Interleukin-2 (IL-2) has been investigated in HIV infection as an immunotherapy tool  for enhancing immune recovery or preserving immune function through its effects on CD4 T-cell production , proliferation  and survival . IL-2, added to antiretroviral therapy (ART), is able to drive a CD4 T-cell restoration quantitatively and qualitatively superior to the one obtained with ART alone [5–9], even in patients with low CD4 cell counts [10–13], although the IL-2 effect is larger with higher current and nadir CD4 cell counts . In untreated patients with mild immunodeficiency, IL-2 increases CD4 cell counts [15,16], sufficiently in one study  to allow significant postponing of ART, indicating that uncontrolled viral replication in itself does not prevent IL-2 to act on its cell targets. However, persistent immune activation may significantly blunt the magnitude of CD4 T-cell restoration , and IL-2 has not been studied in patients with both low CD4 cell counts and high plasma viral loads. The French National Agency for Research on AIDS and Viral Hepatitis (ANRS) 123 Essai d'un Traitement Optimisé et d'IL-2 dans l'Échec (ETOILE) trial evaluated the capacity of IL-2 to limit CD4 T-cell depletion in severely immunocompromised patients highly exposed to ART with few or no antiretroviral options.
Inclusion criteria were the following: HIV-1 infection, age of at least 18 years, CD4 cell count of less than 200 cells/μl at screening visit and at least once during the previous 6 months, plasma HIV-1 RNA of more than 10 000 copies/ml and a genotypic sensitivity score (GSS) showing two or less active drugs. GSS was based on resistance mutations present at screening and on all previously documented mutations. Drugs to which a patient was intolerant were counted out.
The ANRS 123 ‘ETOILE’ study was a prospective, randomized, open-label, multicentre (26 clinics in France) clinical trial. Patients were centrally randomized 1: 1 to receive either optimized background treatment (OBT) and IL-2 (IL-2 group) or OBT alone (control group) and stratified on GSS (0 vs. ≥1) and use of enfuvirtide.
Patients in the IL-2 arm were to receive eight cycles of subcutaneous IL-2 (4.5 MIU twice daily for 5 consecutive days every 6 weeks, from week 2 to 42). In both arms, local investigators were asked to optimize ART as much as possible from week 0, before randomization, following protocol recommendations: use active drugs according to GSS and enfuvirtide in enfuvirtide-naive patients. Optimization was defined by the use of enfuvirtide for at least 24 weeks from week 0 in enfuvirtide-naive patients, use of at least one active antiretroviral drug according to GSS or use of tipranavir/ritonavir (not included in GSS and only if last genotype indicated virus sensitivity). All other patients were considered as nonoptimized.
Follow-up and evaluation
Physical examination and blood tests took place at study visits, planned at weeks 2 and 6, then every 6 weeks to week 42 and at weeks 52, 64, 76 and 96 (extended follow-up visit). CD4 cell count and plasma HIV-1 RNA were measured locally. Genotypic resistance was assessed on plasma HIV RNA using the consensus technique of the ANRS AC11 resistance working group and defined according to the 2004–2005 algorithm (www.hivfrenchresistance.org). Final evaluation at week 52 was based on the proportion of patients with an absolute CD4 cell count of at least 200 cells/μl as primary outcome. Major secondary outcomes were the proportion of patients with a CD4 cell count increase at least 50 cells/μl from week 0 to week 52, median value of plasma HIV-1 RNA, occurrence of HIV-related events and tolerance.
The a priori hypothesis was a success rate (primary outcome) of 30% in the control group. With a type I error α equal to 0.05 and a power (1 − β) equal to 0.80, 49 patients per group were to be enrolled to conclude to a statistically significant intergroup difference of at least 30%. After a 6-month extension of the initial 12-month inclusion period, considering recruitment difficulties and arrival of new drugs in trials, the Scientific Committee decided on 28 October 2005, after a Data Safety Monitoring Board recommendation, to terminate the inclusions, complete patient follow-ups and perform the analysis.
All patients with at least one dose of study treatment were analysed. Proportions were compared using chi-squared or Fischer's exact tests. A missing equal to failure strategy was used for missing CD4 cell count measurements. Differences in plasma HIV-1 RNA between week 24 and week 52 and inclusion in each group were compared using a Wilcoxon test. Baseline data and optimization regimen were included in a univariate logistic regression, and all variables with a P value of 0.25 or less were included in the multivariate model. Global type I error was α equal to 0.05. SAS 9.1 software (SAS Institute, Cary, North Carolina, USA) was used. The study (registered at clinical trials.gov under NCT 00113282) was approved by the institutional ethics review boards.
