AIDS RESEARCH AND HUMAN RETROVIRUSES
Volume 21, Number 8, 2005, pp. 689–695
© Mary Ann Liebert, Inc.
Clinical, Microbiological, and Immunological Characteristics
in HIV-Infected Subjects at Risk for Disseminated
Mycobacterium avium Complex Disease: An AACTG Study
ROB ROY MACGREGOR,1RICHARD HAFNER,2JULIA W. WU,3ROBERT L. MURPHY,4
DAVID C. PERLMAN,5LUIZ E. BERMUDEZ,6CLARK B. INDERLIED,7LOUIS J. PICKER,8
ROBERT S. WALLIS,9,*JANET W. ANDERSEN,3LAURA F. MAHON,10SUSAN L. KOLETAR,11
DOLORES M. PETERSON,12AND THE ACTG PROTOCOL 341 TEAM
The clinical, microbiologic, and immunologic parameters in HIV-infected subjects first presenting with dis-
seminated Mycobacterium avium complex (DMAC) were determined. Four HIV-positive groups not yet on
DMAC treatment were enrolled: 19 subjects with CD4 lymphocyte counts ?50/?l thought to have DMAC on
clinical grounds; 18 subjects newly found to have a positive blood culture for MAC; 25 asymptomatic con-
trols (CD4 cell counts ?50); and 25 asymptomatic controls (CD4 counts 100–250/?l). Outcome measures in-
clude comparisons between groups for clinical characteristics; results of cultures from blood, marrow, and
gastrointestinal and respiratory tracts; immunological markers from staining of marrow and flow cytometry
of circulating lymphocytes; and cytokine production of PBMCs. Only 21% of the 19 patients entered on sus-
picion of having DMAC grew MAC from blood or marrow. Neither clinical presentation nor laboratory tests
differentiated those culture-positive from those culture-negative patients. However, prior PCP or multiple
other opportunistic infections were more common in the DMAC group. MAC was isolated from 82% of mar-
row and 50% of blood specimens from the DMAC group. Respiratory or gastrointestinal colonization was
present in 36% of DMAC subjects, but only 5% of non-DMAC subjects with CD4 counts ?50 cells/?l. CD8?
cells were more frequent in bone marrow, and CD4 cells recognizing MAC antigen were more frequent in
blood from DMAC subjects vs. controls. Results suggest an early stage of tissue dissemination preceding per-
sistent bacteremia, and mucosal entry without persistence of colonization. MAC-specific T cell responses ap-
parently develop and persist during DMAC, but are dysfunctional or too infrequent to prevent persistence.
with very low numbers of CD4-positive lymphocytes.1,2Before
highly active antiretroviral therapy (HAART) became available,
ISSEMINATED INFECTION WITH Mycobacterium avium com-
plex (DMAC) is a late complication of AIDS, associated
a high proportion of individuals with CD4-positive lymphocyte
counts below 50 cells/?l developed DMAC, with risk roughly
doubling with each 10 cell/?l decrement below 50.3At post-
mortem examination, over 50% of patients with AIDS had ex-
tensive tissue involvement with MAC.2However, making a di-
agnosis of DMAC on a clinical basis is difficult, and several
1Infectious Diseases Division, University of Pennsylvania, Philadelphia, Pennsylvania 19104.
2Treatment Research Program, Division of AIDS, National Institute of Allergy and Infectious Diseases, NIH, Bethesda, Maryland 20892.
3SDAC, Harvard School of Public Health, Boston, Massachusetts 02115.
4Northwestern University, Chicago, Illinois 60611.
5Beth Israel Medical Center, New York, New York 10038.
6Department of Biomedical Sciences, College of Veterinary Medicine, Oregon State University, Corvallis, Oregon 97331.
7Department of Pathology, Keck School of Medicine, University of Southern California, Children’s Hospital, Los Angeles, California 90027.
8Vaccine and Gene Therapy Institute, Oregon Health & Science University, Beaverton, Oregon 97005.
9Case Western Reserve University, Cleveland, Ohio 44106.
10AACTG Operations Center, Silver Spring, Maryland 20910.
11Ohio State University Hospitals, Columbus, Ohio 43210.
12University of Texas Southwestern Medical School, Dallas, Texas 75390.
*Current address: Infectious Diseases Division, UMDNJ—New Jersey Medical School, Newark, New Jersey 07103.
aspects of the natural history of the infection are still not clear.
