Background.During acute human immunodeficiency virus (HIV) infection, high viral loads and the induction of host immune responses typically coincide with the onset of clinical symptoms. However, clinically severe presentations during acute HIV type 1 (HIV-1) infection, including AIDS-defining symptoms, are unusual.
Methods.Virus isolates were tested for clade, drug susceptibility, coreceptor use, and growth rate in 2 case reports of sexual transmission of HIV-1 infection. Human leukocyte antigen (HLA) genotype was determined, and HIV-1 specific cytotoxic T lymphocyte responses to an overlapping peptide set spanning the entire HIV clade A and clade B proteome were assayed.
Results.The viruses isolated in the 2 unrelated case reports of severe primary HIV-1 infection showed R5/X4 dual-mixed tropism, belonged to clade B and CRF02-AG, and were highly replicative in peripheral blood mononuclear cell culture. Impaired humoral responses were paralleled by a profound absence of HIV-1 specific cytotoxic T lymphocyte responses to the entire viral proteome in the 2 case reports. In 1 case report for which the virus source was available, there was a remarkable HLA similarity between the 2 patients involved in the transmission event, because 3 of 4 HLA-A and HLA-B alleles had matched HLA supertype for both patients.
Conclusions.The data suggest that concurrence of viral and host factors contributes to the clinical severity of primary HIV-1 infection and that patients infected with highly replicative, dual-tropic viruses are more prone to develop AIDS-defining symptoms during acute infection if they are unable to mount humoral and cellular HIV-1 specific immune responses. The presence of concordant HLA supertypes might facilitate the preferential transmission of HLA-adapted viral variants, further accelerating disease progression.
Symptoms of acute retroviral syndrome typically coincide with high-level viremia and the induction of the host's initial adaptive immune response [1, 2]. However, clinically severe presentations during acute HIV-1 infection, including AIDS-defining symptoms, are considered to occur infrequently . Furthermore, epidemiological studies have revealed that, in the absence of treatment, in <0.5% of HIV-1 infected individuals, the infection progresses to AIDS within 1 year after primary infection . A complex interplay between multiple viral and host factors is most likely to be involved in accelerating disease progression. Among these myriad factors, CXCR4 tropism has been associated with high virulence . Moreover, HLA class I concordance between individuals and the inability to elicit specific cytotoxic T lymphocyte (CTL) responses have been suggested to increase the rate of transmission of HIV-1 infection and disease progression [6, 7].
In our study, we investigated the immunological and virological factors that contributed to the development of AIDS-defining pathogenesis in 2 independent case reports of unusually severe, acute, sexually transmitted HIV infection. The clinical and diagnostic outcomes are shown in figure 1.
Patients, Materials, and Methods
The study participants provided written informed consent to participate in this study, which was approved by the institutional review boards of the hospitals where the participants received medical care. HIV-1 was isolated from the patients' PBMCs, and the viral stocks were titrated in TZM-bl cells . Coreceptor use of primary HIV-1 isolates was assessed by infection of U87.CD4 cells expressing either CCR5 or CXCR4 . Syncytia induction was determined in vitro in MT-2 cells. CCR5 was genotyped in genomic DNA extracted from cryopreserved PBMCs to detect the Δ32 deletion. The growth rate of the viral isolates was determined by infecting phytohemagglutinin-stimulated donor PBMCs . To assess the presence of drug resistance associated mutations, we sequenced the HIV protease region (codons 1 99) and reverse-transcriptase region (codons 40 247) from a plasma sample obtained before the initiation of antiretroviral therapy.
To determine whether the index patient in case report 2 harbored the same virus as the suspected source patient, viral RNA was extracted from plasma. The pol (protease and first 235 codons of the reverse transcriptase) and env (C2 to V5 regions) genes were sequenced [10, 11]. In addition, a total of 46 molecular clones encompassing the env gene were used to estimate diversity in the plasma viral RNA for the source and index patients . Sequence alignments were obtained using Sequencher, version 4.6 (Gene Codes), and ClustalW and were manually edited in the regions of variable length. Genetic distances and evolutionary rates were computed using a Kimura 2 parameter model. Neighbor-joining phylogenetic trees of each patient's pol and env sequences were constructed using MEGA3. The reliability of phylogram clustering was assessed by bootstrapping analyses. Coreceptor use was inferred from env clonal sequences with use of phenotype prediction tools .
