Arch Pathol Lab Med—Vol 129, September 2005Marrow Stroma in HIV/AIDS—O’Malley et al
Evaluation of Stroma in Human Immunodeficiency
Virus/Acquired Immunodeficiency Syndrome–Affected
Bone Marrows and Correlation With CD4 Counts
Dennis P. O’Malley, MD; Joy Sen, MD; Beth E. Juliar, MA, MS; Attilio Orazi, MD, FRCPath
● Context.—Little is known about cellular and extracellular
composition of fibrosis in bone marrows in the context of
human immunodeficiency virus/acquired immunodeficien-
Objective.—To evaluate the stromal composition of
bone marrows affected by human immunodeficiency virus/
acquired immunodeficiency syndrome and to correlate this
with laboratory parameters including CD4 lymphocyte
Design.—We evaluated extracellular matrix and stromal
cell composition in bone marrows and correlated these re-
sults with hematologic parameters. Extracellular matrix,
stromal cells, and smooth muscle differentiation were eval-
uated by immunohistochemistry for collagen type IV ex-
pression and reticulin staining, an antibody directed
against low-affinity nerve growth factor receptor (a marker
of adventitial reticular cells), and actin staining, respec-
tively. Concurrent laboratory information was collected,
including white blood cell count, hemoglobin, platelet
count, CD4 count, CD8 count, CD4/CD8 ratio, and ab-
solute lymphocyte count.
Patients.—Bone marrows of 35 patients with human im-
munodeficiency virus/acquired immunodeficiency syn-
drome were evaluated.
Main Outcome Measures.—Correlation of reticulin,low-
affinity nerve growth factor receptor, actin, and collagen
IV staining with hematologic parameters.
Results.—More than half of the bone marrows showed
moderate to severe reticulin fibrosis. The degree of retic-
ulin fibrosis was correlated with the degree of low-affinity
nerve growth factor receptor expression (P ? .048). Actin
expression was identified in only 3 of 35 cases and collagen
IV in only 5 of 35 cases. No statistical relationship between
degree of fibrosis and CD4 count was identified. Lower
levels of low-affinity nerve growth factor receptor expres-
sion were associated with CD4 counts of ?100 (P ? .04).
Marrow fibrosis was present in almost all cases studied
(97%), and the staining of adventitial reticular cells cor-
related with the degree of reticulin fibrosis.
Conclusions.—There does not appear to be a correlation
between CD4 count and degree of fibrosis, suggesting that
the mechanism of fibrosis is independent of disease status.
(Arch Pathol Lab Med. 2005;129:1137–1140)
drome (AIDS) can be challenging. The constellation of
changes, termed HIV myelopathy, includes a wide range of
abnormal findings and can mimic a variety of primary
bone marrow disorders, including myelodysplastic syn-
dromes and myeloproliferative disorders.1In addition,
HIV/AIDS is associated with an increased incidence of
opportunistic infections, hematologic and nonhematologic
he bone marrow pathology of human immunodeficien-
cy virus (HIV)/acquired immunodeficiency syn-
Accepted for publication May 12, 2005.
From the Department of Pathology and Laboratory Medicine, Divi-
sion of Hematopathology (Drs O’Malley and Orazi), the Department
of Pathology and Laboratory Medicine (Dr Sen), and the Department
of Medicine, Division of Biostatistics (Ms Juliar), Indiana University
School of Medicine, Indianapolis.
The authors have no relevant financial interest in the products or
companies described in this article.
Presented in abstract form at the annual meeting of the United States
and Canadian International Academy of Pathology, Vancouver, British
Columbia, March 2004.
Reprints: Dennis P. O’Malley, MD, Indiana University School of
Medicine, Department of Pathology and Laboratory Medicine, Division
of Hematopathology, 702 Barnhill Dr, Riley 0969, Indianapolis, IN
46202 (e-mail: firstname.lastname@example.org).
