Kaplan-Meier survival analysis was performed for each study outcome and for any combination of them, comparing placebo recipients with vaccine recipients on an intention-to-treat basis. Crude relative risks of vaccine recipients were calculated, taking placebo recipients as the reference population. To adjust estimates for potential prognostic variables (CD4+ cell counts and HIV-1 plasma viremia at enrollment), Cox's model was used. Using the standard error of the regression coefficients, 95% confidence intervals (CI) of relative risks (RR) were calculated. A preintended analysis was planned in the subgroup of patients immunized to IFN-α, that is, those who responded to the vaccine by raising their anti-IFN-α antibodies (antibody responders). This group was compared with vaccinees who did not develop antibodies or placebo recipients. None of the analyses combined these groups. Fisher's exact test was used to evaluate associations between parameters and clinical outcome.
Safety, Tolerability, and Compliance
From December 1995 to July 1996, 242 asymptomatic HIV-1-infected patients were enrolled by eight clinical centers in three countries (Belgium, Israel, and Italy). The characteristics of enrolled patients, according to randomly assigned treatment, are shown in Table 2. Randomization succeeded in balancing the prognostic variables both in placebo and vaccine recipients. The two groups were similar for age, gender, risk factor, number of years from the first HIV-1-seropositive test, number of HCV-coinfected patients, HIV-1 viremia, and CD4+ and CD8+ cell counts at enrollment and for the percentage of patients receiving an antiretroviral treatment either before or electively (63%) during the course of the trial. In all, 4 patients, 2 in the placebo group and 2 in the vaccine group, were lost to follow-up within 1 month from the first injection (9-28 days) and they were accordingly excluded from the analysis.
Vaccination was well tolerated. The most frequently reported side effect was local discomfort at the site of injection and fever. In addition, 68 patients reported local pain (25%), and 39 of those had fever >38°C lasting not longer than 2 days and the therapy for which was then reduced to common analgesics. No patients, whether placebo or vaccine recipients, had to discontinue the immunization program as a result of related complications. The most adverse reaction was the formation of a cold abscess at the site of injection; it was observed in 2 patients in response to the first priming injection of water-in-oil emulsion. Subsequently, these 2 patients were only given injections of i-IFN-α adsorbed onto Ca(OH)PO4 in water; these patients have been included in the analysis. No other complications were found following >700 oil emulsion injections as well as Ca(OH)PO4 water injections performed in this trial. Biochemical and hematologic tests did not reveal any significant variation from preimmunization values, except in 2 patients who showed a marked increase of transaminase levels (>5 times the normal upper limit). These two study subjects were chronically HCV-infected patients and treatment was discontinued, the code was broken, and they were found to be receiving placebos. None of the patients showed an increase of HIV-1 viral load after priming or boosting injections (data not shown).
None of the enrolled patients died during the 18-month follow-up period. Clinical progression of HIV-1 infection from asymptomatic status (CDC stage A) to symptomatic status (CDC stage B or C) was observed in 37 of the 242 patients (15.3%); HIV-1-related signs and symptoms included multidermatomeric herpes zoster (20 cases), oral or systemic candidiasis (7 cases), oral hairy leukoplakia (4 cases), Kaposi's sarcoma (2 cases), Mycobacterium tuberculosis infection (1 case), cytomegalovirus-related enteritis (1 case), aggressive intraepithelial uterine carcinoma (1 case), chronic ulcerative anal herpes simplex infection (1 case), blastocystosis (1 case), atypical mycobacteriosis (1 case), and giant disseminated molluscum contagiosum (1 case).
IFN-α Antibody Response to the Vaccine
Forty patients responded to vaccination with increased anti-IFN-α antibody titers. These antibody responder vaccinees (AbRV) represented 33% of vaccine recipients (Table 3). In most of these study subjects (>90%), the circulating IFN-α antibodies increased more than twofold to 10-fold, following the third injection. These antibody responders were all vaccine recipients and none were placebo recipients, which confirmed the immunogenicity of the vaccine preparation. In the antibody responders, postimmunization serum titers were in the range of 1:2,000 to 1:32,000 whereas these titers were <1:500 in tested antibody nonresponder vaccinees (AbNRV) as well as in placebo (Pla; data not shown).
Some 85% of AbRV (i.e., 30 of 35 patients assayed) exhibited an increase of INC of their sera, as detected by the standard IFN-α neutralization assay. In 75% of these study subjects (22 patients), the serum INC levels highly increased over 3 log2 dilutions compared with levels found at enrollment. By contrast, in AbNRV and Pla groups, serum INC remained unchanged in most patients but declined in a consistent number of study subjects.
