Application of new targeted drugs in relapsed/refractory primary central nervous system lymphoma

ABSTRACT The prognosis of patients with relapsed and refractory primary central nervous system diffuse large B cell lymphoma (PCNSL) is poor and there is no standard salvage treatment currently available. Gene expression profiling and next-generation sequencing have provided massive amounts of data, revealing the pathogenic mechanism of PCNSL and identifying potential mechanisms for therapeutic exploration. Here, we review targeted drugs that may change the current treatment model of PCNSL.


Introduction
Primary central nervous system diffuse large B cell lymphoma (PCNSL) is a rare aggressive non-Hodgkin's lymphoma [1]. The induction regimen based on highdose methotrexate (HD-MTX) improves the remission rate and overall survival (OS) of PCNSL patients; however, nearly 50% of patients experience recurrence, and 10%-15% have primary drug resistance [2].
The prognosis of patients with relapsed and refractory (r/r) PCNSL is extremely poor and there is no standard salvage treatment currently. Whenever possible, participation in a suitable clinical trial should be encouraged. If there is no applicable clinical trial, rechallenge with HD-MTX is the most frequently used treatment regimen despite the lack of supporting data of its effectiveness from prospective studies, especially for patients with a long period of remission after initial treatment. In such cases, the overall response rate (ORR) was 85% and the median OS was 41 months [3]. However, defining a meaningful minimum duration of remission to consider re-treatment with HD-MTX remains controversial [4]. For patients who are not suited to receive HD-MTX or high-intensity chemotherapy, new targeted drug monotherapy, or combined with immunochemotherapy, may be an option.
Finding new therapeutic targets and applying new targeted drugs to improve the prognosis of patients with r/r PCNSL by studying gene mutations and signal transduction pathways has become a research hotspot in recent years. New targeted drugs that have shown clinical activity in diffuse large B-cell lymphoma (DLBCL), such as Bruton tyrosine kinase (BTK), mammalian target of rapamycin (mTOR), immune checkpoint, and phosphatidylinositol 3kinase (PI3 K) inhibitors, and immunomodulators, have been used in the treatment of r/r PCNSL. Relevant research progress (Table 1) and ongoing clinical trials ( Table 2) are summarized in the subsequent sections.

Btk inhibitors
BTK is a key molecule that integrates the B-cell antigen receptor (BCR), Toll-like receptor, and chemokine receptor signaling pathways. Activated BTK can act on the downstream NF-κB signaling pathway that plays a key role in the progression of B-cell diseases [5]. Genomics studies have shown that mutations of the MYD88 and CD79B genes are the most common in PCNSL [6], and both often coexist [7,8]. MYD88 is a key linker molecule in the Toll-like receptor signaling pathway and its mutation can activate the NF-κB signaling pathway. The CD79B mutation can disable the negative regulatory signal of the BCR. This, in turn, leads to chronic activation of the NF-κB signaling pathway [5].

First-generation BTK inhibitors
Ibrutinib is a first-generation selective BTK inhibitor that arrests lymphoma cell growth and induces apoptosis by preventing regulation of the BCR signaling pathway, and shows clinical activity in DLBCL [9]. Preclinical studies in mouse models have shown that the maximum concentrations of ibrutinib in plasma and cerebrospinal fluid (CSF) are similar [10]. In addition, the median time for ibrutinib to pass through the blood-brain barrier is only 0.29 h, which provides a theoretical basis for ibrutinib treatment for central nervous system lymphoma (CNSL).

