| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
1
Calypte Biomedical, Berkeley, CA 94710.
2
Department of Microbiology and Immunology, The
University of Michigan School of Medicine, Ann Arbor, MI 48109-0620.
a Address correspondence to this author at: Calypte Biomedical, 1440 Fourth St., Berkeley, CA 94710. Fax 510-526-5381; e-mail hervdoc{at}aol.com
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: Using a licensed urine-based test for antibody to HIV-1, we performed 25 991 HIV-1 urine antibody enzyme immunoassay (EIA) screening tests [confirmable by HIV-1 Western blot (WB)] on paired urine and blood specimens obtained from high- and low-risk HIV-1 subjects collected at six sites representative of the US population.
Results: Using HIV-1 urine EIA tests confirmed by urine Western blot, a compartmentalized immune response (urine positive/serum negative) occurred in 0.24% of a cohort of 11 896 subjects. In the same cohort, specimens that were urine negative/serum positive occurred in 0.17% of subjects. In a second study of 25 991 subjects that included 859 high-risk individuals, the false-positive urine EIA frequency (urine WB negative or indeterminate) was 1.3%. This false-positive frequency in the high-risk cohort was attributed, in part, to an IgA antibody response. We tabulated urine and serum indeterminate reactivities and examined their possible causes. Data are presented showing that antibodies from a seroindeterminate HIV-1vau group O subject were reactive in urine EIA and urine WB tests. An analysis of the HIV-1vau strain group O env nucleotide sequence disclosed a high frequency of homology with human chromosome 7q31, a fragile site implicated in many human malignancies.
Conclusions: These results demonstrate the utility of urine for alternative HIV-1 antibody testing and provide new insights into the pathogenesis of HIV-1 infection and into potential application of this approach in investigation of other microbial pathogens and toxic compounds.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Urine and serum specimens collected at multiple geographic sites (11 896 paired specimens) were screened by EIA for antibodies to HIV-1 (Calypte HIV-1 Urine Assay) and licensed blood tests. EIA repeatedly reactive (RR) specimens were confirmed for HIV-1 antibodies by WB tests (Calypte Biomedical HIV-1 WB test). The performance characteristics of the EIA screening assays are documented. The frequency of urine-negative/serum-positive (UNSP) reactivities was 0.17%. The frequency of urine-positive/serum-negative (UPSN) reactivities was 0.24%. We attribute such reactivities to a compartmentalized mucosal immune response to HIV-1 antigens, although autoimmune reactivities caused by mimicry between "self-antigens" and HIV-1 antigens are not excluded. We define self-antigens as those either encoded by germ-line sequences or caused by recombinational events attributable to exposure to environmental biohazardous entities (3).
An analysis was carried out to identify the factors that influence indeterminate (ID) WB reactivities in specimens that registered EIA RR. This was done by scoring the number of WB bands to HIV-1 core and env antigens. The results showed that many more bands were found in serum than in urine specimens.
Recent reports of the protective role of HIV-1 IgA in urine and vaginal lavages led us to conduct a multisite study (25 991 paired specimens) to determine the frequency of urine HIV-1 IgA reactivity according to risk factors. The results clearly showed a significantly higher occurrence of HIV-1 IgA EIA reactivity with all high risk factors.
Tests using a urine specimen from a serum HIV-1 ID subject with HIV-1vau strain of group O showed that the urine EIA assay, combined with a confirmatory urine WB test, were effective in detecting HIV-1 antibodies. A mosaic analysis of the HIV-1vau env nucleotide sequence revealed a high frequency of homologies to human chromosome 7q31, a fragile chromosomal site often implicated in human malignancies of diverse types. The new insights provided by these findings in the pathogenesis of HIV-1 infection are discussed.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Samples that are initially reactive are retested in duplicate using the original specimen. If, after repeat testing, one or both of the duplicate tests are reactive, the specimen is considered EIA RR. Before a determination of HIV-1 status can be made, subjects that test EIA RR are confirmed for HIV-1 antibody using only the additional, more specific Calypte Biomedical HIV-1 urine WB kit.
