PD-L1 protein expression is undetectable in normal brain biopsy specimens; therefore, PD-L1-positive cells identified in biopsy specimens from PCNSL patients are likely tumor cells or tumor associated immune cells (23), comprising tumor infiltrating lymphocytes (TILs), myeloid-derived suppressor cells (MDSCs), dendritic cells (DCs), as well as macrophages (24). Moreover, PD-L1 protein is found in TME and the peripheral blood immune system as a soluble or exosomal protein (sPD-L1, exoPD-L1), which engages with PD-1 and exhibits an immunomodulatory role (25). The research identified that PD-L1 and PD-1 upregulation commonly manifested in TME via immunohistochemical analysis of tumor specimens from 20 patients with PCNSL. PD-L1 was predominantly present on neoplastic cells and tumor-associated macrophages (TAMs). PD-1 was predominantly located in TILs and could also usually be present on TAMs, and is rarely expressed on tumor cells (26). The expression of these molecules may be associated with prognosis. summarized in Table 1 . Different definitions about PD-1/PD-L1 positivity across studies resulted in discrepancies in positive expression rates and their prognostic associations. Future research with bigger cohorts is necessary to investigate the positive expression threshold of PD-L1 protein and its correlation with prognosis.

The TME of PCNS-DLBCL predominantly consists of tumor cells, diverse immune cells (including TAMs, microglia, T cells, and a limited number of B cells and DCs, etc), and stromal elements (comprising astrocytes, vascular endothelial cells, and an extracellular matrix rich in hyaluronic acid and proteoglycans). Intricate interactions occur between PD-L1 and tumor microenvironment components, eventually facilitating tumor proliferation and immune evasion. Notably, The TME of HIV⁺/EBV⁺PCNS-DLBCL displays distinct features. In contrast to immunocompetent PCNS-DLBCL, the TME of HIV⁺PCNS-DLBCL has markedly diminished levels of CD4⁺T cells and TAMs, while the EBV⁺ TME is characterized by an abundance of TAMs (32).

Anti-PD-1 mAb inhibit the immunosuppressive influence of PD-L1 on T cells by specifically binding to the PD-1 receptor on T cells surfaces, hence enabling T cells to re-identify and target neoplastic cells ( Figure 1D ). Currently, Camrelizumab, Penpulimab, Sintilimab, and Tislelizumab are sanctioned by the National medical products administration (NMPA) of China for the therapy of R/R classical Hodgkin’s lymphoma (cHL)following at least second-line systemic chemotherapy, while Nivolumab is authorized in the United States Food and drug administration (FDA) for R/R cHL; Cemiplimab lacks any hematologic oncology indication at this time. This section will focus on the efficacy, mechanisms, and potential predictive biomarkers of anti-PD-1 mAb monotherapy and combination therapy.

Many recent research on PD-L1 protein show that it inhibits T cell cytotoxicity, aids neoplasms immune escape through the PD-1/PD-L1 pathway, and modulates tumor growth and proliferation through various mechanisms. anti-PD-L1 monoclonal antibody (anti-PD-L1 mAb) have presented success in several types of lymphoma ( Figure 1D ). A phase 2 trial, single-arm, multicenter of atezolizumab, venetoclax, and obinutuzumab reported a 67.9% ORR (89); The Phase 2 trial of Atezolizumab combined with Obinutuzumab and Lenalidomide for R/R follicular lymphoma (FL) (n=38) demonstrated a CR rate of 71.9% and a 36-month PFS rate of 68.4% (90); Consolidation therapy with atezolizumab significantly extended disease-free survival (DFS) and OS in DLBCL patients who achieved CR following R-CHOP chemotherapy (chemotherapy regimens containing Rituximab, Cyclophosphamide, Doxorubicin, Vincristine, Prednisone) (91, 92). In addition, phase 2 research (n=80) of the anti-PD-L1 mAb Sugemalimab for R/R extra-nodal NK/T cell lymphoma (ENKTL) displayed an 82.5% response rate within 18 months and a 44.9% ORR. Sugemalimab has been approved by the Chinese NMPA and the United States FDA to treat R/R ENKTL (93). To our knowledge, no anti-PD-L1 mAb drugs have been authorized for administering DLBCL or PCNSL. Fortunately, clinical trials investigating anti-PD-L1 mAb for PCNSL are currently in progress (Trial Approval Nos. NCT04899427, NCT04462328, NCT04688151, NCT04462328), and the results will be determined shortly.