For example, is the risk for DMAC infection associated with
the colonization of respiratory and gastrointestinal mucosa? Or
does a period of localized or disseminated MAC tissue infec-
tion precede the phase of sustained bacteremia?4–6
Despite recognition that disseminated disease develops only
after profound CD4?lymphocyte depletion, the specific im-
mune mechanisms that prevent DMAC and are lost in advanced
AIDS patients are not defined. Key defense mechanisms against
mycobacterial infection include CD4-positive and CD8-
positive lymphocytes, macrophages, and various cytokines. In-
terferon-?, interleukin (IL)-2, IL-12, granulocyte colony-stim-
ulating factor (G-CSF), granulocyte-macrophage colony-
stimulating factor (GM-CSF), and tumor necrosis factor-?
(TNF-?) are all believed to enhance the immune response
against MAC.7,8In contrast, IL-6 and TGF-? have been impli-
cated as facilitating MAC growth extracellularly and within
macrophages.9,10Clinical studies to examine these immune ef-
fectors in the tissues of subjects at risk for DMAC may increase
our understanding of the normal mechanisms of immune pro-
This prospective study was designed to examine several of
these issues: (1) identification of clinical predictors for the pres-
ence of DMAC, (2) whether tissue infection with MAC occurs
prior to sustained bacteremia, and (3) determining immune cor-
relates of risk for DMAC by comparing the immune profiles of
DMAC subjects, symptomatic culture-negative subjects, and
asymptomatic HIV-positive control groups.
MATERIALS AND METHODS
The study enrolled several types of subjects: one group had
CD4?cell counts below 50 cells/?l plus one or more signs and
symptoms suggestive of DMAC (chronic unexplained fever,
weight loss, elevated alkaline phosphatase, or anemia). Their
blood and bone marrow MAC culture status was not known at
entry. At analysis, they were divided into Group A—those
whose entry cultures grew MAC, and B—those culture nega-
tive for DMAC. To ensure adequate numbers of subjects with
DMAC, additional subjects newly diagnosed to have DMAC
by blood culture (but not yet on treatment) were recruited for
Group A. Two asymptomatic HIV-infected control groups were
also recruited, one having CD4?cell counts below 50 cells/?l
and a second with CD4?counts between 100 and 250/?l. Group
A subjects were culture positive for MAC in blood and/or bone
marrow. Group B had signs and/or symptoms suggestive of
DMAC and a CD4 count below 50, but had negative entry blood
and marrow cultures for MAC. Groups C and D were asymp-
tomatic HIV-infected controls, stratified by CD4 count. Group
C serves as the primary control group for the immunopatho-
logical analyses because subjects had a CD4 count range sim-
ilar to the DMAC subjects, yet did not have DMAC.
HIV-infected individuals aged ?18 years were eligible for
enrollment if they were receiving either stable or no antiretro-
viral therapy for ?6 weeks, had not taken DMAC prophylaxis
within the prior 6 weeks, and had no clinical or microbiologi-
cal evidence for acute infection or malignancy. Informed con-
sent was obtained from patients participating in this study, and
human experimentation guidelines of the U.S. Department of
Health and Human Services and those of the participant insti-
tutions were followed throughout.
Subjects’ medical histories and physical examinations fo-
cused on risk factors and possible signs and/or symptoms con-
sistent with DMAC. Standard laboratory studies, CD4- and
CD8-positive lymphocyte counts, and plasma HIV-1 concen-
tration were obtained at entry. In addition, the following mi-
crobiological and immunological specimens were obtained on
the day of entry.
Blood, bone marrow, stool, and respiratory specimens (one
from each site) were shipped overnight to the Non-Tuberculous
Mycobacterial Reference Laboratory at USC/Children’s Hos-
pital of Los Angeles. Blood was collected in SPS or ACD va-
cutainer tubes. One bone marrow core specimen was placed in
sterile nonbacteriostatic saline and a second in 10% formalin.
Stool specimens consisted of a rectal swab containing visible
fecal material or approximately 0.5 g of stool. Acceptable res-
piratory specimens included induced sputa, nasopharyngeal
wash, bronchial wash, or expectorated sputum.
Blood and bone marrow core specimens were cultured by
methods previously described.4,11Respiratory and stool speci-
mens were processed using the detergent C18-carboxypropyl-
betaine (CB18).11,12Aliquots of the CB18-treated specimen
were inoculated directly into a Bactec 12B vial with PANTA
supplement and Mitchison 7H11 agar, both in duplicate.4Stool
specimens were first treated with NALC, filtered through a
Spin-X II column (Corning CoStar) containing G-50 Sepharose
beads, treated with CB18, and inoculated as described for res-
All were performed in central laboratories using specimens
shipped overnight in insulated containers.