The Calypte Biomedical HIV-1 urine WB Kit is manufactured from HIV-1 propagated in an H9/HTLV-IIIB T-lymphocyte cell line. Partially purified virus is inactivated by treatment with psoralen and ultraviolet light, and detergent disruption. HIV-1 proteins are fractionated according to molecular weight by electrophoresis on a polyacrylamide slab gel in the presence of sodium dodecyl sulfate. The separated HIV-1 proteins are electrotransferred from the gel to a nitrocellulose membrane, which is then washed, blocked (to minimize nonspecific immunoglobulin binding), and packaged. Individual nitrocellulose strips are incubated with specimens and controls. If HIV-1 antibodies are present, they bind to the viral antigens present on the nitrocellulose strips. The strips are washed again to remove unbound material. Visualization of the human immunoglobulins specifically bound to HIV-1 proteins is accomplished in situ using a series of reactions with goat anti-human IgG conjugated with biotin-avidin conjugated with horseradish peroxidase, and the horseradish peroxidase substrate 4-chloro-1-naphthol. If antibodies to HIV-1 antigens are present in the specimen in sufficient concentration, bands occur corresponding to the position of one or more of the following HIV-1 proteins (p) or glycoproteins (gp) on the nitrocellulose strip: p17, p24, p31, gp41, p51, p55, p66, gp120, gp160 (the number refers to the apparent molecular mass in kilodaltons). The interpretive criterion for a positive result using urine as the sample is the presence of a gp160 band with intensity equal to or greater than the intensity of the gp160 band on the low positive urine control strip. The interpretive criteria for a positive result using serum or plasma specimens (4) are the occurrences of any two or more of the following bands: p24, gp41, and gp120/160. Each band has a reactivity score of "+" or greater. Commonly, the band at gp41 or gp160 is diffuse. Other viral bands may or may not be present. A reactivity score of + is defined as a band with intensity at least as intense as the p24 band on the weakly reactive control strip but less intense than the p24 band on the strongly reactive control strip. The reactivity is scored as ID when WB bands are present that do not meet the criteria for positivity.
patient groups
Medical disorders consisted of autoimmune diseases including
rheumatoid arthritis, Sjogren syndrome, systemic lupus erythematosus,
idiopathic thrombocytopenia purpura, myasthenia gravis, multiple
sclerosis, and autoimmune hemolytic anemia. Malignancies included
chronic lymphocytic leukemia, breast cancer, lung cancer, colon cancer,
Hodgkin disease, and multiple myeloma. The urologic disorders included
acute glomerular nephritis, acute tubular necrosis, acute renal
failure, chronic renal failure (on dialysis), and nephrotic syndrome.
Subjects with medical conditions included subjects that required
medical treatment or hospitalization, those who had attended a clinic
for sexually transmitted diseases (STDs), multiparous and pregnant
women, subjects with exertion dehydration, and subjects who had
received multiple blood transfusions.
clinical studies
A clinical study in support of US licensure of the Calypte HIV-1
urine EIA was performed on 11 344 individuals to compare the accuracy
of HIV-1 urine testing with US-licensed serum testing. These studies
were extended by an additional 552 individuals considered at high risk
for HIV-1 infection (Table 1
, combined total of 11 896). Urine and
serum specimens from low- and high-risk populations, patients with
non-HIV-1-related medical disorders, and known HIV-1-infected
individuals, including AIDS patients, were tested. Subjects at high
risk for HIV-1 infection included intravenous drug users (IDUs),
hemophiliacs, and sexual partners of HIV-1-infected individuals and
commercial sex workers (CSWs). Subjects with medical disorders and
conditions are described above.
|
In a study in support of the Calypte Biomedical urine HIV-1 WB, clinical trial samples were tested with the Calypte Biomedical urine HIV-1 WB Kit regardless of EIA result. For this clinical study, 2159 urine specimens from various populations were tested. These included 109 urine specimens paired with serum HIV-1 WB ID specimens and 114 urine specimens falsely positive on the HIV-1-screening EIA. The urine and serum blots from non-HIV-1 infected subjects were compared.
In another study, urine and serum samples were obtained from low- and high-risk subjects after informed consent. The low-risk site measured EIA reactivity of >25 132 life insurance applicants and was unlinked. Specimens from high-risk individuals were collected at five geographic sites. Subjects (n = 859) belonged to the following risk categories: IDU; seronegative partners of HIV-1-positive individuals (partners); individuals with a sexually transmitted disease (STD); and individuals with multiple sex partners of unknown HIV-1 status (MSP). US Food and Drug Administration-licensed blood and urine HIV-1 EIA tests were used for screening. The HIV-1 urine test measures EIA reactivity using an IgG heavy chain and light chain conjugate. EIA RR specimens then were tested on Food and Drug Administration-licensed confirmation tests. Discordant urine EIA RR specimens from seronegative subjects that were urine WB negative or indeterminate were then tested in duplicate by a research-use-only IgA heavy chain-specific gp160 urine antibody assay.