malignancies, and other hematologic complications, which
typically worsen as the disease progresses.2
Bone marrow stroma is complex; it is a heterogeneous
mixture of cellular and extracellular components that con-
tribute to the unique microenvironment of the marrow
cavity. Bone marrow stromal responses are a common
pathway in both benign and malignant marrow disorders
and account for some of the most striking changes seen
in HIV/AIDS bone marrow. Defects of the bone marrow
microenvironment contribute to the development of he-
matologic abnormalities in HIV/AIDS.2,3It has been
shown that stromal elements are directly infected and may
act as a reservoir for the HIV virus.2–4The HIV virus di-
rectly infects a variety of stromal elements, including
monocytic/macrophage elements, microvascular endothe-
lial cells, and fibroblastic and myoid cells.2–5Infection of
the stromal elements leads to a loss of the hematopoietic
support function.6In addition, megakaryocytes, which
play an active role in the makeup of bone marrow stroma,
are also infected by the HIV virus, as demonstrated by the
presence of CD4 receptors on their surface.7Human im-
munodeficiency virus infection of bone marrow elements
may lead to loss of hematopoietic support function and
Arch Pathol Lab Med—Vol 129, September 2005Marrow Stroma in HIV/AIDS—O’Malley et al
Table 1. Antibodies Used to Characterize Bone
Antibody (Clone)Dilution Main Specificity
LNGFR (ME20-4)1:80 ARC, DRC, stroma cell
Collagen IV (CIV22)1:100
Actin HHF35 (1A4)1:100
* LNGFR indicates low-affinity nerve growth factor receptor; ARC,
adventitial reticular cells; and DRC, dendritic reticulum cells.
may be responsible for some of the hematologic abnor-
malities.3These changes may ultimately lead to cytopenias
and susceptibility to infections, which are significantprob-
lems in the HIV-positive population.
We attempted to characterize the cellular and extracel-
lular stromal reaction seen in the bone marrow in HIV/
AIDS, using an approach similar to that used to evaluate
stromal changes in fibrotic myeloproliferative disorders
and metastatic malignancy.8
The extracellular matrix was evaluated by reticulin
staining and by immunohistochemistry for collagen type
IV expression. Both reticulin fibers and collagen IV are
known to be markedly increased in fibrotic marrow states,
such as chronic idiopathic myelofibrosis.8–10Low-affinity
nerve growth factor receptor (LNGFR) has been recently
described as a specific marker for a subset of bone marrow
stromal cells known as adventitial reticular cells (ARC).11,12
In previous studies, the intensity of bone marrow LNGFR
staining has been found to correlate with the degree of
reticulin fibrosis.8–13Actin staining was used to identify
myofibroblastic differentiation in bone marrow stromal
cells.14Recent evidence has shown that bone marrow my-
oid cells may represent marrow reticular cells undergoing
cytoskeletal remodeling in response to various stimuli,
such as the presence of metastatic carcinoma.15
Because CD4?T-lymphocyte counts are regularly used
as a marker of disease progress and response to therapy
in HIV/AIDS, we attempted to test correlation of the ob-
served stromal responses with CD4?lymphocyte counts
and other hematologic parameters.
MATERIALS AND METHODS
Bone marrow biopsies of patients with HIV/AIDS were ob-
tained from the files of Wishard Memorial Hospital and Clarian
Health Partners (Indianapolis, Ind). In all, formalin-fixed, paraf-
fin-embedded bone marrow biopsies of 35 patients were studied.
In all cases, either significant hematologic abnormalities or pos-
sible involvement by infectious disease prompted bone marrow
examination. Concurrent clinical information was collected in-
cluding age, white blood cell count, hemoglobin, platelet count,
CD4 count, CD8 count, CD4/CD8 ratio, and absolute lymphocyte
count. There was no evidence on chart review of transfusions
within 7 days preceding the bone marrow biopsies.
Standard hematoxylin-eosin sections were evaluated in all cas-
es. Bone marrow stroma was evaluated using immunohistochem-
ical staining for collagen type IV (CIV; DakoCorporation, Car-
pinteria, Calif), actin (HHF-35; Dako), and LNGFR (ME20-4;
Dako) expression (Table 1) and histochemical staining for retic-
ulin. Density of stromal cell staining was graded semiquantita-
tively, as follows: 0? indicates absent or rare; 1?, focal weak
staining; 2?, focally intense or diffuse staining; and 3?, diffuse
Immunoperoxidase stains were performed using an automated
immunostainer (Dako), which employs a standard streptavidin-
biotin-peroxidase complex technique. Endogenous peroxidase ac-
tivity was blocked with 3% hydrogen peroxide in methanol, and
endogenous biotin activity was blocked using avidin and biotin.
The peroxidase activity was developed with 3,3-diaminobenzi-
dine and counterstained with hematoxylin. Controls were eval-
uated and stained appropriately in all cases.