Bone marrow immunohistochemistry. These studies were
performed in Dr. Luiz Bermudez’ laboratory. Antibodies were
mouse antihuman monoclonals except for rabbit anti-BCG an-
tibody.13,14IL-12, tumor growth factor-? (TGF-?), interferon-?
(IFN-?), and TNF-? antibodies were obtained from R&D Sys-
tems (Minneapolis, MN). Anti-CD4, anti-CD8, antimacrophage
marker (NCL-LN5), goat antirabbit, horse antimouse IgG anti-
bodies, and peroxidase substrates DAB and AEC were from
Novacastra Laboratories Ltd. (Newcastle, England).14Anti-
CD56 and anti-CD45ro antibodies were from Pharmingen (San
Diego, CA) and rabbit anti-BCG antibody was from Dako
(Waco, TX). Immunohistochemical staining procedures were
based on the recommended protocol from the Vectastain Elite
ABC Kit mouse IgG (Novacastra Laboratories Ltd.). Marrow
biopsies were sectioned and prepared as previously de-
scribed.15–17Slides were examined at 100? magnification and
20 random fields were scored for the presence of positive cells
by at least two investigators, as follows: 1? ? 1–5 cells, 2? ?
MACGREGOR ET AL.
6–10 cells, 3? ? 11–20 cells, 4? ? 21 or more cells, and
?/? ? the entire slide was examined to find any positively
Peripheral blood cell surface immunophenotyping by flow
cytometry. Peripheral blood mononuclear cells (PBMC) were
isolated using CPT blood collection tubes (Becton-Dickinson
Biosciences, San Jose, CA). PBMC were processed for six-pa-
rameter (four-color) flow cytometric analysis in Dr. Louis
Picker’s laboratory, as previously described.19,20A broad array
of CD4, CD8, and ?/? cell surface markers was chosen for ex-
amination (see Table 1).
Single cell cytokine production following stimulation with
PMA/ionomycin and with specific antigens. Pan-T-cell cytokine
synthesis-defined subsets were determined by Dr. Picker, as
previously described.21Cytokines measured included IFN-?,
TNF-?, Il-2, and IL-4. For determination of antigen-specific
cytokine responses, PBMC were stimulated with or without ap-
propriately titered MAC (crude lysates provided by Dr. L.
Bermudez) and cytomegalovirus (CMV; BioWhittaker, Walk-
ersville, MD) antigen preparations, as previously described.22,23
Induced cytokine synthesis in whole blood cultures. Induced
expression of TNF-? and IFN-? was measured in whole blood
culture by Dr. Robert Wallis, as previously described.24TNF-?
was measured in 20 hr cultures, stimulated by Escherichia coli
026:b6 lipopolysaccharide (LPS, Sigma), 100 ng/ml. IFN-? was
measured in the presence and absence of phytohemagglutinin
A (PHA, Sigma) 5 ?g/ml; candida antigen (CASTA, Greer
Labs) 40 ?g/ml; M. avium sensitin (MAS, Statens Institute,
Copenhagen), or culture filtrate of M. avium LR114 5 ?g/ml,
prepared as previously described.24
Statistical comparisons for categorical data were performed
by Chi square and Fisher’s exact tests as appropriate. The
Kruskal–Wallis test was employed for continuous measure-
ments.25Pair-wise comparisons have been made for each group
vs. control group C (CD4 cells ?50/?l) based on Wilcoxon
tests. The Bonferroni procedure was used to set a significance
level of 0.05/3 ? 0.0167 for each individual test to ensure the
set of three tests has a significant level of 0.05. However, no
adjustment was made for the number of parameters that were
Classification and regression tree (CART) analysis, a form
of recursive partitioning, was used to explore and identify im-
munological factors that might differentiate Group A vs. C, B
vs. C, and D vs. C.26
Eighty-seven subjects were enrolled over a 27-month period
at 10 sites. Three sites (Ohio State—28%, UT-Southwestern—
28%, Penn—20%) accounted for 76% of enrollees. Subjects
were 92% male, 45% were black, and the median age was 37
years. Table 2 displays their baseline characteristics by group.
Group A (culture-proven DMAC) contains 22 subjects. Four of
them were from the group of 19 subjects enrolled because of
signs and symptoms compatible with DMAC and had MAC
isolated from bone marrow (4/4) or blood (1/4) at entry. The
other 18 in Group A were enrolled because of a positive blood
culture reported prior to study entry. Group B contains the 15
subjects from the “suspected DMAC” group who proved to be
The subjects with DMAC (Group A) did not differ from the
asymptomatic control Group C (CD4 ?50) in gender, age, race,
or IV drug use. Roughly half of each group was supposedly
taking antiretroviral therapy (41% of Group A and 52% of C).