blast analysis
A mosaic analysis (3) was applied to HIV-1vau envelope
gene sequence, GenBank no. X80020. Using the Advance Blast program
(http://www.ncbi.nlm.nih.gov/BLAST/), the following parameters were
set: the Homo sapiens database was selected, nr
(all non-redundant GenBank CDS translations + PDB + SwissProt + PIR +
PRF) was selected, the expected threshold was set at 10 and the
HIV-1vau env sequence was queried. Samples of >15 nucleotides (15mer)
or higher with homologies >89% with HIV-1vau were tabulated.
statistical analysis
A 2 x 2 contingency analysis for determination of
P values was performed by using Graphpad InStat (Graphpad
Program Software).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
,
combined total of 11 896). Data are recorded according to risk groups.
Results are summarized in Table 1
. As reported previously (5), the sensitivity of
the urine test was 98.73%. Ten serum samples matched to WB
urine-positive subjects were serum EIA nonreactive. Therefore, the
sensitivity of the serum tests was calculated at 99.15%. The combined
use of urine and serum EIA tests detected a higher frequency HIV-1
antibody-positive subjects than either test alone.
The data in Table 1
bring out several interesting points: of the
11 896 subjects tested, 29 (0.24%) were UPSN, a result consistent
with the compartmentalization of the immune response to HIV-1
infection; the preponderance of UPSN reactivities were in the
high-risk cohort; and 14 of 1111 (1.26%) of the known HIV-1-positive
cohort subjects were either urine negative/indeterminate or urine
negative/serum positive (UNSP). As noted below, we carried out a
systematic study of the factors contributory to ID.
frequency of eia rr in cohorts according to specimen collection
sites
Collated data provided a valuable perspective (Table 1
) of the
frequency of antibodies to HIV-1. However, they did not provide
insights of unique epidemiologic or public health interest. Table 2
summarizes the EIA RR survey data according to specimen
collection site. Of the 25 132 subjects tested in the low-risk cohort,
253 (1.0%) were EIA RR for antibody in urine paired with seronegative
specimens. EIA RR was the most frequent for subjects in site 6 of the
high-risk groups. Specimens from this cohort were collected from
patients in methadone clinics, many of whom presumably were IDUs.
|
frequency of upsn and unsp rr according to hiv-1 risk
categories
These studies were carried out to determine the occurrence of
WB-confirmed UPSN and UNSP reactivities according to category of risk
activity. The results summarized in Fig. 1
show that the only subjects that had discordant reactivities
(UPSN/UNSP) were in IDU or IDU/MSP cohorts. Only one discordant
reactivity (UNSP) was found in the low-risk cohort.
|
IgA HIV-1 EIA RR ACCORDING TO RISK CATEGORIES
While we were conducting the EIA surveys summarized above, a group
of subjects was detected whose urine was repeatedly reactive for HIV-1
IgA antibodies. The corresponding sera were nonreactive for HIV-1
antibodies in licensed tests. The repeatedly reactive EIA urine
tests were not confirmable by urine WB. To clarify these results,
we used a conjugate for the IgA heavy chain to further characterize
urine IgA reactivity. Fig. 2
summarizes the results for the different risk groups. The
frequency of IgA antibody in urine was most common in the seronegative
partners of HIV-1-positive subjects (partners), and STD and MSP
cohorts. These results were consistent with a compartmentalized
immunologic response to such antigens in seronegative subjects. It
would appear that the HIV-1 IgA envelope antibody can be detected in an
EIA format but not by WB. Either the antibodies were expressed as the
result of antecedent exposure to HIV-1 or the reactivities described
may represent autoimmune responses elicited because of antigenic
mimicries between self-antigens and HIV-1.
|
frequency of bands in wb of id specimens
A total of 1096 urine and 994 serum specimens with ID reactivity
were analyzed by WB. This cohort represents a subgroup of the 2159
subjects tested in support of the urine WB clinical trials. Results for
gag, pol, and env antigens were compared. Table 3
shows that bands were found in only 17 urine specimens (1.6%).
