BTKi orelabrutinib improved the outcome of newly-diagnosed PCNS-LBCL by alleviating MyD88L265P-induced PIM1 stability Authors: Yan Yuan, Bo Dai, Haoshu Zhong, Jiajun Zheng, Zhen Fan, Di Chen, Qian Wang, Hao Wu, Yuan Li, Ping Zhu, Tianling Ding, Shu Wang, Kunpeng Wu, Yu Zhang, Zhenhua Wang, Rui Feng, Fengping Zhu, Kang Zheng, Zhiyong Qin, Bojie Yang, Hong Chen, Xiaolei Lin, Yu Yao, Dongxiao Zhuang, Zhifeng Shi, Ying Mao, Tong Chen Published: 2026 ABSTRACT AbstractPrimary central nervous system large B-cell lymphoma (PCNS-LBCL) exhibits the worst prognosis among all extranodal LBCLs, with the enriched genetic mutations in MyD88L265P, CD79B and PIM1. However, the upfront incorporation of targeted therapies remains an unmet need in newly diagnosed (ND) patients. In this study, we found that in the presence of MyD88L265P context, B-cell receptor downstream bruton’s tyrosine kinase (BTK) was significantly overexpressed, which subsequently enhanced the stability of PIM1 oncoprotein. To evaluate the intracranial delivery of BTK inhibitor(BTKi), patient-derived PCNS-LBCL xenografted mice models were treated with highly-selective BTKi orelabrutinib, either monotherapeutically or in combining with methotrexate. Interestingly, orelabrutinib showed excellent efficiency crossing BBB with the functional BTK blockade and promoted PIM1 degradation, providing the molecular basis of incorporating BTKi into PCNS-LBCL therapy. Thus, we determined the efficacy and toxicity of orelabrutinib in combination with high-dose methotrexate in a prospective dose-escalating cohort and real-world practice. The patients being treated with orelabrutinib-included ORMD regimen showed better response and more prolonged survival compared to those with RMD regimen. We further investigated the genetic mutations of PCNS-LBCL across tumor tissue, cerebrospinal fluid (CSF) and plasma. The mutations in CSF circulating tumor DNA (ctDNA) rather than plasma-ctDNA were more consistent to those in tumor tissue, indicating that CSF-ctDNA is a useful tool for monitoring PCNS-LBCL. In summary, our data provided the molecular rationale as well as clinical evidences that incorporation of BTKi into frontline induction therapy is a promising strategy for ND PCNS-LBCL.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13045-026-01812-8.Keywords: Bruton’s tyrosine kinase inhibitor, Orelabrutinib, Primary CNS large B-cell lymphoma, MyD88L265P, PIM1, CSF-circulating tumor DNA. To the editor Primary central nervous system large B-cell lymphoma (PCNS-LBCL) is an extra-nodal large B-cell lymphoma, whose > 80% cases are activated B cell-like (ABC) immunophenotype with the recurrent mutations in MyD88L265P, CD79B and PIM1 [1], [2]. The existence of blood-brain barrier (BBB) claims small molecules with elevated dose to enhance the therapeutic response [3, 4]. Currently high-dose methotrexate (HD-MTX) at 3.5–8 g/m2 has been documented as the backbone for induction therapy [5]. However, the profound response is at the expense of increasing treatment related organ injury [6]. The insufficient pathological knowledge on the above genetic characteristics precludes a valid basis for targeted therapy. Here we investigated the molecular mechanism and clinical response of highly-selective bruton’s tyrosine kinase inhibitor (BTKi) orelabrutinib in PCNS-LBCL (Supplementary Fig. S1 & Supplementary Methods). Firstly, we investigated the molecular function of orelabrutinib on PCNS-LBCL oncogenesis in-vitro and in-vivo. As PIM1 is a kinase relating to lymphoma survival and immune escape [7], we found that it was higher expressed in MYD88L265P-mutated patient derived tissue samples than in MYD88WT ones by immunohistochemical staining (Fig. 1a). The protein levels of PIM1 and BTK were significantly correlated (p = 0.005, Fig. 1b & Supplementary Fig. S2a & Supplementary Table S1). To further investigate the underlying mechanism, we introduced