However, the median CD4 cell count in Group A was 7 vs. 22
cells/?l for Group C (p ? 0.016), and DMAC subjects were
more likely than controls to have a history of prior candidiasis
(p ? 0.01), lower hemoglobin (p ? 0.001), and increased alka-
line phosphatase values (45% vs. 4%, p ? 0.001).
Groups A and B were compared to determine whether or not
clinical data at presentation could discriminate between the two
groups. The frequency of individual symptoms (fever, chills,
night sweats, diarrhea, weight loss, cachexia), mean log HIV-1
RNA values, frequency of hemoglobin ?8 g/dl, thrombocy-
topenia, elevated serum alkaline phosphatase, or history of an-
tiretroviral therapy did not differ between the two groups. In
contrast, 14/22 (64%) Group A subjects had a history of prior
Pneumocystis carinii pneumonia vs. only 3/15 (20%) Group B
subjects (p ? 0.02); similarly, a history of two or more prior
opportunistic infections (OIs) was more common in Group A
(68%) vs. Group B (40%) subjects.
Entry bone marrow cultures were positive in 18 of 22 Group
A subjects (82%); 10 of these marrow-positive subjects also
had positive blood cultures (56%) (Table 3). Of the four sub-
jects whose marrow cultures were negative, one had a positive
blood culture and three were negative. These three had been
positive in their local site laboratory during study screening
and had been granted an entrance exemption for exceeding the
CHARACTERISTICS OF PATIENTS WITH DISSEMINATED MAC
TABLE 1.LYMPHOCYTE CELL SURFACE MARKERS
MEASURED BY FLOW CYTOMETRY
Specific marker measured
CD4 absolute and percent
CD8 absolute and percent
HLA-DR (on monocytes) surface density
Gamma/delta absolute and percent
Naive absolute and percent (out of CD4?)
CD27(?) memory/total memory (out of CD4?)
HLA-DR? absolute and percent (out of CD4?)
CD25 bright? % (out of CD4?), CD25 Dim? %
(out of CD4?)
Naive absolute and percent (out of CD8?)
CD27(?) memory/total memory (out of CD8?)
CD45ro? memory/total memory (out of CD8?)
CD28(?) CD57(?)/total memory (out of CD8?)
HLA-DR? absolute and percent (out of CD8?)
CD38 MFI (mean fluorescent intensity on CD8?cells)
7-day maximum treatment for DMAC prior to entry. They are
included in the DMAC group (A) for analysis purposes because
they were likely to have the immunological profile that permits
the development of DMAC. No control subjects (Groups C and
D) had MAC identified in blood or marrow.
Overall, 8/22 Group A subjects were colonized in respira-
tory and/or gastrointestinal tract (six respiratory, seven gas-
trointestinal, Table 3). In contrast, only 1/15 and 1/25 in Groups
B and C, respectively, were colonized, each in the respiratory
tract only, despite CD4 counts ?50/?l. None of 25 Group D
subjects (CD4 counts 100–250 cells/?l) was colonized.
Bone marrow immunostaining
Subjects with DMAC and asymptomatic controls with CD4
counts below 50 had similar immunologic parameters, including
the generally low frequency of cells producing IL-12, TNF-?,
and TGF-?. However, significantly more CD8-positive cells
(p ? 0.011, Kruskal–Wallis test) and a trend toward more
CD45RO-staining memory T cells (p ? 0.1) were found in mar-
row specimens from DMAC subjects compared to the low CD4
controls (Group C). Classification and regression tree (CART)
analysis of Group A vs. C demonstrated that the combination
of positive staining for CD8?cells, IFN-?, and CD45RO mem-
ory cells differentiated the DMAC group from Group C con-
trols. All 10 subjects who had CD8 greater than a few and
IFN-? greater than none were from Group A, whereas 11 out
of 12 subjects who had zero-to-few CD8 cells and CD45ro ?10
cells/high-power field were from Group C. This analysis cor-
rectly predicted 80% (16/20) of Group A and 71% (17/24) of
Peripheral blood cell surface immunophenotyping by
Twenty-three cell phenotype markers were analyzed (Table
1). Owing to the low numbers of target cells in Groups A, B,
MACGREGOR ET AL.
