Serum Levels of Interleukins 6, 10, and 13 Before and After Treatment of Classic Hodgkin Lymphoma

Multiagent chemotherapy is the standard treatment for Hodgkin lymphoma (HL), and it usually provides high cure rates. However, in some cases, especially in patients with an advanced disease, the disease progresses during treatment or the patient relapses after complete remission.1–6 The identification of patients who would benefit from more- or less-aggressive therapeutic approaches is important, and several investigators have attempted to identify clinical and laboratory markers that predict unfavorable outcomes from HL.

In this setting, 3 large, cooperative groups have independently developed prognostic scoring systems for early Ann Arbor stages of HL using clinical and biologic parameters.7–9 The German Hodgkin Study Group has developed an international prognostic score for advanced stages of the disease, based on 7 adverse laboratorial and clinical parameters.10 Although these prognostic models are useful for identifying better responses to standard therapy, they are of limited use in recognizing patients with a high risk of treatment failure. Indeed, none of the models provide enough information to support changes in the standard treatment that would improve the outcome of HL, even in the early stage of the disease.11,12 In addition, most of the proposed models do not account for important biologic features of HL, notably changes in the immune system during the course of the disease.

It is well established that patients with HL have a defect in the cell-mediated immune response, and that defect may be attributable to cytokines produced by Hodgkin and Reed-Sternberg (H-RS) cells and by reactive inflammatory cells.13–15 Indeed, cytokines are important in the pathogenesis of classic HL (cHL).16–20 Most cHL cases are characterized by an unbalanced production of T_H2 cytokines.15,21–29 

Interleukins (ILs) 6, 10, and 13 are among the cytokines commonly detected in patients with HL. In addition, these cytokines and their respective receptors have been reported to be expressed in H-RS cells.13,26,30–32 Interleukin 6 is an important immunomodulatory cytokine that can induce growth and maturation of B and T cells33,34 and acts as a growth factor for B cells, including those transformed by the Epstein-Barr virus (EBV).35 Elevated IL-6 levels in the serum of patients with HL have been associated with poor prognosis because of lower rates of complete remission.23,27 Interleukin 10 is a cytokine with strong anti-inflammatory properties.13 It supports the growth and differentiation of B cells and inhibits apoptosis of B and T cells induced by glucocorticoids and chemotherapy.36 Elevated serum levels of IL-10 have been associated with poor outcomes in HL.23,37–39 Interleukin 13 regulates the humoral response by driving the proliferation and survival of B cells and triggering immunoglobulin class switching. It has been suggested that IL-13 acts as an autocrine growth factor for the neoplastic cells in HL.24,40–42 

Although the literature suggests that IL-6, IL-10, and IL-13 are important in the pathogenesis of cHL, few studies have analyzed these cytokines for the clinical and prognostic information they may provide about the disease.24–28,39,43 In addition, insufficient data are available on the levels of these cytokines before and after antineoplastic treatment.26 Therefore, the present study aimed to evaluate IL-6, IL-10, and IL-13 serum levels before and after cHL treatment and to investigate the possible associations with clinicopathologic parameters and response to treatment.

Twenty-seven patients were selected for this study. They were all negative for human immunodeficiency virus, and their diagnosis of cHL was confirmed by histopathology and immunophenotyping. None of them had previously undergone cytotoxic or radiation treatments. Patients were treated between 2005 and 2008 at Botucatu Medical School, São Paulo State University, and Amaral Carvalho Hospital, both in São Paulo State, Brazil. Serum samples were collected at the initial medical visit on diagnosis of HL and again 1 month after the end of chemotherapy. In addition, serum samples from 26 healthy, blood donor volunteers were collected and evaluated as a non-HL control group. All samples were stored at −80°C until evaluation. The study was approved by the Research Ethics Committee of both institutions from which subjects originated, and informed consent was obtained from all participants, in agreement with institutional guidelines. All study procedures were in accordance with the Helsinki Declaration of 1975.