At enrollment, circulating IFN-α levels were similar in AbRV, AbNRVs, and Pla recipients. At the end of follow-up, mean values of serum IFN-α levels, even though below critical levels for disease progression (i.e., <30 IU/ml) in the three groups of patients, were significantly lower in AbRVs (8.6 IU/ml) compared with AbNRVs and Pla recipients (mean values, 14.1 and 10.9 IU/ml, respectively) by analysis of variance for repeated measures (F = 3.15; p < .05). Further individual analysis showed that at enrollment, a similar percentage of patients with significantly high levels of serum IFN-α (30 IU/ml) in AbRVs (15%) and nonimmunized patients (AbNRVs and Pla recipients, 20%), and at the end of follow-up, 15% of nonimmunized patients maintained or increased IFN-α levels >30 IU/ml, compared with <3% of antibody responder vaccinees. The difference was statistically significant (p < .05).
Characteristics of AbRVs are reported in Table 4: age, gender, risk factors, and HIV-1 plasma viremia were similar to those of AbNRVs and Pla recipients. Furthermore, circulating antibody levels of various specificities, including tetanus toxoid, HIV-1 gp160, p24, p17, and nef and tat peptides, were at enrollment and following immunization of comparable magnitude in both AbRVs and AbNRVs as ascertained by Student's t-test for paired samples. Finally, no difference in level and increase of anti-IFN-α antibodies was found between patients receiving or not antiretroviral therapy, including protease inhibitors, during the follow-up period (data not shown).
Kaplan-Meier survival analysis was carried out on occurrence of HIV-1-related signs, treatment changes, or initiation due to disease progression and decrease of CD4+ counts to <200 cells/mm3; the latter was analyzed in patients with CD4+ counts at baseline >200 cells/mm3. Differences between the two groups were observed, vaccine recipients having slightly better event-free survival than placebo recipients (Fig. 1; Table 1); however, these differences were not statistically significant. The surveillance committee determined that the rate of clinical and laboratory endpoints was reduced in the two groups, because of the introduction of protease inhibitors, to a level that the likelihood of achieving the study aims was not feasible, resulting in the trial follow-up being terminated. Overall, 154 of 242 enrolled patients were given protease inhibitors during the study period (Table 2), of whom 79 (51%), added drugs because of clinical or laboratory progression.
A Kaplan-Meier survival analysis, comparing patients who were AbRVs with AbNRVs and Pla recipients separately (Fig. 2; Table 4), showed a significantly lower rate of all study endpoints in antibody responder vaccinees. More numerous patients from group B with CD4+ counts >350 cells/mm3 (25 individuals) experienced a rise of anti-IFN-α antibodies than patients from group A with CD4+ counts ≤350 cells/mm3 (15 individuals; Table 5). This result, which confirms previous findings (22), was expected, inasmuch as CD4+ cell counts reflect the immune system's status, and thus confirmed previous findings. As a consequence, CD4+ cell counts of AbRVs were significantly higher than those of AbNRVs, but not those of the Pla group (p = .16) Furthermore, AbRVs at enrollment exhibited significantly lower viral load than AbNRVs (Table 5). Important to note, median CD4+ cell counts and plasma viremia of AbRVs within each stratum were absolutely similar to those of AbNRVs and Pla groups (Table 5). Since during HIV-1 infection CD4+ cell counts as well as plasma viremia are known to be biologic markers of disease progression (28), it was necessary to control the Kaplan-Meier analysis for CD4+ cell count and HIV-1 plasma viremia at baseline, to ascertain whether the rise of IFN-α antibody obtained through active immunization per se accounted for reduction of AIDS progression in AbRVs (Table 4). The statistical adjustment for these parameters at enrollment did not alter the direction, magnitude, or statistical significance of the crude estimates of risk of AbRVs compared with those of AbNRVs and Pla recipients (Table 4). In particular, the RR of occurrence of HIV-1-related clinical signs was reduced in AbRVs by 75% compared with AbNRVs or Pla recipients. Similarly, the RR of occurrence of any of the study outcomes was reduced by 50% in AbRVs. Moreover, the rate of occurrence of all the study endpoints was always lower in AbRVs compared with those of AbNRVs and Pla recipients, as shown by stratification for CD4+ cell counts and plasma viremia in Table 5. Interestingly, the Kaplan-Meier clinical survival analysis comparing the AbNRV subgroup to Pla group showed no significant differences between these two groups even controlled for CD4+ cell counts, which were somewhat lower in AbNRVs compared with findings in Pla recipients (median, 297 and 348, respectively; p = .08; Table 4).