MyD88WT-Flag, MyD88L265P-Flag and BTK-Myc constructions into HEK293T cells. Compared to MyD88WT, MyD88L265P content precipitated more BTK and PIM1 (Fig. 1, c&d), where the PIM1 depletion was significantly delayed in the presence of BTK-Myc plasmid (Fig. 1e), indicating that mutational MyD88L265P can overexpress BTK signaling and subsequently extend PIM1 oncoprotein’s half-life. The co-precipitation and co-localization image also showed a molecular interaction between PIM1 and BTK proteins in the transduced HEK293T cells and primary lymphoma cells, respectively (Fig. 1, f&g), whose PIM1/BTK co-staining rate was at 70.53%±5.38% (Supplementary Fig. S3a). Further investigations revealed that BTK promoted the PIM1-phosphorylation at sites of S36 and S261 (Supplementary Fig. S3, b & c), where S36-phosphorylation protected PIM1 from proteasome-mediated degradation (Supplementary Fig. S3, d & e). After orelabrutinib treatment, a shortened half-life of PIM1 protein with reduced phosphorylation and elevated ubiquitination was observed (Supplementary Fig. S3, f-i). Orelabrutinib suppressed PCNS-LBCL progression by promoting PIM1 degradation. a, The protein level of PIM1 relative to β-actin in MYD88WT PCNS-LBCL tumor tissues and MYD88L265P ones were compared (n = 30) b, Correlation between the protein levels of PIM1 and BTK in PCNS-LBCL tumor tissues (n = 30). Each data point represents the value from an individual patient. Statistical significance was measured by Pearson’s correlation test. c, The interaction between BTK and MYD88 was evaluated by CO-IP assays. The cell extracts of HEK293T cells being transfected with BTK-Myc, MYD88WT-Flag or MYD88L265P-Flag plasmids were immunoprecipitated with IgG or anti-Myc Ab. d, Effect of MYD88WT and MYD88L265P on the BTK/PIM1 interaction. HEK293T cells were transfected with PIM1-HA, BTK-Myc, MYD88WT-Flag or MYD88L265P-Flag plasmids and cell extracts were immunoprecipitated with IgG or anti-Myc Ab. e, Effect of BTK on PIM1 degradation. HEK293T cells were transfected with PIM1-HA, Ctrl-Myc or BTK-Myc plasmids and were incubated with CHX (10 µg/mL) for 0–24 h. The protein levels of PIM1-HA, BTK-Myc and β-actin were detected by western blotting (left panel). Degradation curves of PIM1 level were fitted with 4-parameter logistic model (right panel).f, The interaction between BTK and PIM1 was evaluated by CO-IP assays. HEK293T cells were transfected with PIM1-HA and BTK-Myc plasmids and cell extracts were immunoprecipitated with IgG or anti-Myc Ab. g, Colocalization of PIM1 and BTK was detected by immunostaining (blue: DAPI, red: PIM1, green: BTK. Scale bar, 1 μm). h-i, The primary PCNS-LBCL cells were treated with vehicle or orelabrutinib in vitro. The protein level (h) and mRNA expression (i) of PIM1 were detected by western blotting and qRT-PCR respectively (n = 30). j, Effect of BTK inhibition on PIM1 degradation. HEK293T cells were transfected with PIM1-HA plasmids and treated with or without orelabrutinib (1 µM) for 12 h. The cells were incubated with CHX (10 µg/mL) for additional 0–24 h. The protein levels of PIM1-HA and β-actin were detected by western blotting (left panel). Degradation curves of PIM1 level were fitted with 4-parameter logistic model (right panel). k, Strategy for constructing PDX models from PCNS-LBCL patients and validating efficacy of BTK inhibitor. l, The concentrations of orelabrutinib in PB and BTs were measured after the PDX mice were treated for 3 weeks monotherapeutically (n = 10 per group). m, The percentage of apoptotic lymphoma cells derived from PDX models which were treated with MTX and orelabrutinib, being assessed as Annexin V-positive cells by flow cytometry and CytExpert software (n = 6 per group). n, The MFI of Ki67 in lymphoma cells derived from PDX models which were treated with MTX and orelabrutinib, being assessed by flow cytometry and FlowJo software (n = 6 per group). o, The infiltration of U2932-luc cells in vivo was quantitatively monitored by a biophotonic imaging system in mice treated with MTX and orelabrutinib. The infiltration of U2932-luc cells was monitored on days 7, 14, 21 and 28 after transplantation and representative images of indicated groups were shown (n = 10 per group). Bars, mean ± S.D. