Introduction
Glioblastomas (GBMs) develop resistance to therapy, primarily due to their cellular heterogeneity, their local invasive capacities, and challenges of drug delivery over the blood-brain-barrier (BBB) [
1‐
3]. Therefore, GBM cells effectively evade current therapeutic strategies leading to a poor prognosis [
4].
GBMs are known to have tumor cells with stem cell like properties (GSCs), that show an ability to self-renew. GSCs will to a large extent recapitulate the original tumor when orthotopically xenotransplanted in animals [
5]. Therefore, specifically targeting GSCs, represents a potential way to improve treatment and patient outcomes.
Despite advances in RNA single-cell sequencing, there is still a limited mechanistic understanding of how individual GBM cells invade the brain. This may be explained by difficulties in isolating single invasive tumor cells from the complex brain microenvironment [
6]. Recently, it has been shown that tumors display direct paracrine and electrochemical communication with neurons. Such interactions may regulate oncogenesis, growth, invasion and metastatic spread, treatment resistance, and more [
7]. Progress within this field may therefore represent an important new research avenue related to GBM therapy. Yet, despite these significant advances [
8,
9], it remains unclear how tumor cellular/host interactions regulate tumor cell proliferation and invasion, ultimately determining patient outcome.
Here, we used a differentiated brain organoid model confronted with GSCs and isolated individual invasive tumor cells for molecular analysis [
10]. The brain organoid model contains abundant astrocytes, myelinated neurons, microglia, oligodendrocytes and other stromal cells within a complex neuropil. The brain organoids secrete neurotransmitters, and metabolism-related receptors, which closely simulate the brain microenvironment [
11]. Co-cultured with patient-derived human GSC spheres, we isolated individual invasive GSCs and performed ultra-low input RNA sequencing. Differentially expressed genes between invasive and non-invasive cells were imported into the Connectivity MAP database to screen for potential drugs targeting their expression. We selected the FDA-approved drug SKF83566 for further studies. SKF83566 is an inhibitor of dopamine type 1 receptors (DRD1 and DRD5) which has the ability to cross the BBB. We show that SKF83566 suppress GBM progression and invasion via the DRD1-c-Myc-UHRF1 axis in vitro as well as in an orthotopic tumor model in mice. Our results add a new facet to current knowledge on tumor neural interactions, highlighting DRD1 as a target within the GBM invasive compartment.
Materials and methods
Ethics statement
The research involving human participants was reviewed and approved by the Scientific Research Ethics Committee of Qilu Hospital, Shandong University (approval number: 2,015,063). Individuals provided written informed consent for their participation and the use of relevant tissues for research purposes. The experiments conformed to the principles set out in the WMA Declaration of Helsinki and the Department of Health and Human Services Belmont Report. Animal procedures were approved by the Scientific Research Ethics Committee of Qilu Hospital, Shandong University (approval number DWLL-2021-087; Shandong, China) and the Institutional Animal Care and Use Committee (IACUC) of Shandong University.
Cell culture
Patient-derived human glioma stem cells (GSCs), GG16, P3, BG5 and BG7, were established from GBM surgical specimens at the Department of Biomedicine, University of Bergen (Bergen, Norway; P3, BG5, BG7) and University Medical center Gröningen, The Netherlands; GG16). Short tandem repeat (STR) analysis was performed to confirm the identity of GSCs derived from the original patient material as described in previous studies by us [
12]. Cells from human tumors were validated as GSCs through neurosphere formation assays and detection of the expression of GSC markers such as SOX2 and c-Myc. Cells were cultured in Neurobasal™ medium (Gibco/Thermo Fisher Scientific; Waltham, MA, USA) supplemented with 2% B-27 supplement (Invitrogen; Carlsbad, CA, USA), 10 ng/mL bFGF (PeproTech; Rocky Hill, NJ, USA) and 20 ng/mL EGF (PeproTech). Accutase (ThermoFisher Scientific) was used to digest tumor spheroids to expand human GSCs. Serum-cultured GBM cell lines, LN18, LN229, U251 and A172, were purchased from the American Type Culture Collection (Manassas, VA, USA) and cultured in Dulbecco’s modified Eagle’s medium (DMEM; Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (FBS; Clark Bioscience; Richmond, VA, USA). HEK293T cells were cultured in the same media as for the GBM cell lines. Normal human astrocytes (NHAs) were obtained from Lonza (Walkersville, MD, USA) and cultured in an astrocyte growth medium supplemented with rhEGF, insulin, ascorbic acid, GA-1000, L-glutamine, and 5% FBS.
Clinical specimens
Archived paraffin-embedded glioma tissues (WHO grade IV) were collected from patients (n = 32) who underwent surgery in the Department of Neurosurgery, Qilu Hospital of Shandong University. Normal brain tissue samples (n = 8) were taken from trauma patients who underwent partial resection of normal brain as decompression treatment for severe head injuries.
Cell viability
Cell viability was determined indirectly by measuring the intracellular levels of ATP using the CellTiter-Glo Luminescent Cell Viability Assay (Promega; Madison, WI, USA). Luminescence was measured with a Mithras LB 940 multimode microplate reader (Berthold; Bad Wildbad, Germany).
Flow cytometry
For cell cycle analysis, cells were harvested by Accutase, rinsed three times in PBS, fixed in 75% ethanol. The cells were then dehydrated and RNAse treated before staining with propidium iodide (concentration: 50 µg PI/ml in PBS; BD Biosciences; San Jose, CA, USA) at room temperature for 15 min. To detect apoptosis, cells were rinsed with PBS, resuspended in 500 µL Annexin V binding buffer (1X concentrate; BD Biosciences), and incubated with Annexin V-FITC and PI (BD Biosciences) for 15 min at room temperature. Cell cycle distribution and apoptosis were analyzed on a C6 flow cytometer (BD Biosciences; San Jose, CA, USA). Data were post-processed by FlowJo-V10 software (ACEA Biosciences; San Diego, CA, USA).