All patients had their disease stage determined clinically according to the Ann Arbor system.44 Each patient underwent a complete medical history interview and physical examination and had blood cell counts; biochemical profiles; computed tomographies of the chest, abdomen, and pelvis; bone marrow biopsy and aspiration. The presence of B symptoms was characterized by one or more of the following symptoms: unexplained fever of 38°C or higher, drenching night sweats, or unexplained loss of more than 10% of body weight in the preceding 6 months. Bulky disease was defined as the presence of a nodal mass greater than 10 cm in diameter. Hemoglobin, white blood cell counts, serum albumin, erythrocyte sedimentation rate, and serum lactate dehydrogenase were measured by standard assays. Anemia was defined as a hemoglobin level of less than 12 g/dL. In both institutions, the lower reference range cutoff for albumin was 3.5 g/dL. Serum β2 microglobulin was measured by radioimmunoassay, and its upper reference range cutoff was 3.5 mg/dL. All patients with advanced HL were categorized according to their international prognostic score.10 

Immunohistochemical staining was performed on histologic sections of paraffin-embedded tumor biopsies. The sections were stained with the monoclonal antibodies anti-CD43 (DF-T1 clone, 1∶150), anti-CD45 (2B11 + PD7/26 clone, 1∶200), anti-CD20 (L26 clone, 1∶200), anti-CD30 (Ber-H2 clone, 1∶40), anti-CD15 (C3D-1 clone, 1∶150), and the polyclonal anti-CD3 (1∶150) antibody, all from DakoCytomation (Carpinteria, California). All cases were reviewed by an experienced hematopathologist. Epstein-Barr virus infection was assessed by in situ hybridization with a biotinylated probe against EBV-encoded RNA 1, as previously described.45 Validated EBV-positive Burkitt lymphoma biopsies were used as positive controls for EBV assessment, using EBV-encoded RNA 1 in situ hybridization.

The levels of IL-6, IL-10, and IL-13 were determined in serum of patients and controls. Samples were stored frozen in small aliquots (500 µL) and thawed only once. Quantification of cytokines was performed using the following commercially available, sandwich enzyme-linked immunosorbent assays (ELISA), according to manufacturer instructions: human IL-13 Quantikine (R&D Systems, Inc, Minneapolis, Minnesota) and human IL-6 ELISA MAX and human IL-10 ELISA MAX (BioLegend, Inc, San Diego, California). Each serum sample was assayed twice, and cytokine levels were determined with the Bio-Rad microplate reader model 680 (Bio-Rad Laboratories Inc, Hercules, California) set at a wavelength of 450 nm, with (IL-13) or without (IL-6 and IL-10) corrections at 570 nm. According to the information provided by the manufacturer of the ELISA kits, the lower detection limits for IL-6, IL-10, and IL-13 were 4 pg/mL, 2 pg/mL, and 32 pg/mL, respectively.

Treatment of patients with HL was based on adriamycin, bleomycin, vinblastine, and dacarbazine or on bleomycin, etoposide, adriamycin, cyclophosphamide, vincristine, procarbazine, and prednisone regimens. Involved-field radiation therapy was considered consolidation treatment for patients in early and intermediate stages or for those presenting with bulky disease. Complete remission was defined as absence of the disease for at least 1 month, as assessed by physical examination and appropriate laboratory and imaging studies. Partial remission was defined as more than 50% reduction of tumor mass measurable in 2 dimensions. Progressive disease was defined as enlargement (>25%) of a tumor at preexisting disease site or development of disease at a previously uninvolved site. Primary treatment failure was defined as progressive disease during initial treatment, failure to achieve complete remission or partial remission after initial therapy, or progressive disease within 5 months after complete remission.

Data were grouped as medians and lower and upper quartiles. Categorical variables were compared by the χ² test and continuous variables by the Mann-Whitney U test or a 1-way analysis of variance. The Wilcoxon test was used to compare serum levels of both IL-6 and IL-10 before and after treatment. The Spearman rank correlation test was conducted to assess correlations between serum levels of the studied ILs and the storage time of the serum samples to investigate any detrimental effect of this parameter on the cytokines levels detected by the ELISA.