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 Then, we investigated the function of BTKi orelabrutinib in-vitro and in-vivo for verifying the translational potential of inhibiting BTK function in PCNS-LBCL (Supplementary Fig. S1). After PIM1-HA-transduced HEK293T cells were cocultured with orelabrutinib for 24 h, PIM1 degradation was significantly advanced on the protein level even though no transcript difference was observed (Fig. 1, h-j, & Supplementary Fig. S2b). Interestingly, the orelabrutinib-induced pro-apoptotic and anti-proliferative effects were significantly attenuated after PIM1 over-expressed (Supplementary Fig. S3, j-l). Patient-derived xenografted (PDX) and cell-line derived xenografted (CDX) models were further applied to determine the clinical relevance by respectively implanting single cell suspensions of fresh PCNS-LBCL specimens and U2932-Luc cells stereotactically into the brains of 6–8-week-old female NOD.Cg-PrkdcscidIl2rgtm1Sug/JicCrl (NOG) mice. The BBB-crossing efficiency of orelabrutinib was at 11.78%±7.82%, providing the possibility of applying orelabrutinib to control intra-cranial lymphoma disease (Fig. 1, k & l). The parameter of cell apoptosis, Ki67 and tumor burden in CDX mice collectively confirmed the function of orelabrutinib either monotherapeutically or in combining with MTX (Fig. 1, m-o, & Supplementary Fig.S4). Then we determined the therapeutic outcome of orelabrutinib on immunocompetent newly-diagnosed (ND) PCNS-LBCL. Considering the high relapsed rate of BTKi monotherapy in PCNS-LBCL [8], a BOIN design employing isotonic regression and toxicity probability interval methods was applied to determine the optimal doses of orelabrutinib and MTX in ND patients [9]. A total of 13 ND-patients were enrolled as DL1 (n = 3), DL2 (n = 5) and DL3 (n = 5), respectively (Supplementary Methods & Fig. S1). The demographic characteristics of the patients are presented in Supplementary Table S2. Only 1 case of grade 3 lung infection, 1 case of grade 3 alanine aminotransferase (ALT) and aspartate aminotransferase (AST) increased were reported at DL2 and DL3, respectively (Supplementary Table S3). The concentrations of orelabrutinib in CSF and plasma were measured in 12/13 patients on C3D1. The mean concentration of orelabrutinib in CSF was 16.4 ng/mL, 36.2 ng/mL and 54.8 ng/mL at DLs 1, 2 and 3. Being corrected for protein binding of 93.5%, the CSF/plasma ratio was 51.5% (DL1), 53.8% (DL2) and 65.5% (DL3), respectively (Supplementary Table S4). All 13 patients were included in the efficacy analysis (Fig. 2a). During the induction therapy, the best responses included 8 complete response (CR), 2 CR unconfirmed (CRu), 1 partial response (PR), 1 stable disease (SD), and 1 progressive disease (PD), with an overall response rate (ORR) of 84.6% (11/13) and CR rate (CRR) of 76.9% (10/13). The CRR at dose levels 1, 2, and 3 was 66.7%, 80.0%, and 100.0%, respectively (Fig. 2a, Supplementary Table S2 & S5). Ten CR/CRu patients were continued with orelabrutinib maintenance. At the date of cutoff, 4/10 patients with maintenance therapy developed PD, 2 of whom were during the maintenance and 2 were after finishing one-year of maintenance. Three of 13 patients died of progression during the median follow-up of 24.97 (range, 11.8–33.13) months, with a median progression-free survival (PFS) of 25.3 months (95% CI: 7.7-NE) and the median overall survival (OS) not reached (Fig. 2a). Therapeutic outcome and mutation profile in newly diagnosed PCNS-LBCL. a, Dynamic clinical response assessment, following-up outcomes, CSF-ctDNA monitoring for 13 ND PCNS-LBCL patients