HIV-1 and HIV-2 are closely related viruses with distinct epidemiologic and biologic properties. In west Africa many populations of high-risk individuals are at risk for both HIV-1 and HIV-2. As the viruses share a common cellular receptor and risk determinants for infection, the potential for dual infection with both viruses has been considered, and several case reports have confirmed that this occurs. However, current studies have not evaluated the frequency of this event nor, more importantly, the clinical significance of dual infection in disease pathogenesis and transmission.
Since 1986 a number of west African countries have reported significant rates of individuals possessing antibody responses to both HIV viruses, referred to as a dual serologic profile. Strong seroreactivity to the envelope antigens of HIV-1 or HIV-2 by immunoblot and/or radioimmunoprecipitation assay has aided in the distinction between HIV-1 and HIV-2 infection [1–3]. However, some degree of cross-reactivity can be demonstrated, and sera with equivalent envelope reactivity to both viruses are serologically classified as dually reactive. The use of synthetic peptide assays based on short, type-specific regions of the transmembrane envelope antigens have also been useful in HIV type diagnosis [4–6]. Variation in type-specific peptides used and qualitative aspects of the assay format have only partially resolved the diagnostic dilemma posed by HIV dual-reactive sera. Infection with both viruses has been confirmed in a few individuals by virus isolation or polymerase chain reaction (PCR) assays [7–11]. Previous investigations using PCR to detect proviral sequences of both viruses have reported that 33–62% of the dually seroreactive individual can be confirmed as dually infected [9,10]. It is unclear whether the dual seropositives not confirmed as true dual infection by PCR or viral isolation are due to extensive HIV-1 cross reactivity as suggested, misclassification of samples based on serodiagnosis, or insensitivity of the PCR assays.
We conducted this study to further investigate the rate of HIV dual infection in populations that were exposed to both viruses. Using a well-defined serologic algorithm, we used a PCR assay consisting of two sets of nested primers for HIV-1 and HIV-2 optimized for detection of proviruses in singly infected individuals. These primers were found to be highly specific and sensitive, and further evaluation with a dilution series consisting of known HIV proviral copy number demonstrated equivalent sensitivity per primer set. We then evaluated the ability of PCR to diagnose serologically defined dually reactive samples.
Patients and methods
Serum and peripheral blood mononuclear cells (PBMC) DNA samples were obtained from registered female commercial sex workers and hospitalized patients in Dakar, Senegal. We have previously reported on various epidemiologic and clinical aspects of our cohort of registered female commercial sex workers [12,13]. During 1990–1993, all hospitalized patients at the Infectious Disease Service at Fann Hospital were evaluated for clinical-immunologic status and HIV infection (unpublished data). All study participants gave informed consent prior to enrollment and sample collection. All sera were evaluated for antibody reactivity to the major viral antigens for HIV-1 and HIV-2 by immunoblot on disrupted whole viral lysates; a screening assay was not employed. Sera were considered potentially dual-seropositive if they showed antibodies to at least two envelope proteins of each virus and if they reacted with strong and equal intensity to the gp41 of HIV-1 and the gp32/36 of HIV-2 by immunoblot [3,12]. All samples showing dual reactivity on immunoblot were confirmed with the sensitive and specific recombinant peptides from the envelope regions of HIV-1 and HIV-2, designated 566 and 996 respectively [14,15].
The lymphocyte subset typing was performed by flow cytometry (FACSCAN; Becton & Dickinson, San Jose, California, USA) and the total lymphocyte count was performed by an automated blood analyzer (Coulter Counter; Coultronics, Margancy, France) . Of the 35 individuals with monotypic HIV-2 infection, six met the Centers for Disease Control (CDC) revised classification for AIDS, which includes CD4+ lymphocyte counts < 200 × 106 cells/l , whereas 20 of the 35 individuals with monotypic HIV-1 infection were diagnosed with AIDS. Mean CD4+ lymphocyte counts for the monotypic HIV-2-infected individuals was 913 ± 346 CD4+ lymphocytes × 106/l and 425 ± 378 CD4+ lymphocytes × 106/l for monotypic HIV-1 infected individuals.