HLA class I and class II genotypes were identified by high-resolution sequencing in an approved clinical laboratory. HLA class I supertype assignment was based on functional classification for the many different 4-digit, high-resolution HLA alleles that overlap in their peptide-binding specificities .
Cellular immunity to HIV and Epstein-Barr virus (EBV) was assessed by IFN-γ enzyme-linked immunospot assays. T cell responses to an overlapping peptide set spanning the entire HIV clade A and clade B protein sequence were detected . In addition, optimal epitopes known to be presented by the patients' HLA class I alleles were either included in their clade-specific consensus version or based on sequence variants identified in the index or source patient. To assess general immune reactivity, 3 peptide pools containing a previously described set of EBV-derived optimal CTL epitopes were also tested . Positive responses were determined on the basis of specific cutoffs that are defined elsewhere .
Case report 1. A 26-year-old Spanish man with a history of recurrent conjunctivitis and mild bronchial asthma who reported having had only 1 heterosexual partner during the previous 2 years was admitted to a hospital in Ferrol, Spain, for progressive dyspnea, tachypnea, tachycardia, fever, and nonproductive cough. Four months before hospital admission, the patient had received a diagnosis of leukopenia but had negative results of HIV serological examination. Four weeks before hospital admission, he developed maculopapular rash, conjunctivitis, balanitis, and oral thrush. At hospital admission, laboratory testing revealed leukocytosis, bilateral alveolar-interstitial lung infiltrates, and severe hypoxia. Two days later, he fulfilled the criteria for acute respiratory distress syndrome and required orotracheal intubation, mechanical ventilation, and broad-spectrum antibiotic therapy. An EIA was weakly positive for HIV-1 antibodies, and the Western blot results were undetermined. Pneumocystis jiroveci was detected in a bronchoalveolar lavage specimen, and Candida albicans was isolated from oropharyngeal exudates. The patient's plasma HIV-1 RNA level was 31,600 copies/mL, and his CD4+ T cell count was 108 cells/mm3. Antiretroviral therapy with tenofovir, emtricitabine, lopinavir-ritonavir, and enfuvirtide was initiated. The acute respiratory distress syndrome was refractory to treatment, and the patient died of multiorgan failure 3 weeks after hospital admission.
Case report 2. Case report 2 involved an index patient and the presumed source patient. The index patient was a 15-year-old girl from Ecuador who was admitted to a hospital in Badalona, Spain, with fever, severe headache, fatigue, malaise, severe odynophagia, and a 4-day history of continuous retrosternal chest pain. She had a 1-month history of asthenia, a 2-kg weight loss, and fever. Her previous medical history was unremarkable. The index patient reported having had only 1 regular sexual partner during the previous 2 years. At hospital admission, she was febrile and had enlarged bilateral cervical lymph nodes, oral thrush, and soft, nonswollen hepatomegaly (1.5 cm). C. albicans was subsequently isolated from oropharyngeal and vaginal exudates. Haemophilus influenzae (biotype 2) was isolated from a sputum sample. Serological testing was reactive for IgG against cytomegalovirus and Toxoplasma gondii. An EIA for HIV-1 antibodies was weakly reactive, and a line immunoassay for detection of IgG antibodies against HIV-1/2 disclosed only 1 band for gp41 in 2 independent samples obtained 5 days apart. The index patient's HIV-1 RNA level was 2×106 copies/mL, and her CD4+ T cell count was 0 cells/mm3. The patient initiated therapy with tenofovir, emtricitabine, and lopinavir-ritonavir. Three days later, she developed cognitive deficits. Cranial MRI revealed ischemic and hemorrhagic supratentorial and infratentorial lesions in the cortical-subcortical interface and basal ganglia. Cryptococcus neoformans was isolated from blood and CSF cultures. The patient had a favorable clinical course with appropriate treatment. Two months later, her HIV RNA level was undetectable, and her CD4+ T cell count was 51 cells/mm3 (CD4+ T cell percentage, 3%). Despite repetitive antibody testing, complete seroconversion did not occur until 9 months after initial presentation.
The index patient's partner, presumably the source patient, was a 23-year-old man who was also from Ecuador. He confirmed having had several sexual partners, but he had never been tested for HIV-1 infection. EIA and line immunoassay were reactive for HIV-1. His plasma HIV-1 RNA level was 40,000 copies/mL, and his CD4+ T cell count was 32 cells/mm3 (4%). The source patient initiated therapy with tenofovir, emtricitabine, and lopinavir-ritonavir. Nine months later, his HIV RNA level was undetectable, and his CD4+ T cell count was 130 cells/mm3.