in the prospective dose-escalating cohort. Abbreviation “C1-8” represents the induction therapy cycle1-8 and “MF1-4” represents the maintenance phase 1-4. The “Not Evaluated” represented the subject being terminated because of disease progression or death. The bar indicates the response level at each phases, while the line graph shows the frequency of CSF-ctDNA mutations, where dots represent CSF-ctDNA positivity. b-d, Forest plots of odds ratio (OR) with 95% confidence interval (CI) in multivariable logistic regression for overall response rate (b), hazard ratios (HRs) with 95% CI in multivariable Cox proportional hazards regression for progression-free survival (c) and overall survival (d) after propensity score matching. e, Oncoplots of the top genomic alterations in tumor tissues, CSF-ctDNA and plasma-ctDNA from 47 patients. The diameter of each pie chart was proportional to the mutation frequency across these patients. f, Tumorous driver genes identified by OncodriveCLUST algorithm in tumor tissues, CSF-ctDNA and plasma-ctDNA respectively. g, Cohen’s Kappa Coefficient analysis of top 30 genomic alterations detected in tumor tissue, CSF-ctDNA and plasma-ctDNA. h, The mutation frequencies of PCNS-LBCL-associated key driver geneset were analyzed by McNemar’s test. ***p < 0.001 We further compared the efficacy and toxicity of orelabrutinib-incorporated ORMD regimen with RMD regimen in real-world practice. The HIV-negative patients were enrolled from December 1 st, 2020 to October 1 st, 2023 at single center of Huashan Hospital of Fudan University. Compared with RMD regimen, the patients being treated with ORMD regimen achieved a higher response rate, and prolonged PFS and OS (Supplementary Fig.S5 & Supplementary Table S6). With propensity score matching method, 27 patients in ORMD cohort and 54 patients in RMD cohort were yielded under 1:2 matching ratio (Supplementary Table S7). Front-line incorporation of orelabrutinib (ORMD vs. RMD, OR:4.838, CI:1.358–22.189) was significantly associated with patients’ ORR (Fig. 2b). Multivariate Cox proportional hazards regression analyses demonstrated that the orelabrutinib-incorporative regimen (ORMD vs. RMD, HR:0.421, CI:0.223–0.798), COO subtype, and surgery strategy were significantly associated with PFS (Fig. 2c & Supplementary Fig.S6), while COO subtype, surgery strategy and transplantation consolidation exhibited significant association with OS, all of which requires long-term follow-up (Fig. 2d). The frequencies of AE in each regimen were listed in Supplementary Table S8&9. We additionally analyzed the consistency of detected mutations across various sources from 47 pairs of tumor tissue, plasma circulating tumor DNA (ctDNA) and CSF-ctDNA with a published 475-gene panel (Fig. 2e & Supplementary Table S10) [10]. Notably, both MyD88 and CD79B were identified as the driver genes in tumor tissue and in CSF-ctDNA by OncodriveCLUST algorithm analysis (Fig. 2f) [11]. The cohen’s kappa coefficiency for tissue-plasma, CSF-plasma, and tissue-CSF was − 0.20, −0.29, and 0.48 respectively (Fig. 2g). More substantial similarity of mutation-positive rates were indicated between tumor tissue and CSF-ctDNA (Fig. 2h) [12]. The patients in dose-escalating portion were periodically followed-up with CSF-ctDNA analysis. All the patients with CR/CRu and PR showed negative CSF-ctDNA detection. Among the 10 patients having orelabrutinib maintenance, Pts. I-01, I-06, I-10 and I-13 experienced one or more fluctuations in CSF-ctDNA and were ultimately confirmed of PD. In the rest patients with long-term remission, only 1/6 patient (Pt.I-04) experienced one-spot of unstable CSF-ctDNA detection (Fig. 2a). Piling up with the consistency of mutation genes from different samples, our data indicated that CSF-ctDNA is a surrogate source for monitoring PCNS-LBCL disease status. In summary, our results for the first time provide preliminary clinical evidence