Lymphoma and Epstein−Barr virus DNA in blood during interleukin-2 therapy in antiretroviral-naïve HIV-1-infected patients: a substudy of the ANRS 119 trial

Authors


Abstract

Objectives

Interleukin-2 (IL-2) therapy increased CD4 cell counts and delayed antiretroviral therapy (ART) initiation in HIV-infected patients in the Agence Nationale de Recherche sur le SIDA et les Hépatites Virales (ANRS) 119 trial. However, four cases of lymphoma were reported. Epstein−Barr virus (EBV) replication is associated with an increased risk of lymphoma in immunocompromised patients. We assessed whether IL-2 had an impact on EBV replication and the development of lymphoma.

Methods

A total of 130 ART-naïve patients were randomized to receive IL-2 therapy (n = 66) or no treatment (n = 64). Clinical data for patients with lymphomas were reviewed and tumours assessed for evidence of EBV infection and CD25 (the IL-2 receptor) expression. EBV DNA levels were measured in whole blood and plasma in both arms using real-time polymerase chain reaction (PCR), up to 48 weeks after baseline (BL).

Results

Four lymphomas occurred, a median of 61 weeks [range 40−94 weeks] after randomization at a median CD4 cell count of 396 cells/μL (IQR 234–536 cells/μL). In the IL-2 arm, two patients developed EBV-positive Hodgkin's lymphoma, and one developed EBV-negative Burkitt-type lymphoma. One patient in the control group developed EBV-positive non-Hodgkin's lymphoma. CD25 was negative in all cases. Among the 41 of 55 (control arm) and 44 of 58 (IL-2 arm) patients with detectable EBV DNA in whole blood at both BL and week 48, the median change in EBV DNA between BL and week 48 was +0.04 log10 copies/ml in both arms (P = 0.7). In plasma, EBV was detected at least once in 22 of 52 controls and 21 of 54 IL-2-treated patients (P = 0.8).

Conclusions

IL-2 therapy had no significant effect on EBV replication over 48 weeks in these ART-naïve patients. The occurrence of lymphomas did not seem to be associated with IL-2 therapy.

Introduction

Interleukin-2 (IL-2) is a cytokine produced in vivo by T lymphocytes which regulates the proliferation, differentiation and survival of lymphocytes, mainly CD4 T cells. The use of IL-2 in combination with antiretroviral therapy (ART) in patients with HIV infection has been demonstrated to increase CD4 T-cell counts in a number of randomized clinical trials [1-3]. However, no clinical benefit of this rise in CD4 cell count was found in two recent randomized trials, suggesting that these IL-2-induced CD4 cells have a low immunoprotective ability [1]. The main effect of IL-2 therapy is the expansion of regulatory T cells (Tregs), a subset critically important for the maintenance of self-tolerance. The increase of the peripheral Treg pool in IL-2-treated HIV-infected patients may account for the clinical inefficiency of these expanded CD4 cells [4].

Concerns have also been raised about the possibility that IL-2 therapy may induce lymphoma. Indeed, IL-2 is also involved in B-cell differentiation into plasmocytes and can directly stimulate the proliferation of B cells from HIV-infected patients [5]. Additionally, IL-2 has been reported to induce replication of human herpesvirus 8 [6], which, like the Epstein−Barr virus (EBV), is associated with HIV-related malignancies. B-cell non-Hodgkin's lymphoma (NHL) and Hodgkin's lymphoma (HL) are the most common malignancies in HIV-infected patients and are in most cases associated with EBV [7]. Several IL-2-induced cytokines (e.g. IL-6 and tumour necrosis factor) may stimulate the proliferation of B cells, thus leading to the expansion of EBV-positive B cells [8, 9].

Intermittent IL-2 therapy was shown to increase CD4 cell counts and delay ART initiation in 130 naïve HIV-infected patients in the Agence Nationale de Recherche sur le SIDA et les Hépatites Virales (ANRS) 119 trial [10]. However, four cases of lymphoma occurred during follow-up, which translated into a particularly high incidence rate compared with other studies assessing IL-2 treatment in patients with HIV infection [1, 11, 12]. The purpose of this work is to report these cases of lymphoma. As EBV replication is associated with an increased risk of lymphoma in immunocompromised patients, we assessed whether IL-2 had an impact on EBV DNA loads in patients' peripheral blood and plasma.

