CXCR4 Mutations in Lymphoplasmacytic Lymphoma Lead to Altered CXCR4 Expression

In addition to the MYD88 p.(L265P) mutation, approximately 30% of patients with lymphoplasmacytic lymphoma (LPL) carry a mutation of the CXCR4 gene (Treon et al, 2014), resulting in diminished response to therapy with Bruton tyrosine kinase (BTK) inhibitors (Cao et al, 2015; Treon et al, 2015) as well as more aggressive disease (Treon et al, 2014). The mechanism responsible for this resistance has not yet been elucidated and protein expression of CXCR4 in LPL has not been investigated. CXCR4 leads to activation of the nuclear factor (NF)κB pathway (Han et al, 2001; Hideshima et al, 2002; Chatterjee et al, 2014) whereas BTK inhibitors eventually target the NFκB pathway, thereby decreasing cell survival in B-cell lymphomas (Figure S1). As these inhibitors show less efficacy in CXCR4-mutated LPL patients, we hypothesized that CXCR4 mutations influence CXCR4 as well as NFκB protein expression. Therefore, we investigated CXCR4 expression and mutation status in LPL patients, as well as (classical) NFκB (p105/p50) expression, in order to further decipher the mechanism responsible for BTK inhibitor resistance. A total of 30 LPL bone marrow biopsies obtained between 2004 and 2017 were selected from the archives of the University Medical Centre Utrecht, the Netherlands. Slides of the bone marrow biopsies were stained for CXCR4 (Biolegend, San Diego, CA, USA; 12G5, 1:800) and NFκB p105/p50 (Abcam, Cambridge, UK; E381, 1:500), using a Ventana BenchMark Ultra (Roche, Almere, the Netherlands). A total of 3 normal bone marrow biopsies were included to analyse protein expression of CXCR4 and NFκB in normal haematopoiesis. After staining, slides were scored independently by two investigators. CXCR4 expression was scored as being primarily cytoplasmic or nuclear. For NFκB, staining pattern (heterogeneous versus diffuse) was combined with strength of staining (weak versus strong staining), with a cut-off of 70%: if ≥70% of the lymphoma cells expressed diffuse and strong staining of NFκB, the case was described as “diffuse and strong”. In addition, next generation sequencing (NGS) was performed with the Ion Chef™ on the PGM™ and S5™ Systems (Thermo Fisher Scientific, Waltham, MA, USA) using an in-house NGS haematology panel, custom-designed for leukaemias and lymphomas (Table SI). The following genes were reported in this study and their variants annotated according to these transcripts: MYD88 (NM _002468), CXCR4 (NM_001008540) and CD79B (NM_001039933). In normal bone marrow biopsies, CXCR4 showed weak nuclear staining of erythropoiesis and megakaryopoiesis (Fig 1A). No staining was seen in granulopoiesis. In contrast, NFκB showed intense cytoplasmic staining of granulopoiesis, while erythropoiesis was negative and megakaryopoiesis was either negative or showed weak cytoplasmic staining (Fig 1B). Of the 30 LPL bone marrow cases, 29 (97%) had a MYD88 p.(L265P) mutation, 9 (30%) of which also carried a CXCR4 mutation (Tables SII and SIII). In addition, one CXCR4 mutated case had a CD79B mutation. Seven out of 8 (88%) CXCR4-mutated cases showed a CXCR4 cytoplasmic staining pattern (P = 0·001; Fig 1C); CXCR4 staining was not interpretable in the remaining case. In contrast, the vast majority of cases without a CXCR4 mutation showed a nuclear staining pattern of CXCR4 (17 out of 21 cases: 81%; Fig 1D; Table 1). Furthermore, cytoplasmic CXCR4 expression was associated with a characteristic heterogeneous, weak staining pattern of NFκB, occurring in 7 out of 10 cases with cytoplasmic CXCR4 staining (70%; P = 0·001; Fig 1E; Table 1). In contrast, nuclear CXCR4 expression was associated with a diffuse and strong staining pattern of NFκB in 15 out of 16 cases with nuclear CXCR4 staining (94%; Fig 1F; Table I). No significant correlation was detected when comparing CXCR4 mutation status with NFκB staining (P = 0·206; Table SIV). As LPL patients with a CXCR4 mutation are less responsive to targeted treatment with BTK inhibitors, it is important to decipher the mechanism underlying this resistance in order to develop and improve (targeted) treatment