In this observational, phase I, pilot study conducted at Chelsea and Westminster Hospital, London, 36 antiretroviral-naive patients were initiated on ART at week 0, which was continued for the duration of the study. ART comprised 2 nucleoside analogues and one protease inhibitor or non-nucleoside reverse transcriptase inhibitor. At week 17 patients were randomized to receive immunotherapy with IL-2 and/or therapeutic immunisation with a gp120-depleted whole inactivated HIV-1 immunogen. Sufficient Remune™ was donated for use in 20 patients by Immune Response Corporation (IRC), Carlsbad, CA, USA. Patients were randomized at week 17 only if their viral load was <50 copies ml/plasma and CD4 T-cell count was ≥ 300 cells/μl blood at week 16. Treatment groups for randomization were as follows: A) ART alone (n = 9); B) ART plus IL-2 (Proleukin™) (n = 11); C) ART plus IL-2 and Remune™ (n = 7); and D) ART plus Remune™ (n = 9). IL-2 (5 × 10⁶U) was administered S/C, twice daily, for 5 days at weeks 17, 21 and 25. 100 μg Remune™ was administered I/M at weeks 17, 29, 41 and 53. Laboratory analysis was conducted at weeks- -6, -3 and 0 before ART, and weeks 1, 2, 4, 8, 16, 17, 21, 25, 29, 41, 53 and at study completion at week 65. The primary outcome was induction of positive changes in lymphocyte proliferative responses to HIV-1 antigens. In addition to the main study time points, a further sub-study of viral loads and lymphocyte subset numbers was conducted in a sub-set of patients (n = 15) receiving IL-2 in groups B and C on the 5^th day of each IL-2 cycle, i.e. at weeks 18, 22 and 26. This sub-study was initiated after the main study had begun and included all patients receiving IL-2 in groups B and C from the date of its inception. Appropriate regulatory approval was granted by Riverside Ethics Committee and patients gave written informed consent.

PBMC were separated from whole blood by density gradient centrifugation and cultured in RPMI-1640 with NaCO₃ (Sigma, Poole, UK) and 100 IU/ml penicillin, 100 μg/ml streptomycin and 2 mM L-glutamine supplemented with 10% human AB plasma (all Sigma).

Plasma viral load was measured using the Bayer HIV-1 RNA 3.0 assay (bDNA) (lower limit 50 copies/ml) (Bayer Diagnostics, Newbury, UK). Whole blood lymphocytes were counted with monoclonal antibodies to: CD3, CD4, CD8, and CD45 (Tetra One, Beckman Coulter, High Wycombe, UK) on an Epics XL-MCL flow cytometer (Beckman Coulter).

Recombinant HIV-1 and recall antigens (Medical Research Council Centralised Facility for AIDS Reagents, National Institute for Biological Standards and Controls, Potters Bar, UK) were used as described previously 2, at a final concentration of 10 μg/ml. Remune™ and its native p24 (IRC, Ca) were used at 3 μg/ml 28.

Proliferation assays and supernatant collection for IL-4 assessment were conducted as previously described 2. Stimulation indices (SI) for triplicates (standard error <15%) were calculated as antigen-stimulated β-particle counts per minute (cpm)/no antigen cpm. A positive response was regarded as a SI >5.

IL-4 bioassays were carried out as previously described 2 using IL-4 (CT.h4S) dependent cell lines. Briefly, IL-4 was measured in culture supernatants of HIV-1 antigen stimulated proliferation assays. Proliferation of the IL-4-dependent cell line CT.h4S was measured by incorporation of tritiated thymidine. Results are presented as IL-4 driven cpm.

In vivo delayed-type hypersensitivity (DTH) skin tests to Remune™ antigen were performed to assess HIV-1-specific cell-mediated immune responses as described elsewhere 29 in all patients at weeks 17, 29 and 53.

HLA haplotypes of patients were assessed by PCR-SSP 30.

IFN-γ ELIspots were conducted as described previously 31. Briefly, 2.5 × 10⁵ PBMC from individual patients who had been tissue typed were cultured with or without 10 μg/ml of appropriate HLA-restricted peptides or PHA (positive control) in 96-well anti-IFN-γ (Mabtech, Stockholm, Sweden) coated PVDF-backed plates (Millipore, Watford, UK). After overnight incubation IFN-γ spot-forming cells (SFC) were detected according to the manufacturer's instructions (Mabtech).