1) Ibrutinib monotherapy
The multi-center, single-arm, prospective phase II study of 'ibrutinib in the treatment of r/r PCNSL' by LYSARC [11] in France is the largest study to date. Fifty-two patients with a median age of 70 years were enrolled in the study. In the first assessment after two months of treatment, 10 (19%), 17 (33%), and 5 (10%) cases showed complete response (CR), partial response (PR), and stable disease, respectively. The median follow-up, median progression-free survival (PFS), and OS times were 25.7, 4.8, and 19.2 months, respectively. Thirteen patients received ibrutinib treatment for more than one year. Mutations in the BCR signaling pathway were detected in 18 patients. Among seven patients with wild-type (wt)CARD11, wtCD79B, and wtMYD88, two cases showed CR and two showed PR, suggesting that patients without mutations in the BCR signaling pathway may also be included in clinical trials using BTK inhibitors to treat PCNSL.
Grommes et al. [12] reported a non-randomized, single-center, dose-escalation study aimed at determining the maximum tolerated dose for ibrutinib monotherapy. Thirteen patients with r/r PCNSL and seven with secondary CNS lymphoma (SCNSL) were enrolled. Pharmacokinetic data obtained from 18 patients showed that the CSF concentration of ibrutinib was dose-dependent. The ORR of 13 cases of PCNSL was 77% (five cases of CR, five cases of PR, two cases of stable disease, and one case not evaluated). At 16 months (median follow-up), the median PFS and OS times were 4.6 and 15 months, respectively. In this study, a comprehensive analysis of the mutational landscape in PCNSL showed only one missense mutation of CARD11 that was detected in one patient completely resistant to ibrutinib. CARD11 mutations have been shown to promote BTK-independent activation of NF-κB and are found clinically in ibrutinib-resistant patients with systemic DLBCL and mantle cell lymphoma [9,13]. None of the six patients with MYD88 and CD79B mutations reached CR. The relationship between the MYD88/ CD79B mutation and the efficacy of ibrutinib remains controversial because it is difficult to establish a clear correlation between the PCNSL mutation profile and efficacy of ibrutinib in a limited number of patients.
Chamoun et al. [14] reported a retrospective study of 13 patients with r/r PCNSL and one patient with primary testicular lymphoma. All cases underwent heavy prior treatment, and 12 patients had refractory diseases. Eleven patients were treated with ibrutinib for less than five months and only two used ibrutinib when the paper was published.  (32) Abbreviations: RCT = randomized clinical trial; HD-MTX = high-dose methotrexate; R2 = lenalidomide + rituximab; POM = pomalidomide; DEX = dexamethasone; r/r = relapsed/refractory; PCNSL = primary central nervous system lymphoma; PVRL = primary vitreoretinal lymphoma; SCNSL = secondary central nervous system lymphoma; ORR = overall response rate; mPFS = median progression free survival; mOS = median overall survival. 2) Ibrutinib combined with other regimens In a phase Ib study by Grommes et al. [15], a combined regimen of ibrutinib (560 or 840 mg/d), HD-MTX (3.5 g/m 2 q2w), and rituximab (500 mg/m 2 q2w) was used for the treatment of r/r CNSL (PCNSL, n = 9; SCNSL, n = 6). All patients received an HD-MTX-based regimen as initial treatment; seven patients were at first relapse during enrollment and three patients were refractory to HD-MTX. The regimen was well-tolerated, no dose-limiting toxicity was found, and only three cases of grade 4 adverse events were observed (two of them were hematopoietic system toxicities). Eight out of nine PCNSL cases (89%) achieved PR or above. At 19.7 months (median follow-up), the median PFS reached 9.2 months and the median OS was not reached, with the OS at 1 year being 71.1%. The median duration of remission for all patients was 12.8 months, and that for the six patients who underwent maintenance treatment was 14.3 months. The response rates in this study appear to be similar to those of HD-MTX salvage treatment for recurrent PCNSL [3]. It is worth mentioning that, in the retrospective study, the median time from the end of induction treatment to the first relapse was more than 2 years, [15] indicating that most of the included patients were sensitive to MTX. Moreover, the PFS time in this study was longer than that in previous studies using ibrutinib as a single agent [12]. This suggests that ibrutinib combined with HD-MTX is a potentially effective regimen for the treatment of r/r PCNSL.
Lionakis et al. [7] combined ibrutinib and the DA-TEDDI-R (temozolomide, etoposide, liposomal adriamycin, dexamethasone, rituximab, and intrathecal arabinos glycoside) regimen to treat PCNSL. Eighteen patients were enrolled with a median age of 66 years. Five of these patients were newly diagnosed and 13 had r/r PCNSL. Among the 13 r/r PCNSL cases, six progressed or died. At 15.5 months (median follow-up), the median PFS for the 13 r/r patients was 15.3 months, the median OS was not reached, and the 1-year OS was 51.3%. Despite the remarkably durable treatment response, this regimen resulted in a high rate of Aspergillus infections (n = 7, 39%). Eight of the 18 (44%) patients died and three of these deaths were reported as treatment-related. Thus, future studies with ibrutinib combination protocols should be based on pre-proven chemotherapy regimens.
Fourteen heavily pretreated patients with r/r PCNSL from the French LOC database treated with R 2 I (rituximab, lenalidomide, and ibrutinib) were analyzed retrospectively [16]. The R 2 I regimen resulted in a high response rate with manageable toxicity and allowed three patients to proceed to consolidation. These results support the use of R 2 I for r/r PCNSL patients failing conventional chemotherapies, and a prospective trial is ongoing (NCT03703167).