DNA was purified from frozen uncultured PBMCs, remaining after T-cell subset analysis, using proteinase-K phenol extractions. Briefly, the cells were resuspended in sodium chloride-Tris-EDTA/sodium dodecyl sulfate (SDS) 0.5%/Proteinase-K (100 µg/ml) and incubated at 56°C for 2 h. The cell lysate was then subjected to four successive phenol-chloroform extractions and ethanol precipitated. The DNA was resuspended in Tris-EDTA and the DNA concentration was determined by the optical density at 260 nm.
Proviral HIV-1 and HIV-2 DNA were amplified using two sets of nested PCR primers. Two different regions of env and gag were amplified for each virus type as previously described [17,18]. For each set of primers an appropriate probe was chosen to the respective intervening sequence for Southern blot hybridization. Each set of primers was specifically optimized for MgCl2 concentration, annealing temperature, cycling parameters, and primer concentrations. The PCR reactions were performed in a 50 µl volume mixture containing: 1 µg of DNA, 5 µl of 10 × PCR buffer II and 1.25 units of Taq polymerase (Perkin-Elmer PCR reagents; Roche Molecular Systems, Branchburg, New Jersey, USA), 0.2 mM of each deoxy-nucleotide, and 3 mM MgCl2. Each reaction was subjected to 30–40 cycles of denaturation (1 min at 94°C), annealing (45 sec) with a specific temperature for each set of primers, and extension (45 sec at 70°C). This was followed by a final extension of 3 min at 70°C in an automatic thermal cycler. The second round was performed under the same conditions using 2 µl of the first round PCR product as a template.
As negative controls, DNA from uninfected T-cell lines, PBMCs from an HIV-negative blood donor, and deionized water were included in each step of the manipulation. The PCR products were visualized by electrophoresis of 20% of the second round product followed by ethidium bromide staining and ultraviolet transillumination. The gel was transferred overnight to a nylon membrane 0.22 µm (MSI, Westboro, Massachusetts, USA), cross-linked, and prehybridized for 2 h at 68°C in a solution containing 6 × sodium chloride-sodium citrate (SSC), 5 × Denhardt's, and 0.5% SDS. Hybridization was performed by adding 106 c.p.m./ml of 32P end-labeled oligonucleotide probe to the same buffer, which was incubated overnight at room temperature. The membranes were washed twice: 5 min in 2 × SSC/ 0.5% SDS at room temperature, and 5 min in 2 × SSC/ 0.1% SDS at 10°C below the melting temperature of the probe, followed by autoradiography overnight at 70°C. A sample was judged positive when a PCR fragment of the expected size was positive after Southern blot hybridization.
The sensitivity and specificity of the different primer pairs were tested using a panel of PBMC DNAs from 35 HIV-1- and 35 HIV-2-seropositive individuals from the same study groups as described. In order to determine the limit of detection of the PCR assay, plasmid DNA prepared from full length molecular clones of HIV-1 (HXB2) and HIV-2 (pGH-123, kindly provided by Dr A. Adachi) were diluted in HIV negative PBMC DNA, ranging from 1 to 106 copies per sample, and subjected to PCR and Southern blot hybridization [19,20]. Mixing of diluted HIV-1 and HIV-2 plasmid DNA samples was performed and tested to evaluate the effect of heterotypic template on the detection limits of the assay.