Patients and methods

ANRS 119 was a multicentre, randomized, 96-week open-label trial which included 130 HIV-1-infected adult patients. The trial design has been reported elsewhere [10]. Briefly, 64 patients were assigned to no treatment (control) and 66 to IL-2 therapy. Eligible patients were ART-naïve and were not currently receiving therapy. CD4 cell counts were 300–500 cells/μL. The baseline characteristics of the patients, including baseline CD4 cell count, baseline HIV RNA viral load, Centers for Disease Control and Prevention (CDC) stage, and hepatitis B virus (HBV) and hepatitis C virus (HCV) coinfections, were well balanced between the study arms [10]. In the IL-2 arm, patients underwent three cycles of treatment with 4.5 MIU recombinant IL-2 (Aldesleukine; Chiron, Emeryville, CA) subcutaneously twice-daily for 5 consecutive days at weeks 0, 8 and 16 and could receive up to three additional cycles depending on CD4 counts. Patients were assessed at baseline (BL) and every 8 weeks until week 96. ART was initiated if CD4 counts dropped to < 300 cells/μL or in the case of an AIDS-defining event. During follow-up, four patients developed lymphoma (three in the IL-2 arm and one in the control arm). The data security management board recommended that IL-2 therapy be discontinued and that follow-up be extended for 1 year after week 96. Among the 130 study participants, 121 completed the 96-week follow-up and 100 (77%) completed the extended follow-up: 46 in the control group and 54 in the IL-2 group. No subsequent case of lymphoma was reported.

Lymphoma histology and immunochemistry

Biopsy specimens from the four patients who developed lymphoma were examined using haematoxylin and eosin-stained sections by the same expert pathologist (JB). Appropriate immunophenotypic markers were used, including CD15 (Becton-Dickinson Immunocytometry Systems, San Jose, CA), CD30 (clone BerH2), BcL2 (clone 124) and Ki-67 (clone MIB-1; all from DAKO Cytomaton, Carpinteria, CA). CD25 (clone 4C9), the IL-2-alpha chain receptor, was assessed in all specimens. Intratumoral EBV expression was assessed by mRNA in situ hybridization for EBV-encoded small nonpolyadenylated RNA 1 (EBER1) sequences and indirect immunofluorescence staining for latent membrane protein 1 (LMP1) and latency-associated nuclear antigen (LANA) antibodies (clone NL3; Advanced Biotechnologies, Columbia, MD).

EBV DNA measurement

The number of EBV genomes was determined by real-time polymerase chain reaction (PCR) using primers and labelled probes designed for the fifth leftward open reading frame (BALF5) of the EBV genome, as previously described [13]. Absolute quantification was performed by using serial dilution of a plasmid containing one target copy. The lower limit of detection was 100 genome-equivalents (gEq) of the target sequence/mL. Plasma EBV load was measured at BL and weeks 8, 16, 32 and 48 and whole blood was measured at BL and week 48.

Statistical analysis

The incidence of lymphoma was calculated as the number of cases per 100 patient-years of follow-up. In both plasma and whole blood, we compared the numbers of patients with detectable levels of EBV DNA in the control and IL-2 groups; for patients with detectable EBV DNA levels, we compared the EBV loads between the control and IL-2 groups using the nonparametric Wilcoxon test. A P-value < 0.05 was considered statistically significant.

Results

Incidence of lymphoma

In the IL-2 arm, two patients developed HL and one developed Burkitt-type lymphoma (BTL). In the control group, one patient developed non-Hodgkin's lymphoma (NHL). The total duration of follow-up in the trial was 341.73 patient-years. The overall incidence of lymphoma was 1.170 cases/100 patient-years. The incidence of lymphoma was 1.68 cases/100 patient-years and 0.61 cases/100 patient-years in the IL-2 and the control arms, respectively (p = 0.63). Lymphomas occurred a median of 61 weeks [range 40–94 weeks] after randomization and 40 weeks (range 20–74 weeks) after the last IL-2 injection for the three patients in the IL-2 arm. The median CD4 cell count at BL for the four patients in whom lymphoma occurred was 355 cells/μL (IQR 334–375 cells/μL), similar to that for the rest of the population (p = 0.56). The main features of the four cases of lymphoma are shown in Figure 1 and presented in chronological order of the occurrence of lymphoma.

