IQGAP2 acts as an independent prognostic factor and is related to immunosuppression in DLBCL

Diffuse large B-cell lymphoma (DLBCL) is the most common form of lymphoma and accounts for 25–35% of non-Hodgkin lymphomas [1]. Cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) chemotherapy have been the predominant therapies for several decades. In the past two decades, the results of several phase III trials have established a regimen that utilizes a combination of rituximab and CHOP (R-CHOP) to achieve a better result. Therefore, it has become the standard-of-care therapy for patients with DLBCL, with 50–70% of patients typically being cured using this approach [2, 3]. However, the prognoses of the remaining patients who require alternative therapy are poor, and the majority of them will eventually succumb to their disease [4].

The identification of biologically distinct subtypes of DLBCL is a milestone, and gene expression profiling studies have distinguished three molecular subtypes of DLBCL, which are known as germinal centre B-cell-like (GCB) DLBCL, activated B-cell-like (ABC) DLBCL, and unclassified (UC) DLBCL [5, 6] and are correlated with a differential response to therapy. Patients with these DLBCL subtypes have significantly different gene expression profiles and overall survival rates [6‐9]. Therefore, we believe that there must be some biological characteristics in DLBCL associated with poor prognosis. Further investigation into the prognostic biomarkers in refractory DLBCL is essential for identifying patients who are unlikely to be cured by R-CHOP and for developing potential therapeutic targets. The development of monoclonal antibody (mAb) therapy directed against CD20 has had the biggest clinical impact on the treatment of DLBCL [10]. This indicates that targeted therapy is promising for treatment of DLBCL. Gene risk scores based on the expression of 6 genes have been proposed to predict the outcome of DLBCL patients [11]. However, no further application or new therapy has been developed that utilizes these genes. Further insight into prognostic biomarkers is urgently needed to develop new therapeutic strategies for refractory DLBCL.

IQ motif-containing GTPase activating protein (IQGAP) 2 is a protein coding gene, and IQGAP2 is a member of the IQGAP family. Mammals express three isoforms of IQGAP: IQGAP1, IQGAP2 and IQGAP3. Their compositions are all similar, and they regulate the cytoskeleton, cytokinesis and carcinogenesis [12]. Despite their 62% sequence identity, IQGAP1 is an oncogene, while IQGAP2 has been reported as a tumour suppressor in hepatocellular, prostate, and gastric carcinomas [13]. However, the role of IQGAP2 in DLBCL remains unclear.

In this study, we explored the mRNA expression of IQGAPs in 1457 cell lines. We also collected a total of 614 samples with mRNA microarray data to identify the relationship between the mRNA expression pattern of IQGAP2 and the clinical outcome. Furthermore, we examined the correlation between IQGAP2 mRNA and immune processes in DLBCL.

It is widely accepted that IQGAP2 is predominantly expressed in the liver rather than in the heart, brain, spleen, lung, kidney, testis, and skeletal muscle. This result was detected by an RNA blot in 1996 [16]. In the last decade, IQGAP2 was reported to be found in cancer tissues of prostate [20], lung, breast, liver, kidney, and colorectal cancer [21]. IQGAP2 is also expressed in podocytes [22] and glomerular endothelial cells [23]. However, these studies ignored the expression of IQGAP2 in leukocytes. Our study shows that IQGAP2 mRNA levels are increased in liver cancer cell lines compared to those in glioma, lung, and kidney cancer cell lines (Fig. 1), which is consistent with previous reports. Despite this, we found that cell lines derived from haematologic malignancies exhibit higher IQGAP2 levels than liver cancer cell lines (Fig. 1), which does not occur for IQGAP1 and IQGAP3 (Supplementary Fig. 1). Furthermore, single-cell sequencing showed that IQGAP2 is predominantly expressed on leukocytes but not melanoma cells (Supplementary Fig. 1). It has been reported that mice lacking IQGAP2 are resistant to chemically induced colitis and experience diminished neutrophil and macrophage production and recruitment [24]. This shows that IQGAP2 is predominantly expressed in leukocytes but not in the liver. More work is needed to confirm the expression pattern of IQGAP2 in haematologic cells.

