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

Clinical samples were collected from Harbin Medical University (Harbin, China). All enrolled patients gave written informed consent for their tissues to be used for scientific research. The study was approved by the Institutional Review Board (IRB) of the Harbin Medical University (Harbin, China), consistent with the recommendations of the declaration of Helsinki for biomedical research (Harbin Medical University, Harbin, China) and followed standard institutional protocol for human research. Moreover, the animal study protocol was approved by the Animal Care and User Committee at Harbin Medical University (Harbin, China) (Affidavit of Approval of Animal Use Protocol # Harbin Medical University).

Forty cancer tissues from the patients diagnosed with glioma (with different grades, please refer to Additional file 1: Table S1 for clinicalpathological features) were collected for this study. All tissue samples were pathologically confirmed and immediately snap-frozen in liquid nitrogen until RNA extraction. Written informed consent to the use of the tissue samples for research purposes was obtained from each patient. All procedures were conducted in accordance with the principles outlined in the Declaration of Helsinki, and all applicable international, national and/or institutional guidelines for the care and use of animals were followed. The study protocol was approved by the Ethics Committee of Harbin Medical University (Harbin, China). TMZ-resistant (termed TMZ-R) and TMZ-sensitive (termed TMZ-S) clinical samples were collected and cultured for further analyses in this study. The procedures used to isolate and culture TMZ-R and TMZ-S cells were according to a previously published study [9]. The human microglial cell line, HMC3 used in our study was obtained from American Type Culture Collection (ATCC, Manassas, VA, USA). HMC3 cells were cultured according to the recommendations by ATCC, where EMEM (ATCC® 30–2003™) was used as the base medium and completed by adding 56 mL FBS (ATCC® 30–2020™) to a 500 mL of base EMEM.

Co-culture assays of tumor cells with macrophages TMZ-R and TMZ-S GBM cells were co-cultured with HMC3 microglia (using Boyden Chamber) at a density of 1:10 or 1:5 in a 6-well plate. After 72 h, HMC3 were analyzed for their M1, M2 phenotypes using real-time PCR technique. This system was then used for testing palbociblic’s influence on GAM polarization, palbociblic (0.5 μM) was added into the TMZ-R and HMC3 co-culture system after the seeding and cultured for 72 h. The same experimental conditions were used for co-culturing lncSNHG15-silenced or overexpressed TMZ-R cells with HMC3 cells. Primer sequences for M1 M2 markers can be found in Additional file 2: Table S2. M1 M2 cytokines secreted into the culture medium were determined using M1/M2/MDSC Cytokines, ELISA Kit (Cat# MBS590066, MyBiosources.​com). The procedures were performed according to vendor’s protocols.

The isolation of total RNA and qRT-PCR were carried out as previously described [10]. All experiments were repeated in triplicates. Please refer to Additional file 2: Table S2 for primer sequences used in this study.

For gene silencing experiments, shRNA (Santa Cruz Biotechnology, USA) was constructed with sequences specifically against lncSNHG15. miR-627-5p mimic and inhibitor molecules and negative controls (NC) were purchased from GenePharma (Shanghai, China). For overexpression experiments, pcDNA3.1 (+) vector (GenePharma, Shanghai, China) was obtained to construct a pcDNA3.1-SNHG15 overexpressing plasmid. Please refer to Additional file 2: Table S2 for siRNA sequences.

Expression analyses in this study were all carried out sing standard SDS-PAGE (10–12%). Cellular protein lysates were subsequently transferred to nitrocellulose membranes (Sigma), washed and incubated with primary antibodies in cold overnight, followed by washes and secondary antibodies incubation. Primary antibodies used were all purchased from Cell Signaling Technology unless otherwise specified. Please refer to Additional file 3: Figure S1. Full-size blots of Fig. 2e. Figure S2. Full-size blos of Fig. 4b. Figure S3. Full-size blots of Fig. 6a. 

The colony formation assay was performed to determine the proliferation and evaluation the effects of gene manipulation and/or drug treatment of GBM cells. Briefly, GBM cells (500 cells per well) were incubated in 6-well plates. One week to 10 days, formed colonies were fixed with methanol and stained with 0.1% crystal violet. Colonies were quantified under a light microscope (Olympus Corp.). The experiments were done in triplicates.

Cell viability assay or (MTT assay) was used to determine the viability, drug and/or gene manipulation effects on the GBM cells. GBM cells (5000 cells per well) were seeded in 96-well plates. Control and cells with different treatment were incubated with 20 μL MTT (5 mg/mL) in each well at the indicated time points (cells were collected at 0, 24, and 48 h). DMSO (Sigma) was added (150 μL/well) to each well to dissolve the crystals. OD was read at 490 nm on a microplate reader (Molecular Devices, Sunnyvale, CA, USA). The experiments were carried out in triplicates.