TABLE 2.SUBJECTS’ DEMOGRAHIC, CLINICAL, AND LABORATORY CHARACTERISTICS
(n ? 87)
(n ? 22)(n ? 15)(n ? 25)(n ? 25)
Mean log viral load
80 (92%) 19 (86%)15 (100%) 24 (96%)22 (88%)0.328c
18 (21%) 10 (45%)5 (33%)1 (4%) 2 (8%)0.001c
5 (6%)3 (14%)2 (13%) 0 (0%)0 (0%)0.001d
aGroup A ? proven DMAC. Group B ? symptomatic, MAC culture-negative. Group C ? asymptomatic control, CD4 ? 50.
Group D ? asymptomatic control, CD4 100–250.
bGroup A vs. C; bonferroni significance cut-off ? 0.0167; unknowns excluded.
cFisher’s exact test.
BLOOD AND MARROW (GROUP A) AND
CULTURES GROWING MAC (ALL GROUPS)
MAC CULTURE RESULTS FROM
Marrow (Group A)
and C, it was not possible to perform complete flow analysis
for some subjects, thus reducing the power to detect statistical
differences between groups. Comparing Groups A and C, few
differences were found: the absolute number of circulating
CD8?lymphocytes was significantly lower among the DMAC
group (A) compared to the Group C controls (p ? 0.009,
Kruskal–Wallis test), and there was a trend toward a lower ab-
solute number of circulating ?/? T cells in the subjects with
DMAC (p ? 0.024). A trend was found for a higher number
and percent of CD8?HLA-DR?activated cells in the DMAC
group compared to Group C (p ? 0.045).
Single cell cytokine production following stimulation
with PMA/ionomycin or antigens
Groups A and C had similar numbers of CD4-positive, CD8-
positive, and ?/? lymphocytes responding to PMA/ionomycin
with IFN-?, IL-2, and IL-4 secretion. There was a trend among
TCR-?/? T cells toward diminished TNF-? secretion in the
DMAC group (Group A median ? 20.6%, Group C ? 54.5%,
p ? 0.1). The DMAC group had significantly more CD4?cells
secreting IFN-? in response to MAC antigen than did the con-
trol Group C (p ? 0.006); in contrast, the response to cy-
tomegalovirus (CMV) antigen did not differ between Groups A
Induced cytokine synthesis in whole blood cultures
High levels of TNF-? were produced in LPS-stimulated cul-
tures of blood of all subjects (Table 4). In contrast, the pro-
portion of subjects with detectable responses to the nonspecific
T cell stimulus PHA differed among groups, being greatest in
Group D. The frequency of this response in Group A (12%)
was less than in control Group C (40%, p ? 0.08). In contrast,
the proportions of MAC antigen responses in these two groups
were similar (36% vs. 57%, p ? 0.40), even though the level
of IFN-? production was significantly less in Group A.
Clinical suspicion was a poor predictor of DMAC. Only 4
of the 19 subjects (21%, CI ? 6–46%) enrolled because of
signs, symptoms, or routine laboratory findings suggestive of
DMAC had a positive culture of blood or bone marrow. Had
all suspected cases been treated empirically, 79% would have
received unnecessary therapy. The infected group (A) appeared
to be already highly vulnerable to opportunistic infections as-
sociated with profound loss of circulating CD4-positive cells
(68% had CD4 cell counts ?10/?l. vs. 40% of the symptomatic
but culture-negative Group B). Group A subjects more fre-
quently had prior PCP (64% vs. 20%), candidiasis (91% vs.
56%), and two or more prior OIs (68% vs. 40%) than Group B
Compared to the asymptomatic control group with CD4 cells
below 50/?l, the DMAC group had significantly more can-
didiasis, lower CD4 cell counts and hemoglobin, and higher al-
kaline phosphatase values, further indicating their profound de-
gree of functional immunosuppression.
Bone marrow biopsies and blood were obtained for culture
from subjects with suspected or recently diagnosed MAC dis-
ease to determine whether tissue infection occurs prior to sus-
tained bacteremia. Marrow cultures were more often positive
(82%) than concurrent blood cultures (50%) among DMAC
subjects, suggesting the greater utility of marrow culture in doc-
umenting DMAC. In a prior study comparing MAC burden in
blood and bone marrow in HIV-infected subjects, the bacterial
concentration in blood did not correlate with that in bone mar-
row, and the burden in bone marrow was substantially greater
than in blood.4In the current study, only one MAC blood-cul-
ture-positive subject was bone marrow negative. This finding
is compatible with a previous report of the infrequent occur-
rence of transient MAC bacteremia,27which perhaps occurs
only during the earliest phase of MAC infection. Such episodes
may seed bone marrow and other tissues, where MAC burden
then steadily increases. Although the MAC tissue burden may
be variable at this stage, most patients have substantial infec-
tion of the bone marrow, even though blood cultures may not
yet be consistently positive. In the final stage of infection, bac-
teremia becomes sustained and the tissue burden may increase
Colonization of gastrointestinal or respiratory tracts was
more common among subjects with DMAC than among those
whose CD4 counts were below 50 but whose marrow and blood
cultures were negative for MAC (36% in Group A vs. 5% in
Groups B and C). While the respiratory and gastrointestinal
tracts are the likely portals of invasion for MAC, colonization
CHARACTERISTICS OF PATIENTS WITH DISSEMINATED MAC
TABLE 4.INDUCED CYTOKINE SYNTHESIS IN WHOLE BLOOD CULTURES
LPS 17/17 (4050)14/14 (5655)24/24 (4193)24/24 (6853)
aNumber of responders/number tested (geometric mean of response, in pg/ml, excluding nonresponders from calculation).