The sensitivity and specificity of the HIV-1 and HIV-2 primer sets were evaluated using uncultured PBMC DNA prepared from individuals with monotypic HIV infection; 35 HIV-1- and 35 HIV-2-seropositives (Table 1). These individuals were randomly selected from the same study groups as the HIV dual reactive individuals and their serostatus was based on the identical serologic algorithm employed for the HIV dual reactives. All samples that demonstrated the correct size fragment by ethidium bromide staining also demonstrated a positive Southern blot signal. Southern blot hybridization increased the sensitivity of the PCR assay 11% for HIV-1 and 25% for HIV-2 (data not shown). Although other studies have shown greater than 95% detection of HIV-1 proviral signals in the absence of Southern blot hybridization, these studies were not conducted on small volume samples collected in field settings . Of the 35 HIV-1 samples, 32 reacted with gag primers, 31 with env primers, and all 35 were amplified with gag and/or env primers (Table 1). The HIV-1 primers did not amplify the HIV-2 samples. Of the 35 HIV-2 samples, 31 were amplified with gag alone, 34 with env primers, and all 35 were amplified with gag and/or env primers. The HIV-2 primers did not cross-react to amplify the HIV-1 samples. Thus, using two primer sets per virus and Southern blot hybridization to confirm the amplified product, the assay yielded 100% type-specific detection of HIV-1 and HIV-2 (Table 1).
The sensitivity of the different HIV primer sets was further evaluated using a dilution series of HIV plasmid in HIV negative genomic DNA, with concentrations ranging from 1 to 106 copies. Both the HIV-2 gag and env primer pairs detected samples with ≥ 10 copies. The HIV-1 gag primer pair also detected ≥ 10 copies, while the HIV-1 env primer pair detected ≥ 100 HIV-1 copies. Thus, if the results from both primer pairs were considered together, the detection limits for HIV-1 and HIV-2 would be the same (Table 2). Of note, all samples were positive in both ethidium bromide staining and Southern blot hybridization.
A mixing study was performed to anticipate the evaluation of samples that would contain proviral sequences of both HIVs. The results from mixing diluted HIV-1 and HIV-2 plasmids demonstrated that the limit of detection for each primer set was not affected by the presence of heterologous plasmid DNA, ranging from 102 to 106 copies (Table 2). However, we recognize that detection limits might differ with integrated proviral targets; HIV-2 cell lines with known proviral copies were not available to allow this determination.
We tested 34 uncultured PBMC DNA samples from individuals with dual serologic profiles; 21 were asymptomatic female commercial sex workers and 13 were hospitalized patients. Among all of the 34 dually seropositive individuals, both HIV-1 and HIV-2 proviral DNA sequences were found in 26 of 34 samples (76%; Table 3). HIV-1 proviral DNA was detected in 33 of 34 samples (97%) amplified with the gag primers, and 30 of 34 samples (88%) which were amplified with env primers. If the results of either primer pair were used together, HIV-1 proviral DNA was detected in 100% of the samples (34 of 34). In contrast, only 26 of the 34 samples (76%) were amplified with HIV-2 primers using gag or env primer results together. As demonstrated with amplification of monotypic HIV PBMC DNAs, Southern blot hybridization increased the sensitivity of all primer sets. For the detection of HIV-1 provirus in individuals with dual serologic profiles, Southern blot hybridization was required in five of 34 (14.7%) of the cases. Whereas the detection of HIV-2 provirus required Southern blot hybridization in 10 of 26 (38.5%) of the same individuals.
The demonstration of HIV-1 and HIV-2 proviral sequences in serologically diagnosed dual reactives differed between the two groups of subjects studied (Table 3). Whereas all 21 asymptomatic female commercial sex workers were confirmed dually infected using PCR (21 out of 21, 100%), only five out of the 13 hospitalized patients showed evidence of dual infection by PCR (38.5%). In the hospitalized patients, all 13 samples were positive for HIV-1 proviral sequences, but eight of the 13 lacked HIV-2 proviral sequences. Due to the systematic discordance of PCR results with serology by study group, we considered differences in the two groups that might account for this.