The extracellular matrix was evaluated by reticulin staining,
using the Gomori technique, and by immunohistochemistry for
collagen type IV expression.
Stromal Cell Composition
Low-affinity nerve growth factor receptor staining was used to
identify bone marrow stromal ARCs. Actin staining was used to
evaluate myofibroblastic differentiation in bone marrow stromal
Statistical analyses were performed using LogXact (version 2.1)
(Cytel Software Corporation, Cambridge, Mass) and SAS (version
8.2) (SAS Institute Inc, Cary, NC) software packages using a P of
.05 for all tests performed. Comparison of CD4 (?100 or ?100)
and LNGFR/reticulin staining was performed using exact logis-
tic regression. Comparison of CD4 count as a continuous variable
and LNGFR/reticulin staining was performed using linear re-
gression. For analyses of CD4 count, LNGFR and reticulin stain-
ing were tested separately in univariate and together in multiple
regressions. The association between LNGFR and reticulin stain-
ing was analyzed using the Cochran-Mantel-Haenszel test.
Demographics and Hematologic Parameters
The patients included in the study consisted of 9 women
and 26 men. The patients’ ages ranged from 18 to 57 years,
with a mean age of 38 years. The mean hematologic pa-
rameters were as follows: white blood cell count, 4.3 ?
103/?L; hemoglobin, 10.6 g/dL; and platelet count, 140 ?
103/?L. The mean absolute lymphocyte count was 1014/
?L (normal range, 1400–3800/?L), with all but 1 patient
having a decreased CD4/CD8 ratio (?0.9) and a mean
CD4/CD8 ratio of 0.3. The CD4?lymphocyte count had
a wide range (1–928/?L) (normal range, 550–1600/?L),
with 21 of 35 patients having an absolute CD4 count less
than 100. We chose to compare patients with CD4?lym-
phocyte counts of less than 100 CD4?cells/?L with pa-
tients with 100 or more CD4?cells/?L. The CD8 counts
also varied widely (19–1961/?L) (normal range, 200–700/
?L), with an average of 572/?L. The results are summa-
rized in Table 2.
Hematoxylin-eosin–stained sections were
reviewed in all cases (Figure, A).
Reticulin Fibrosis and LNGFR.
all cases showed at least 1? reticulin fibrosis (0?, 1 case;
1?, 15 cases; 2?, 15 cases; and 3?, 4 cases). A major pro-
portion of cases (19/35, or 54%) showed moderate or se-
vere reticulin fibrosis (Figure, B). All cases but one showed
an increase in LNGFR staining (0?, 1 case; 1?, 15 cases;
2?, 15 cases; and 3?, 4 cases) (Figure, C). In cases with
increased reticulin fibrosis, the expression of LNGFR in
ARC was typically upregulated. The degree of reticulin
fibrosis paralleled the degree of LNGFR expression (P ?
Actin and Collagen IV .
was identified focally in only 9% (3/35) of cases (Figure,
With one exception,
Actin (HHF-35) expression
Arch Pathol Lab Med—Vol 129, September 2005 Marrow Stroma in HIV/AIDS—O’Malley et al
Table 2. Summary of Results*
(n ? 14)
(n ? 21)
* WBC indicates white blood cell; ALC, absolute lymphocyte count;
and LNGFR, low-affinity nerve growth factor receptor.
† Mean values of graded scores on a scale of 0? to 3?, as follows:
0? indicates absent or rare; 1?, focal weak staining; 2?, focally in-
tense or diffuse staining; and 3?, diffuse intense staining.
A, Human immunodeficiency virus/acquired immunodeficiency syndrome bone marrow. Hypercellular bone marrow with increased megakaryo-
cytes and plasma cells mimicking a myeloproliferative disorder (hematoxylin-eosin, oil immersion, original magnification ?500). B, Reticulin
staining. Increased number and diffuse distribution of reticulin fibrils corresponding to moderate to severe reticulin fibrosis (2? to 3?) (reticulin,
oil immersion, original magnification ?500). C, Low-affinity nerve growth factor receptor staining. Adventitial reticular cells with delicate cyto-
plasmic processes are highlighted. Some nonspecific staining is seen in plasma cells (immunohistochemistry, oil immersion, original magnification
?500). D, Actin staining. Large vessels were highlighted by actin staining, with no actin staining seen in adjacent marrow elements. The presence
of actin staining would suggest myofibroblastic differentiation of the marrow stromal elements (immunohistochemistry, oil immersion, original
magnification ?500). E, Collagen IV staining. Large vessels were outlined by collagen IV staining, highlighting basement membrane with some
weak staining seen in adjacent sinusoids. No staining was seen in marrow interstitial spaces (immunohistochemistry, oil immersion, original
D). Collagen IV demonstrated weak staining of vessel
walls in only 5 of 35 cases of HIV/AIDS marrow biopsies
(Figure, E). Linear staining was noted only in large-caliber
vessel walls. Focal staining of small-caliber vessels was
seen only rarely.