bM. avium sensitin, Statens Institute.
is usually not detected at the time of presentation. These find-
ings indicate that the organisms may attach, invade the mucosa,
and enter the bloodstream in a relatively rapid manner, but do
not persist as mucosal colonizers.28,29
In situ immunostaining showed that DMAC subjects had
higher numbers of CD8?and CD45RO?cells in bone marrow
specimens than did asymptomatic low CD4-count controls.
Conversely, the absolute number of circulating CD8?lym-
phocytes was significantly lower among the DMAC group (A)
compared to the Group C controls. This suggests that these im-
mune effector cells were concentrating in infected tissue, but
in small numbers and of questionable effectiveness. Both the
DMAC Group and the low-CD4 controls had a low but simi-
lar frequency of cells staining for IL-12, TNF-?, and IFN-? (cy-
tokines associated with active antimicrobial cellular defense),
as well as for TGF-?, a possibly suppressive cytokine. Increased
IFN-? might be expected with the presence of infection but was
not observed despite an increased number of bone marrow CD8
T cells in Group A subjects. This observation suggests that the
function of CD8?lymphocytes was impaired, possibly related
to the extremely low numbers of CD4 cells in this group. CART
analysis indicated that positive staining for CD8?cells, IFN-?,
and CD45RO memory cells differentiated the DMAC group
from Group C controls. All 10 subjects with CD8 more than a
few and IFN-? more than none were from Group A. Produc-
tion of IFN-? was present in the bone marrow of the Group A
subjects with CD8 cells more than a few, although at a very
low level. Cell-associated IL-12 was not increased in the bone
marrow samples from DMAC patients. These results agree with
recent findings both in vitro and in mice that MAC infection
suppresses IL-12 production by macrophages.15,30
Analysis of single T cell cytokine production following non-
specific pan T cell stimulation showed that cells from all pa-
tient groups were capable of cytokine production, particularly
by CD8 cells. Groups A and C had similar frequencies of CD4-
positive, CD8-positive, and ?/? lymphocytes responding with
IFN-?, IL-2, and IL-4 secretion. The DMAC group had more
CD4-positive cells producing IFN-? in response to MAC anti-
gen than did controls, but no greater production induced by
CMV antigen. MAC-specific effector T cells secreting IFN-?
either develop with early infection or are maintained from pre-
vious exposure and appear to remain in small numbers at the
time of disseminated infection.
The whole-blood culture technique assessment of cytokine
production showed that TNF-? was well expressed in response
to LPS by cells from all subjects, regardless of clinical group.
Subjects in the control groups produced relatively high levels
of IFN-? following stimulation with MAC antigen, although
only approximately half of those in either group had responses.
Of interest, the proportion of patients in Groups A and C who
responded to MAC antigen did not differ statistically, but the
magnitude of response was significantly lower in Group A.
These low-level MAC IFN-? responses in some DMAC sub-
jects indicate that very low numbers of antigen-responsive
CD4?T cells persist in these patients, and are consistent with
the results of their MAC antigen-induced IFN-? expression
measured by flow cytometry.
HIV-infected individuals with advanced immunosuppression
maintain small populations of T cells (especially CD4?T cells)
that produce cytokines in response to MAC antigen. MAC-spe-
cific clinical immune recovery and the occurrence of episodes
of immune reaction inflammatory syndrome occur relatively
rapidly in subjects with DMAC after successful initiation of
HAART.31–33This restoration of effective immunity against
MAC infection is consistent with the prompt expansion and
functional improvement of the residual MAC-reactive T cells
in response to antiretroviral therapy.