Figure 1.

Serial HIV plasma viral loads, CD4 cell counts, IL-2 treatment cycles, initiation of highly active antiretroviral therapy (H), onset of lymphoma and histological findings in the four patients who developed lymphoma. IL-2, interleukin-2.

Patients with lymphoma

Patient 1 was a 37-year-old male ART-naïve patient diagnosed HIV-1 positive 9 years before enrolment. He was randomized to the IL-2 arm and received three cycles of IL-2. CD4 cell count and HIV viral load at BL were 348 cells/μL (CD4 percentage 13%) and 3.96 log10 HIV-1 RNA copies/mL, respectively. At week 40 the patient presented with fever, sweats, asthenia and weight loss. The CD4 cell count was then 288 cells/μL (CD4 percentage 15%) and the HIV viral load was 4.3 log10 copies/mL. Type III HL was diagnosed in an axillary lymph node (Reed-Sternberg cells CD30+, CD15+, CD20+, CD3−). Immunofluorescence staining for the EBV protein LMP1 was positive. IL-2 receptor-specific antibody staining (CD25) was negative. ART was initiated; the patient received eight cycles of dose-modified antitumoral chemotherapy. He achieved complete remission, and had not relapsed at the end of the 3-year follow-up period.

Patient 2 was a 37-year-old male patient diagnosed with HIV-1 infection 1 year before enrolment. He was ART-naïve and randomized to the IL-2 group. He received four cycles of IL-2. The BL CD4 cell count and HIV viral load were 362 cells/μL (CD4 percentage 23%) and 4.1 log10 copies/mL, respectively. At week 42, ART was started because he reached the immunological endpoint (CD4 cell count 265 cells/μL); the HIV viral load was 5.22 log10 copies/mL. At week 45, a unilateral cervical lymph node was found on physical examination. The CD4 cell count was 234 cells/μL and the HIV viral load was 2.6 log10 copies/mL. A diagnosis of nodular sclerosis stage IIb HL was made (Reed-Sternberg cells CD30+, CD15+, CD20+, CD3−). Immunofluorescence staining for LMP1 was positive and CD25 staining was negative. The patient received six cycles of antitumour chemotherapy and cervico-axillary radiotherapy. He was in complete remission at the end of follow-up.

Patient 3 was a 36-year-old ART-naïve female patient diagnosed with HIV infection 1 year before enrolment. She was randomized to the control group. At BL, the CD4 cell count was 375 cells/μL (CD4 percentage 28%) and the HIV viral load was 5.4 log10 copies/mL. At week 48, she reached the immunological endpoint (CD4 count 227 cells/μL); the HIV viral load was 6 log10 copies/mL. ART was started. Thirty weeks later, she presented with abdominal pain, vomiting and dyspnoea. The CD4 cell count was 505 cells/μL and the HIV viral load was < 20 copies/mL. A multinodular liver was found on computed tomography (CT) scan. A diagnosis of diffuse large-cell lymphoma with plasmacytic differentiation (CD138+, CD20−) was made on a liver biopsy. CD25 staining was negative, as was staining for LMP1. However, in situ hybridization revealed the presence of EBER-positive cells, confirming that tumour cells were infected with EBV. Chemotherapy was started but the patient died 2 weeks later of septic complications.

Patient 4 was a 49-year-old ART-naïve male patient. HIV-1 infection had been diagnosed 12 years before enrolment. He was randomized to the IL-2 group; the CD4 cell count was then 334 cells/μL (CD4 percentage 12%) and the HIV viral load was 5.83 log10 copies/mL. He received four cycles of IL-2. At week 48, he reached the immunological endpoint (CD4 count 226 cells/μL; HIV viral load 594 000 copies/mL) and ART was initiated. At week 94 he presented with abdominal pain and fever. An abdominal CT scan showed multiple large coelio-mesenteric nodes. BTL was diagnosed. Immunostaining showed CD20+, BcL2− and Ki67+, and EBV staining was negative (EBER1–, LMP1−). CD25 was negative. The CD4 cell count was 534 cells/μL and the HIV viral load was < 20 copies/mL. Chemotherapy was started resulting in complete remission, and he was well at the end of follow-up.