Clinical and laboratory data were used in 2 models of multiple linear regression to predict serum levels of IL-6 and IL-10. Before performing the statistical tests, values for cytokine levels were transformed to their Naperian logarithms to normalize their distribution. Collinearity among variables was tested using the χ² test. The models were checked for linearity, independence, and normality. All statistical calculations were performed with the SigmaStat 3.5 Software (Systat Software Inc, San Jose, California), and P values were considered significant if lower than .05.

Serum IL-6 levels in the blood donor (control) group ranged from less than 4 pg/mL to 20.2 pg/mL (median, <4 pg/mL), and serum IL-10 levels from less than 2 pg/mL to 50.1 pg/mL (median, <2 pg/mL). Serum IL-13 levels were always less than the detection limit of the assay (<32 pg/mL). Eighteen out of 26 controls (69%) had serum IL-6 levels less than the assay detection limit (<4 pg/mL), whereas the levels of 19 controls (73%) were less than the detection limit for IL-10 (<2 pg/mL).

Age and sex distributions were similar between patients with HL and control volunteers. During the follow-up (median, 29 months), 25 of the 27 patients (93%) survived. Five patients (19%) failed to respond to treatment: 4 of the 5 patients (80%) responded only partially to chemotherapy, and 1 of the 5 patients (20%) presented with disease progression 4 months after complete remission. Among the 5 patients, 2 patients (40%) had advanced-stage disease. Deaths were due to disease progression in 2 of the 5 patients (40%) in the treatment-failure group.

The median value for the IL-6 serum levels at diagnosis was 20.7 pg/mL (range, <4–434.2 pg/mL). Only 5 patients (19%) had serum IL-6 levels less than the detection limit of the assay. In the case of IL-10, the median value of the pretreatment serum levels was 14.6 pg/mL (range, <2–250 pg/mL), and only 4 patients (15%) had serum levels less than the assay detection limit. For both IL-6 and IL-10, the serum levels of the cytokines were significantly higher in patients with HL before treatment compared with volunteers in the control group (P < .001). After treatment, the median serum IL-6 level was less than 4 pg/mL in patients with HL—significantly less than their pretreatment median (P < .001). The same was observed for IL-10 (median, 4.6 pg/mL after treatment; P < .001). Before and after treatment, all patients with HL had serum IL-13 levels less than the assay detection limit.

The median time samples that were stored before ELISA experiments were conducted was 18.6 months for IL-6 (range, 4.4–40.7 months), 18.8 months for IL-10 (range, 13.3–30.5 months), and 18.6 months for IL-13 (range, 13.2–30.3 months). The storage duration had no influence on the levels of cytokines detected (IL-6, R  =  −0.15, P  =  .44; IL-10, R  =  −0.15, P  =  .46).

When analyzed as continuous variables, serum IL-6 levels before treatment were higher in patients with abdominal HL involvement, hepatomegaly, B symptoms, and anemia (Figure 1). Serum IL-6 was also compared with age and gender, Ann Arbor stage, mediastinal involvement, splenomegaly, extranodal disease, bone marrow involvement, bulky disease, leukocytosis, lymphocytopenia, albumin serum levels, lactate dehydrogenase, serum β2 microglobulin, EBV-encoded RNA 1, HL histologic subtype, CD15 and CD20 positivity and response to treatment, but no association was found (Table 1). Higher pretreatment levels of IL-6 were observed in patients with treatment failure (P  =  .03), independent of the initial response. Otherwise no association was observed between the posttreatment levels of IL-6 and treatment response or treatment failure. The 2 patients who died from disease progression presented with significantly higher posttreatment levels of IL-6 (P  =  .02).

Serum IL-10 levels were higher in patients with low serum albumin. In addition, patients with advanced Ann Arbor stage, abdominal involvement, hepatomegaly, B symptoms, and anemia also tended to present with higher IL-10 levels (Figure 2). There was no association between IL-10 levels and the remaining aforementioned variables. In advanced HL, no association was found between serum IL-6 or IL-10 and the prognostic categories defined by international prognostic score.