Correlation of Hematologic and Histologic Parameters
Statistical analysis did not reveal any correlation be-
tween CD4 count and the degree of reticulin fibrosis.
However, lower levels of LNGFR expression were associ-
ated with CD4 counts of 100 or more (P ? .04 in univariate
and multiple regression). In analyses of continuous CD4
count, decreased LNGFR was associated with increased
CD4 count (P ? .03 and .02 in univariate and multiple
A wide variety of hematologic abnormalities occur in
HIV/AIDS patients, and many of these arise as a result
of direct and indirect effects of the disease on the bone
marrow. Stromal changes in the bone marrow and their
relationship to hematologic parameters have not been
closely examined in previous studies of HIV/AIDS-af-
fected bone marrows.
We found that marrow fibrosis, as measured by reticulin
staining, was present in all cases studied. As in other stud-
ies, we found a statistically significant association between
the staining of ARCs and the degree of reticulin
fibrosis.11–13No statistical correlation between CD4?lym-
phocyte count and degree of reticulin fibrosis was iden-
tified. This appears to suggest that the mechanism of mar-
row fibrosis is independent of disease status, as measured
by CD4?lymphocyte counts.
No significant expression of collagen IV was seen in the
HIV/AIDS-affected bone marrows, except staining nor-
mally seen in the walls of large vessels. Also, no signifi-
cant actin staining was seen in the bone marrows studied.
This suggests that the development of a ‘‘myofibroblastic’’
phenotype is not a common pathway of stromal response
in HIV/AIDS-affected bone marrow.
Megakaryocytes, ARCs, and monocyte/macrophages
may all play a role in the development of stromal respons-
es in bone marrow. With such diverse components con-
Arch Pathol Lab Med—Vol 129, September 2005Marrow Stroma in HIV/AIDS—O’Malley et al
Table 3. Staining of Stromal Elements in Human
Immunodeficiency Virus/Acquired Immunodeficiency
Syndrome (HIV/AIDS) Versus Other Disorders*
* LNGFR indicates low-affinity nerve growth factor receptor; CIMF,
chronic idiopathic myelofibrosis.
† Material adapted from Fang et al.8
tributing to the development of stromal changes, it is pos-
sible that different marrow disorders develop different
‘‘stromal phenotypes.’’ Our previous studies of bone mar-
row stroma in myeloproliferative disorders and metastatic
malignancies suggest significant differences in the stromal
response to these disorders.8
Our findings in HIV/AIDS-affected bone marrows sug-
gest that the stromal responses are similar to those seen
in other hematologic disorders in which the stromal re-
sponse is characterized by an increase in the number of
ARCs (as assessed by LNGFR expression), with a concom-
itant increase in reticulin fibrosis (Table 3). In contrast to
metastatic malignancies, there was no significant increase
of actin expression, suggesting a lack of myofibroblastic
differentiation. These are most likely ‘‘common pathway’’
effects of cytokine production by hematopoietic and stro-
mal cells in response to changes in the marrow microen-
We did not attempt to extend our findings to include
type or duration of therapy in this study. However, several
of the patients were on combination antiretroviral thera-
pies (eg, highly active antiretroviral therapy). In spite of
these therapies, there were still clinically significant stro-
mal alterations in these marrows. We speculate that the
stromal alterations in the bone marrow may persist in
spite of highly active antiretroviral therapy, perhaps be-
cause of the effects of the nucleoside analogues.2It is pos-
sible that bone marrow changes and hematologic abnor-
malities may persist in spite of treatments that improve
Our evaluation of stroma in bone marrows of patients
with HIV/AIDS suggests that fibrosis with increase of
marrow ARCs is common in HIV/AIDS patients. Further,
these appear to be independent of CD4?lymphocyte
counts. Further research into stromal responses to differ-
ent marrow disorders may provide insight into the path-
ophysiology of these diseases and may improve our un-
derstanding of these disorders.