Immunoglobulin (Ig)G was purified from patient and HIV-1 negative plasma using the MAbTrap kit (Amersham Biosciences, Little Chalfont, UK), and quantified by Protein Assay (Bio-Rad, Munich, Germany). Two-fold serial dilutions of purified IgG, from 1 mg/ml, were incubated with 100 focus forming units (FFU) of HIV-1_SF162 for 1 hr at 37°C and then plated onto NP2/CD4⁺/CCR5⁺ cells. Infection was detected after 48 hrs by p24-immunostaining, as detailed before 32. The percentage reduction of infection was calculated: 100 × (1-(mean FFU with patient IgG)/(mean FFU with seronegative IgG at 1 mg/ml)).

Changes in lymphocyte subsets and viral load between week 0 and 16, weeks 0 and 65 and weeks 16 and 65, were assessed using the Wilcoxon Signed Rank test. Proliferation and IL-4 data were assessed with repeated measures analysis of variance using MIXED procedure in SAS statistical software. For log-transformed antigen-specific proliferative stimulation indices, between and within subject weeks 0 to 17, weeks 17 to 65 and weeks 0 to 65, adjusted separate slopes were estimated for each study arm with 95% confidence interval. Differences in overall viral load from week 18–26, between groups B and C were assessed using the Pearson Chi-squared test.

Fifty two anti-retroviral naïve patients were screened for this study which was carried out at the Chelsea and Westminster Hospital, London over a six year period. Of the screened patients 16 dropped out prior to randomisation at week 17. All the remaining 36 patients who were randomised completed the study. The mean age of patients at study entry was 38.75 years. There was one female patient (2.78%) and 100% of patients were of white European ethnicity.

Median viral loads and CD4 T-cell counts for each group are depicted in Figure 1. The median week 0 viral load for all patients was 88,699 copies/ml plasma (range 50–779,254). The median absolute CD4 T-cell count was 294 cells/μl whole blood (range 76–551). Nineteen of 36 patients (53%) had CD4 T-cell counts below 300 cells/μl blood at pre-ART time points.

By week 17 the median viral load in all groups was less than 50 copies. All patients, except 5 (one in Group A, three in Group B, one in group D), maintained viral suppression at the main study time points from week 17 onwards while receiving ongoing ART. One patient in Group C (patient 11) elected to discontinue ART at week 47 causing subsequent rebound in viraemia. At week 17 the median CD4 T-cell count in each complete group was as follows: Group A, 456 cells/μl (range 261–673) (Figure 1A); Group B, 377 cells/μl (range 81–742) (Figure 1B); Group C, 468 cells/μl (range 306–680) (Figure 1C); Group D, 337 cells/μl (range 289–702) (Figure 1D). IL-2-induced increases in absolute CD4 T-cell counts are also apparent in groups B and C. By week 65 these values were 602 cells/μl (range 300–918) for Group A, 731 cells/μl (range 253–1025) for Group B, 619 cells/μl (range 358–1404) for Group C and 505 cells/μl (range 345–768) for Group D.

HIV-1-specific proliferative responses increased transiently for many patients. Regression analysis of these responses revealed no significant changes in group A (Fig. 2A). In group-B from week 0 to 65 a significant increase in the np24 response (p = 0.005) was observed (Fig. 2B). In group C no increase in HIV-1-specific proliferative responses were seen (Fig. 2C). Of note was an outstanding response to Remune™ with an SI of 95 at week 65 for patient 11 in this group (data not shown) coincident with rebound viraemia (152,536 copies/ml) following self-imposed ART discontinuation. In group D a positive regression was seen in responses to Nef (p = 0.004) and whole Remune™ antigen (p = 0.005) from week 0 to 65. (Fig. 2D). The only response which significantly increased between week 17 and 65 in any group was the p24 response in group D (p = 0.039). Patient 8 in this group demonstrated a very large response to Remune™ (SI = 261) and np24 (SI = 144) at week 29 following breakthrough of resistant virus at week 25 (6,315 copies/ml) (data not shown). No other substantial increases in responses occurred between week 17 and 65 for any group or at any time point between the groups.

Induration size of hypersensitivity reactions to Remune™ did not become positive (>9 mm diameter) in any patient at any of the time points tested (weeks 17, 29 and 53).