Multiple linear regression was performed for Ann Arbor stage, mediastinal involvement, bulky disease, hepatomegaly, bone marrow infiltration, B symptoms, leukocytosis, and lymphocytopenia, all well-known clinical parameters in cHL. Gender and age were included, independent of association with IL-6 or IL-10 levels, because they are also relevant prognostic parameters for HL. Abdominal involvement, splenomegaly, extranodal disease, and anemia were excluded from the model because of collinearity. Serum β2 microglobulin, lactate dehydrogenase, serum albumin, and EBV-encoded RNA 1 were not included because data on these parameters were not available for some patients. B symptoms and lymphocytopenia were identified as potent predictors of serum IL-6 levels before treatment, as shown in Table 2. Although hepatomegaly was found to be associated with high IL-10 levels (P  =  .01), none of the variables evaluated could predict the serum levels of this cytokine.

Classic HL is characterized by well-defined histologic features, notably H-RS cells surrounded by heterogeneous and abundant inflammatory infiltrates, which provide a favorable microenvironment for H-RS cell proliferation and survival. A complex signaling network, shaped by cytokines, chemokines, and cell-cell interactions, seems to be crucial for the development and progression of HL tumors.16,46 Several aspects of the cytokines that contribute to HL pathogenesis provide a better understanding of immune dysfunction and the symptomatology associated with this disease.13,14,16,17 Nevertheless, the exact role of cytokines in the cause-effect relationship between H-RS cells and the microenvironment remains to be properly elucidated.

In the present study, serum levels of IL-6 and IL-10 were frequently elevated in patients with cHL at diagnosis, when compared with the levels in healthy, volunteer blood donors, and after treatment, those levels decreased substantially. Interleukin 13, however, was always undetectable in serum. Serum IL-6 and IL-10 were associated with some clinical and laboratory features of HL. Remarkably, in the patients with HL evaluated in this study, the level of serum IL-6 could be predicted by the presence of B symptoms (high levels) and lymphocytopenia (low levels).

Besides its activity as an activator of acute phase response and as a stimulatory factor for growth and proliferation of lymphocytes, IL-6 has important roles documented in both innate immunity and in the development of acquired immune response.47,48 In inflammation, IL-6 acts as a regulator of the transition between humoral and cellular responses, which may explain some features of HL pathobiology and its clinical manifestations.

Clinical and laboratory correlates of IL-6 levels in HL have been previously studied in both untreated and relapsed patients, with some controversial results.19,22,25,49 In agreement with previously reported data, the present study verified an association between higher IL-6 serum levels before treatment and the presence of B symptoms.25,26 Additionally, patients with higher IL-6 levels more often presented with hepatomegaly, abdominal disease, and anemia. On the other hand, we found no association with gender, Ann Arbor stage, or bulky disease.

Interleukin 6 is known to be important in the pathophysiology of cancer-induced cachexia because of its catabolic properties.34,50,51 Other systemic changes that might be associated with elevated IL-6 levels include the fever, night sweats, and weight loss observed in patients with HL, which are relevant for HL diagnosis and treatment design. To date, the predominant source of IL-6 in patients with HL and its activity in the tumoral microenvironment remain elusive. Of note, it has been reported that reactive inflammatory cells, especially T lymphocytes, adjacent to H-RS cells may stimulate the growth of H-RS cells both by direct production of IL-6 and by stimulation from secreted IL-6 producing other stimulatory cytokines.30 

Interestingly, lower serum IL-6 levels in the present study were predicted by lymphocytopenia, which has previously been reported, likewise B symptoms, as an adverse prognostic factor associated with advanced.18 One possible explanation for this phenomenon is that the low lymphocyte count might reflect a histologic change in HL toward a pattern with more proliferation of H-RS cells, rather than a pattern of reactive inflammatory cell stimulation. Taken together, these data suggest that a subset of HL may exhibit a less-prominent inflammatory component, along with a higher number of H-RS cells, and consequentially, a worsening evolution of the disease.