Changes in EBV DNA loads in whole blood and plasma during the study

Results from EBV measurements in whole blood and plasma are presented in Figure 2.

Figure 2.

Median Epstein−Barr virus (EBV) DNA viral loads in whole blood at baseline and week 48 in patients with detectable EBV DNA viral loads at both times. EBV DNA viral loads for the four patients with lymphoma are shown. IL-2, interleukin-2.

Whole blood

The EBV DNA load was measured in whole blood at BL and week 48 for 55 (86%) patients in the control arm and 58 (88%) in the IL-2 arm. Of these, 41 (74%) and 44 (76%), respectively, had detectable EBV DNA at both BL and week 48 (P = 0.9). For patients with detectable EBV DNA, median EBV loads were not significantly different between the control and IL-2 arms at BL (3.91 and 3.89 log10 copies/ml, respectively; P = 0.62). At week 48, median EBV DNA loads were 3.97 and 3.98 log10 copies/mL in the control and IL-2 arms, respectively (P = 0.7). The median change in EBV DNA in whole blood between BL and week 48 was +0.04 (range −0.31 to +0.22) and +0.04 (range −0.32 to +0.20) log10 copies/ml in the control and IL-2 groups, respectively (Fig. 2). Of the four patients with lymphoma, all had positive EBV DNA detected in whole blood at BL and week 48, with no trend between BL and week 48.

Plasma

EBV DNA load was measured in plasma every 8 weeks from BL to week 48 in 52 (81%) and 54 (82%) patients in the control and IL-2 arms, respectively. EBV was detected at least once during follow-up in 22 of 52 (42%) controls and 21 of 54 (39%) patients in the IL-2 group (P = 0.35). EBV was detected in plasma in every sample in only one and two patients, respectively (P = 0.6). A low proportion of patients (four in each group) with undetectable EBV DNA in plasma at BL had a positive detection at week 48 (P = 0.8). In the two patients with EBV-related HL (patients 1 and 2), a significant increase in EBV plasmatic DNA was seen after week 20, when IL-2 therapy had been discontinued. Patient 3, who had EBV-positive NHL, had undetectable or low EBV plasmatic DNA (< 2 log10 copies /mL) which did not vary significantly over time.

Discussion

The high incidence of lymphoma found in this trial compared with other cohorts prompted us to report the clinical and histological characteristics of the four cases of lymphoma and to assess EBV DNA load in all the available samples of the trial population to determine whether EBV DNA load differed between those receiving and not receiving IL-2 therapy [1, 11, 12]. Our findings suggest that the possible involvement of IL-2 treatment in the occurrence of lymphoma is unlikely be through increasing EBV replication.

IL-2 has now been tested in several large cohorts of HIV-infected patients. Results from these trials have shown no differences in the risk of lymphoma between IL-2-treated patients and controls. The large Subcutaneous recombinant, Interleukin-2 in HIV-Infected Patients with Low CD4+ Counts under Active Antiretroviral Therapy (SILCAAT) trial and the Evaluation of Subcutaneous Proleukin in a Randomized International Trial (ESPRIT), which included approximately 6000 ART-treated HIV-infected patients receiving or not receiving IL-2, found similar rates of HL and NHL in the IL-2 and control groups [1]. The incidence of lymphoma in these trials was relatively low (between 0.3 and 0.21 cases/100 patient-years) compared with the incidence found in the ANRS 119 trial (1.170 cases/100 patient-years), probably because in the latter trial patients were ART-naïve. Although two patients were receiving ART and had controlled HIV replication at the onset of lymphoma, they had been diagnosed with HIV infection 2 and 13 years before initiating ART, respectively. We believe that prolonged viraemia and decreasing CD4 cell counts associated with untreated HIV infection may explain the high incidence rate of lymphoma found in this trial, although this may not be the sole explanation. Cumulative viraemia and low CD4 cell counts have indeed been found to be independently associated with the risk of lymphoma, and this association was stronger for BTL [14, 15]. Additionally, no tumour in the trial was positive for the IL-2 receptor, CD25, which may also be reassuring regarding the potential impact of IL-2 through direct contact with, or signalling to, lymphoma cells.