Immune compromised NOD/SCID females (6 weeks old) were supplied by Animal Care facility of the Harbin Medical University (Harbin, China). The animal study protocol was approved by the Animal Care and User Committee at Harbin Medical University (Harbin, China) (Affidavit of Approval of Animal Use Protocol # Harbin Medical University). Patient-derived TMZ-R cells (500,000 cells per injection) were injected subcutaneously. Mice were monitored weekly for tumor development using a standard caliper. Mice were subdivided into 4 groups, vehicle control, TMZ only (4 mg/kg, p.o., 5 times a week), palbociclib only (75 mg/kg, i.p injection, 5 times a week) and combination. At the end of experiments, mice were humanely euthanized or when tumor burden became symptomatic. Tumor samples were harvested for further analyses. Tumor growth were calculated using the following formula where tumor volume = (a × b²)/2, where a is the long axis and b is the short axis of the tumor.

Previously, increased level of lncSNHG15 has been linked to malignant characteristics of cancer cells [11‐13] but its role in GBM has not been fully appreciated. Here, we first compared the level of lncSNHG15 between normal brain tissue and GBM (database GSE4290) [14] and found that lncSNHG15 was significantly higher in the GBM samples (approximately 3.4 fold higher) as compared to that in the normal brain (Fig. 1a). In addition, analysis performed on another database (GSE16581) [15] showed that a higher expression of lncSNHG15 is associated with a high risk for GBM (Fig. 2b). Kaplan-Meier survival curve of the same cohort indicated that patients with a higher lncSNHG15 expression had a shorter survival time as compared to ones with a lower level of lncSNHG15 (Fig. 1c). Moreover, we examined a larger cohort of patients with glioma (GSE108476, N = 524) [16], a similar observation was made where patients with a higher lncSNHG15 was predicted to have a significantly shorter survival time (Fig. 1d). These observations strongly suggest that an increased level of lncSNHG15 plays an important role in GBM tumorigenesis.

Next, we examined the amount of lncSNHG15 from 5 pairs of non-tumors versus GBM grade 4 clinical samples, previously collected from our neurosurgical department. Comparatively, the level of lncSNHG15 was found significantly higher in the TMZ-resistant samples as compared to their TMZ-sensitive counterparts (Fig. 2a). Comparatively, TMZ-resistant GBM cells formed a significantly higher number of colonies as compared to their TMZ-sensitive counterparts (Fig. 2b). Subsequently, we cultured both TMZ resistant and sensitive cells under serum-derived conditions to examine the neurosphere-generating ability. TMZ-resistant cells formed a significantly higher number of neurospheres as compared to the sensitive counterparts (Fig. 2c). Data from flow cytometric analysis supported this observation where TMZ-resistant GBM samples contained a significantly higher percentage of CD133+ cells as compared to their TMZ-sensitive counterparts (Fig. 2d). In support, comparative expression analysis revealed TMZ-resistant GBM cells contained a higher level of stemness markers including Sox2, β-catenin and oncogenic markers, EGFR and CDK6 (Fig. 2e) in both Western blot and q-RT-PCR detection.

We collected and cultured TMZ-resistant (TMZ-R) and TMZ-sensitive (TMZ-S) from patients (please refer to Additional file 1: Table S1 for clinicopathological information) and found that TMZ-R cells significantly induced the M2 markers, CD163 and CD206 in GAMs; TMZ-S cell also promoted M2 polarization in HMC3 cells but with a less extent as compared to TMZ-R cells (Fig. 3a); the level of M1 markers, INF-γ and TNF-α were slightly decreased but not statistically significant. Next, we examined and compared the secreted M2 cytokines in GAMs in the presence of TMZ-R and TMZ-S. In accordance, M2 cytokines, IL-6 and TGF-β secreted by GAMs co-cultured with TMZ-R were higher than naïve GAM counterparts (Fig. 3b). For instance, IL-6 and TGF-β level was approximately 2-fold and 3-fold respectively higher in the TMZ-R co-cultured GAMs (left panel, Fig. 3b). A similar observation was made in the GAMs co-cultured with TMZ-S cells but with lesser extent (right panel, Fig. 3b). On the front of cancer cells, both TMZ-R and TMZ-S cells showed a significant increase in the M2 cytokines mRNA level of IL-10, IL-4, IL-13 and CCL2 as compared to their parental counterparts (Fig. 3c).

