Elevation of serum interleukins 8, 4, and 1β levels in patients with gastrointestinal low-grade B-cell lymphoma

Proinflammatory cytokines that are produced by helper T cells (Th) regulate immune reactions, facilitate class switching of B cells, and prolong the lifespan of B and T cells. Eradication therapy using antibiotics is sometimes effective against gastrointestinal (GI) malignant lymphoma, suggesting that the tumor development or progression is affected by the inflammatory microenvironment. In the present study, serum samples from 148 patients with various subtypes of malignant lymphoma were tested for 11 proinflammatory Th1/Th2 cytokines. In the comparison by subtype or GI lesions, serum interleukin (IL)-8 (P = 6.7E−05), IL-4 (P = 7.5E−05), and IL-1β (P = 0.0043) levels showed significant differences among subtypes, being particularly elevated in follicular lymphomas (FL) and mucosa-associated lymphoid tissue (MALT) lymphomas. Serum IL-8 levels were elevated in GI-FL and MALT lymphomas, and serum IL-4 and IL-1 β levels were elevated in MALT lymphomas. These findings show that GI low-grade B-cell lymphoma could develop against the background of an inflammatory microenvironment. Thus, these cytokines may be useful as diagnostic markers and could provide new insights into tumor development.


Results
Serum interleukin (IL)-8, IL-4 and IL-1β levels were different depending on a lymphoma subtype. Among the 11 analyzed cytokines, IL-8, IL-4, and IL-1β levels were significantly different from control depending on the lymphoma subtype using one-way analysis of variance among control and lymphoma subtypes (P = 6.69E− 05, 7.46E− 05, and 0.0043, respectively; Table 1). These three cytokines were also significantly different among the lymphoma subtypes (P = 1.63E− 04, 1.15E− 04, and 0.0040, respectively; Table 1). When we limited the analysis to three major subtypes (DLBCL, FL, and MALT lymphoma) and subdivided the cases by the presence or absence of GI lesions, these three cytokines also showed serum levels significantly different from controls using one-way analysis of variance among control and lymphoma subtypes (P = 0.0039, 0.00019, and 0.024, respectively; Table 2). These three cytokines were also significantly different among the lymphoma subtypes (P = 0.021, 0.00076, and 0.045, respectively; Table 2). Serum IL-8 level was elevated in GI low-grade B-cell lymphoma. The serum IL-8 level was elevated in FL and MALT lymphomas with GI lesions (P = 0.0039; Fig. 1A,B), and there was no difference between these two subtypes with GI lesions (P = 0.25; Fig. 1A). In FL, the serum IL-8 level in cases with GI lesions was higher than that in cases without GI lesions, but in MALT lymphomas, there was no difference between cases with and without GI lesions (Fig. 1A). Although the IL-8 value was higher in GI(− )DLBCL than in GI(+ )DLBCL, it was not significantly different (P = 0.138).