The majority (11 of 17) occurred in the medical disorders and
conditions cohort. The results for serum specimens stood in marked
contrast. Bands were found in 336 of 994 specimens (33.8%). Bands to
env antigens occurred in only 12 of 994 serum specimens (1.20%). Of
these, 11 occurred in sera of the high-risk cohort. On the whole, these
results suggest that band occurrence is a not random event; that it is
much more common in serum than in urine; that it occurs most commonly
to core antigens, particularly p24; and that there is reasonable
evidence that specimens from patients with medical disorders and
conditions are a major contributor to WB ID reactivities. The
latter is consistent with the autoimmune reactivity of HIV-1-infected
subjects to self-antigens that have homologies with HIV-1 antigens.
|
occurrence of eia rr in subjects with non-hiv-1 medical disorders
This survey was performed to determine the frequency of EIA RRs
not confirmable or ID by WB in urine and serum specimens obtained from
375 patients (Table 1
) with medical disorders and conditions other than
HIV-1 infection. Table 4
shows that 69 of 375 (18.4%) of urine tests were EIA RR. In
the autoimmune and STD cohorts, urine EIA RR responses occurred at
frequencies of 20.4% and 17.9%, respectively. In the corresponding
serum assays, four (1.1%) were EIA RR. EIA RR responses were confined
to the STD cohort. Of these four EIA RR responses, two were confirmable
by WB. The other two were ID.
|
upsn reactivity in a patient positive for a group o hiv-1 isolate
A 41-year-old French woman originally clinically diagnosed in 1986
with leukoneutropenia associated with cervical carcinoma progressed to
CD4 T-cell depletion with opportunistic infections and eventually died
in 1992. Her serum WB reactivity was consistently ID (lack of env serum
WB bands). The initial diagnosis presumed to be idiopathic CD4+ T-cell
lymphocytopenia (ICL) became a diagnosis of AIDS upon isolation of a
group O strain of HIV-1, referred to as HIV-1vau. Therefore, it was of
interest to determine whether the urine and serum antibody tests
described here were suitable for detecting HIV-1 antibodies. Urine EIA
tests were performed using both recombinant gp160 antigen derived from
an HIV-1 M strain (H9/HTLV-IIIB) and urine WBs
using HIV-1 Group M viral lysate. The results presented in Table 5
show that both serum and urine specimens reacted with HIV-1
group M gag proteins (p24 and p55). The serum reacted with pol (p31 and
p66), but the urine did not. On the other hand, urine reacted with
gp160, but the serum did not. This result provides further evidence for
the compartmentalized immune response to HIV-1 infection.
|
genetic analysis of HIV-1vau env SEQUENCE
Studies of genetic abnormalities in veterans with Persian Gulf War
syndrome (3) led to analogous investigation of the patient
described above who had no known risk factors for HIV-1 infection. Our
central interest based on reactivity of urine to HIV-1 gp160 was in the
HIV-1vau env nucleotide sequence. Basically, a mosaic analysis involved
the determination of signature segment homologies to known human
chromosomal sequences. Mosaic analysis of the HIV-1vau env sequence was
carried out using the Advance Blast program described in
Materials and Methods. Table 6
shows that signature segments of HIV-1vau env had homologies
with 14 different human chromosomes. Segments from chromosome 7
predominated. Of these, six segments had homologies with 7q31. This is
of interest because 7q31 is a fragile chromosomal site involved in
mutations and recombinations. Mutations in 7q31 have been implicated in
several human malignancies (6)(7)(8)(9). In addition, a 37mer
segment of African green monkey simian immunodeficiency virus (SIVagm)
env was found.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The occurrence of UPSN reactivity in urine tests for antibodies to
HIV-1 (5) required further analysis of the EIA assay used to
screen specimens. A recombinant gp160 antigen was used because it had
superior performance characteristics. In our initial survey
(5), parallel urine and serum tests were performed on
11 344 subjects. EIA reactivities that were repeatedly positive were
confirmed by urine WB. The serum test detected 99.15% of all
antibody-positive individuals. The urine test detected 98.73% of the
same cohort. Twenty-five assays had discordant results. Of these, 10
were urine positive but serum nonreactive or indeterminate (UPSN).
Fifteen were urine negative or indeterminate but serum positive (UNSP).
Table 1
summarizes the results of a more extended study in which
11 896 paired urine and serum specimens were tested. Of these, 0.24%
were UPSN and 0.17% were UNSP. As pointed out previously
(5), the combined use of urine and serum tests detected a
greater number of antibody-positive subjects than either test alone.