There was a correlation between lower CD4+ count and the absence of the HIV-2 PCR signal. Among the 34 dually seroreactive individuals, the mean CD4+ cell count was higher among individuals with confirmed dual PCR reaction (548 ± 302 CD4+ lymphocytes × 106/l) versus those that were only HIV-1 PCR-positive (235 ± 170 CD4+ lymphocytes × 106/l), and this was statistically significant (Student's t test, P value = 0.009). The ability to detect HIV-2 provirus by PCR decreased in individuals with < 400 CD4+ lymphocytes × 106/l, while HIV-1 provirus by PCR remained detectable (Pearson χ2 = 4.04, P value = 0.04). In patients with > 400 CD4+ lymphocytes × 106/l, PCR confirmed dual-serologic profiles in 89.5% of cases, whereas only 60% of subjects with < 400 CD4+ lymphocytes × 106/l could be confirmed by PCR. In all cases of serologic and PCR discordance it was the absence of HIV-2 provirus that resulted in the discordance observed, despite the fact that the HIV-1 and HIV-2 PCR assays were considered equally sensitive for proviral detection.
The aim of this study was to genetically characterize HIV-1 and HIV-2 dual infection in individuals demonstrating a dual-serological profile. Two optimized sets of nested PCR primers for each virus type were characterized and found to be of high specificity and sensitivity in the detection of proviral copies. The HIV-1 and HIV-2 PCR assays as developed were also found to be equally sensitive for the detection of the respective targets, and this was not impaired by the presence of heterologous target sequences. The overall prevalence of dual infection, defined as the simultaneous presence of proviral HIV-1 and HIV-2 sequences detectable by PCR in uncultured PBMC DNA, was 76% (26 of 34). In individuals with greater than 400 CD4+ lymphocytes × 106/l, the PCR detection rate was 90%. This was a significantly higher rate of serologic and PCR concordance for dual infection compared to previously reported data. In studies reported from the Ivory Coast, 33% (12 out of 36) of serologically diagnosed HIV-duals were confirmed by PCR, whereas a second study reported 62% (21 out of 34) serology and PCR concordance [9,10]. There may be multiple explanations for the differences in PCR confirmation of dual infection from our study compared to others. The most obvious explanation has been the differences in serologic algorithms that defined dual-reactive individuals, where unusually high prevalence rates of dual-reactives have been reported in some studies [8,10] but not in others [12,15,22].
Differences in the PCR assays employed also may have been responsible for the disparate results. Optimization of the PCR assay is necessary when high sensitivity, specificity, and reproducibility are required. We demonstrated in this study that the use of nested PCR and Southern blot hybridization can improve the sensitivity of the assay significantly. Further, the documented genetic variability of HIV-1  and HIV-2 [24,25] may necessitate the use of different sets of primers from multiple conserved regions in order to maximize the sensitivity of the assay. Hence, the co-amplification of multiple regions of both HIV-1 and HIV-2 in a single reaction for greater efficiency and sensitivity has been proposed [26,27]. The previously published studies of dual infection, used one set of primers per HIV virus type and these were non-nested [8–11]; only one study utilized ethidium bromide and Southern blot hybridization for confirmation of the amplified product . The low rate of PCR confirmation of dual HIV infection may also have been due to the health status of the patients. Our study showed a lower rate of HIV dual PCR confirmation in individuals with low CD4+ lymphocytes counts.
Despite improved concordance between PCR results and serologic diagnosis of dual-reactive individuals, some discordance remains. Although our primer pairs for both HIV-1 and HIV-2 were found to have similar detection limits, it is likely that the range of proviral loads in HIV-1 and HIV-2 infection may differ. Viral isolation from cells and plasma have suggested lower viral load in HIV-2 infection than in HIV-1 infection, both being correlated to health status . Peeters et al., studying seven dual-seropositive individuals, have shown that the sensitivity of HIV-2 PCR detection varied from four of seven for primary lymphocytes, to five of seven for three-week-old cultures and zero of seven for six-week-old cultures, while HIV-1 proviral DNA could be detected at all timepoints . In contrast, quantitative DNA PCR studies have indicated similar proviral loads in HIV-1 and HIV-2 infection . Further studies are clearly needed to determine if viral loads in these distinct HIVs truly differ, as has been suggested.
