Overexpression of sphingosine-1-phosphate receptor 1 and phospho-signal transducer and activator of transcription 3 is associated with poor prognosis in rituximab-treated diffuse large B-cell lymphomas

Diffuse large B-cell lymphoma (DLBCL) is a biologically and clinically heterogeneous entity that accounts for 30-50% of non-Hodgkin lymphomas, depending on geographical area [1, 2]. Germinal center B-cell-like (GCB) and activated B-cell-like (ABC)/non-GCB subgroups were previously identified as two distinct subgroups of DLBCL that showed differentially activated signaling pathways [1, 3‐6]. Typically, the NF-κB pathway is constitutively activated in ABC-like DLBCLs and cooperates with the STAT3 pathway to promote cell survival [7‐9], while dependency on the PI3K/Akt pathway has been demonstrated in GCB-type DLBCL [10]. Recently, high-throughput techniques have revealed more complex features of genetic alterations and identified novel therapeutic pathways in DLBCL [9, 11, 12]. One of the promising candidate pathways for targeted therapy in DLBCL is the STAT3 pathway [13].

Unlike inflammatory conditions with transient STAT3 activation, STAT3 is aberrantly and constitutively activated in many cancers, including hematolymphoid malignancies [14]. Activated STAT3, i.e., phospho-STAT3 (pSTAT3), is transported into the nucleus, functioning as a transcription factor for various genes involving cellular apoptosis, proliferation, and survival [15]. In lymphomas, the expression and activation of STAT3 have previously been investigated in human lymphoma tissues and cell lines [16‐18]. The nuclear expression of STAT3 or pSTAT3 alone, as detected by immunohistochemistry, was shown to be a poor prognostic factor in all DLBCL patients, including the GCB and non-GCB/ABC subgroups [16, 17].

S1PR1 is a member of the G-protein-coupled receptor for sphingosine-1-phosphate (S1P), a chemokine mediating immune cell migration [19, 20]. S1P is produced intracellularly by sphingosine kinase (SPHK) 1/2; it is released from the cells and then binds to the S1P receptors (S1PR1-S1PR5) of target cells in an autocrine and/or paracrine manner [20, 21]. S1PR1 transduces intracellular signals, leading to various biologic effects, including cell proliferation, survival and migration via the ERK, Akt, and Rac pathways, respectively. Recently, it has been reported that S1PR1 is also transcribed by pSTAT3, and enhanced S1PR1 subsequently and reciprocally activates STAT3, thus building a positive feedback loop that involves the S1PR1/pSTAT3 pathway, which is important for consistent STAT3 activation in mouse and human solid tumors and tumor-associated myeloid cells [22].

At present, only a few studies on S1PR1 in malignant lymphoma are available. Hodgkin lymphoma and mantle cell lymphoma showed S1PR1 expression in cell lines or tissues, suggesting potential biologic roles for S1PR1 in this context [23, 24]. Furthermore, co-activation of S1PR1 and STAT3 was observed in ABC-DLBCL cells and tissues, and S1PR1 was suggested as a potential target for blocking STAT3 activation [18]. However, there have been no integrated studies on the clinicopathologic and prognostic implications of S1PR1 and STAT3 activation in DLBCL patients. We hypothesized that S1PR1, STAT3, and/or the co-activation of S1PR1/STAT3 pathway might be useful prognostic markers in DLBCL. In this study, we comprehensively investigated the expression of S1PR1 and pSTAT3 and analyzed their correlation with clinicopathologic features and impacts on clinical outcomes in rituximab-treated DLBCL patients.

In the present study, we demonstrated for the first time that 1) S1PR1 is an independent prognostic factor in rituximab-treated DLBCL patients, and 2) the S1PR1/pSTAT3 risk category is useful for risk stratification of DLBCL patients.