Second-generation BTK inhibitor
Tirabrutinib is a second-generation, highly selective oral BTK inhibitor. It was approved in Japan in March 2020 for the treatment of r/r PCNSL. Narita et al. [17] reported the results of a phase I/II clinical trial of tilarutinib for the treatment of r/r PCNSL (n = 44). The dosage of tilarutinib was 320 or 480 mg/d, and no dose-limiting toxicity was observed. A grade 5 adverse event (Pneumocystis pneumonia) occurred in only one patient. At 9.1 months (median follow-up), the median PFS was 2.9 months and median OS was not reached. Patients with CARD11, MYD88, and CD79B mutations had similar ORRs to the corresponding wild-type patients.
To date, clinical evidence of ibrutinib effectiveness in PCNSL is limited to single-arm studies. More than 50% of r/r PCNSL patients are expected to have a significant chance of regression of lymphoma by ibrutinib monotherapy. This seems higher than that for patients with r/r DLBCL outside the central nervous system (CNS) (25%) [9], suggesting that both genetic factors and the lymphoma microenvironment may play important roles in the response of lymphoma to BTK inhibitors. Moreover, remission does not persist. The median PFS was approximately five months. These results call for further assessment of the benefits of ibrutinib in combination with chemo/immunotherapies in r/r PCNSL.

Pi3 K/mTOR inhibitors
The PI3 K/AKT/mTOR signaling pathway regulates cell growth and proliferation by integrating signals generated using growth factors, hormones, nutrition, and energy metabolism [18]. Preclinical data [19] show that PI3 K/mTOR inhibitors have anti-tumor activity in lymphoma cells. Clinical studies [20,21] have confirmed that PI3 K/mTOR inhibitors alone or combined with chemotherapy can be used for r/r mantle cell lymphoma. In a previous study, high concentrations of temsirolimus were found in brain tumor specimens with brain/blood ratios of 1.43 and 0.84 for temsirolimus and its metabolite, sirolimus, respectively [22].
Korfel et al. [23] initiated a multicenter phase II clinical trial to evaluate the effectiveness and safety of temsirolimus (mTOR inhibitor) as a single agent in the treatment of patients with r/r PCNSL. Thirty-seven patients were enrolled (median age of 70 years) who received temsirolimus (25 mg, n = 6; 75 mg, n = 29) once a week. CR and PR were achieved by eight (three unproven) and 12 patients, respectively. The ORR, median PFS, and OS were 54%, 2.1, and 3.7 months, respectively. Five patients (13%) died because of treatment-related complications. Fourteen blood/CSF pairs from nine patients were collected and drug was detected in the CSF of only one patient from the 75 mg cohort. There is a clear discrepancy between the activity of temsirolimus and its undetectable concentration in the CSF, possibly because the drug concentration in CSF differs from that in the brain parenchyma.
Grommes et al. [24] reported the results of a phase II clinical trial at the 2016 ESMO meeting. This trial recruited patients with r/r PCNSL and SCNSL to treat with single-agent buparlisib (pan-PI3 K inhibitor). The trial was closed prematurely because of the limited clinical response that was observed; the concentration of buparlisib in the CNS was lower than the meaningful IC50. Thus, it might not have single-agent activity in CNSL.
Though almost 50% of patients respond to temsirolimus, the response is short-lived and the median PFS is only 2.1 months. Most importantly, treatment-related toxicity is substantial. However, preclinical studies have shown that the combination of BTK and PI3 K/ mTOR inhibitors has a synergistic anti-lymphoma effect, especially in CD79B mutant cells [12], suggesting that the feasibility of combining BTK and PI3 K/mTOR inhibitors in the treatment of PCNSL may be further explored.