Rise of IFN-α Antibodies Is Inversely Associated to AIDS-Related Clinical Signs Rate
Because previous reports showed that the level of circulating IFN-α was highly correlated to clinical manifestations (23) and to fast-progressor HIV-1-infected patients (29), we questioned whether the rise of IFN-α antibodies, enhancing the capacity to neutralize circulating IFN-α in AbRVs could significantly contribute to the reduction of AIDS occurrence observed in this group (Tables 4 and 5). To eliminate any bias due to differences of CD4+ cell counts and plasma viremia at enrollment, the adjustment for these parameters, as mentioned previously, did not alter the statistical differences of risk in AbRVs versus AbNRVs and Pla recipients (Table 4).
Furthermore, the results of the Fisher's exact test showed the following:
- As expected, in vaccinees, a significant correlation was found between high CD4+ cell count and low plasma viremia and anti-IFN-α response (p = .03 and p = .01, respectively). Indeed, patients of vaccine group B (CD4+ counts >350 cells/mm3) had an antibody response rate better than those of the vaccine group A (CD4+ counts < 350 cells/mm3).
- As anticipated, a positive correlation of lower occurrence of clinical manifestations with higher CD4+ cell counts (RR = 0.55) and lower plasma viremia (RR = 0.91) at enrollment was found (even though not statistically significant in these series of patients: p = .16 and p = .85, respectively).
- Importantly, the risk of occurrence of AIDS-related symptoms was significantly reduced in patients with rise of anti-IFN-α antibody (RR = 0.25, p = .05).
This international, multicenter, randomized, double-blind, placebo-controlled trial was designed to evaluate safety, immunogenicity, and efficacy of an active immunization against IFN-α. The purpose of anti-IFN-α immunization was to counteract IFN-α overproduction, which is characteristic of HIV-1 infection (5,6) and plays a central role in HIV-1-induced immunosuppression (8). The safety of the immunization against this cytokine, reported by previous studies (21-23), was confirmed by this study, during which only 2 cold abscesses occurred following >700 priming (water-in-oil) injections and none after booster (water) injections. In previous studies on anti-IFN-α, immunization of patients untreated with antiretroviral therapy, 6 to 7 monthly repeated injections of i-IFN-α elicited an anti-IFN-α antibody response rate of >80% of patients (21-23). In the current trial, in which only three oil injections were given for priming, only 33% of patients exhibited an anti-IFN-α antibody response, showing that the immunization scheme applied was not optimal. Even though CaOHPO4 water suspension for priming injections, as used in the ANRS trial, did not allow expression of the immunogenicity (22), previous trials showed that CaOHPO4 adjuvant was effective after priming for booster injection (23). Because in this trial only a low percentage of patients (33%) developed IFN-α antibodies after the three oil injections used for priming, we concluded that optimization of IFN-α immunization regimen was not achieved in this trial, and an optimal immunization schedule using six to seven oil priming injections should increase the percentage of AbRVs to >80% (21). Such an optimal immunizing regimen should enhance the study's power to show an effect.
As previously reported (22,23), in AbRVs, increased serum INC and lower serum IFN-α levels were also observed following immunization, by contrast to AbNRVs and Pla recipient groups.
The recent availability of protease inhibitors for HIV-1 disease treatment and their use by enrolled patients according to those patients' request and treatment guidelines was an important confounder of trial analysis, because they lowered the rate of endpoints that would have been achieved based on our power calculation. Although no significant difference in clinical endpoints could be detected in vaccine and placebo recipients, on subgroup analysis those who had a positive antibody response to active immunization against IFN-α (i.e., AbRVs) had a statistically significant lower risk of occurrence of clinical and/or biologic signs of progression (Fig. 2) compared with AbNRVs and Pla recipients (Table 4).