Enrolment, baseline characteristics and follow-up
From June 2004 to December 2005, 69 patients were assessed for eligibility, and 57 were randomized (29 in the IL-2 group and 28 in the control group). The Scientific Committee decided to exclude from the analysis one patient who died at week 0 and never received IL-2. Baseline characteristics are shown in Table 1. Ten follow-up discontinuations were reported (seven in the IL-2 group and three in the control group) between week 0 and 52, and three during extended follow-up (IL-2 group), principally due to treatment inefficacy, patient decision and inclusion in other trials.
Optimization and interleukin-2 cycles
ART was deemed not optimized in 33 patients (17 in the IL-2 group and 16 in the control group), either because it was not modified or consisted of drugs that were all inactive according to GSS or because enfuvirtide was used in an enfuvirtide-experienced patient. ART was optimized with enfuvirtide only in six patients and with one or more active drug(s) other than enfuvirtide in seven patients (Table 1). Seventeen patients (61%) did not complete the eight IL-2 cycles, the main reasons for interruption (after one to seven IL-2 cycles) being failure (n = 7), patient's and investigator's decisions (n = 4) and serious adverse events (n = 4).
At week 52, the proportion of patients reaching a CD4 cell count of at least 200 cells/μl was 14% (4/28) in the IL-2 arm compared with 18% (5/28) in the control arm (P = 1.00). The proportion of patients with a CD4 cell count increase of at least 50 cells/μl at week 52 was 25% (7/28) in the IL-2 group and 32% (9/28) in the control group (P = 0.77).
The primary endpoint of reaching a CD4 cell count at least 200 cells/μl at week 52 was met in nine patients, including 50% (8/16) of those receiving an enfuvirtide-optimized regimen, 2.5% (1/7) of those with one other active drug and 0% of nonoptimized patients. Baseline CD4 cell count (≥100 vs. <100 cells/μl), time from HIV diagnosis and optimization with enfuvirtide with or without other active drugs were associated with this endpoint (all P ≤ 0.25). In the multivariate analysis, only optimization with enfuvirtide (8/16 vs. 1/40, adjusted P = 0.003) and baseline CD4 cell count (P = 0.01) were statistically associated with success. Figure 1 shows the evolution of CD4 cell counts according to randomization group and enfuvirtide optimization.
The secondary endpoint of a CD4 cell count increase of at least 50 cells/μl at week 52 was met by 16 patients, including 75% (12/16) of enfuvirtide-optimized patients, 2.5% (1/7) of those receiving a regimen optimized with one other active drug and 9% (3/33) of nonoptimized patients. Baseline CD4 cell count, optimization with enfuvirtide with or without other active drugs and time from diagnosis were associated with this endpoint in the univariate analysis, whereas only optimization with enfuvirtide remained significant in the multivariate model (12/16 vs. 4/40, P < 0.0001).
On available data, median plasma HIV-1 RNA in the IL-2 group was 4.7 log10 copies/ml [interquartile range (IQR) 3.7–5.3] at week 24 and 4.5 log10 copies/ml (IQR 3.5–5.2) at week 52, compared with 4.7 log10 copies/ml (IQR 4.1–5.2) and 4.6 log10 copies/ml (IQR 2.8–5.1) in the control group. Only 11% of patients (3/28) in each group reached plasma HIV-1 RNA of less than 50 copies/ml at week 52. Median differences in plasma HIV-1 RNA were −1.61 log10 copies/ml (IQR −2.87, −0.15) for enfuvirtide-optimized patients compared with 0.04 log10 copies/ml (IQR −0.28, 0.31) for nonenfuvirtide-optimized patients at week 24 (P = 0.003) and −1.53 log10 copies/ml (IQR −2.51, −0.23) vs. 0.10 log10 copies/ml (IQR −0.51, 0.46) at week 52 (P = 0.02).
