The qualitative nature of the primary immune response to HIV infection is a prognosticator of disease progression independent of the initial level of plasma viremia

Twenty-one subjects with well-documented historical, clinical, and laboratory parameters of primary HIV infection were studied serially at various times following infection. All patients included in this study experienced an acute, mononucleosis-like viral syndrome of variable severity. The clinical histories of patients 1–6 have been previously reported (12, 18). Patients 1, 2, 3, 4, 10, and 16 had a severe clinical syndrome, and the first five patients were hospitalized. Only patient 22 had mild symptoms (pharingitis and diarrhea) that did not require medical attention, whereas the remaining 14 patients had a clinical syndrome of moderate severity, which brought them to the attention of physicians. Only patients 15, 20, and 22 received mono-(zidovudine) or combination-(zidovudine plus didanosine) antiviral therapy during the first year of infection.

The Vβ repertoire in freshly unfractionated peripheral blood mononuclear cells (PBMC) collected at different time points from the onset of symptoms was analyzed by a semi-quantitative polymerase chain reaction (PCR). At least a doubling in the relative expression of Vβ families among sequential time points by PCR was required for the change to be considered significant. Briefly, 2 μg of total RNA, extracted by the RNAzol method (23), were reverse transcribed using random hexamers (20 μg ml⁻¹; Promega), avian myeloblastosis virus reverse transcriptase (60 units; Life Science, Arlington Heights, IL) and dNTPs (Boehringer Mannheim). PCR analysis for the different Vβs used a 5′-specific Vβ primer and a common 3′ constant-domain Cβ primer. 5′ and 3′ Cα primers were included in each PCR reaction as internal controls. To monitor the migration of Vβ and Cα bands, 3′ Cβ and 3′ Cα primers were radiolabeled with [³²P-γ]ATP (Amersham). Primer sequences of Vβ-specific oligonucleotides and of the control Cβ and Cα oligonucleotides have been reported (24). One aliquot of each PCR reaction was loaded and separated on 10% polyacrylamide gels containing 7 M urea. Gels were exposed overnight on Kodak storable phosphor screens, and the radioactive signal for each Vβ was quantified using a PhosphorImager (Molecular Dynamics).

Determination of viremia has been performed by competitive PCR in patients 1–6 (25–27), and branched-DNA has been used in the other patients included in this study (28). The same method has been used for the first and the last time points.

With regard to the slopes of CD4⁺ T cell decline, the results were graphed in a bivariate scattergram plot, and differences between the curves submitted to unpaired t test analysis. Similar statistical analysis was performed to determine the differences in the levels of viremia observed in the patients belonging to the three patterns of Vβ perturbations, either during primary infection or 1 year after primary infection.

Analysis of the Vβ repertoire was performed in PBMC samples collected at different time points by using a semi-quantitative PCR assay (21, 24, 29). The first PBMC sample was generally collected within 1–8 weeks from the onset of symptoms, and longitudinal analysis was performed on at least three PBMC samples collected during the first 16–20 weeks from the onset of symptoms. Perturbations (i.e., increments and/or declines) of a Vβ family were considered to be significant only if there was at least a 2-fold change in the percentage of the Vβ family in question over several time points. However, 2-fold changes were not considered to be significant if they involved Vβ families that were expressed at very low levels (i.e., 1–2%). In fact, when the Vβ families are expressed at very low levels, the bands corresponding to these Vβs are generally very faint, and may be very difficult to measure accurately on the PhosphorImager. For these reasons, other authors (30) prefer to assign an arbitrary value of <1% to those families expressed at very low levels. The rationale for considering a 2-fold increase to be significant is based on the findings that: (i) changes of this magnitude have always been confirmed by cytofluorometric analysis when antibodies specific for the expanded Vβ families were available (21, 31), (ii) changes of this magnitude have always been found to be associated with the oligo-monoclonal nature of the expansions of the Vβ families in question (21, 31), and (iii) the TCR repertoire is very stable in the absence of antigenic stimulation. This latter point is supported by the observation that the TCR repertoire was consistently found to be identical in healthy HIV-negative homozygotic twins (29, 32, 33). It is obvious that these results would not be compatible with a high rate of spontaneous perturbations in the TCR repertoire.

Longitudinal analysis of the Vβ repertoire was performed on PBMC samples collected from 21 subjects during primary HIV infection and during and after the transition to the chronic phase. Three major patterns of Vβ perturbations were observed: (i) expansion of a single Vβ family (Type 1), (ii) expansion of two Vβ families (Type 2), and (iii) expansion of multiple Vβ families or no detectable expansions with polyclonal activation of multiple Vβ families (Type 3).

