Modulating lncRNA SNHG15/CDK6/miR-627 circuit by palbociclib, overcomes temozolomide resistance and reduces M2-polarization of glioma associated microglia in glioblastoma multiforme

Glioblastoma (GBM) is one of the most lethal brain tumors. The average survival is 12–18 months and the 5-year overall survival is approximately 5% [1]. Even with aggressive interventions including surgery, or combination of radiotherapy and chemotherapies, the prognosis for the patients remains poor. One of the major underlying contributing factors to GBM’s malignancy is the resistance to both traditional and targeted therapies. Due to the inherent heterogeneous nature of GBM, the signaling pathways involved in the acquired drug resistance represents a complicated task to decipher. The role of tumor microenvironment adds another layer of complexity. Glioma-associated microglial cells (GAMs) are functionally similar to that of tumor-associated macrophages in the peripheral system and interact with GBM cells intimately via intracellular communications [2, 3]. GAMs have been found to secrete a spectrum of cytokines and signaling molecules to promote tumor proliferation, anti-apoptosis and angiogenesis/metastasis [4, 5]. Thus, a better understanding of the intracellular molecular communications between GBM and GAMs will provide foundation for therapeutic development.

Emerging evidence indicates the importance of abundantly transcribed non-coding genes termed long non-coding RNAs (lncRNAs) in virtually every aspect of cell biology including glioma tumorigenesis. Recently, a lncRNA, termed SNHG15 has gained much attention. SNHG15 was shown to be elevated and associated with tumorigenic functions including proliferation and metastasis in a variety of cancer types including breast, lung, and liver [6–8]. Notably, lncRNA SNHG15 level was positively linked to the histological grade, tumor node metastasis stage (TNM) and the poor overall survival in hepatocellular carcinoma [8]. However, the potential role of SNHG15 in GBM tumorigenesis has not been fully investigated. Initially, we searched public databases and identified that an elevated level of lncSNHG15 in GBM cells as compared to normal brain tissues, and this elevation of lncSNHG15 was associated with a significantly higher risk of developing GBM and a shorter survival time in the patients.

In this study, we employed in vitro and in vivo assays to demonstrate the tumorigenic roles of lncSNHG15. First, we found that a significantly higher level of lncSNHG15 in TMZ-resistant clinical GBM samples and was associated with GBM malignant properties. Increased lncSNHG15 level was associated with an increased expression in markers of oncogenesis such as EGFR, CDK6 and stemness including Sox2 and β-catenin. More importantly, TMZ-resistant (TMZ-R) GBM cells were more capable of promoting M2-polarization of glioma associated microglia (M2-GAMs) than TMZ-sensitive (TMZ-S) counterparts. Down-regulation of lncSNHG15 resulted in reduced tumorigenesis, self-renewal and increased TMZ sensitivity and the reverse was true when increased. Interestingly, CDK6 inhibitor, palbociclib treatment suppressed GBM tumorigenesis as well as the generation of M2 GAMs. Palbociclib’s anti-GBM effects were associated with a reduced level of lncSNHG15 and increased level of tumor suppressor miR-627. Mechanistically, miR-627 could target not only CDK6 but also Sox2 and β-catenin. Finally, we used patient-derived xenograft model to demonstrate that palbociclib treatment alone significantly reduced GBM tumorigenesis and with a greater extent when combined with TMZ.

In summary, we provided preclinical insights into the functional roles of lncSNHG15/CDK6/miR-627 regulatory circuit in the development of GBM and polarization of GAMs. More importantly, the feasibility of employing palbociclib for treating TMZ-resistant GBM was examined and supported by both in vitro and in vivo models.

Glioblastoma multiforme (GBM) is one of the most malignant cancer types and currently there is no effective regiments for combating this deadly disease. Despite advancement in the development of chemotherapeutic agents including targeted therapeutic agents, the overall survival time often does not extend beyond 2 years post diagnosis. Recent experimental and clinical insights indicate the tumor microenvironment (TME) of GBM also play instrumental roles in promoting tumorigenesis and the development of drug resistance. Glioma-associated microglia (GAM) represents one of the key cell types that have been shown to contribute significantly towards GBM tumorigenesis and progression [17, 18]. In this study, we examined the potential signaling networks involved in promoting TMZ resistance and the feasibility of using CDK6 inhibitor, palbociclib as a candidate agent for treatment.