In comparison to anti-PD-1 mAb, anti-PD-L1 mAb not only can disrupt the immunosuppressive effects of tumor cells on T cells by obstructing the PD-1/PD-L1 pathway but also partially inhibit the activity of atypical PD-L1, thereby reducing tumor growth and proliferation (94). It has to be mentioned that the anti-PD-L1 mAb Avelumab not only inhibits the PD-1/PD-L1 axis but also activates NK cell-mediated ADCC, resulting in the direct elimination of PD-L1-expressing neoplasm-cells (95). Liu et al. discovered that PD-L1 protein is also present on T-cell surfaces, where it inhibits T cell development and the excretion of critical cytokines, such as IL-2, TNF-α; however, anti-PD-L1 mAb obstruct this mechanism (96). In preclinical lung, kidney, pancreatic, and melanoma models, anti-PD-1 antibodies stimulated the PD-L1-NLRP3 inflammatory signaling in cancer-cell. The activation recruited PMN-MDSCs to the TME, causing acquired resistance in anti-PD-1 mAb patients (97). In contrast, anti-PD-L1 mAb directly target the PD-L1 protein, reducing the likelihood of resistance development. Regarding side effects, anti-PD-1 mAb impede the binding of both PD-1 protein and its other ligand, PD-L2, perhaps leading to an increased frequency of adverse events of grade 3 or higher (OR = 1.58) (98). These have displayed us the importance of anti-PD-L1 mAb in PCNSL therapeutic research and the necessity for more. Nonetheless, considering that the efficacy of anti-PD-1 versus anti-PD-L1 may also be related to different tumor types, additional studies with larger cohorts are still required to analyze this issue comparatively in the future. The subsequent table ( Table 2 ) summarizes clinical trials involving anti-PD-1/PD-L1 mAb treatment protocols.

It is widely believed that, cases respond better to anti-PD-1/PD-L1 monoclonal antibodies when TME cells express PD-L1 (99, 100). Barfi et al. demonstrated that 4T1-breast and CT26-colon cancer cell lines responded favorably to anti-PD-L1 mAb in conjunction with Ibrutinib despite expressing minimal PD-L1 levels (101); Additionally, a patient with CD20^-PCNSL demonstrated negative PD-L1 immunohistochemistry and FISH results, with PD-L1-expressing TAMs <10% and minimal PD-1⁺TILs. After developing resistance to HD-MTX, the patient underwent two cycles of induction therapy with zanubrutinib plus tislelizumab combination therapy and subsequently achieved complete remission, maintaining disease stability for 20 months (102). This proposes that the effectiveness of anti-PD-1 mAb cannot be accurately predicted based on a patient’s pre-treatment PD-L1 levels. Concurrently, it suggests that immunotherapy that Targeting PD-L1 may be advantageous for those who are PD-L1-negative patients as well (101). Some research explored potential predictive markers of anti-PD-1/PD-L1 mAb efficacy, as shown below:

The Blood-Brain Barrier (BBB) makes it arduous to transport many medications to the peri-lesion region, and anti-PD-1 antibodies have a substantial molecular weight and rarely cross it (102, 109), which constrains therapeutic efficacy and constitutes the primary significant challenge in the management of PCNSL. In 2022, induction therapy that integrated intrathecal nivolumab with systemic chemotherapy resulted in full remission in a patient with multifocal, parenchymal recurrent PCNSL, without any associated toxicity or adverse effects, indicating that intracerebroventricular administration alone facilitates the penetration of monoclonal antibodies into the deeper brain parenchyma, thereby enhancing efficacy. Furthermore, intracerebroventricular injections utilize a smaller dosage compared to systemic administration, resulting in reduced toxicities, which is crucial for elderly patients who cannot endure high doses of systemic immunotherapy (110).