The authors thank Lee Ann Baldridge, HT(ASCP) QIHC, and
Cecelia Dodson, HT(ASCP) QIHC, for their excellent technical
1. Thiele J, ZirbesTK, Bertsch HP,Titius BR, Lorenzen J, Fischer R.AIDS-related
bone marrow lesions: myelodysplastic features or predominant inflammatory-re-
active changes (HIV-myelopathy)? A comparative morphometric study by immu-
nohistochemistry with special emphasis on apoptosis and PCNA-labeling. Anal
Cell Pathol. 1996;11:141–157.
2. GillV, Shattock RJ, Scopes J, et al. Human immunodeficiency virusinfection
impairs hemopoiesis in long-term bone marrow cultures: nonreversal by nucle-
oside analogues. J Infect Dis. 1997;176:1510–1516.
3. Kulkosky J, Bouhamdan M, Geist A, Nunnari G, Phinney DG, Pomerantz
RJ. Pathogenesis of HIV-1 infection within bone marrow cells. Leuk Lymphoma.
4. Ercoli L, Sarmati L, Parisi SG, et al. Human immunodeficiency virus infec-
tion of human bone marrow stromal myoid cells. Scand J Infect Dis. 1996;28:
5. Moses AV, Williams S, Heneveld ML, et al. Human immunodeficiency virus
infection of bone marrow endothelium reduces induction of stromal hematopoi-
etic growth factors. Blood. 1996;87:919–925.
6. Bahner I, Kearns K, Coutinho S, Leonard EH, Kohn DB. Infection of human
marrow stroma by human immunodeficiency virus-1 (HIV-1) is both required and
sufficient for HIV-1-induced hematopoietic suppression in vitro: demonstration
by gene modification of primary human stroma. Blood. 1997;90:1787–1798.
7. Sato T, Sekine H, Kakuda H, Miura N, Sunohara M, Fuse A. HIV infection
of megakaryocytic cell lines. Leuk Lymphoma. 2000;36:397–404.
8. Fang W, An C, Jiang J, Czader MB, O’Malley DP, Orazi A. The stromal
composition of myelofibrosis: differences between chronic myeloproliferativedis-
orders and metastatic malignancies [abstract]. Mod Pathol. 2003;16:232A.
9. Thiele J, RompcikV, Wagner S, Fischer R.Vascular architecture and collagen
type IV in primary myelofibrosis and polycythaemia vera: an immunomorphom-
etric study on trephine biopsies of the bone marrow. Br J Haematol. 1992;80:
10. Thiele J, Kvasnicka HM, Fischer R, Diehl V. Clinicopathological impact of
the interaction between megakaryocytes and myeloid stroma in chronic myelo-
proliferative disorders: a concise update. Leuk Lymphoma. 1997;24:463–481.
11. Thomson TM, Rettig WJ, Chesa PG, Green SH, Mena AC, Old LJ. Expres-
sion of human nerve growth factor receptor on cells derived from all three germ
layers. Exp Cell Res. 1988;174:533–539.
12. Orazi A, Cattoretti G, Schiro R, et al. Recombinant human interleukin-3
and recombinant human granulocyte-macrophage colony-stimulating factor ad-
ministered in vivo after high-dose cyclophosphamide cancer chemotherapy: ef-
fect on hematopoiesis and microenvironment in human bone marrow. Blood.
13. Cattoretti G, Schiro R, Orazi A, Soligo D, Colombo MP. Bone marrow
stroma in humans: anti-nerve growth factor receptor antibodies selectively stain
reticular cells in vivo and in vitro. Blood. 1993;81:1726–1738.
14. Bonanno E, Ercoli L, Missori P, Rocchi G, Spagnoli LG. Homogeneous
stromal cell population from normal human adult bone marrow expressing alpha-
smooth muscle actin filaments. Lab Invest. 1994;71:308–315.
15. Papadopoulos N, Simopoulos C, Kotini A, Lambropoulou M, Tolparidou I,
Tamiolakis D. Differential expression of alpha-smooth muscle actin molecule in
a subset of bone marrow stromal cells, in B-cell chronic lymphocytic leukemia,
autoimmune disorders and normal fetuses. Eur J Gynaecol Oncol. 2001;22:447–