Serum IL-4 and IL-1β levels were elevated in GI-MALT lymphoma cases. Serum IL-4 and IL-1β levels were elevated in MALT lymphoma cases compared with healthy controls (P = 0.00019 and 0.024, respectively; Figs 3A and 4A). Both cytokines were also elevated when we analyzed only serum samples from MALT lymphoma cases with GI lesions (Figs 3B and 4B). In MALT lymphoma cases, the serum levels of both cytokines were not linked to the presence of GI lesions (P = 0.39 and 0.40, respectively; Figs 3A and 4A). Similarly, IL-4 and IL-1β were higher in GI− DLBCL than in GI+ DLBCL, but they were not significantly different (P = 0.410, 0.045, respectively). Discussion IL-8 (CXCL8) was first identified as a neutrophil chemoattractant that belongs to the CXC chemokine family 10 . IL-8 is a proinflammatory chemokine and an angiogenic factor that mediates various inflammatory responses and is associated with tumor progression 11 . In non-small cell lung carcinoma, IL-8 is overexpressed under the influence of KRAS and enhances the stromal response by inducing inflammation and angiogenesis, which are associated with tumor proliferation and negatively correlate with patients' survival 12 .   After antigen stimulation in the Peyer's patch, B cells undergo class switching to IgA expressing α 4β 7 integrin and then home to the intestine via MAdCAM-1, which is expressed on endothelial cells of intestinal blood vessels 13 . Judging by the gene expression analysis reported previously, MAdCAM-1 and CCL20 were overexpressed  IL-4 is produced by Th2 cells, mast cells, and basophils and activates the JAK/STAT and PI3K pathways. In addition to the association of IL-4 with inflammatory responses and allergic reactions, B cells express IL-4 receptor, which promotes their proliferation as a co-mitogen, controls class switching to IgE or IgG4, and prolongs the cellular lifespan. IL-4 also makes B cells express THY1 by associating with lipopolysaccharides 8,15 . IL-1β is a component of the IL-1 complex and is linked to several inflammatory processes by activating macrophages. In addition to IL-8, we examined IL-4 receptor α (IL-4RA) immunostaining on 30 cases (10 GI FL, 10 nodal FL, and 10 GI MALT). In contrast to IL-8RB, IL-4RA was weakly positive for more than half of the cases regardless of the subtypes (P = 0.63) and was not reflected high serum IL-4 value in MALT lymphoma. (data not shown) It is suggested that IL-4 acts indirectly on tumor cells. Unfortunately an adequate level of IL-1 receptor antibody was not available.
HP is known to be a major factor in the MALT lymphoma development. According to some researchers who performed cytokine expression analysis of HP-related and HP-unrelated gastritis, several Th1/Th2 cytokine combinations participate in the HP-related inflammation 9 . Elevation of IL-4, IL-8, IL-1β , IL-6, IL-10, tumor necrosis factor (TNF)-α , and interferon (IFN)-γ levels was also observed. In HP-related gastritis caused by the CagA-positive strain, elevation of IL-1β and IL-18 levels was reported 9 .
In the present study, elevation of IL-4, IL-1β , and IL-8 levels was detected in MALT lymphomas. This change may be caused by the HP-related inflammation. In another report, IL-4 production was found to be increased when a patient's peripheral blood mononuclear cells (PBMCs) were stimulated by HP hsp60 or HP lysate 16 . Gene pleomorphism of the IL-4 gene and IL-1 receptor antagonist (IL-1RN) gene was reported to negatively correlate with HP-related gastric cancer 17 , suggesting that reactivity to such cytokines is linked to tumor development. In the present study, the correlation of HP infection with changes in cytokine production was not observed probably because of the insufficient sample size (Supplementary Figure).
When confined to low-grade B-cell lymphoma, most of the cases were primary GI cases and a few cases were those of secondary GI involvement in this study. We could not examine if there were any clinicopathological differences between the primary and secondary cases because the number of secondary cases was a few. However, it is suggested that the cytokine profiles shown in this study reflected those of primary GI low-grade B-cell lymphoma.
Although the main lesion of primary GI FL is often in the duodenum, total GI tract screening using double-balloon endoscopy revealed that 85% of the cases had lymphoma lesions that were broadly located in the GI tract 7 . In addition, some cases of MALT lymphoma of the stomach have concomitant lesions in the duodenum or colon. In our series, there was small number of patients whose GI lesion restricted to single organ were 6 cases (all stomach cases) for MALT, 13 (duodenum: 10, ileum: 3) for FL, 7 (stomach: 5, ileum: 1, colon: 1) for DLBCL. Considering these facts, HP + /− gastritis patients were not necessarily fit as adequate controls for this study as well as the inflammatory disease patients affecting corresponding organs, such as duodenitis, ileitis and colitis, because comparison among other subtypes, would be difficult. Therefore, we included healthy subjects as the controls in this study.