We thank Agouron, Inc. for supplying the nelfinavir,
Boehringer-Ingelheim Pharmaceuticals, Inc. for providing the
nevirapine, and Pfizer, Inc. for providing the azithromycin used
in the second stage of the study (the focus of a subsequent re-
port). Most of all, we are grateful to all of the subjects who
gave their time, effort, and specimens (including bone marrow)
so that the study could be accomplished. Funding support: Na-
tional Institutes of Allergy and Infectious Diseases, NIH, AIDS
Clinical Trials Group, Grant U01 AI38858.
Investigators of record: Diane Havlir, M.D. (University of
California, San Diego); Richard Reichman, M.D. (University
of Rochester); Michael Dube, M.D. (University of Southern
California); Peter Tsang, M.D. (St. Vincent’s Hospital);
Mitchell Goldman, M.D. (Indiana University); John Phair, M.D.
(Northwestern University); Clifford Gunthel, M.D. (Emory
Site acknowledgments: Linda Meixner, R.N. and Gary Dyak
(UCSD); Mary Shoemaker, R.N. (University of Rochester);
Connie Funk, R.N., B.S.N. (University of Southern California);
Diane Gochnour, R.N. (Ohio State); Jean A. Craft, R.N.C.S.,
M.S.N. (Indiana University); Pam Donath, R.N. (Northwestern
University); Ericka Patrick, R.N., B.S.N. and Tedra Flynn, B.S.
(Emory University); Doris Shank, R. N. (University of Penn-
sylvania); Robert Barber, R. N. (University of Texas South-
western Medical School).
1. Jacobson MA, Hopewell PC, Yajko DM, et al.: Natural history of
disseminated Mycobacterium avium complex infection in AIDS. J
Infect Dis 1991;164:994–998.
2. Hawkins CC, Gold JWM, Whimbey E, et al.: Mycobacterium
avium complex infections in patients with the acquired immuno-
deficiency syndrome. Ann Intern Med 1986;104:184–188.
3. Nightingale SD, Byrd LT, Southern PM, Jockusch JD, Cal SX, and
Wynne BA: Incidence of Mycobacterium avium-intracellulare
complex bacteremia in human immunodeficiency virus-positive pa-
tients. J Infect Dis 1992;165:1082–1085.
4. Hafner R, Inderlied CB, and Peterson DM, et al.: Tissue burden
in disseminated MAC infection. J Infect Dis 1999;180:438–447.
5. Torriani FJ, McCutchan JA, Bozzette SA, Grafe MR, and Havlir
DV: Autopsy findings in AIDS patients with Mycobacterium avium
complex bacteremia. J Infect Dis 1994;170:1601–1605.
6. Nichols L, Florentine B, Lewis W, Sattler F, Rarick M, and Brynes
RK: Bone marrow examination for the diagnosis of mycobacterial
and fungal infections in AIDS. Arch Pathol Lab Med 1991;115:
7. Bermudez LE: Potential role of cytokines in disseminated my-
cobacterial infections. Eur J Clin Microbiol Infect Dis 1994;(Suppl.
MACGREGOR ET AL.
8. Bermudez LE: Immunobiology of Mycobacterium avium infection.
Eur J Clin Microbiol Infect Dis 1994;November:1000–1006.
9. Shiratsuchi H, Toossi Z, Mettler MA, and Ellner JJ: Colonial mor-
photype as a determinant of cytokine expression by human mono-
cytes infected with Mycobacterium avium. J Immunol 1993;150:
10. Denis M and Gregg EO: Recombinant tumour necrosis factor-al-
pha decreases whereas recombinant interleukin-6 increases growth
of a virulent strain of Mycobacterium avium in human macro-
phages. Immunology 1990;71:139–141.
11. Thornton CG, MacLellan KM, Brink TL, et al.: Novel method for
processing respiratory specimens for detection of mycobacteria by
using C18-carboxypropylbetaine: Blinded study. J Clin Microbiol
12. Thornton CG, MacLellan KM, Brink TL Jr, and Passen S: In vitro
comparison of NALC-NaOH, Tween 80, and C18-carboxypropyl-
betaine for processing of specimens for recovery of mycobacteria
J Clin Microbiol 1998;36:3558–3566.
13. Wiley EL, Mulhollan TJ, Beck B, Tyndall JA, and Freeman RG:
Polyclonal antibodies raised against Bacillus Calmette-Guerin, My-
cobacterium duvalii, and Mycobacterium paratuberculosis, used to
detect mycobacteria in tissue with the use of immunohistochemi-
cal techniques. Am J Clin Pathol 1990;94:307–312.
14. Barbolini G, Bisetti A, Colizzi V, Damiani G, Migaldi M, and Vis-
mara D: Immunohistologic analysis of mycobacterial antigens by
monoclonal antibodies in tuberculosis and mycobacteriosis. Hum
15. Athanasou NA, Quinn J, Heryet A, Woods CG, and McGee JO’D:
Effect of decalcification agents on immunoreactivity of cellular
antigens. J Clin Pathol 1987;40:874–878.