As a prognostic biomarker, IL-6 has been associated with lower survival rates in pretreated and refractory patients,25 and IL-6 immunoexpression in HL biopsies has been correlated with decreased complete response.22 In the present study, an association between higher pretreatment serum levels of IL-6 in newly diagnosed patients with HL was observed, as well as an increased likelihood of treatment failure. These data strengthen the idea that IL-6 may have a relevant role as a biomarker of treatment response and outcome in HL.

Higher IL-10 serum levels before treatment of HL, however, were associated with low serum albumin and hepatomegaly and were also associated with advanced Ann Arbor stages, abdominal involvement, B symptoms, and anemia, which is in agreement with what has been found in previous studies.23,27 However, the previously reported association between high IL-10 levels and EBV infection status was not observed in this study. Herling et al (2003) reported that higher serum IL-10 levels before treatment were associated with EBV LMP-1 immunostaining, mainly in the mixed-cellularity cHL subtype.28 One limitation of the present study is that the EBV status could not be assessed for all patients, and the lack of association between IL-10 levels and EBV infection might be due to the limited number of patients with EBV-positive HL, or even the low number of HL cases evaluated.

For all subjects evaluated, serum IL-13 levels were systematically undetectable. Similarly, a low frequency of detectable levels of this cytokine and no association with clinical features of HL were previously described.52 Interestingly, several studies have suggested a key role of IL-13 as an autocrine growth factor in HL,13,24,40,41 and IL-13 mRNA is frequently expressed in the tumors, almost exclusively by H-RS cells.40 The IL-13 receptor transcripts are expressed not only by H-RS cells but also by a large proportion of other cells within the reactive infiltrate, including fibroblasts.41 These data suggest that IL-13 may also support the maintenance of the reactive infiltrate in HL. In addition, treatment of an HL-derived cell line with an antibody that neutralizes IL-13 resulted in a dose-dependent inhibition of H-RS proliferation.40–42 The accumulating evidence indicates that coexpression of IL-13 and its receptor is a common feature of H-RS cells that boost IL-13 activity as an autocrine growth factor in HL. Although the lack of detectable serum IL-13 does not refute this hypothesis, it suggests that IL-13 might act predominantly at the tumor microenvironment level, possibly at very low concentrations and with a short range of action.

To the best of our knowledge, serum IL-6 levels after HL treatment had been evaluated previously in only one study,26 and there is no report on the IL-10 levels after treatment. Seymor and coworkers (1997) reported that only 10 of 65 patients (15%) evaluated had detectable levels of serum IL-6 after treatment, and they were significantly lower than the levels measured before chemotherapy.26 In the present study, a significant decrease of IL-6 and IL-10 was observed in serum after the antineoplastic treatment, which demonstrates that the chemotherapy disrupts the microenvironment network of cytokine signaling. In addition, it was also observed that patients who died because of disease progression had higher IL-6 levels in their serum posttreatment, suggesting that IL-6 may be involved in mechanisms of tumor resistance. A longer follow-up with a more significant number of patients with HL would be valuable to conclusively define whether the reduction of IL-6 with treatment is a clear-cut predictor of good response to treatment.

In summary, it was observed that pretreatment serum levels of IL-6 and IL-10 were associated with relevant clinical and laboratory findings in patients with cHL and that these cytokines decreased significantly after standard treatment for the disease. Higher IL-6 serum levels before chemotherapy seem to be associated with treatment failure, although this finding requires further validation with a larger series of cHL cases.

We thank Ms Celene Maria Gandin, Mr Marcos Roberto Franchi, and Mr Luis Fernando Franchi for their technical assistance in histologic and immunohistochemical techniques; Carlos Eduardo Bacchi, MD, PhD, and Francisco Carlos Quevedo, MD, for providing us access to HL biopsies for analysis; Sergio Alberto Rupp de Paiva, MD, PhD, Marcos Ferreira Minicucci, MD, PhD, and Suzana Erico Tani Minamoto, MD, PhD, for their support with the statistical analysis; and Alice de Oliveira Gonçalves, BSc, and the staff of the Molecular Pathology Laboratory at Botucatu School of Medicine for help with samples and experiments. This study was funded by the Fundação de Amparo à Pesquisa do Estado de São Paulo, grant AP 2006/00591-5 (Dr Oliveira).