Certain researchers propose that a possible mechanism by which peripheral intravenous co-infusion of anti-PD-1/PD-L1 mAb improves brain tumor efficacy relative to standard first-line therapy is the local production of IFN-γ by CD4⁺T cells upon traversing the BBB. IFN-γ enhances the expression of vascular cell adhesion molecule 1 (VCAM-1) and intercellular adhesion molecule 1 (ICAM-1), resulting in the disruption of tight junctions between brain microvascular endothelial cells, which in turn augments BBB permeability and enables the infiltration of more lymphocytes and circulating therapeutic agents into the CNS ( Figure 1E ) (111, 112);Taggart et al. thought that anti-PD-1 mAb or CTLA-4 inhibitors activated and released CD8⁺T cells in melanoma with central and peripheral organ involvement, enhancing intracranial CD8⁺T cell trafficking via upregulation of VCAM-1and ICAM-1, leading to intracranial antitumor effects ( Figure 1E ). However, if the tumor arises in the CNS without peripheral organ involvement, intracranial tumors may evade treatment with ICIs, for example, monoclonal antibodies that target PD-1 (113), indicating that anti-PD-1/PD-L1 mAb may augment BBB permeability to a certain degree. It also reveals that we must be careful to avoid antioxidants such as Semaglutide and other drugs that downregulate VCAM-1 in treating PCNSL.

Additionally, some research has found lymphatic channels that connect the CNS to the deep cervical lymph nodes. T cells in these lymph nodes can infiltrate the cerebrospinal fluid and brain parenchyma. Additionally, both T cells and antigens located in the CNS have the ability to enter the peripheral lymphatic system, where they can activate T cells in lymphatic tissues. This may offer a perspective on circumventing the BBB (111). Recently, a novel biopolymer drug combines three components: anti-PD-1mAb, agents against c-Myc, and AP-2 to aid brain delivery. This drug has demonstrated strong brain penetration and efficacy against A20 mouse brain lymphoma by inhibiting c-Myc and the PD-1/PD-L1 pathway (114); moreover, anti-PD-L1 nanobodies (C7, 5DXW) administered locally by the adoptive cellular transfer (ACT) method demonstrated enhanced tumor infiltration and sustained efficacy in preclinical models (115), representing a successful fusion of nanotechnology and immune-targeted therapy, as well as a novel approach to surmount the BBB.

Patients initially negative for PD-L1 via IHC exhibited elevated PD-L1 expression upon relapse (35), indicating temporal heterogeneity in PD-L1 expression. Factors such as tumor clonal evolution, loss of CD58 surface proteins (116, 117) and the induction of chemotherapeutic agents like Temozolomide (118) may enhance PD-L1 expression, facilitating evasion of immune surveillance and resulting in standard treatment failure. Moreover, recent studies have determined that PD-L1 expression exhibits intracellular spatial heterogeneity. Currently, the predominant focus of scholarly research on PD-L1 in PCNS-DLBCL is based on IHC, which makes it difficult to distinguish membrane-bound PD-L1 on tumor-cell surfaces from atypical PD-L1 in the cytoplasm, nucleus, and other cellular areas. This represents a potential explanation for instances of elevated PD-L1 expression in immunohistochemistry while exhibiting primary resistance to anti-PD-1 mAb therapy. To sum up, the temporal and spatial heterogeneity of PD-L1 expression, coupled with the unique immune microenvironment response of PCNS-DLBCL, renders its targeted therapeutic efficacy in PCNS-DLBCL markedly variable. Future investigations must focus on elucidating the regulatory process governing PD-L1 levels of expression in PCNS-DLBCL and the interactions within the immune microenvironment, alongside multi-homology typing (CS1-4), to optimize PD-L1 protein-targeted therapy and enhance patient outcomes.

Moreover, numerous PD-L1-overexpressing tumors exhibit initial resistance to PD-1/PD-L1 monoclonal antibodies, and PCNS-DLBCL may be included. Future research is required for this cohort of patients to investigate whether primary resistance to PD-1/PD-L1 monoclonal antibodies can be surmounted and tumor eradication accomplished through combinations with other ICIs agents such as CTLA-4, LAG-3, or TIM-3 inhibitors, as well as targeting immunosuppressive cells (e.g., MDSCs, Tregs) and enhancing the expression of tumor-associated antigens, among other approaches. Furthermore, studies focused on NSCLC have demonstrated that a combination therapy involving an anti-PD-L1 mAb and anti-VEGF (e.g., bevacizumab) significantly extends PFS (119); concurrently targeting PD-L1 and other pro-tumor signaling pathways (TGFβ, CD47, VEGF, CTLA4, etc.) with bispecific antibodies (BsAbs) can yield synergistic anti-tumor effects and reduce the incidence of drug resistance (120); this approach may represent a promising therapeutic avenue for patients with PCNS-DLBCL, but the ability of this novel immunotherapy to overcome the blood-brain barrier, along with its efficacy and safety, remains an area requiring further investigation.