Given that S1PR1 and pSTAT3 closely co-operate during inflammatory and immunologic processes and neoplasms, the S1PR1/pSTAT3 risk category was developed to help stratify the risk of rituximab-treated DLBCL patients using immunohistochemistry for S1PR1 and pSTAT3. This category has prognostic value in a stage- and IPI-independent manner. Notably, DLBCLs with S1PR1 expression exhibited worst prognosis regardless of pSTAT3 expression and represented the high risk (S1PR1+) group. Meanwhile, DLBCLs without S1PR1 expression could be divided into a low risk group (S1PR1-/pSTAT3-), which is assumed to have an inactive form of S1PR1 and pSTAT3, and an intermediate risk group (S1PR1-/pSTAT3+), which might reflect pSTAT3 activation via an alternative non-S1PR1 pathway. These data suggest that aside from the S1PR1/pSTAT3 positive feedback loop, S1PR1- or pSTAT3-associated alternative signaling pathways might also be involved in the biology of DLBCL. In fact, in the present study, S1PR1 and pSTAT3 expression did not correlate significantly with each other (p = 0.265). Because S1P-S1PR1 signaling is known to be associated with several important pathways, including the mTOR pathway in T cells [26] and the Akt pathway in non-lymphoid cells [27, 28], it is possible that S1PR1-mediated oncogenic signaling pathways other than STAT3 might underlie the aggressive behavior of S1PR1-positive DLBCLs. As such, the possible roles of S1PR1 in the biology of DLBCLs via mechanisms other than STAT3 signaling remain to be clarified.

It was previously reported that STAT3 or pSTAT3 was a prognostic factor in DLBCLs [16, 17]. In Wu’s series of 74 patients, approximately half of whom had been treated with R-CHOP therapy, the prognosticator was not pSTAT3, but STAT3, although the expression of these two molecules was highly concordant [16]. STAT3 activation has previously been described as a main mechanism of ABC-DLBCL [18], and in Huang’s series of 185 patients who had undergone R-CHOP therapy, pSTAT3 was a significant prognostic factor for event-free survival in ABC-DLBCL [17]. In the present study, the nuclear expression of pSTAT3 was much higher in the non-GCB type and tended to be associated with shorter overall survival in patients in the early stages or in those who had low IPIs. However, overall, pSTAT3 expression alone was not an independent prognostic factor in DLBCL patients treated with rituximab and showed no prognostic significance in either the GCB or non-GCB/ABC subgroup. Based on these observations, pSTAT3 expression itself is thought to have an influence on the prognosis of DLBCLs in a more complicated way, which remains to be further investigated.

S1PR1 signaling can be blocked by an effective inhibitor, FTY720, which was developed as an immunosuppressant and is being used in the treatment of patients with multiple sclerosis [20]. Moreover, other selective S1PR1 modulators such as Syl930 have been developed to reduce side effects including bradycardia [29]. These molecules may be useful for functional studies to clarify the role of S1PR1 signaling in the biology of DLBCL. In a previous study by Liu, STAT3 was co-localized with S1PR1 in the tumor cells of a few ABC DLBCL tissues, and the inhibition of S1PR1 with FTY720 or S1PR1 shRNA successfully suppressed STAT3 activity and tumor cell growth in vitro and in an in vivo murine lymphoma model [18]. SPHK1, which catalyzes S1P production within the cells, was more frequently expressed in B-cell non-Hodgkin lymphomas with higher clinical grade [30]. Considering that the immunohistochemical expression of S1PR1 was independently associated with poor clinical outcome of DLBCL patients in the present study, the S1P/S1PR1 pathway is suspected to play a role in contributing to the aggressive behavior of DLBCL and, in addition to STAT3, is considered to be a promising therapeutic target in DLBCL.

To measure the expression level of proteins in this study, we used IHC method and graded according to the intensities and proportions of the stained tumor cells in DLBCL tissues. To more validate our data, a proteomic study using spectral abundance in a shotgun study or a more quantitative multiple reaction monitoring might also be applicable in future studies [31, 32].

Another notable finding in this study is that S1PR1 and pSTAT3 are frequently expressed by DLBCLs primarily occurring in the extranodal and UAT areas. Although UAT has been described as a unique site of extranodal NK/T cell lymphoma, no specific features have been recognized for DLBCLs of the primary UAT lesion. Interestingly, a recent study revealed that S1P/S1PR1/STAT3 signaling was an important link between chronic intestinal inflammation and colitis-associated cancer [33]. Considering the important role of the S1PR1/STAT3 pathway in the inflammatory reaction and inflammation-associated carcinogenesis, this anatomic predilection of S1PR1 and STAT3-expressing DLBCLs suggests that the S1PR1/STAT3 pathway may be involved in lymphomagenesis in inflammation-prone areas, such as UAT.

We demonstrate here that S1PR1 is a new immunohistochemical prognostic marker and that the S1PR1/pSTAT3 risk category can be used for risk stratification in rituximab-treated DLBCL patients. We also suggest that the S1PR1 may be a potential therapeutic target in a subset of DLBCLs.