options. Currently, the mechanism of this diminished response is poorly understood. In our study, immunohistochemical expression of CXCR4 in LPL showed two dominant patterns: cytoplasmic and nuclear staining. The cytoplasmic staining pattern occurred in 37% of LPL patients and significantly correlated with CXCR4 mutation status. This is in accordance with a previous study which measured CXCR4 expression by flow cytometric instead of immunohistochemical analysis (Poulain et al, 2016). In this study, although both CXCR4-mutated and non-mutated cases showed CXCR4 expression, the 12 mutated cases showed higher expression of CXCR4, which was independent of the CXCR4 mutation type (Poulain et al, 2016). In addition, CXCR4 expression in other lymphomas has been studied by immunohistochemistry, but the localization of the protein was not described (Shin et al, 2014; Cai et al, 2017). Interestingly, Poulain et al (2016) found that CD79B mutations were mutually exclusive with CXCR4 mutations. In contrast, the one case with a CD79B mutation in our study also had a CXCR4 mutation. In this case, the mutation frequency of CXCR4 (15%) was clearly less compared to MYD88 and CD79B (44% and 39% respectively). This might indicate that the CXCR4 mutation is present in a subclone, corresponding to the 4 cases reported by Poulain et al (2016), demonstrating mutation frequencies of CXCR4 at the subclonal level. Nonetheless, this case still displayed a cytoplasmic staining pattern of CXCR4. Four cases in our cohort displayed cytoplasmic CXCR4 staining while no CXCR4 mutation was detected, which might suggest that the CXCR4 protein expression pattern is not influenced solely by mutation status. As 3 of these cases also showed heterogeneous, weak NFκB expression, it might be hypothesized that pattern of CXCR4 protein expression (cytoplasmic versus nuclear) might have greater influence on NFκB regulation than mutation status alone. Because BTK inhibitors eventually target the NFκB pathway, it might be worthwhile to investigate this in more detail in future studies. In conclusion, our results suggest that CXCR4 mutations in LPL patients result in altered CXCR4 protein expression, i.e. cytoplasmic instead of nuclear. This altered protein expression does not seem to activate the NFκB pathway as it normally does, displayed by heterogeneous, weaker NFκB expression in cases with cytoplasmic expression of CXCR4. Thus, it might be hypothesized that altered activation of NFκB might be the mechanism responsible for diminished response to BTK inhibitors. To further unravel this mechanism, future studies with larger cohorts will need to focus on both protein expression and functioning, thereby bridging the gap between the molecular level, epitope expression and response to therapy of the individual patient. The authors thank Onur Karademir and Paula de Vree for technical assistance in the NGS analysis, Dorine van den Brink and Jojanneke Renes for assistance with CXCR4 and NFκB immunohistochemistry and Petra van der Weide for assistance with specimen collection. We thank Paul van Diest for scoring the immunohistochemistry results. Hiemcke-Jiwa had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Concept and design: Hiemcke-Jiwa, Leguit, Jiwa, Huibers, Minnema; Acquisition, analysis or interpretation of data: Hiemcke-Jiwa, Leguit, Huibers; Drafting of the manuscript: Hiemcke-Jiwa, Leguit, Huibers; Critical revision of the manuscript for important intellectual content: Leguit, Huibers, Minnema; Statistical analysis: Hiemcke-Jiwa. This research was rewarded by patient donation via ‘Vrienden van het UMC Utrecht’. The authors have no potential conflicts of interest to report. Table SI. Ion AmpliSeqTM hematology panel (custom design, University Medical Center Utrecht, the Netherlands).Table SII. Immunohistochemistry results compared to NGS results for n = 30 LPL bone marrow cases.Table SIII. NGS results of n=9 LPL cases with a CXCR4 mutation.Table SIV. NFκB staining pattern in correlation with CXCR4 mutation status.Figure S1. Schematic representation of the CXCR4-BTK-NFκB pathway. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.