Any decision to use both tests simultaneously rests on the
circumstances under which they are to be used, a clinical or public
health judgment. Whether dual testing should be performed for blood
banking remains an open question. Nonetheless, it is noteworthy that
urine tests for antibody (Table 5
) were effective in detecting
antibodies to the HIV-1vau group O strain.
The occurrence of ID results in WB tests for antibody confounds the
interpretation of results. Accordingly, we carried out studies to
determine which HIV-1 core, pol, or env antigens were involved. Urine
(1096) and serum (994) specimens obtained from subjects of different
risks for HIV-1 infection were tested for the frequency of WB bands.
Twelve of 994 serum specimens (1.2%) had bands to env antigens. Of
these, 11 were from high-risk subjects. These results show that ID
reactivities for the most part were not attributable to HIV-1 env
antigens. The results with core antigens stood in marked contrast. When
sera were tested, 336 of 994 subjects (33.8%) had bands. Only 17
subjects (1.6%) had bands when 1096 urine specimens were tested. Thus,
when band number is used as a parameter of ID reactivity, urine
specimens had a minimal response. When they occurred, they were found
primarily in the medical disorder and conditions group. It is of
interest that the frequency of bands in sera from the same cohort (73
of 199; 36.7%) and from the high-risk group (123 of 371; 33.2%) were
not different. In serum tests, ID reactivity could frequently be
attributed to p24. A reasonable explanation of the described findings
is that in many cases ID results are caused by autoimmune reactivity to
self-antigens (11)(12). This notion is supported
by the finding that many subjects in the medical disorders and
conditions group with chronic autoimmune diseases often gave ID
reactivity in WB tests. Considered together, Tables 1
and 3
suggest
that urine assays for antibodies to HIV-1 are sensitive and accurate.
Bands to env antigens in urine WB confirmatory tests appear to be a
more reliable indicator of exposure to HIV-1 antigens than core
antigens.
There appear to be at least three indicators of immunologic exposure to HIV-1 antigens in high-risk subjects who, nonetheless, remain seronegative: UPSN reactivity that we report here, cell-mediated immune (CMI) reactivity to HIV-1 antigens in subjects at high risk for HIV-1 infection, and antibodies in mucosal lavages of seronegative subjects. CMI reactivity to HIV-1 antigens in seronegative subjects is well documented. Clerici and co-workers (13)(14)(15) reported HIV-1-specific T-cell reactivity to env antigens in seronegative healthcare workers exposed to HIV-1 contaminated blood. Pinto et al. (16) reported env-specific (gp160) cytotoxic T-cell reactivity in seronegative subjects who had been exposed to accidental needle sticks contaminated with HIV-1-positive blood. In their publication, Pinto et al. (16) cited 10 publications reporting CMI responses to HIV-1 antigens in high-risk seronegative subjects. Mazzoli et al. (17) reported the occurrence of serum CMI reactivity, and IgA (reactivity to an HIV-1 gp160) in urine and vaginal wash specimens obtained from seronegative sexual partners who were HIV-1 positive.
The report by Rowland-Jones et al. (18) deals with a different type of cohort. Six seronegative CSWs in Gambia, West Africa, who were at high risk for HIV-1 infection, demonstrated specific cytotoxic T-cell reactivity to HIV-1 antigens. The nef, pol, p24, and p17 peptides were used for cytotoxicity tests. The strongest responses were to pol and nef. Kaul et al. (19) studied three cohorts of Kenyan women of different risks for HIV-1 infection: 21 HIV-1-resistant CSWs, 19 HIV-1-infected CSWs, and 28 low-risk women. Resistance was defined as persistent seronegative reactivity to HIV-1 antigens and negative reverse transcription-PCR assays for HIV-1 sequences over 3 or more years of commercial sex work. Cervicovaginal lavage specimens from all subjects were tested for HIV-1-specific IgA and IgG. T-helper lymphocyte responses (CMI) to HIV-1 antigens also was determined. HIV-1-specific IgA was present in cervicovaginal lavages of 16 of 21 (76%) of HIV-1-resistant CSWs, in 5 of 19 (26%) HIV-1-infected women, and in 3 of 28 (11%) low-risk women. CMI reactivity was present in 11 of 25 (55%) of HIV-1-resistant women, in 4 of 18 HIV-1-infected women, and in 1 of 25 (4%) low-risk women. The authors suggest a protective mucosal immune response to HIV-1 infection that can be independent of CMI responsiveness.