Although it is generally believed that HIV viral load increases with progression to disease, it is difficult to predict how two different virus infected cell populations might interact in a dually infected individual. HIV-2 is known to be less transmissible and significantly slower in its rate of disease progression compared to HIV-1 [13,22,31]. Thus, it is reasonable to hypothesize that HIV-2 viral infection would be at a replicative disadvantage in the dual infected state. In our study there was a correlation between lower CD4+ count and the loss of the HIV-2 PCR signal. It is conceivable that during the natural history of dual infection, there is an overgrowth of some highly replicating HIV-1 strains which may overwhelm the number of HIV-2 infected cells, thus limiting their detection by the PCR assay.
Alternatively, levels of HIV-2 may decrease over the course of infection and/or virus infected cells may sequester outside of the peripheral blood system. Based on the detection limits of our assay, it is possible that these discordant individuals still harbour HIV-2 provirus, but below the detection limit of 10 copies of HIV-2 per 150 000 PBMC.
Many regions of the world may be affected by both HIV-1 and HIV-2, including most of west Africa and south-western India . Our understanding of how these viruses may interact in vivo is still quite limited. The results from this study demonstrate that most individuals with a carefully defined serologic profile of HIV-dual reactivity also harbour both proviruses. The demonstration that some immunosuppressed individuals may lack or have lost HIV-2 provirus suggests that dual infection may not be a static event. Therefore, over time the dual-infected individual may only harbour significant HIV-1 virus that would be pathogenic and capable of transmission. This would have important ramifications for the future predictions of HIV-1 and HIV-2 spread in many parts of the world. Further prospective studies are clearly warranted to identify and characterize the mechanisms involved in this complex in vivo interaction.
We thank B. Chaplin, C. Mullins and E. Zhang for technical support. This work was conducted under the Inter-University Convention for the Prevention of AIDS and Other Sexually Transmitted Disease, including Cheikh Anta Diop, Harvard, Tours, and Limoges Universities.
1. Barin F, MBoup S, Denis F, et al.
: Serological evidence for virus related to simian T-lymphotropic retrovirus III in residents of west Africa
2. Tedder RS, O'Connor T, Hughs A, Corrah T, Whittle H: Envelope cross-reactivity in Western blot for HIV-1 and HIV-2 may not indicate dual infection
3. World Health Organization: Proposed WHO criteria for interpreting results from Western blot assays for HIV-1, HIV-2, and HTLV-I/HTLV-II
. Wkly Epidemiol Rec
4. Gnann Jr JW, McCormick JB, Mitchell S, Nelson JA, Oldstone MBA: Synthetic peptides immunoassay distinguishes HIV type 1 and HIV type 2 infections
5. Broliden PA, Ruden U, Ouattara AS, et al.
: Specific synthetic peptides for detection of and discrimination between HIV-1 and HIV-2 infection
. J Acquir Immune Defic Syndr
6. Baillou A, Janvier B, Leonard G, Denis F, Goudeau A, Barin F: Fine serotyping of human immunodeficiency virus serotype 1 (HIV-1) and HIV-2 infections by using synthetic oligopeptides representing an immunodominant domain of HIV-1 and HIV-2/simian immunodeficiency virus
. J Clin Microbiol
7. Evans LA, Moreau J, Odehouri K, et al.
: Simultaneous isolation of HIV-1 and HIV-2 from an AIDS patient
8. Rayfield M, DeCock K, Heyward W, et al.
: Mixed human immunodeficiency virus (HIV) infection in an individual: demonstration of both HIV type 1 and type 2 proviral sequence by using polymerase chain reaction
. J Infect Dis
9. Peeters M, Gershy-Damet GM, Fransen K, et al.
: Virological and polymerase chain reaction studies of HIV-1/HIV-2 dual infection in Cote d'Ivoire