Representative examples of the above patterns are shown in Fig. 1. With regard to the distribution of the 21 patients into the 3 patterns of Vβ perturbations, 4 patients had a Type 1 pattern, 4 had a Type 2 pattern, whereas most patients (13 of 21) manifested a Type 3 pattern (Fig. 2). It is important to underscore that with regard to the criteria used for grouping the patients, the division was made strictly on the basis of the patterns of Vβ perturbations, which have been blindly analyzed, several months prior to the evaluation of the subsequent clinical outcome. The magnitude of the changes, i.e., fold increase or decline, in the relative percentage of the perturbed Vβ families observed over time were extremely variable among the different Vβs, and were not related to the type of Vβ perturbation observed (Fig. 2A). This was not surprising since there are substantial differences in the steady-state percentage of expression of the different Vβ families, and thus the number of fold increase or decline observed during an ongoing immune response can be, at least in part, dependent upon the initial steady-state level, and does not necessarily reflect the magnitude of the cell population that has been mobilized during the immune response. In addition, there were substantial differences in the percentage of the single Vβ families found perturbed in Type 1, Type 2, and Type 3 patterns (Fig. 2B). The percentage of the Vβ families found perturbed in the four patients with Type 1 pattern was higher (ranging between 18.5% and 37%) compared with that in Type 2 and Type 3 patterns (Fig. 2B). Despite the differences in the percentage of the different Vβ families among the three Types of Vβ perturbations (higher percentage of the perturbed Vβs in Type 1 versus Type 2 and Type 3), it is of note that if one totals the percentages of the different Vβs found perturbed in each patient, the overall percentage of the cell population potentially involved in the ongoing primary immune response to HIV infection generally ranged between 16% and 38% regardless of the type of Vβ perturbation observed (Fig. 2B). This is also the case in those patients (patients 4, 9, 10, 19, 21, and 25) in whom no significant Vβ perturbations were detected (Type 3); it has been consistently observed that, similar to those patients with significant Vβ perturbations, a large percentage (up to 50%) of CD8⁺ T cells expressed the HLA-DR activation marker (data not shown; refs. 21 and 31), which generally defines the activated cell population potentially involved in the ongoing immune response (34). In these patients, HLA-DR expression is distributed among different CD8⁺ Vβ cell subsets, whereas in patients with clear-cut Vβ expansions, HLA-DR antigen is preferentially expressed on the expanded Vβs (21, 31). Therefore, because multiple CD8⁺ Vβ cell subsets are activated even when Vβ perturbations were below 2-fold, patients showing these features were included in the Type 3 pattern.

It was of interest to determine whether the levels of viremia detected in the first sample of plasma collected during primary infection were significantly different among the three groups of HIV-infected subjects experiencing different patterns of Vβ perturbations. In fact, there were no significant differences in the levels of viremia detected during primary infection among the three groups of patients with Vβ perturbations (Type 1 versus Type 2, P = 0.53; Type 1 versus Type 3, P = 0.29; and Type 2 versus Type 3, P = 0.77) (Table 1). Similarly, in a study including 72 patients with primary HIV infection, no relationship has been found between the levels of viremia detected during the first 6 months after infection and disease progression (T.S. and L.C., personal communication).

An important issue to investigate was the relationships among the three patterns of Vβ perturbations observed during primary infection and the rate of disease progression. In a preliminary analysis, CD4⁺ T cell counts determined approximately 1 year after primary infection were used as an indicator of the status of progression of HIV disease. The four HIV-infected subjects (patients 1, 5, 14, and 16) with a Type 1 pattern of Vβ perturbation had already progressed to AIDS 1 year after primary infection as indicated by CD4⁺ T cell counts below 200/μl (Table 1); the mean CD4⁺ T cell count in patients with Type 1 pattern was 101 ± 34 μl (Table 1). CD4⁺ T cell counts detected in patients with Type 2 pattern (patients 2, 3, 8, and 22) were significantly higher (456 ± 103 μl) than those found in patients with Type 1 pattern (Table 1). In 10 of 13 patients with Type 3 pattern, CD4⁺ T cell counts were above 500/μl, and the overall mean CD4⁺ T cell count in this group of patients was 651 ± 226 μl (Table 1). A more precise analysis of disease progression was performed by analyzing the slopes of decline of CD4⁺ T cells (Fig. 3). The values of CD4 T cell counts shown for each time point are the means of CD4 T cell counts detected in the patients (4 for type 1, 4 for type 2, and 13 for type 3) with different patterns of Vβ perturbations. The minimal estimates of CD4⁺ T cell loss per week obtained from this analysis were 8.76 cells for Type 1 pattern, 6.14 cells for Type 2, and 3.34 cells for Type 3. These slopes of decline of CD4⁺ T cells were significantly different among the three groups (Type 1 versus Type 2, P = 0.01; Type 1 versus Type 3, P = <0.0001; and Type 2 versus Type 3, P = 0.03) (Fig. 3). Therefore, these results indicate that the rate of disease progression was substantially different in HIV-infected individuals who had different patterns of Vβ perturbations during primary infection.

The patterns of Vβ perturbations were compared with the levels of viremia detected approximately 1 year after seroconversion. Levels of viremia at this time point were significantly higher in patients with Type 1 pattern compared with those with Type 3 pattern (Type 1 versus Type 3, P = 0.04); in addition, there was a trend toward a difference in levels of plasma viremia in Type 1 and 2 versus Type 3 that did not reach statistical significance (P = 0.09). These results were consistent with previous studies showing that the levels of viremia observed after 6 months following primary HIV infection and during chronic infection are a strong predictor of the rate of disease progression (11, 35) (T.S. and L.C., personal communication).