Safety and tolerance
Twenty-three clinical serious adverse events occurred in 14 patients of the IL-2 group and 19 occurred in 12 patients of the control group. Diagnoses and causalities were reviewed by an event review board. Eight events were related to ART. Two cases of non-Hodgkin's lymphoma (NHL) were considered as possibly related to IL-2 but occurred in patients with CD4 cell counts of less than 10 cells/μl. One case of renal failure was considered as possibly related to both ART and IL-2. Five patients died, two with NHL in the IL-2 group and three with pneumonia (two with severe sepsis) in the control group. AIDS-defining events occurred in eight of 28 patients in each group. Cumulative probability over time of meeting an AIDS event or death was not different (log-rank P = 0.96) in the IL-2 and control groups. In IL-2 patients, most events (90%) were of grade 1 or 2 and were expected (fever, muscle aches, nausea–vomiting and fatigue). The proportion of patients having an enfuvirtide-related event (principally injection-site reactions) was similar in the two groups. Among five grade 4 biological events (two creatine phosphokinase elevations, two neutropenias and one lipase increase), none was considered as related to study drug.
The occurrence of two NHL cases (not unexpected in immunocompromised patients) in the IL-2 group led the Scientific Committee to recommend stopping IL-2 in nonresponding patients on 2 May 2006. In order to survey any new NHL case, follow-up was extended until week 96. One AIDS-defining event (cachexia) was reported in the control group during this period.
IL-2 failed to increase CD4 cell counts in the severely immunocompromised patients with advanced failure included in the ANRS 123 ‘ETOILE’ trial. Two main obstacles to IL-2 efficacy were met. First, baseline CD4 cell counts were very low, a factor known to decrease IL-2 efficacy, although not to annihilate it, as immunocompromised patients with controlled replication may experience immune recovery on IL-2 [10–13]. Second, viral replication was uncontrolled, a factor that does not in itself block IL-2 effects in mildly immunodepressed patients [15–17] but probably prevented any effect of IL-2 on these patients' deeply altered naive and central memory CD4 cell pools through maintaining an intense inflammation/activation state [18,19]. There was no protective effect of IL-2 on the occurrence of HIV-related events. Our results are consistent with those of the Study of IL-2 in people with low CD4+ T-cell counts on Active Anti-HIV Therapy (SILCAAT) trial  showing that IL-2 did not provide any benefit in advanced patients of whom, however, more than 80% had a plasma RNA viral load of less than 500 copies/ml at entry and the majority exhibited controlled viral load throughout the study.
On the contrary, IL-2 and the administration of IL-2 with enfuvirtide were relatively well tolerated in these fragile patients. Two cases of NHL occurred in the IL-2 group (none in the control group) and were considered as possibly related to IL-2 administration, but this complication is expected precisely in the severely immunodepressed population targeted by this trial , and recent results indicate that IL-2 does not increase the risk of NHL .
Multivariate analysis showed that baseline CD4 cell count and use of enfuvirtide were associated with immunological success defined by reaching a CD4 cell count of at least 200 cells/μl at week 52, and that the use of enfuvirtide was the only factor associated with a CD4 cell count gain of at least 50 cells/μl during the study. This emphasizes the importance of including antiretrovirals from new classes in salvage regimens [23,24]. Surprisingly, this was observed even in the setting of poorly (or not at all) optimized ART, a situation approaching de-facto enfuvirtide monotherapy. As a matter of fact, 43% of patients had a GSS equal to 0, ART could be optimized in only 41% and the addition of enfuvirtide was the sole therapeutic improvement in 11%. Optimization with enfuvirtide was associated with immunological success independently of the use of other active drug(s). In addition to a significant but modest reduction in viral load, this could suggest other effects of enfuvirtide, such as loss of fitness or cytopathogenicity in viruses with enfuvirtide resistance mutations, some mutation patterns being associated with a CD4 cell count decrease and others with a CD4 cell count increase in the presence of ongoing replication , or prevention of CD4 cell apoptosis [26,27].
In conclusion, IL-2 failed to improve immunological status in patients with advanced ART failure presenting with both low CD4 cell counts and high viral loads. In contrast, adding enfuvirtide was associated with immunological success, underlining that CD4 T-cell gain in advanced patients is mainly associated with the improvement, even if it is only moderate, of viral replication control. The ETOILE trial provides new data on the effect of IL-2 in patients not eligible in large phase III studies (patients with CD4 cell count <50 cells/μl), in whom the balance between expected benefit (CD4 cell count increase) and toxicity of IL-2 was a major concern. Our data point out that IL-2 therapy is not a treatment option for patients with limited ART possibilities.