These results are in agreement with results from the French Hospital Database on HIV (FHDH) showing no increased risk of lymphoma in HIV-infected patients treated with IL-2 [12]. Indeed, in the FHDH study, IL-2 treatment was even associated with a nonsignificant decrease in the risk of both lymphomas (−36% for NHL and −67% for HL), after adjustment for sex, age, period, CD4 cell count, plasma HIV RNA and AIDS status [12]. Also, among the 267 untreated HIV-infected patients enrolled in the study of aldesleukin with and without antiretroviral therapy (STALWART) trial who received IL-2 treatment or no treatment, only one case of NHL occurred in a patient receiving IL-2 with pericycle ART. The incidence of lymphoma was not significantly different in the three arms, although IL-2 administration led to a delay in the initiation of ART, which was associated with a trend towards a greater incidence of opportunistic diseases or death and a statistically significant increase in adverse events [11]. Over the extended follow-up period, two more cases of NHL occurred in the same group; the event rates in the IL-2 groups remained higher than in controls, but differences were not statistically significant [16]. The limited number of lymphomas in this ART-naïve HIV-infected population receiving IL-2 is reassuring.

Furthermore, we found that changes in EBV DNA loads in both plasma and whole blood were not significantly different between patients treated with IL-2 and those not receiving IL-2 treatment. Indeed, after 48 weeks, no difference was observed in EBV DNA load between the control and IL-2 arms. In particular, no increase in EBV DNA load was detected in the IL-2 arm over time. Changes in EBV replication would be especially relevant, given the well-documented increased risk of EBV-associated lymphoma in HIV-infected patients. The patients in this trial were already at high risk for lymphoma because of their underlying HIV infection, prolonged time without therapy, and relatively low CD4 cell counts. In contrast to our results, a randomized study, including patients receiving 4 different IL2 treatment regimen, suggested that higher doses of IL-2 therapy may increase circulating EBV load, while low-dose IL2 therapy tended to decrease EBV load, suggesting that depending on the dose, IL2 therapy may affect the dynamics of EBV replication [17].

Additionally, concerning the four patients who developed lymphoma, no clear pattern of tumour EBV positivity or EBV plasmatic or whole-blood DNA was identified. Only the two patients with HL had EBV-positive Reed-Sternberg cells and an increase in EBV DNA in plasma. It has been well documented that HIV-related HL cells are typically EBV positive [18, 19]. Circulating EBV DNA can be detected in about 20% of asymptomatic HIV-infected patients [20]. The level of EBV DNA, however, is not predictive of progression to NHL. Moreover, EBV DNA can be detected in the vast majority of patients with EBV-associated HL, as in these two cases [21]. However, because of the limited number of cases in this study, it is not possible to completely rule out the possibility of lymphoma occurring in the IL-2 group by other mechanisms than an increase in EBV replication.

Taken together, these results provide reassurance about the safety of IL-2 regarding EBV reactivation.

The use of IL-2 in HIV-infected patients does not seem to be clinically beneficial, despite an interesting rise in CD4 levels [1], which has led to the discontinuation of the use of this drug in HIV-infected subjects. However, low-dose IL-2 is being evaluated for its immunosuppressive effects, as a consequence of the increase in peripheral Tregs, in autoimmune diseases such as cryoglobulinaemia complicating HCV infection and graft vs. host disease in bone-marrow recipients, in both of which cases there is a high risk of lymphoma [22, 23]. Moreover, other cytokines, in particular IL-7, are being investigated for their ability to increase T-cell recovery and expansion in HIV-infected patients with persistently low CD4 counts despite ART [24, 25].

In conclusion, no effect of IL-2 on EBV DNA load was found in the ART-naïve HIV-infected patients in the ANRS 119 trial. The occurrence of lymphoma does not seem to be related to IL-2 therapy, but rather to the deferred use of ART with prolonged viral replication and decreasing CD4 cell counts. These findings do not provide support for an increased risk of lymphoma associated with IL-2 treatment.

Ethical approval

The initial protocol for the ANRS 119 trial was approved by the ethics committee of Saint Louis Hospital, APHP, Paris. All patients provided written informed consent.

Acknowledgements

Conflicts of interest: All authors declare that they have no conflict of interest concerning this work.

Financial disclosure: Funding for the ANRS 119 trial was provided by the French Agence Nationale de Recherche sur le SIDA et les Hépatites Virales (ANRS).

Ancillary