The lower risk of disease progression in AbRVs versus AbNRVs and Pla recipients could be ascribed either to the effect of the anti-IFN-α immunization, to a selective bias at enrollment, or to both. The AbRV group exhibited higher CD4+ cell count and lower plasma viremia at enrollment, as anticipated by our previous study showing lower antibody response when CD4+ counts were <350 cells/mm3 (22). Because CD4+ cell count and plasma viremia are known to be important predictors of the risk to disease progression and to evaluate whether the rise of IFNa antibodies was correlated to reduction of AIDS progression, it was necessary to control for these major markers of AIDS progression, that is, CD4+ cell count and plasma viremia in the Kaplan-Meier statistical analysis. Adjustment for CD4+ cell count and viremia did not alter the statistical significance of reduced risk of progression in AbRVs compared with that of AbNRVs and Pla recipients. This statement is further confirmed by the results of Fisher's exact test. This test showed that low levels of CD4+ cell counts at enrollment influence, but not significantly, AIDS progression, whereas the rise of IFN-α antibodies reduces significantly AIDS-related symptoms (p = .05). This result is consistent with previous clinical reports that demonstrated that a high circulating IFN-α level is strongly correlated to AIDS evolution (23) and fast progression (28). Furthermore, recent experimental data stressing that IFN-α (the well-known immunosuppressive cytokine  overproduced by infected patients [6,8]) mediates AIDS immunosuppression (30), and thus may provide an explanation of why the rise of IFNa antibodies is inversely correlated to AIDS progression. These data further showed that the cellular immune response of PBMCs inhibited in vitro by HIV-1 can be restored in presence of anti IFN-α antibodies. Therefore, the rise of specific antibodies neutralizing the overproduction of IFN-α, found at early stages in infected lymphoid foci and at later stages in sera of HIV-1-infected patients (6), should block the spontaneous evolution of HIV-1 replication toward immune collapse and AIDS.
Nevertheless, we recognize it remains possible that more subtle yet unidentified differences between the subgroups at enrollment may have contributed to the better behavior of AbRV patients. In any event, statistical analysis of the results of this trial supports our assumption that reduction of AIDS-related symptoms in AbRVs may be secondary to the rise of IFN-α antibody in this subpopulation of patients. It supports our decision to continue to evaluate the clinical efficacy of this approach in a larger efficacy trial.
In conclusion, although the power of this trial has been reduced by two unforeseen confounders, that is, introduction of triple therapy delaying/modifying the kinetics of endpoints and suboptimal immunizing regimen reducing the number of AbRVs, this double-blind, placebo-controlled clinical trial carried out on 242 patients divided into groups receiving or not receiving antiretroviral therapy supports the safety of the IFN-α vaccine and suggests those with a rise in anti-IFN-α antibodies, have improved prognosis.
The EURIS Study Group is represented by: F. Adorni (Coordinating Center, Milan, Italy), J-M. Andrieu (Paris, France), A.G. Angius (Milan, Italy), A. Barath (Brussels, Belgium), G.P. Cadeo (Brescia, Italy), M. Carcagno (Buenos Aires, Argentina), G.P. Carosi (Brescia, Italy), C. Chiaro (Coordinating Center, Milan, Italy), C.M. Farber (Brussels, Belgium), M. Feldman (Rehovot, Israel), K. Kabeya (Brussels, Belgium), A. Lachgar (Paris, France), H. Le Buanec (Paris, France), W. Lu (Paris, France), P.M. Mannucci (Coordinating Center, Milan, Italy), M. Morfini (Florence, Italy), M. Muça-Perja (Coordinating Center, Milan, Italy), R. Naaman (Rehovot, Israel), E. O Doherty (Brussels, Belgium), L. Palvarini (Brescia, Italy), A. Rocino (Naples, Italy), S. Sant (Coordinating Center, Milan, Italy), M. Schiavoni (Bari, Italy), S. Sprecher (Brussels, Belgium), M. Stein (Rehovot, Israel), A. Turano (Brescia, Italy), J.F. Zagury (Paris, France), R. Zerboni (Milan, Italy).
Acknowledgments: This work has been supported scientifically and financially by Neovacs (Paris, France). We acknowledge Biosidus (Buenos Aires, Argentina) and Seppic (Paris, France) for generously supplying the immunizing and adjuvant reagents, the patients for their great collaborative participation, and Robert C. Gallo for his continued advice, scientific discussions, and encouragement.