16. Macon WR and Salhany KE: T-cell subset analysis of peripheral
T-cell lymphomas by paraffin section immunohistology and cor-
relation of CD4/CD8 results with flow cytometry. Am J Clin Pathol
17. Shi SR, Key ME, and Kalra KL: Antigen retrieval in formalin fixed
paraffin embedded tissues: An enhancement method for immuno-
histochemical staining based on microwave over-heating of tissue
sections. J Histochem Cytochem 1991;39:741–748.
18. Macon WR and Salhany KE: T-cell subset analysis of peripheral
T-cell lymphomas by paraffin section immunohistology and cor-
relation of CD4/CD8 results with flow cytometry. Am J Clin Pathol
19. Collins Jr RH, Sackler M, Pitcher CJ, et al.: Immune reconstitu-
tion with donor-derived memory/effector T cells after orthotopic
liver transplantation. Exp Hematol 1997;25:147–159.
20. Terstappen LWMM, Huang S, and Picker LJ: Flow cytometric as-
sessment of human T-cell differentiation in thymus and bone mar-
row. Blood 1992;79:666–677.
21. Picker LJ, Singh MK, Zdraveski Z, et al.: Direct demonstration of
cytokine synthesis heterogeneity among human memory/effector T
cells by flow cytometry. Blood 1995;86:1408–1419.
22. Waldrop SL, Pitcher CJ, Peterson D, Maino VC, and Picker LJ:
Determination of antigen-specific memory/effector CD4? T cell
frequencies by flow cytometry: Evidence for a novel, antigen-spe-
cific homeostatic mechanism in HIV-associated immunodefi-
ciency. J Clin Invest 1997;99:1739–1750.
23. Waldrop SL, Davis KA, Maino VC, and Picker LJ: Normal human
CD4? memory T cells display broad heterogeneity in their acti-
vation threshold for cytokine synthesis. J Immunol 1998;161:5284–
24. Wallis RS, Lederman HM, Spritzler J, et al.: Measurement of in-
duced cytokines in AIDS clinical trials using whole blood: A pre-
liminary report. Clin Diagn Lab Immunol 1998;5:556–560.
25. Neter J, Kutner MH, Nachtsheim CJ, and Wasserman W: In: Ap-
plied Linear Statistical Models, 4th ed. Richard D. Irwin, Burr
Ridge, IL, 1996, pp. 1096–1097.
26. Breiman L, Friedman JH, Olshen RA, and Stone CJ: Classifica-
tion and Regression Trees. Wadsworth International Group, Bel-
mont, CA, 1984.
27. Kemper CA, Havlir D, Bartok AE, et al.: Transient bacteremia due
to Mycobacterium avium complex in patients with AIDS. J Infect
28. Chin DP, Hopewell PC, Yajko DM, et al.: Mycobacterium avium
complex in the respiratory or gastrointestinal tract and the risk of
M. avium complex bacteremia in patients with human immunode-
ficiency virus infection. J Infect Dis 1994;169:289–295.
29. Doherty TM and Sher A: Defects in cell-mediated immunity affect
chronic, but not innate, resistance of mice to Mycobacterium avium
infection. J Immunol 1997;158(10):4822–4831.
30. Wagner D, Sangari FJ, Kim S, Petrofsky M, and Bermudez LE:
Mycobacterium avium infection of macrophages results in pro-
gressive suppression of interleukin-12 production in vitro and in
vivo. J Leukoc Biol 2002;71:80.
31. Zeller V, Truffot C, Agher R, et al.: Discontinuation of secondary
prophylaxis against disseminated Mycobacterium avium complex
infection and toxoplasmic encephalitis. Clin Infect Dis 2002;34:
32. Shafran SD, Mashinter LD, Phillips P, et al.: Successful discon-
tinuation of therapy for disseminated Mycobacterium avium com-
plex infection after effective antiretroviral therapy. Ann Intern Med
33. Kirk O, Reiss P, Uberti-Foppa C, et al.: Safe interruption of main-
tenance therapy against previous infection with four common HIV-
associated opportunistic pathogens during potent antiretroviral
therapy. Ann Intern Med 2002;137:239–250.
Address reprint requests to:
Rob R. MacGregor
University of Pennsylvania
502 Johnson Pavilion/6073
Philadelphia, Pennsylvania 19104
CHARACTERISTICS OF PATIENTS WITH DISSEMINATED MAC