J.-P.V., C.F., Y.L., L.W. and G.C. designed the study; C.F., J.-P.V. and G.C. supervised the overall conduct of the study; V.B., C.F. and G.C. conducted the statistical analysis; M.-L.C. and C.R. were responsible for the virologic coordination; J.-P.V., Y.L., L.W., N.C.dV., M.B. and G.P. enrolled the largest numbers of patients; J.-P.V., C.F., Y.L., L.W., V.B. and G.C. drafted the manuscript; all authors contributed to the critical revision of the manuscript.
The French National Agency for Research on AIDS and Viral Hepatitis (ANRS) sponsored the trial; Chiron laboratories provided interleukin-2. The authors wish to thank the patients, Sophie Tabuteau and Caroline Roy (Clinical Research Assistants), Christine Broissand (Pharmacy Coordinator), Marie-Josèphe Commoy (ANRS), Jade Ghosn, Jacques Gilquin and David Zucman (Event Review Committee), Jean-Pierre Aboulker, Corinne Amiel and Pierre-Marie Girard (Data Safety Monitoring Board) and investigators and clinical study technicians at study sites.
1. Lévy Y. Cytokine-based modulation of immune function in HIV infection. Curr Opin HIV AIDS 2006; 1:69–73.
2. Carcelain G, Saint-Mézard P, Altes HK, Tubiana R, Grenot P, Rabian C, et al
. IL-2 therapy and thymic production of naive CD4 T cells in HIV-infected patients wit severe CD4 lymphopenia. AIDS 2003; 17:841–850.
3. Natarajan V, Lempicki RA, Sereti I, Badralmaa Y, Adelsberger JW, Metcalf JA, et al
. Increased peripheral expansion of naive CD4+ T cells in vivo after IL-2 treatment of patients with HIV infection. Proc Natl Acad Sci U S A 2002; 99:10712–10717.
4. Kovacs J, Lempicki RA, Sidorov IA, Adelsberger JW, Sereti I, Sachau W, et al
. Induction of prolonged survival of CD4+ T lymphocytes by intermittent IL-2 therapy in HIV-infected patients. J Clin Invest 2005; 115:2139–2148.
5. Lévy Y, Capitant C, Houhou S, Carrière I, Viard JP, Goujard C, et al
. Comparison of subcutaneous and intravenous interleukin-2 in asymptomatic HIV-1 infection: a randomized trial. Lancet 1999; 353:1923–1929.
6. Gougeon ML, Rouzioux C, Liberman I, Burgard M, Taoufik Y, Viard JP, et al
. Immunological and virological effects of long-term IL-2 therapy in HIV-infected patients. AIDS 2001; 15:1729–1731.
7. Lévy Y, Durier C, Krzysiek R, Rabian C, Capitant C, Lascaux AS, et al
. Effects of interleukin-2 therapy combined with highly active antiretroviral therapy on immune restoration in HIV-1 infection: a randomised controlled trial. AIDS 2003; 17:343–352.
8. Durier C, Capitant C, Lascaux AS, Goujard C, Oksenhendler E, Poizot-Martin I, et al
. Long-terme effects of intermittent interleukin-2 therapy in chronic HIV-infected patients (ANRS 048+-079 trials). AIDS 2007; 21:1887–1897.
9. Mitsuyasu R, Gelman R, Cherng DW, Landay A, Fahay J, Reichman R, et al
. The virologic, immunologic, and clinical effects of interleukin 2 with potent antiretroviral therapy in patients with moderately advanced human immunodeficiency virus infection. Arch Intern Med 2007; 167:597–605.
10. Arnó A, Ruiz L, Juan M, Jou A, Balagué M, Zayat MK, et al
. Efficacy of low-dose subcutaneous interleukin-2 to treat advanced human immunodeficiency virus type 1 in persons with <250/microL CD4 T cells and undetectable plasma virus load. J Infect Dis 1999; 180:56–60.
11. David D, Naït-Ighil L, Dupont B, Maral J, Gachot B, Thèze J. Rapid effect of interleukin-2 therapy in human immunodeficiency virus-infected persons whose CD4 cell counts increase only slightly in response to combined antiretroviral treatment. J Infect Dis 2001; 183:730–735.
12. Marchetti G, Meroni L, Varchetta S, Terzieva V, Bandera A, Manganaro D, et al
. Low-dose prolonged intermittent interleukin-2 adjuvant therapy: results of a randomized trial among human immunodeficiency virus-positive patients with advanced immune impairment. J Infect Dis 2002; 186:606–616.
