In this study, AKT, mTOR, S6 and 4E-BP1 were selected as targets to explore the role of the PI3K/AKT/mTOR signaling pathway in PCNSL. AKT is one of the major carcinogenic effectors in this signaling pathway, which can be indirectly activated by upstream PI3K and activate downstream mTOR. mTOR comprises mTOR complex 1 (mTORC1) and mTORC2. Activated mTORC1 can continue to activate its downstream S6 and 4E-BP1, while activated mTORC2 can directly phosphorylate its upstream target AKT so that AKT activity is fully activated (32, 33). After phosphorylation by mTORC1, 4E-BP1 and S6 are able to promote mRNA transcription and protein translation (34).
Our results suggested that the protein expression rates of p-AKT, p-mTOR, p-S6 and p-4E-BP1 in PCNSL samples were significantly higher than those in reactive hyperplastic lymph nodes. The correlation analysis indirectly confirmed the association of their expression levels. In addition, the relative mRNA expression of MTOR gene was also abnormally increased. Therefore, we speculated that the PI3K/AKT/mTOR signaling pathway is abnormally activated in PCNSL.
However, Marosvari et al (35) proposed a different point of view. The positive immunohistochemical staining rate of p-S6 was 58.1% in PCNSL specimens, while that of p-mTOR was only 25.8%, suggesting that the expression of p-S6 might not be induced by mTOR activation. However, another study (36) showed that Rheb, p-4E-BP1 and p-S6 were overexpressed in more than half of PCNSL cases. The results indirectly suggested aberrant activation of mTORC1. In addition, copy number variations (CNVs) in the related genes were detected in PCNSL samples (37), and the frequency of CNVs in PCNSL samples was significantly higher than that in systemic DLBCL, representing greater genomic instability (11). These results provide evidence for the activation of this pathway in PCNSL.
To further explore the pathological mechanism of the aberrant activation of the PI3K/AKT/mTOR pathway, we detected the loss of the PTEN gene in PCNSL samples. PTEN is an important negative regulator in this pathway, and its loss can result in the occurrence and development of various malignant tumors (38, 39). In our study, PTEN loss occurred in 18.9% of PCNSL cases and was associated with the expression of p-AKT. Currently, there is no relevant study on the detection of PTEN loss by FISH in PCNSL. In systemic DLBCL, some studies found that PTEN loss often occurred in the GCB subtype (40). Furthermore, PI3K inhibitors selectively induced the death of PTEN-loss GCB subtype DLBCL cell lines. Although the majority of PCNSLs are non-GCB subtypes, mutations and CNVs in PTEN can also be found in the second-generation sequencing analysis (37, 41, 42). Todorovic et al (41) found that the mutation rate of PTEN was up to 37% and that PTEN mutation was related to shorter OS in PCNSL patients. Unfortunately, the effect of PTEN loss on the treatment response and prognosis of PCNSL was not further investigated.
We also analyzed the relationship between abnormal activation of the PI3K/AKT/mTOR signaling pathway and the clinical features and prognosis of PCNSL. The results suggested that the positive expression of p-mTOR and p-S6 was significantly correlated with disease recurrence in patients with PCNSL. A study on systemic DLBCL indicated that abnormal activation of this pathway might be correlated with drug resistance to rituximab (43). Takashima et al. (44) revealed that the PI3K/AKT/mTOR signaling pathway might induce resistance to methotrexate. In addition, Grommes et al. (45) found that PI3K inhibitors combined with ibrutinib could overcome the resistance of PCNSL cells to ibrutinib. Therefore, the mechanism by which abnormal activation of this pathway leads to disease relapse in PCNSL might be related to the acquisition of drug resistance in tumor cells. However, due to the lack of PCNSL cell lines, our experiments on the effects of relevant inhibitors on the activity of PCNSL cells were limited.
Moreover, we found that the loss of PTEN was more common in the GCB subtype of PCNSL and that patients with PTEN loss showed a worse OS than those with normal PTEN. These findings were consistent with the conclusion of Todorovic et al. (41). Although the difference in OS between the PTEN loss group and the normal group was not statistically significant (P = 0.072) in our study, further expansion of the sample size or extension of the follow-up time might increase the significance, and the results would be more convincing.
In the multivariate Cox regression analysis, we found that p-mTOR expression was an independent risk factor in terms of PFS in patients with PCNSL. This result also confirmed the previous assumption that high expression of p-mTOR may be the crucial factor affecting the prognosis of PCNSL. Therefore, for newly diagnosed patients with PCNSL, the detection of p-mTOR is extremely important. On the one hand, we can estimate the risk of relapse to guide the administration of stronger combined immunochemotherapy approaches as needed and strengthen the monitoring of the disease. On the other hand, it also provides a basis for individualized treatment strategies.
We previously reviewed the application prospects of inhibitors targeting the PI3K/AKT/mTOR signaling pathway in PCNSL and found that the efficacy of inhibitors as single agents in the treatment of patients with PCNSL might be limited, and it might be difficult to achieve long-term control of the disease with this approach (46). Thus, rational combination strategies should be considered, especially for patients with relapsed and refractory PCNSL and newly diagnosed patients who cannot tolerate intensive chemotherapy.