Second-generation immunomodulators
Studies have shown that lenalidomide interferes with the survival and proliferation of lymphoma cells through various mechanisms of action [25]. Singlearm clinical trials have confirmed that lenalidomide shows clinical activity as a monotherapy for r/r non-Hodgkin's lymphoma [26,27].
The REVRI study, conducted by the French LOC network, is the largest prospective multicenter study of lenalidomide-containing regimens in r/r PCNSL and a cornerstone for further exploration of this therapy [28]. Lenalidomide combined with rituximab was used to treat 55 patients with r/r PCNSL. The median age of the patients was 69 years (46-86 years) and most (49%) had an ECOG score of 2-4. The best ORR was 67% (18/45 CR and 12/45 PR); however, it dropped to 36% by the end of the induction therapy. At 19.2 months (median follow-up), the median PFS and OS were 7.8 and 17.7 months, respectively.
Rubenstein et al. [29] reported the results of a phase I clinical trial that enrolled 14 patients with r/r CNSL (PCNSL, n = 6; SCNSL, n = 8) with a median age of 66 years and administered lenalidomide (10, 15, or 20 mg/d). Among these patients, nine achieved remission (ORR: 64%), of which six had remission durations greater than nine months, and four had remission durations greater than 18 months. Lenalidomide was detected in the CSF samples of 10/13 patients. The CSF concentration of lenalidomide was highest (7.37 ng/mL) at the 15 mg dose level. The investigators also reported that 12 CNSL patients received maintenance treatment with the rituximab regimen following salvage treatment. Five patients remained in remission for more than two years. However, the value of lenalidomide as a first-line maintenance therapy for DLBCL is uncertain [30,31]. Specifically, lenalidomide maintenance therapy does not seem to reduce the risk of CNS recurrence in patients with DLBCL [32]. The difference in results may be partly due to the different characteristics of the patient populations enrolled.
Houllier et al. retrospectively analyzed the outcomes of six patients with r/r PCNSL receiving lenalidomide monotherapy. The median age of the cohort was 73.5 years (64-78 years). All patients had received second-line and above treatment in the past. Of the six patients, two achieved CR and one continued with remission after 24 months [33]. The results of this study suggest that lenalidomide exhibits moderate activity in highly pretreated r/r PCNSL and is welltolerated by elderly patients.