Case report 1. Laboratory assessment of the patient indicated a change in HIV-1 antibody reactivity near the time of presentation. The results of 3 previous assays (HIV-1 antibody , nucleic acid , and antigen-based assays) performed within 9 months before presentation were negative. Antibody and antigen tests and Western blots became partially reactive, and plasma HIV-1 RNA was present at the time of presentation, suggesting HIV-1 primary infection (table 1).
The replication-competent virus isolated from PBMCs was able to infect and replicate in both CCR5 and CXCR4-U87.CD4 cells, as concluded from the p24 antigen production and the formation of syncytia in the cell cultures (figures 2A and 2B). The patient did not have a Δ32 genotype in the CCR5 chemokine receptor gene that might have explained an early selection of CXCR4-tropic viruses . The production of p24 antigen in growth kinetics cultures of donor PBMCs was similar to the production seen with the laboratory-adapted viral strain HIV-1>NL4 3 (figure 2C). Phylogenetic analyses with boot-scanning methods for the genetic subtyping of pol indicated the presence of a subtype B virus. The HIV-1 genotype had no drug resistance associated mutations. The results of HLA typing are shown in table 1.
Case report 2. Clinical symptoms and analytical results for the index patient were consistent with a diagnosis of advanced HIV-1 infection and AIDS. However, the patient denied HIV risks other than sexual contact with her partner during the previous 2 years. The patient's mother tested negative for HIV-1 infection, thus excluding potential vertical transmission. Moreover, the viruses isolated from the source and index patients were similar both phenotypically and genotypically (figure 3 and table 1). Laboratory assessment of the index patient's original sample provided clear reactivity data on the presence of HIV-1 antigens, but despite high levels of immunoglobulins, antibody-based tests showed partial reactivity, indicating a lack of HIV-1 specific antibodies (figure 1 and table 1).
Bootstrap analysis of pol and env sequences from the index and source patients revealed values of ⩾99% in 1000 replicates (figure 3A and 3B), indicating that sequence clustering was unlikely to have occurred by chance. The genetic distance between the pol sequences from the index and source patients was <0.1%, whereas the mean genetic diversity between randomly selected sequences from local, epidemiologically unrelated HIV-infected individuals with the CRF02-AG subtype was 3.0%. The high degree of similarity between viral sequences indicates a likely viral transmission from one patient to the other. Clonal analysis of env sequences indicated that all sequences from the index patient were closely related (mean diversity, 1.8%), and the source patient's sequences had a mean diversity of 2.4% (figure 3C and 3D).
After 5 days of culture, the viruses isolated from the source and index patients were able to infect and replicate in both CCR5 and CXCR4-U87.CD4 cells, as indicated by the p24 antigen production and the formation of syncytia in the cell cultures (figure 3E and 3F). Neither of the individuals showed a Δ32 genotype in the CCR5 chemokine receptor gene. Phenotypic inference of the V3 amino acid sequence in multiple clones from each individual suggests that all clones from the index patient could use CXCR4 (R5X4 tropism) for viral entry, whereas the source patient had clones that could use only CCR5 and clones that could use CXCR4 (R5X4 tropism) (figure 3C). The production of p24 antigen in growth kinetics cultures in PBMCs was equal in the 2 viral isolates and comparable to that in the laboratory-adapted viral strain HIV-1>NL4 3 (figure 3G).
Genetic subtyping of the gag, pol, and env genes in the patients' viruses indicated that both patients had the AG circulating recombinant form 02 (CRF02-AG). Drug-resistance genotyping revealed no drug resistance associated mutations in the reverse transcriptase. Several polymorphisms were detected in the protease gene; this might have been associated with possible tipranavir resistance in non subtype B viruses (table 1).
Cellular immune responses. The patient in case report 1 showed a single weak response against 1 overlapping peptide that was not subsequently observed in the reconfirmation test and only a borderline response to 1 EBV-peptide pool. This atypical lack of EBV-specific CTL responses suggests a widespread impairment of the ability to mount adequate CTL responses .