To determine the functional connection between lncSNHG15 and GBM tumorigenesis, gene-silencing experiment was performed. First, expression profiles were performed and showed that post lncSNHG15 silencing, stemness markers including Sox2, β-catenin, and oncogenic markers EGFR and CDK6 were significantly reduced (qPCR analysis, left panel; western blots, right panel, Fig. 4a). In support, western blots of TMZ-R cells with lncSNHG15 silenced (KD) or overexpressed (OE) demonstrated that expression of lncSNHG15 was positively associated with the expression of oncogenic markers including Sox2, β-catenin, EGFR and CDK6 (Fig. 4b). Next, SNHG15 knockdowned GBM cells (TMZ-R1 KD cells) showed a significantly lower ability to form colonies and neurospheres as compared to its parental counterpart (Fig. 4c) whereas opposite observations were made in TMZ-R cells overexpressing lncSNHG15 (Fig. 4d). More importantly, lncSNHG15 level was positively correlated with TMZ sensitivity. For instance, silencing lncSNHG15 led to an increased TMZ sensitivity in TMZ-R cells as compared to its parental counterpart (Fig. 4e), as the IC₅₀ value decreased from approximately 310 μM to 225 μM post lncSNHG15 silencing and increased to more than 550 μM in the lncSNHG15-overexpressing cells (Fig. 4e).

CDK6 has been shown to be overexpressed in GBM and according to our data, CDK6 level is higher in the TMZ-R GBM cells. Here, we test the potential of using CDK6 inhibitor, palbociclib for the treatment of TMZ resistant GBM cells and associated effects on the tumor microenvironment. First, we showed that palbociclib treatment significantly suppressed the proliferation of TMZ-R and TMZ-S cells (Fig. 5a). We also observed palbociclib treatment significantly suppressed GBM cells to form colonies (Fig. 5b) and neurospheres (Fig. 5c). Equally important, we examined the effects of lncSNHG15 silencing and CDK6 inhibition on TAM polarization. LncSNHG15-silenced TMZ-R cells showed a significantly decreased ability to generate M2 GAMs and appeared to promote M1 polarization, as demonstrated by the increased M1 markers IFN-γ and TNF-α while decreased M2 markers CD163 and CD206 (left panel, Fig. 5d); the M1 cytokines, IFN-γ and TNF-α were also increased in the GAMs co-cultured with lncSNHG15-silenced TMZ-R cells while M2 cytokines, IL-6 and TGF-β were prominently suppressed (left panel, Fig. 5e). Consistently, palbociclib treatment resulted in similar observations where M1 markers and cytokines were upregulated while M2 suppressed (right panel, Fig. 5d and right panel, 5E).

Western blots of TMZ-R cells treated with palbociclib showed the decreased expression of oncogenic markers including CDK6, β-catenin, EGFR, and stemness marker, Sox2 (Fig. 6a). To further explore the regulatory signaling pathway(s) associated with CDK6, we employed 3 different algorithms (miRmap, PITA and microT) for predicting the microRNAs with the potential to target CDK6. We identified miR-627-5p could potentially bind to the 3′ untranslated region (3UTR) of CDK6 as well as Sox2 and β-catenin (Fig. 6b). We then overexpressed and silenced miR-627 to determine the consequential effect on CDK6 expression. TMZ-R cells transfected with miR-627-5p mimic molecules resulted in the significantly decreased CDK6 expression along with Sox2, β-catenin and lncSNHG15 (Fig. 6c). In addition, exogenous miR-627-5p (mimic molecules) resulted in the increased TMZ sensitivity of TMZ-R cells whereas inhibitor of miR-627-5p further increased TMZ resistance (Fig. 6d). For instance, the IC50 value of TMZ-R cells transfected with miR-627-5p mimic molecule was estimated 200 μM as compared to the NC’s (negative control) 300 μM and greater than 500 μM in the case of TMZ-R cells transfected with miR-627-5p inhibitor (Fig. 6d). In support, we observed a strong negative correlation (Pearson r = − 0.008) between CDK6 mRNA level and miR-627-5p in the clinical TMZ-R samples.

Finally, we examined the in vivo efficacy of palbociclib using TMZ-R patient-derived xenograft mouse model. We found that the tumor growth curve of TMZ-treated group did not significantly differ from the vehicle control group; mice treated with palbociclib alone showed significantly delayed tumor growth and the combination of TMZ and palbociclib yielded the most significantly delayed tumorigenesis in vivo (Fig. 7a). Tumor samples were harvested and compared for their ability in generating neurospheres. In accordance, the combination treatment resulted in in the least neurospheres generated followed by palbociclib alone group, while TMZ and vehicle control showed the highest number of neurospheres generated (Fig. 7b). In addition, comparative qPCR analysis of the tumor samples showed that the combination treatment resulted in the lowest level of CDK6, β-catenin, Sox2 and lncSNHG15 (Fig. 7c). The degree of M2 TAM polarization was also demonstrated in the immunohistochemical analysis of the tumor samples. Consistently, the combination treatment group showed the least M2 TAM staining (CD206) and the lowest expression of CDK6 and β-catenin (Fig. 7d). The IHC staining was quantified using H-score and was represented by the dot plots (Fig. 7d).