Considered together, the described findings document an important role
of mucosal immunity (compartmentalization) mediated by IgA in
resistance to HIV-1 infection. We found, as shown in Fig. 2
, that
subjects of all high-risk groups had higher frequencies of HIV-1 IgA
EIA reactivity than the low-risk cohort. Because such HIV-1 IgA EIA RRs
were not confirmable by WB, it appears that WB tests for HIV-1 IgA were
ineffective or that the EIA RR occurred because of autoreactivity
(20)(21)(22)(23) to self-antigens elicited by HIV-1 or other
environmentally acquired agents. Such results suggest that
inappropriate immunizations with HIV-1 antigens may provoke deleterious
results.
The occurrence of antibodies, particularly IgA, in mucosal tissues is a well-recognized barrier to bacterial or viral infections (24)(25). Two recent publications (26)(27) presented in detail a histologic description of the local mucosal immune response in the reproductive tissues of both sexes. The mucosal barrier to HIV-1 infection has been discussed (28). Here we direct attention to interesting experiments in nature (29) in which subjects repeatedly exposed to infection by HIV-1 remain seronegative. Four representative reports are discussed. Beyrer et al. (30) studied a cohort of Thai prostitutes designated as HIV-1 "highly exposed but persistently seronegative". In this cohort, sera were persistently negative for IgG and IgA. However, cervicovaginal lavages from 6 of 16 subjects were HIV-1 IgA EIA RR. In seropositive controls, 11 of 11 and 8 of 11 vaginal lavages were positive for IgG and IgA antibodies, respectively. These results showed that high-risk seronegative subjects had IgA in their genital mucosa, that it likely was produced locally, and that it probably served as a barrier to systemic HIV-1 infection. In a similar investigation, Belec et al. (31) studied 150 paired serum and vaginal secretions obtained from HIV-1 seronegative women who resided in West Africa. Antibody (IgG) was detected in 2.5% of the vaginal secretions. Antibodies in such specimens were broadly reactive with HIV-1 core and env antigens. The authors made two interesting comments: in the subjects studied, the immune response to HIV-1 appeared to be restricted to the vaginal mucosa; and the cervicovaginal mucosa is an immunocompetent tissue possessing antigen-processing cells and lymphocytes. We discussed above (17)(19) the occurrence of mucosal IgA in highly exposed but persistently seronegative subjects. Considered together, the results of the studies discussed above suggest that in a small number of subjects, exposure to HIV-1 infection can lead to resistance to infection, that in these subjects, it does not depend on serum antibody, that there is a compartmentalized barrier to infection, and that in the described women, the barrier to infection was the cervicovaginal mucosa.
Based on the findings we report here and on the literature summarized above, we propose that antibody found in UPSN subjects is of local origin, i.e., is derived compartmentally. Functionally, a substantial portion of it may be IgA. Studies of locally produced IgA are important because IgA neutralizes viruses extra- and intracellularly (32)(33). IgA also has an important excretory function, i.e., IgA may bind HIV-1 antigens within the mucosal lamina propria and then excrete it through the epithelium into the lumen (34). However, these dynamics do not explain the seronegative responses to HIV-1 that we and others have found. There must be a powerful active mechanism that suppresses the serum antibody response. Shearer and Clerici (29)(35) propose a Th1-Th2 switch mediated by cytokine cross-regulation.
HIV-1 is an insidious virus that infects humans: It replicates in a wide variety of tissues, i.e., is pantropic; it permanently integrates its gene sequences into the host cell genome; it has an extraordinarily high mutation/recombination rate; multiple subtypes are known (see below) to infect a subject at the same time in a single transmitting event; HIV-1 subtypes not only recombine with each other but may also recombine with human endogenous retroelements (11) or human chromosome segments (3)(36); HIV-1 infection may activate human endogenous retroelements with or without recombination; and because of compartmentalization, the population dynamics of HIV-1 subtypes differ in different tissue compartments (see below). This latter phenomenon was documented recently by Kiessling et al. (37).
Katz and Skalka (38) recently reviewed the recombination genetics of retroviruses. Here we focus on the population dynamics of HIV-1 subtypes and restrict consideration to representative publications that bring out valuable principles. Broadly speaking, three types of virus dynamics can be considered: monotypic, in which a single subtype causes infection; multitypic, in which more than one subtype causes infection; and infections by "mosaics" that are recombinants between or among single subtypes. All of these dynamics are known to occur in humans and in the individual tissue compartments. Moreover, mosaic subtypes generated in a given tissue compartment are known (see below) to differ from those in a different tissue compartment in the same host.