The index and source patients in case report 2 expressed HLA class I alleles that were highly related. In fact, 3 of the 4 HLA-A and -B alleles had matched HLA supertype for both 2 patients (table 2). The source patient showed weak responses to 3 different regions of the virus (table 3), which represented an overall weak response rate, compared with a median of 17 responses among >300 previously tested patients with chronic HIV infection (authors' unpublished data) . The index patient showed an even weaker HIV-specific T cell response to only a single peptide, which was detected before the patient received treatment but was subsequently lost 9 months after infection. Importantly, the index patient was able to mount a T cell response to a peptide pool containing EBV-derived CTL epitopes, which indicated that the absence of HIV-specific T cells was not attributable to poor cell viability or a general immune incompetence in this individual.
Although some responses to autologous sequence variants that were not tested may exist, the data are in line with a remarkable absence of HIV-specific T cell immunity in both the patient in case report 1 and the index patient in case report 2. This absence of immunity may be related to the extraordinarily fast disease progression in these individuals.
The interplay between the viral and host factors influencing accelerated disease progression is complex and poorly understood. The 2 temporarily coincident case reports described here suggest immediate progression to AIDS from primary HIV-1 infection after sexual transmission. In both cases, the diagnosis of primary HIV-1 infection was supported by nucleic acid and antigen-based screening tests, with an evolving antibody pattern. The patient in case report 1 tested negative for HIV several times before presentation, and subsequent HIV-1 Western blot was only partially reactive. The index patient in case report 2 lacked previous negative test results but presented with a very high plasma HIV RNA level, which is consistent with acute HIV infection . In addition, HIV-1 specific antibody tests were only partially reactive and did not become positive until 9 months after the initiation of antiretroviral treatment. Although detection of HIV-specific antibodies after the onset of symptoms of primary infection can take 5 15 days , complete seroconversion may occasionally be delayed until 12 months after identification of infection by antigen testing, when virological control has been achieved with effective antiretroviral therapy [19, 20]. The fact that serum levels of IgG, IgM, and IgA (in the index patient in case report 2) were within the reference range or higher suggests polyclonal B cell activation . Moreover, positive IgG responses to cytomegalovirus, T. gondii, and hepatitis A virus indicate the ability of antibodies to maintain an appropriate response against microorganisms that cause persistent latent infection. Although plasma viral load had greatly decreased with treatment, CD4+ T cell recovery increased slowly from total absence at presentation, which could have delayed HIV seroconversion.
The diversity of the viral population in HIV-1 env increases in parallel with divergence at a rate of 1% per year for a few years after seroconversion, before reaching a peak and then leveling off or decreasing . Nevertheless, the rates of diversity are higher among patients who have a sharp decrease in their CD4 T cell count . In case report 2, the mean HIV-1 diversity was lower in the index patient than it was in the source patient, indicating that viral evolution took longer in the latter. Although viral diversity tends to decrease in the later stages of infection, most of the genetic distances would remain >2% , thus supporting the direction of transmission in this pair and the theory of a very early presentation after viral transmission to the index patient.
The development of acute retroviral syndrome typically coincides with high-level viremia and the host's initial immune response. However, these 2 case reports reveal primary HIV-1 infection with unusually severe clinical symptoms. Other reports have described severe presentations during primary HIV-1 infection, including acute myopericarditis, renal failure, acute liver failure, and opportunistic infections [24,27], but viral and host factors have not been addressed in detail.
In both case reports, the virus isolated from the patients' PBMCs was able to use CCR5 and/or CXCR4 as entry coreceptors and to replicate very efficiently in phytohemagglutinin-stimulated donor PBMCs. These data indicate that both viral isolates were either dual-tropic viruses or a mixed population of CCR5-tropic and CXCR4-tropic viruses with high replication capacity. This observation would suggest that the transmitted virus had the ability to deplete CCR5+ and CXCR4+/CD4+ T lymphocytes, which may help to explain the total loss of the CD4+ T cell population and rapid clinical progression observed in the index patient at the time of transmission. Infection with dual-tropic HIV-1 variants in injection drug users has been associated with an immediate and rapid decrease in total T cell count and progression to AIDS within 4 years after the estimated time of infection . Furthermore, patients who experienced seroconversion of CCR5Δ32/Δ32 and showed the uncommon pattern of early syncytia-inducing virus and rapid decrease in CD4+ T cell count had a uniformly high viral load and dual-tropic coreceptor use . A link between the detection of syncytia-inducing variants and a rapid decrease in CD4+ T cell count in vivo has already been established .