1. Zagury D, Bernard J, Leonard R, et al. Long-term cultures of HTLV-III-infected T cells: a model of cytopathology of T-cell depletion in AIDS. Science
2. Clerici M and Shearer GA. Th1→Th2 switch is a critical step in the etiology of HIV infection. Immunol Today
3. Fan J, Bass HZ, Fahey JL. Elevated IFN-gamma and decreased IL-2 gene expression are associated with HIV-1 infection. J Immunol
4. Clerici M, Lucey DR, Berzofsky JA, et al. Restoration of HIV-specific cell-mediated immune responses by Interleukin-12 in vitro. Science
5. Ambrus JL, Poiesz BJ, Littie MA, et al. Interferon and interferon inhibitor levels in patients infected with varicella-zoster virus, acquired immunodeficiency syndrome, acquired immunodeficiency syndrome-related complex, or Kaposi's sarcoma and in normal individuals. Am J Med
6. Francis ML, Meltzer MS, Gendelman HE. Interferons in the persistence, pathogenesis and treatment of HIV infection. AIDS Res Hum Retroviruses
7. Taylor-Papadimitriu J. Effects of interferons on cell growth and function. Interferon
8. Orenstein JM. Ultrastructural markers in AIDS. Lancet
9. Fall L-S, Chams V, Le Coq H, et al. Evidence for antiviral effect and interferon neutralizing capacity in human sera; variability and implications for HIV infection. Cell Mol Biol
10. Tovey MG, Lebon P, Meyer F, et al. The role of interferon in the development of AIDS. The 1992 ISIR meeting in the IFN system [abstract]. J Interferon Res
11. Ferbas JJ, Toso JF, Logar AJ, Navratil JS, Rinaldo CR Jr. CD4-blood dendritic cells are potent producers of IFN-alpha in response to in vitro HIV-1 infection. J Immunol
12. Capobianchi MR, Ameglio F, Fei PC, et al. Coordinate induction of interferon α and γ by recombinant HIV-1 glycoprotein 120. AIDS Res Hum Retovir
13. Gulick RM, Mellors JW, Havlir D, et al. Treatment with indinavir, zivodine and lamivudine in adults with human immunodeficiency virus infection and prior antiretroviral therapy. N Engl J Med
14. Finzi D, Hermankova M, Pierson T, et al. Identification of a reservoir for HIV-1 in patients on highly active antiretroviral therapy. Science
15. Chun TW, Engel D, Berrey MM, Shea T, Corey L, Fauci AS. Early establishment of a pool of latently infected, resting CD4(+
) T cells during primary HIV-1 infection. Proc Natl Acad Sci USA
16. Volberding P, Moody DJ, Beardslee D, Bradley EC, Wofsy CB. Therapy of acquired immune deficiency syndrome with recombinant interleukin-2. AIDS Res Hum Retroviruses
17. Kovacs JA, Baseler M, Dewar RJ, et al. Increase in CD4 T lymphocytes with intermittent courses of interleukin-2 in patients with human immunodeficiency virus infection. A preliminary study. N Engl J Med
18. Emery S. Lane HC. Immune based therapies in HIV infection: recent developments. AIDS
19. Zagury D, Bernard J, Halbreich A, et al. One-year follow-up of vaccine therapy in HIV-infected immunodeficient individuals: a new strategy. J Acquir Immune Defic Syndr
20. Carelli C, Halbreich A, Bernard J, et al. Immunogenicity of combined anti-HIV and anti-immunosuppressive vaccine preparations. Biomed Pharmacother
21. Gringeri A, Santagostino E, Mannucci PM, et al. A randomized, placebo-controlled, blind anti-AIDS clinical trial
: safety and immunogenicity of a specific anti-IFN alpha immunization. J Acquir Immune Defic Syndr
22. Gringeri A, Santagostino E, Mannucci PM, et al. Anti-alpha interferon immunization: safety and immunogenicity in asymptomatic HIV positive patients at high risk of disease progression. Cell Mol Biol
23. Gringeri A, Santagostino E, Cusini M, et al. Absence of clinical, virological and immunological signs of progression in HIV-1 infected patients receiving active anti-alpha interferon immunization: a 30-month follow-up. J Acquir Immune Defic Syndr Hum Retrovirol
24. Lachgar A, Bizzini B. Involvement of α-interferon in HIV-1 induced immunosuppression. A potential target of AIDS prophylaxis and treatment. Biomed Pharmacother
25. Bizzini B, Achour A. Kinoids: the basis for anticytokine immunization and their use in HIV infection. Cell Mol Biol
26. Coulaud IP, Gougeon ML, Gomard E, et al. A placebo-controlled clinical phase I trial with combined anti-HIV and anti-interferon alpha immunization. AIDS
27. Lu W, Han DS, Yuan J, Andrieu JM. Multi-target PCR analysis by capillary electrophoresis and laser-induced fluorescence. Nature
28. Mellors JW, Munoz AM, Giorgi JV, et al. Plasma viral load and CD4- lymphocytes as prognostic markers of HIV-1 infection. Ann Intern Med
29. Petricoin EF III, Ito S, Williams BL, et al. Antiproliferative action of IFNa requires components of T-cell-receptor signalling. Nature
30. Zagury D, Lachgar A, Chams V, et al. Interferon a and Tat involvement in the immunosuppression of uninfected T cells and C-C chemokine decline in AIDS. Proc Natl Acad Sci USA
Keywords:© 1999 Lippincott Williams & Wilkins, Inc.
AIDS clinical trial; Anti-interferon-α immunization; Anti-cytokine vaccination