IQGAP2 is a well-known tumour suppressor. IQGAP2 is decreased in several cancers, including hepatocellular [25], prostate [20], and gastric [26] carcinomas. Of note, IQGAP2 deficiency results in hepatocellular carcinoma in mice, but IQGAP1 and IQGAP2 double disrupted mice have less hepatocellular carcinoma and normal survival. This suggests that IQGAP1 and IQGAP2 have inverse functions [27]. Our results verify that IQGAP2 mRNA expression is correlated with the overall survival of patients with hepatocellular and kidney carcinoma (Supplementary Fig. 3). In contrast, we found that IQGAP2 mRNA expression is inversely related to the survival of patients with DLBCL (Fig. 2 and Fig. 3). Given that IQGAP2 mRNA is also related to immunosuppression and leukocyte infiltration (Fig. 6 and Fig. 7), we propose that the high expression of IQGAP2 in leukocytes and various leukocytic components in DLBCL may contribute to the immunosuppressive role of IQGAP2 in DLBCL.

Immune evasion plays a key role in DLBCL oncogenesis. The development of DLBCL requires both deregulated expression of the oncogene as well as escape from T cell-mediated tumour surveillance [28]. Most DLBCL cells fail to express both β2-microglobulin and CD58, which are required for immune recognition of malignant cells [29]. The expression of PD-L1 is related to poor outcomes [30]. In addition, nontumour cell infiltration into DLBCL also leads to a poor prognosis [31]. When the tumour immune microenvironment was impaired by high PD-1 expression on CD8+ T cells or PD-L1 expression on T cells and macrophages, patients had significantly worse outcomes after R-CHOP therapy [32]. Immunotherapy provides additional choices for patients with DLBCL that is refractory to chemotherapies or that has relapsed after stem cell transplantation. The chimeric antigen receptor T-cell (CAR-T) therapeutic tisagenlecleucel produced a 52% overall response rate in these patients (40% of the patients had complete responses, and 12% had partial responses) [33]. Although nivolumab (anti-PD-1) monotherapy showed a low overall response rate among these patients, two patients survived for 30 and 38 months after the first treatment [34]. Our data show that IQGAP2-associated genes are associated mostly with immune processes (Fig. 6. A and B) and that IQGAP2 mRNA is correlated with immunosuppressive genes (Fig. 6. C and D). Samples with relatively higher IQGAP mRNA levels consistently exhibited higher infiltration of leukocytes (Fig. 7). These results indicate that IQGAP2 might be involved in immunosuppression in DLBCL.

IQGAP2 is mainly proposed as a cell motility modulator. IQGAP2 recruits active Rac1 to promote membrane ruffling by interacting with AKAP220 [35]. IQGAP2 regulates actin assembly downstream of thrombin stimulation by interacting with Arp2/3 and F-actin [36]. IQGAP2 also inhibits GTPase activity by interacting with Cdc42 and Rac1 [16]. In addition, IQGAP2 is required for the glomerular filtration barrier by maintaining podocyte structure and function [23]. Moreover, IQGAP2 is an IFNα antiviral effector gene acting nonconventionally through the NF-κB pathway in HCV-infected hepatocytes [37]. Our results show that IQGAP2-associated genes are related to chemotaxis. The effect of IQGAP2 on cell motility in DLBCL needs further investigation.

A previous study showed that IQGAP2 is involved in regulating IFN-stimulated gene (ISG, with known antiviral properties) induction by IFN in an NF-κB-dependent manner in hepatic cells [37]. This indicates that IQGAP2 may be necessary for the activation of the NF-κB pathway. NF-κB is a key oncogenic pathway in DLBCL [38‐40]. The activation of the NF-κB pathway dependent on IQGAP2 is probably the main reason for immunosuppression in DLBCL because of its induction of immune checkpoint molecules such as PD-L1 [41]. This speculation is worth further verification.

Our work indicated that IQGAP2 is an independent prognostic factor and is related to immunosuppression in DLBCL. This indication was mainly derived from the results of silico microarray and RNA sequencing. The protein expression of IQGAP2 was compared among different subtypes of DLBCL (Fig. 5). But the prognostic value and immunosuppressive role of IQGAP2 protein in DLBCL need to be further validated in fresh specimens.

In conclusion, our data show that IQGAP2 mRNA is more highly expressed in several cell lines derived from haematologic malignancies compared to most other cancer cell lines. We also demonstrate that IQGAP2 is an inversely survival-related marker in DLBCL, which differs from its role in other tumours. Furthermore, we propose that IQGAP2 may play a role in immunosuppression in DLBCL. Our work provides evidence for further investigation of IQGAP2 in DLBCL.