Third-generation immunomodulators
Preclinical data show that the penetration rate of pomalidomide (a third-generation immunomodulator) in the CNS is nearly 39%. Pomalidomide has significant anti-tumor activity and the ability to regulate the tumor microenvironment in CNS lymphoma [34]. Its effect on the tumor microenvironment may help overcome the immune evasion that characterizes PCNSL [8]. Tun et al. reported the results of a phase I study of pomalidomide in the treatment of r/r PCNSL and ocular lymphoma [35]. At 16.5 months (median follow-up), the median PFS and remission times were 5.3 and 9 months, respectively. The ORR of 25 patients was 48%, of which eight and four were CR and PR, respectively. The median duration of remission was 4.7 months. CNS pharmacokinetic analysis was performed on a patient in the 3 mg group. The CSF/ plasma ratio of pomalidomide was comparable to lenalidomide with 19% and 17% on days 1 and 14 of the first cycle, respectively.
There is insufficient clinical evidence to support the use of single agent immunomodulators in r/r PCNSL. The existing results come from phase I/II, small sample size, and single-arm studies with short remission periods. However, lenalidomide or pomalidomide are still candidates for combining with other agents for r/r PCNSL therapy, based on their novel mechanisms of activity, blood-brain barrier penetration efficiencies, and modest toxicity profiles. Two ongoing studies have aimed to combine R2 regimen with BTK inhibitors for the treatment of r/r PCNSL (NCT03703167, NCT04938297).

Immune checkpoint inhibitors
Genomic analysis of PCNSL patients shows that the frequency of copy number gains in chromosome 9p24.1, which contains the programmed cell death ligand-1 (PD-L1) and PD-L2 genes, is high [8]. Overexpression of PD-L1/PD-L2 in PCNSL represents a mechanism of immune escape. Therefore, PD-1 blockade may be a potential therapeutic strategy.
Existing reports using antibodies against PD-1/PD-L1 to treat PCNSL are retrospective analyses of small samples [36][37][38][39]. In one study, five patients with r/r PCNSL that received nivolumab monotherapy all responded (4 CR and 1 PR) [36]. All patients were alive at 17 months (median follow-up), and three patients remained radiographically and clinically progression-free at 13-17 months following the initiation of nivolumab therapy. Ambady et al. [39] reported the results of combining a PD-1 monoclonal antibody with rituximab in the treatment of r/r CNSL (PCNSL, n = 3; SCNSL, n = 3). Three patients achieved CR, of which two remained in CR after seven months of treatment. One patient progressed after stopping the drug; however, this patient again achieved CR after restarting treatment.
Studies have found that, in Hodgkin's lymphoma, the increased copy number of 9p24.1 on R-S cells, and the high expression of PD-L1, are associated with a good prognosis from nivolumab treatment [40]. However, it is unclear whether this principle also applies to PCNSL. The observed clinical responses in the small case series mentioned above, and the reported 9p24.1 alteration in PCNSL, supports the prospective evaluation of PD-1 blockade in this disease. A prospective multicenter single-arm clinical trial (NCT04609046) conducted by the National Cancer Institute is ongoing. These results require further investigation to be conclusive.

Conclusions
The advent of new targeted drugs has provided novel treatment options for elderly or frail patients, and improved the outcomes of patients with r/r PCNSL to a certain extent. To date, there is a lack of evidence from multicenter randomized controlled studies on r/ r PCNSL. The combination of new targeted drugs with different mechanisms of action, or combining them with chemotherapy regimens to further improve the efficacy, extend the duration of remission, and reduce treatment-related side effects, are issues that need to be further explored. Clinical trial enrollment should be pursued whenever possible to help ensure that standards of care continue to improve as these and other agents become more readily available. For r/r PCNSL, the heterogeneous patient population and subgroups selected for special research should be further refined when designing clinical studies. Chimeric antigen receptor T-cell (CAR-T) immunotherapy is a new approach that has recently attracted attention and shown impressive results in lymphomas. CAR-T clinical trials for r/r PCNSL patients are currently underway (NCT04134117, NCT04608487). The combination of new targeted drugs with CAR-T is also worth exploring in the future.

Disclosure statement
No potential conflict of interest was reported by the author (s).

Funding
This work was supported by the Nantong Science and Technology Project under grant number HS2019003.