First, we found that an emerging tumorigenic lncRNA, SNHG15 is elevated in the GBM clinical samples and associated with a significantly shorter survival time (Fig. 1). We also provide clinical validation using our own clinical samples demonstrating a significantly higher lncSNH15 level in the TMZ-R cells than TMZ-S cells, implicating lncSNHG15’s potential involvement in drug resistance (Fig. 2). We then showed that TMZ-R cells with a higher level lncSNHG15 possessed a significantly higher self-renewal capacity, coincident with the higher expression of Sox2, β-catenin and EGFR, all of which have been attributed to GBM’s ability to fend off therapeutics [19–21]. In addition, our observation was supported by a recent study where lncSNHG15 level was found elevated in the endothelial cells and involved in promoting angiogenesis [22]. In addition, an increased level of lncSNHG15 was linked to the increased metastatic potential of non-small cell lung cancer [23] and liver metastasis in colon cancer [24]. Results from this study and others support the tumorigenic role of lncSNHG15.

Notably, we added another layer of complexity in terms of lncSNHG15’s role in GBM tumorigenesis where the increased lncSNH15 level was associated to the propensity to promote M2 polarization in the GAM. This association was supported by the observations where lncSNHG15 silencing significantly reduced M2 polarization while towards M1 polarization (Fig. 3). Moreover, we showed a positive correlation between the level of lncSNHG15 and oncogenic/stemness markers such as EGFR, CDK6, SOX-2 and β-catenin (Fig. 4). This observation added another possible target for drug development for treating GBM since these markers not only have been documented to be elevated in GBM but also contribute towards malignancy. For instance, β-catenin signaling was linked to the generation of CD133+ glioma stem cells [25] as well as linked to the increased oncogenic activity of EGFR [26]; more importantly, CDK6 has been shown to be an emerging target for drug development for GBM where CDK6 expression/activity is elevated in GBM cells and inhibitor of CDK6 could significantly suppressed GBM in vitro and in vivo [27]. Based on these studies and our observations, we have further supported the tumorigenic role of lncSNHG15 in GBM.

To further explore the connection between CDK6 and lncSNHG15, we showed that treating TMZ-R cells with CDK6 inhibitor, palbociclib also led to the decreased level of lncSNHG15; palbociclib-treated TMZ-R cells showed a significantly decreased ability to generate M2 GAM (Fig.5). Consistently, palbociblic treatment (suppression of CDK6) led to decreased expression of EGFR, Sox2 and β-catenin, similarly seen in the lncSNHG15-silenced TMZ-R cells. More importantly, lncSNHG15 knockdown and palbociclib treatment both led to an increased sensitivity towards TMZ. Mechanistically, we identified that a tumor suppressor, miR-627-5p targeted CDK6 at its 3’UTR and palbociclib treatment led to an increased level of this tumor suppressor (Fig. 6). Interestingly, in a previous report, miR-627 was found to target a histone demethylase (JMJD1A) and show anti-cancer function in colon cancer [28]. Since JMJD1A is an epigenetic regulator, it was logical that palbociclib-induced miR-627 level could suppress tumorigenesis via a myriad of targets. Finally, our PDX study showed that the addition of palbociblic could overcome TMZ-resistance in vivo (Fig. 7).

Palbociclib (trade name: Ibrancea, approved by FDA in February 2015 to treat HR+/HER2- breast cancer), amebaciclib and ribociclib (trade name: Kisqali, approved by FDA in March 2017 to treat HR+/HER2- breast cancer) are recently developed CDK4/6 inhibitors currently undergoing clinical testing as potential chemotherapeutics for the treatment of primary or secondary brain tumors [29]. Palbociclib was shown to promote survival in a genetic mouse model of brain stem glioma, but its unbound brain-to-plasma partition coefficient was only 5 min after intravenous administration of 1 mg/kg [30]. There is preliminary evidence and preclinical data that ribociclib, abemaciclib and palbociclib cross the brain–blood barrier with data suggesting that albemaciclib may be more efficient in crossing the blood–brain barrier compared to palbociclb. Several clinical trials are ongoing to evaluate CDK4/6 inhibitors in patients with brain metastases.

In a Phase I dose escalation safety study, palbociclib was examined in 3 weeks on 1 week off schedule in 41 patients with advanced malignancies. Therecommended phase II dose, at which neutropenia was the sole significant toxicity was determined to be 125 mg once daily. As dose limiting toxicities, a reversible neutropenia was identified. Grade 3 hematological toxicities included neutropenia (12%) and anemia (7%). Non-hematological toxicity was mild, including fatigue, nausea and diarrhea. There was a clear signal for clinical activity of the drug: Thirty-seven patients were evaluable for tumor response; 10 (27%) had stable disease for ≥4 cycles of whom six derived prolonged benefit (≥10 cycles) [31]. Beyond breast cancer, new indications for palbociclib are currently being evaluated in other malignancies such as sarcoma, pancreatic cancer, head & neck cancer, NSCLC, brain tumors, or even hematological malignancies.