The report by Robertson et al. (39) provides a perspective of the global distribution of M and O HIV-1 subtypes and the relative frequency of recombinants. One hundred fourteen HIV-1 stains were compared by analyzing gag and env sequences. Phylogenetic trees then were constructed. Of the 114 strains, 10 appeared to be recombinants among just the M subtypes. Recombinants between M and O subtypes also were found. All of the recombinants were from geographic areas where multiple subtypes were known to circulate. The authors note that a proportion of subjects were co-infected with strains of different subtypes. This can occur singly or sequentially. The authors discuss the importance of the global frequency of such recombinants, their biological significance, their putative impact on vaccine design, and anti-HIV-1 drug effectiveness. Ramos et al. (40) carried out molecular epidemiologic studies of HIV-1 strains in Rio de Janeiro. Seventy-nine subjects were screened at the molecular level over a period of 1 year. Heteroduplex analysis, restriction fragment length polymorphism analysis, and gene sequencing were used to distinguish between cotransmission and superinfection. Their results showed that 88–89% of isolates were monotypic, 7.6% were recombinants, and 3.8% were dual infections. Recombination between gag and pol sequences was more common than between env and pol sequences. The authors direct attention to the heterogeneous nature of the strains active in the epidemic. The generation of new subtypes with altered infectivity is confounded by the higher concentrations of chemokine receptors per cell in the mucosal tissue compared with the blood lymphocyte counterparts (41).
Fisher et al. (42) analyzed molecular variants within a single HIV-1 isolate. In brief, proviral clones were derived from tissues of a single HIV-1-infected patient, and their biological properties were compared. The authors report that hybrid genomes (env sequences) generated viruses that differed in their ability to replicate in various human cell lines in vitro. The authors suggest that variation exists in vivo among HIV-1 viruses endogenous to host tissues. When such viruses are propagated serially in vitro, heterogeneity is diminished, probably through a process of selection. The authors point out the importance of clarifying how well in vitro-adapted HIV-1 strains reflect their in vivo counterparts.
One hallmark of HIV-1 infection is the diversity of strains found in different tissue compartments in the same individual. Kiessling et al. (37) compared cloned sequences of virus populations derived from semen and peripheral blood mononuclear cells (PBMCs). Protease gene sequence analysis revealed differences in the virus populations derived from different tissue compartments. Further evidence for the diversification of HIV-1 populations in tissue compartments resulted from studies of HIV-1 strains derived from the blood and genital compartments. Poss et al. (43) analyzed proviral env gene sequences (V1, V2, and V3) in infected cells obtained from cervical secretions and PBMCs. Specimens were obtained from six women who had just seroconverted. Three patterns of diversity were found: homogeneity between cervical and PBMC-derived strains, variants of cervical and PBMC origin had modest heterogeneity within each genotype, and multiple variants occurred within each compartment. Zhu et al. (44) carried out a similar study by comparing HIV-1 gp160 sequences in longitudinal specimens obtained from five "acute seroconverters" and their corresponding sexual partners (transmitters). These investigators reported that the quantitative homoduplex tracking assay used, combined with selective sequencing, was superior in detecting genetic diversity of HIV-1 strains than conventional PCR techniques; that virus populations of transmitters were compartmentalized, i.e., HIV-1 variants in genital populations were different from those derived from PBMCs or the blood plasma; that the virus populations of both transmitters and recipients were diverse; that there was a strong selective pressure by recipients on variants in the transmitted virus population; that up to the time of seroconversion by recipients, the variants that propagated were relatively homogeneous genetically; and that the selection made by the recipient on the donor virus populations was for a minor subpopulation.
Epstein et al. (45) reported on the tissue-specific evolution (brain and spleen) of HIV-1 variants in children with AIDS. The authors used the term "quasispecies" to describe HIV-1 variants. In brief, V3 domain sequences were analyzed in variants that were derived from the brain and spleen. The authors found that brain and spleen HIV-1 populations differed from each other (tissue compartmentalization and selection) and that each population evolved independently. The authors discuss "immune pressure" and "escape mutants" in the selection of variants in the two tissue compartments. In studies of Thai subjects, Artenstein et al. (46) reported "the first evidence of dual HIV-1 infection of humans by subtypes B and E".