Despite the fact that the characterized viral subtype CRF02-AG in case report 2 is rather unusual in our area (1.1% of pol sequences tested for antiretroviral resistance during 1999 2007), it is still the second most common non-B subtype. CRF02-AG is the predominant and most rapidly spreading HIV strain in West Africa and western Central Africa [31, 32], which raises concerns about its superior replication fitness and/or transmission efficiency. In fact, primary HIV-1 CRF02-AG isolates from Cameroon exhibited higher ex vivo replicative fitness than did subtype A and G viruses from the same geographic region [33, 34]. These observations are consistent with the high replication rate that we observed in primary phytohemagglutinin-stimulated PBMCs, although we compared these rates with the rate of replication of the laboratory-adapted subtype-B HIV-1>NL4 3 strain.
In our case reports, concurrent host factors may have also contributed to higher susceptibility to HIV-1 infection or disease progression. For example, specific HLA haplotypes have been proposed as an important risk factor in this context . Among these, HLA-B*35, which is in high linkage disequilibrium with HLA-Cw*04, has been consistently associated with rapid progression to AIDS [36,38]. Specifically, the allele HLA-B*3503, present in the patient in case report 1, has been reported to increase the risk of progression to AIDS by 2.7-fold (95% CI, 1.7 4.3; P<.001) . In case report 2, the source patient expressed the HLA-A*68, HLA-B*53, and HLA-Cw04 alleles, which have been associated with rapid disease progression . Although, none of the alleles in the index patient have been associated with accelerated disease progression, the donor and recipient expressed 3 of 4 HLA-A and HLA-B alleles that were in the same HLA supertypes (i.e., clusters of functionally related, 4-digit, high-resolution HLA class I alleles) . This may have facilitated the transmission of viruses with CTL escape mutations, thus diminishing the number of epitopes recognized in the newly infected individual [6, 7, 40, 41]. Remarkably, optimal epitope variants representing autologous sequence diversity did not elicit a response, which suggests effective CTL escape (table 2). This hypothesis would fit with the fact that the index patient in case report 2 showed almost complete absence of HIV-1 specific CTL responses; this is rather unusual during primary HIV-1 infection. In a previous study, only 1 of 5 patients who presented with primary HIV-1 infection showed absence of precursor CTL specific for cells expressing viral proteins . Another study involving acute and early HIV-infected patients reported a slightly higher breadth and magnitude of HIV-1 specific CTL responses . However, that study used a less comprehensive pool of overlapping peptides, and contrary to our case reports, in which there was a persistent absence of responses at month 9, CTL responses increased after 6 12 months of treatment. Although we could not identify the source patient in case report 1, the lack of HIV-1 specific CTL responses in the index patient might allow us to speculate a potential HLA class I concordance at transmission. Nonetheless, the inability to elicit HIV-1 specific CTL responses at the time of primary infection was paralleled in these 3 patients with AIDS-defining pathogenesis and severe clinical presentation. Although HIV-1 specific CTL responses have been considered to be a crucial factor in HIV disease progression, we had limited experimental and clinical evidence of the detrimental effect that the inability to elicit these responses might have in symptomatic primary HIV-1 infection [1, 43]. Moreover, the coincident inability in these 2 case reports for the patients to mount an effective adaptive immune response against HIV-1, albeit not to other pathogens, might be a consequence of a potential defect in the innate immunity. Clearly, further studies involving these and other patients with accelerated disease progression will be needed to address these factors.
In conclusion, we describe 2 case reports of sexual transmission of highly replicative, dual-tropic HIV-1 of subtypes B and CRF02-AG that resulted in an aggressive clinical progression to severe symptomatic AIDS in young patients. Adaptive cellular and humoral immune responses in the host might have simultaneously failed to control the virus.
We thank T. Puig, E. Grau, R. Ayen, J. R. Santos, V. González, M. Juan, E. Palou, and B. Ortiga, for expert medical and technical assistance in this study; A. Gladden, for sequencing analysis; and R. Haubrich, for helpful discussion and critical manuscript review.
Financial support. Spanish Ministry of Education and Science (SAF2004-06991 and SAF2007-64696), Spanish AIDS Network “Red Temática Cooperativa de Investigación en SIDA” (RD06/0006), “Ciber en Epidemiología y Salud Pública,” Program to HIV Vaccine Development in Catalonia, Fundación para la Investigación y Prevención del Sida en España (36523/05, 36356/05, and 36621/06), and Departament d'Universitats, Recerca i Societat de la Informació from the Generalitat de Catalunya (to N.I.-U.).
Potential conflicts of interest. All authors: no conflicts.