It was reported that the clinical behavior of the GI FL cases was very indolent, similar to MALT lymphoma but in contrast to that of the nodal type 7,18 . Moreover GI FL was reported to have similar characteristics with MALT lymphoma, both in the immunohistochemical and molecular aspects including gene expression profiles 14,19 . In the present study, GI FL and MALT lymphoma had similar elevations in serum IL-8; to discriminate MALT from GI FL, IL-4 and 1β elevations could be used. In this study, the functions of the cytokines indicated for the specific lymphoma subtypes were not investigated and warrant further study. For example, lymphoma development under the influence of specific cytokines using transgenic mice holding t(14;18) needs to be examined. To our knowledge, there is no report analyzing the proinflammatory cytokine profiles that are specific for GI lymphoma, and we believe that the relevance of this study is in the discovery of important serum cytokines, which could be the factors involved in the pathogenesis of lymphoma.
In summary, analysis of Th1/Th2 cytokine levels in serum samples shows that IL-8 levels were elevated in GI FL and MALT lymphomas, whereas IL-4 and IL-1β levels were elevated in MALT lymphomas. These results show that GI low-grade B-cell lymphoma probably develops in an inflammatory microenvironment. This finding needs confirmation with a larger sample size, particularly for the utility of these cytokines as auxiliary diagnostic biomarkers. Further studies of the mechanisms involving these cytokines in the tumor microenvironment would be warranted and may lead to new therapeutic targets.  Table 3. The patient age had a median age of 62 years (range 41-83 years), and the male to female ratio was 16:15. Serum samples for all cases, except three, were taken before treatment. The three cases were as follows: 1 mantle cell lymphoma (MCL), 1 Hodgkin lymphoma (HL), and 1 MALT lymphoma (HP eradication therapy alone). The other clinical characteristics were obtained from 31 cases. The numbers of serum samples corresponding to each lymphoma subtype were as follows: 55 cases of DLBCLs including 2 intravascular large B-cell lymphomas (IVL), 36 FLs, 15 MALT lymphomas, and 42 other lymphomas (2 angioimmunoblastic T-cell lymphomas, 2 adult T-cell leukemias/ lymphomas, 2 acute lymphoblastic leukemias, 2 non-HLs, 2 Burkitt lymphomas, 1 hairy cell leukemia, 1 hepatosplenic T-cell lymphoma, 5 HLs, 3 MCLs, 4 natural killer (NK)/T-cell lymphomas, 15 peripheral T-cell lymphomas, 2 small lymphocytic lymphomas, and 1 cutaneous follicle center lymphoma). Serum samples from 21 healthy volunteers served as controls. Approval for the study protocol was obtained from the institutional review board at Okayama University Graduate School of Medicine, Dentistry, and Pharmaceutical Sciences and similar boards at the collaborating hospitals. Informed consent was obtained from all subjects. All study procedures were conducted in accordance with the guidelines of the Declaration of Helsinki. In our hospital and other collaborating centers, diagnosis was based on the current WHO classification using appropriate immunostaining, such as CD20, 10, 5, BCL2, and cyclinD1 for low-grade B-cell lymphoma, and CD20, 10, and 5 for DLBCL 6 . In the analysis comparing the presence of GI lesions in DLBCL, FL, and MALT, cases of IVL (n = 2) and composite MALT and DLBCL (n = 1) were excluded. The behavior of composite MALT and DLBCL could not be investigated with a low sample number. Relevant information concerning involved sites was obtained from 81 out of 106 cases of major 3 subtypes (DLBCL, FL, and MALT). Among them, 37 had GI lesions and 28 of these were primary sites of lymphoma (case numbers of "primary GI cases" and "cases with secondary involvement or suspicious for secondary involvement of the GI tract" were 1 and 7 for DLBCL, 17 and 1 for FL, and 10 and 1 for MALT lymphoma, respectively).

Statistical analysis.
One-way analysis of variance was used. For statistical tests P < 0.05 was set as significant.   Table 3. Cases included in this study. DLBCL: diffuse large B-cell lymphoma, FL: follicular lymphoma, MALT: mucosa-associated lymphoid tissue lymphoma.