and a molecular basis for the application of BTK inhibitors in frontline combination therapy. Aiming at deeper remission and improved long-term outcomes, it offers the rationale to modify a BTKi-based targeted regimen with more efficiency and less toxicity for PCNS-LBCL. Supplementary Information Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Acknowledgements We would like to thank all the patients and their families for participating the study and donating samples. Abbreviations Activated B cell-like Adverse event Alanine aminotransferase Aspartate aminotransferase Blood-brain barrier B-cell receptor Brain tissues Bruton’s tyrosine kinase Bruton’s tyrosine kinase inhibitor Cell-line derived xenograft Confidence interval Cell of origin Complete response Complete response rate Complete response unconfirmed Cerebrospinal fluid Dose level Circulating tumor DNA High-dose methotrexate Hazard ratio Newly-diagnosed NOD.Cg-PrkdcscidIl2rgtm1Sug/JicCrl Odds ratio Overall response rate Overall survival Peripheral blood Primary central nervous system large B-cell lymphoma Progressive disease Patient-derived PCNS-LBCL xenografted Progression-free survival Partial response Stable disease Author contributions T. C., Y. M., Z. S., D. Z. designed, conceived the study, recruited the clinical participant and wrote the manuscript. T. C., Y. Y., B. D., H. Z conducted the experiments, collected the data and wrote the manuscript. J. Z., Q. W., P. Z., T. D., S.W., K.W., Y. Z., D. C., F. Z., Z. W., R. F., F. Z., K. Z., Y. Y., Z. Q., B.Y., D. Z., H. C. helped with participant recruitment and sample collection. H. W., Y. L. helped with evaluation of eMRI image. X. L. helped with the statistical analysis. Funding This study was supported by the National Natural & Science Foundation of China (82330004), National Key R&D Program of China (2022YFA1103500), Science and Technology Innovation Plan of Shanghai Science and Technology Commission (23Y21900300), and Program of Shanghai Municipal Human Resources for Outstanding Leadership (Pujiang Project) to Dr. Tong Chen. Data availability All data are available in the main text or the supplementary materials. Raw sequence data were deposited into the National Genomics Data Center (https://ngdc.cncb.ac.cn/) as of the project number (PRJCA021654). No custom algorithms were used in this study. The R code used to perform analysis of sequencing data can be obtained from the corresponding author upon reasonable request. Declarations Ethics approval and consent to participate The protocols of clinical cohorts were conducted in accordance with the Declaration of Helsinki guidelines and approved by the Institutional Review Board of Huashan Hospital, Fudan University (KY2021-726, KY2023-677), and were registered at Clinicaltrials.gov (NCT05036577, NCT06115824) respectively. Informed consent was obtained individually from the included patients. The animal models were approved by the Animal Experimentation Ethics Committee of Fudan University (2023-HSYY-241-JZS). Consent for publication All authors agree to publication of the article. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Yan Yuan, Bo Dai, Haoshu Zhong and Jiajun Zheng contributed equally to this work. Contributor Information Dongxiao Zhuang, Email: zhuangdx@fudan.edu.cn. Zhifeng Shi, Email: shizhifeng@fudan.edu.cn. Ying Mao, Email: maoying@fudan.edu.cn. Tong Chen, Email: chentong@fudan.edu.cn. References Associated Data Supplementary Materials Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Data Availability Statement All data are available in the main text or the supplementary materials. Raw sequence data were deposited into the National Genomics Data Center (https://ngdc.cncb.ac.cn/) as of the project number (PRJCA021654). No custom algorithms were used in this study. The R code used to perform analysis of sequencing data can be obtained from the corresponding author upon reasonable request.