L. S., L Wiviott, M. Wu, P. Kolonosk, R. Bolan, and C. B. Inder/ied: Azithromycin for treatment of Mycobacterium avium-intracellulare complex infection in patients with AIDS. Lancet 338 (1991) 1107-1109

Dr. Georg Plum, Institut fur Medizinische Mikrobiologie und Hygiene der Universitiit zu Koln, Goldenfelsstrage 21, D-50935 Koln, Germany