Baseline to week 16 T-cell responses have been described in detail previously 1. Regression analysis revealed significant increases in recall responses in all groups between week 0 and 65 (Fig. 2E–H), predominantly for the persistent recall antigens: For group A (Figure 2E) responses to CMV and Candida significantly increased (p = 0.027 and p = 0.006 respectively); for group B (Figure 2F) HSV responses significantly increased (p = 0.007), CMV responses increased p = 0.002 and Candida responses increased (p = 0.0006); in group C (Figure 2G) CMV and Candida responses increased (p = 0.002 and p = 0.007 respectively); in group D (Figure 2H) HSV and CMV responses significantly increased (p = 0.002 and p = 0.001). Responses to transient antigens were less apparent. In group A PPD and tetanus responses showed significant positive regressions over the 65 weeks (p = 0.018 and p = 0.018 respectively); in group B only the PPD response regression curve was significant (p = 0.044); and in group C only the tetanus response was significant (p = 0.0003). There were no significant differences in recall responses between treatment groups.

Antigen-specific IL-4 production may be associated with suppression of proliferative responses and dampening of inflammatory immune responses. Therefore we measured IL-4 production in proliferation assay culture supernatants in response stimulation with each HIV-1 antigen, as previously described 1. Significant increases in HIV-1-specific IL-4 production from weeks 17 to 65 were observed in response to p24 in group-B (p = 0.023), and group C (p = 0.032) and in response to gp120 in group D (p = 0.037) (Fig. 3A). HIV-1-specific IL-4 production did not significantly increase in group A (data not shown).

We observed an apparent relationship between IL-4 production and loss/lack of proliferation in 17 of 28 patients tested, distributed evenly between groups, evident from week 0 to 16 and week 17 to 65. Patient 11 (group C) is shown as an example, where HIV-1-specific proliferative responses to p24, gp120, nef and Remune™ are high at week 29 (Fig. 3B) with absence of IL-4 (Fig. 3C), while at week 41 IL-4 secretion increases as proliferation declines.

Where possible we investigated CD8 T-cell IFN-γ production in response to characterised HLA class I restricted peptides (listed in Table 1), by ELISpot assay using cryo-preserved PBMCs. As reported previously 15 Remune™ did not induce increases in CD8 T-cell responses. Although fluctuations occurred in CD8 T-cell IFN-γ responses these were not sustained and could not be attributed to ART or immunotherapy. Summary data is presented from week 0, following ART alone at week 16 and at week 65 in Table 2. Assays in which responses were not observed to the postive control, PHA, were considered to have failed, as denoted by "F". Of note is the observation that no patient in groups B, C or D demonstrated an increase in responses between week 16 and week 65. Patient 11 is an exception, although this patient had high baseline responses and by week 65 had discontinued ART causing considerable rebound viraemia, concomitant with strong HIV-1-specific lymphocyte proliferative responses as detailed above.

Secondary to the main study protocol, further viral load and lymphocyte subset analysis was conducted at additional time points – week 18, week 22 and week 26 – for a sub-group of 15 patients receiving IL-2 in groups B (n = 9) and C (n = 6) on the last day of each IL-2 cycle. Small, transient elevations in HIV-1 RNA load were apparent on 12 occasions for 7 of 9 (78%) patients in group-B, and 2 of 6 patients (33%) in group-C. All these viral load elevations had spontaneously resolved by the next viral load test 3 weeks later at weeks 21, 25 and 29. The difference in occurrence of these viral load elevations remains evident when comparing the median viral loads for all the patients in groups B and C as shown in Fig. 1B and 1C. The median viral load change (from 50 copies/ml) for the patients who had day 5 viral load measurements in group B were 23.5 copies/ml, 53 copies/ml and 0 copies/ml at the end of each IL-2 cycle. In comparison the median change in viral load for patients in group C who had day 5 viral load measurements was 0 at each time point. Change in viral load was not significantly different when comparing the sub-study patients in group B with those in group C, at any of the 3 individual IL-2 day 5 time points. However when taking all three day 5 time points together for each group, comparison between the two groups revealed a trend towards significance (p = 0.071, Pearson Chi-Squared).

As Remune™ displays the transmembrane (gp41) envelope antigen with strong neutralization epitopes 333435, it may prime neutralizing antibodies to new epitopes. We tested the ability of plasma IgG from patients in Group C who did not experience day 5 viral blips to neutralize HIV-1_SF162 at weeks 0, 16 and 65. There was no effect of immunotherapy or ART on virus neutralization (data not shown).