10. George JR, Ou C-Y, Parekh B, et al.
: Prevalence of HIV-1 and HIV-2 mixed infections in Cote d'Ivoire
11. Léonard G, Chaput A, Courgnaud V, Sangaré A, Denis F, Brechot C: Characterization of dual HIV-1 and HIV-2 serological profiles by polymerase chain reaction
12. Kanki P, MBoup S, Marlink R, et al.
: Prevalence and risk determinants of human immunodeficiency virus type 2 (HIV-2) and human immunodeficiency virus type 1 (HIV-1) in west African female prostitutes
. Am J Epidemiol
13. Kanki PJ, Travers K, MBoup S, et al.
: Slower heterosexual spread of HIV-2 than HIV-1
14. Guèye-2diaye A, Clark RJ, Samuel KP, et al.
: Cost-effective diagnosis of HIV-1 and HIV-2 by recombinant-expressed env peptide (566/966) dot blot analysis
15. Travers K, MBoup S, Marlink R, et al.
: Natural protection against HIV-1 infection provided by HIV-2
16. Centers for Disease Control: Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults
17. Simmonds P, Balfe P, Peutherer JF, Ludlam CA, Bishop JO, Brown AJ: Human immunodeficiency-infected individuals contain provirus in small numbers of peripheral mononuclear cells and at low copy numbers
. J Virol
18. Grankvist O, Bredberg-Raden U, Gusstafsson A, et al.
: Improved detection of HIV-2 DNA in clinical samples using nested primer-based polymerase chain reaction
. J Acquir Immune Defic Syndr
19. Fisher AG, Collalti E, Ratner L, Gallo RC, Wong-Staal F: A molecular clone of HTLV-III with biological activity
20. Shibata R, Miura T, Hayami M, et al.
: Mutational analysis of the human immunodeficiency virus type 2 (HIV-2) genome in relation to HIV-1 and simian immunodeficiency virus SIV AGM
. J Virol
21. Albert J, Fenyo E-M: Simple, sensitive and specific detection of human immunodeficiency virus type 1 in clinical speciments by polymerase chain reaction with nested primers
. J Clin Microbiol
22. Kanki, P: Epidemiology and natural history of HIV-2
. In AIDS: Etiology, Diagnosis, Treatment and Prevention, 4th Edn
. Edited by DeVita Jr V, Hellman S, Rosenberg S. Philadelphia: JB Lippincott Co.;1997:127–135.
23. Goudsmit J, Back NK, Nara PL: Genomic diversity and antigenic variation of HIV-1: links between pathogenesis, epidemiology and vaccine development
. FASEB J
24. Schulz TF, Whitby D, Hoad JG, Corrah T, Whittle H, Weiss RA: Biological and molecular variability of human immunodeficiency virus type 2 isolates from the Gambia
. J Virol
25. Sankalé JL, Sallier de la Tour R, Renjifo B, et al.
: Intra-patient variability of the human immunodeficiency virus type-2 (HIV-2) envelope V3 loop
. AIDS Res Hum Retroviruses
26. Kumar U, Heredia A, Soriano V, Bravo R, Epstein JS, Hewlett IK: Enhanced diagnostic efficiency of polymerase chain reaction by co-amplification of multiple regions of HIV-1 and HIV-2
. J Virol Methods
27. Loussert-Ajaka I, Simon F, Farfara I, et al.
: Comparative study of single and nested PCR for the detection of proviral HIV-2 DNA
. Res Virol
28. Simon F, Matheron S, Tamalet C, et al.
: Cellular and plasma viral load in patients infected with HIV-2
29. Peeters M, Fransen K, Gershy-Damet GM, et al.
: Effect of methodology on detection of HIV-1/HIV-2 infections in Cote d'Ivoire
. J Virol Methods
30. Berry N, Ariyoshi K, Jobe O, et al.
: HIV type 2 proviral load measured by quantitative polymerase chain reaction correlates with CD4+ lymphopenia in HIV type 2-infected individuals
. AIDS Res Hum Retroviruses
31. Marlink R, Kanki P, Thior I, et al.
: Reduced rate of disease development after HIV-2 infection as compared to HIV-1