From this brief review, it is clear that genetic diversity and tissue compartmentalization is a hallmark of HIV-1 infection. The dynamics of the pathogenesis of AIDS, the effects of antiviral drugs, the epidemiology of HIV-1 infections, and vaccine design and administration all are impacted by the described genetic diversity and tissue compartmentalization.
The report by Charneau et al. (47) raises the important question of whether the occurrence of CD4+ T-cell lymphocytopenia before the onset of an aberrant AIDS syndrome is causally related to HIV-1 infection. The parallel question is whether HIV-1 infection serves as an accelerant to CD4+ T-cell lymphocytopenia that leads to a terminal illness. In their study, the patient at the time of death had no CD4 T cells; therefore, it is reasonable to assume that the cervical carcinoma in the same patient was not a confounding medical condition with regard to these questions. We previously suggested (48) that a study of ICL patients provides an important approach to understanding the underlying host molecular mechanisms in the pathogenesis of AIDS. Although HIV-1 is considered the sole and primary cause of AIDS, it is apparent that HIV-1 infection may accelerate underlying disease processes. This hypothesis has implications for both antiretroviral therapies and HIV-1 vaccine program development.
Our mosaic studies of the HIV-1vau env sequence suggest additional mechanisms by which HIV-1 infection may have pathologic consequences. For example, the subject studied by Charneau et al. (47) was an agricultural worker exposed to farm-related toxic materials. In this study, we describe homologies between HIV-1vau env sequences and chromosomal segments in which 7q31 is prominently involved. The possibility that there is interaction between precursor sequences of HIV-1vau, 7q31 sequences, and farm-related toxic materials is suggested by our recent studies of Persian Gulf War veterans (3). In our earlier study, we found genetic abnormalities that possibly arose from exposure to environmental toxins.
The homologies between HIV-1vau env sequences and segments of 7q31 suggest the possibility that sequences of the latter may have been implicated in the genetic evolution of HIV-1vau. As described in our recent report on the pathogenesis of Persian Gulf War syndrome (3), environmental biohazardous materials may have participated in such evolutionary events, i.e., there is an interaction at the molecular level between antecedents to HIV-1vau sequences, environmental toxins, and fragile sites in 7q31. The activation, induction, and recombination of endogenous retroelements is a well-recognized phenomenon. The mosaic analysis also disclosed a occurrence of SIVagm homology with the HIV-1vau env nucleotide sequence. Possibly, such sequences were of poliovirus vaccine origin and may have been involved in recombinational events during the evolution of the HIV-1vau genome. (49). On the other hand, the complex viral isolation techniques involved in the isolation of HIV-1vau adds to the uncertainty of its biological and molecular origins.
The considerations outlined above provide a framework for future studies that developed from the EIA surveys summarized here and provide new insights into the pathogenesis of HIV-1 infection and AIDS.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  M. Guan Frequency, Causes, and New Challenges of Indeterminate Results in Western Blot Confirmatory Testing for Antibodies to Human Immunodeficiency Virus Clin. Vaccine Immunol., June 1, 2007; 14(6): 649 - 659. [Full Text] [PDF]
|
|
|
|
 |
|
 M. L. Santiago, M. Lukasik, S. Kamenya, Y. Li, F. Bibollet-Ruche, E. Bailes, M. N. Muller, M. Emery, D. A. Goldenberg, J. S. Lwanga, et al. Foci of Endemic Simian Immunodeficiency Virus Infection in Wild-Living Eastern Chimpanzees (Pan troglodytes schweinfurthii) J. Virol., July 1, 2003; 77(13): 7545 - 7562. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Ling, M. L. Santiago, S. Meleth, B. Gormus, H. M. McClure, C. Apetrei, B. H. Hahn, and P. A. Marx Noninvasive Detection of New Simian Immunodeficiency Virus Lineages in Captive Sooty Mangabeys: Ability To Amplify Virion RNA from Fecal Samples Correlates with Viral Load in Plasma J. Virol., February 1, 2003; 77(3): 2214 - 2226. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. M. R. Oelemann, C. M. Lowndes, G. C. Verissimo da Costa, M. G. Morgado, L. R. R. Castello-Branco, B. Grinsztejn, M. Alary, and F. I. Bastos Diagnostic Detection of Human Immunodeficiency Virus Type 1 Antibodies in Urine: a Brazilian Study J. Clin. Microbiol., March 1, 2002; 40(3): 881 - 885. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||