13. Katlama C, Carcelain G, Duvivier C, Chouquet C, Tubiana R, De Sa M, et al
. Interleukin-2 accelerates CD4 cell reconstitution in HIV-infected patients with severe immunosuppression despite highly active antiretroviral therapy: the ILSTIM study-ANRS 082. AIDS 2002; 16:2027–2034.
14. Markowitz N, Bebchuk JD, Abrams DI. Nadir CD4+ T cell count predicts response to subcutaneous recombinant interleukin-2. Clin Infect Dis 2003; 37:e115–e120.
15. Youle M, Emery S, Fisher M, Nelson M, Fosdick L, Janossy G, et al
. A randomised trial of subcutaneous intermittent interleukin-2 without antiretroviral therapy in HIV-infected patients: the UK-Vanguard study. PloS Clin Trials 2006; 1:e3, doi: 10.1371/journal.pctr.0010003.
16. Herzman C, Cuthbertson Z, Fosdick L, Fisher, Nelson M, Perry N, et al
. Long-term clinical and surrogate marker effects of subcutaneous intermittent interleukin-2 without antiretroviral therapy in HIV-infected patients
. J Antimicrob Chemother
17. Molina JM, Lévy Y, Fournier I, Boulet T, Bentata M, Beck-Wirth G, et al. Predictors of slow disease progression in ART-naïve HIV-1-infected patients treated with IL-2:3 year extended follow-up of the INTERSTART ANRS 119 trial [abstract #702]
. 15th Conference on Retroviruses and Opportunistic Infections
; 3–6 February 2008; Boston, Massachusetts; 2008.
18. Sereti I, Sklar P, Ramchandani S, Read SW, Aggarwal V, Imamichi H, et al
. CD4+ T cell responses to interleukin-2 administration in HIV-infected patients are directly related to the baseline level of immune activation. J Infect Dis 2007; 196:677–683.
19. ESPRIT Research Group. Predictors of CD4 count change over 8 months of follow up in HIV-1-infected patients with a CD4 count ≥300 cells/μl who were assigned to 7.5 MIU interleukin-2
. HIV Med
20. Levy Y, SILCAAT Scientific Committee. Effect of Inteleukin-2 on clinical outcomes in patients with C4+ cell count 50 to 299/mm3: primary results of the SILCAAT study [abstract #90bLB]
. 16th Conference on Retroviruses and Opportunistic Infections
; 8–11 February 2009; Montreal, Canada; 2009.
21. Grogg KL, Miller RF, Dogan A. HIV infection and lymphoma. J Clin Pathol 2007; 60:1365–1372.
22. Fontas E, Kousignian I, Pradier C, Duvivier C, Poizot-Martin I, Durier C, et al
. Interleukin-2 therapy does not increase the risk of Hodgkin or non-Hodgkin lymphoma in HIV-infected patient, results from FHDH ANRS CO4. J Acquir Immune Defic Syndr 2009; 50:206–214.
23. Lalezari JP, Henry K, O'Hearn M, Montaner JS, Piliero PJ, Trottier B, et al
. Enfuvirtide, an HIV-1 fusion inhibitor, for drug-resistant HIV infection in North and South America. N Engl J Med 2003; 348:2175–2185.
24. Lazzarin A, Clotet B, Cooper D, Reynes J, Arastéh K, Nelson M, et al
. Efficacy of Enfuvirtide in patients infected with drug-resistant HIV-1 in Europe and Australia. N Engl J Med 2003; 348:2186–2195.
25. Svicher V, Aquaro S, D'Arrigo R, Artese A, Dimonte S, Alcaro S, et al
. Specific enfuvirtide-associated mutational pathways in HIV-1 gp 41 are significantly correlated with an increase in CD4+ cell count, despite virological failure. J Infect Dis 2008; 197:1408–1418.
26. Barretina J, Blanco J, Armand-Ugón M, Gutiérrez A, Clotet B, Esté JA. Anti-HIV-1 activity of enfuvirtide (T20) by inhibition of bystander cell death. Antivir Ther 2003; 8:156–161.
27. Barretina J, Blanco J, Bonjoch A, Llano A, Clotet B, Esté JA. Immunological and virological study of enfuvirtide-treated HIV-positive patients. AIDS 2004; 18:1673–1682.