Brain organoid co-cultures
The preparation and culture of brain organoids have been described previously [
10]. In brief, rat fetal brains at the 18th day of gestation were dissected out aseptically, cut into small pieces, and rinsed three times with PBS. Tissue pieces were digested with Accutase for 20 min to obtain single cells that were seeded into agar-coated culture flasks and incubated for 21 days to obtain mature organoids.
For co-culture, the brain organoids were confronted with GSC spheres of equal size as the brain organoids. Images of the co-cultures were obtained, after 24 and 72 h, were acquired by confocal microscopy (Leica, TCS SP8). The co-cultures were treated with SKF83566 (ApexBio, Cat. B6797; Houston, Texas, USA). We selected 100µM as a treatment concentration, which is less than the IC50 of all cell lines, GSCs and NHA (IC50: NHA 986µM; U251 163.2µM; LN18 119.4µM; LN229 138µM; A172 204.7µM; BG5 370.1µM; BG7 381.7µM; P3 201.4µM). Co-cultures were also established using DRD1, or UHRF1 knockdown GSC spheres.
GSCs (200 cells/100 µL/well) were seeded into 96-well plates (Corning Inc., Corning; NY, USA) and cultured for 14 days. An inverted phase contrast microscope (Nikon; Tokyo, Japan) was used to count and acquire GSC sphere images.
21-day mature rat brain organoids were co-cultured with GFP-labeled tumor GSC spheres in round well low-attachment 96-well plates for 24 h, 48 or 72 h. The co-cultures were cut into two parts (a brain organoid part containing invasive tumor cells and one part representing the main tumor mass) under a Nikon WD70 microscope, (C-DSD230, Nikon). The two parts were dissociated using the ACS Neural Tissue Dissociation Kit(p) (#130-092-628, Miltenyi; Bergisch Gladbach, Germany). GFP-labeled tumor cells derived from the two parts were sorted into sterile 96-well plates using a cell sorter (BD FACSAriaTM IIu Cell Sorter, San Jose, CA, USA), and 30 cells were placed into each well. GFP-labeled cells obtained from brain organoid part were considered to be invasive tumor cells, whereas those obtained from tumor sphere lysates were considered to be non-invasive or potentially invasive tumor cells. An additional movie file shows this in more detail [see Additional file
1]. Four replicates were analyzed for each sample. The SMART-Seq v4 Ultra Low Input RNA Kit (TaKaRa; Shiga, Japan) was used to prepare libraries for ultra-low input RNA sequencing (ui-RNA seq). Paired-end sequencing (2 × 75 bp) was performed using a NextSeq 500 (Illumina; San Diego, CA, USA) with an average of 9.7 Mio reads per sample. After demultiplexing with bcl2fastq 2.20.0.422, raw reads were trimmed for adapters and read quality with Trimmomatic [
13] 0.36 (LEADING:15, TRAILING:15, SLIDINGWINDOW:4:15, and MINLEN:36). Reads were 2-pass mapped on the genome of the 1000 Genomes Project using STAR [
14]. Reads were first mapped against an index created from the genome sequence and gene annotation (Gencode GRCh38.p7). All detected splice junctions were then used as the guide for the second mapping pass. The following parameter sequence was used: --alignIntronMax 500,000 --alignMatesGapMax --outSAMprimaryFlag OneBestScore --outFilterMultimapNmax 100 --outFilterMismatchNmax 2 --alignSJstitchMismatchNmax 5 − 1 5 5. Read counts were determined with featureCounts (Rsubread) [
15]. Genes with 0 counts for all samples and rRNA were discarded. DESeq2 [
16] was used to find differentially expressed genes, which had multiple-testing adjusted p-values < 0.05.
RNA sequencing for SKF83566-treated GSC cells
RNA-Seq libraries were prepared using the Illumina TruSeq RNA sample preparation kit (Illumina) and sequenced through paired-end (150 base paired-end reads) sequencing performed on the Illumina NovaSeq 6000 platform. Raw data were then quality filtered to generate “clean reads” for further analysis. The clean reads were aligned to the human genome reference (hg19) with STAR software and the reference-based assembly of transcripts was conducted with HISAT2. Picard was used to compare the results and to remove redundancy, and Sentieon software was used to detect single-nucleotide variations and InDels. All previously identified single-nucleotide variations and InDels were determined with the dbSNP database. Gene expression values were expressed as reads per kilobase of exon per million fragments mapped with kallisto software. To identify true differentially expressed genes, the false discovery rate was used for the rectification of the P values. The differentially expressed genes (
P value ≤ 0.05, |Log2FC|≥1) were subjected to enrichment analyses for gene ontology and Kyoto Encyclopedia of Genes and Genomes for pathways. The genome-wide transcriptome analysis was performed on a set of 3 separate experiments (SKF83566 and vehicle control treatment groups). Expression2Kinases [
17] was used to perform transcription factor enrichment analysis, and data visualization was accomplished with R software. Gene set enrichment analysis (GSEA [
18]) was performed to find differential phenotypes between SKF83566 and vehicle control treatment groups.
Extreme limiting dilution assay
P3, BG5 and BG7 human GSCs (treated with SKF83566 (50/100µM) or DMSO; si-DRD1 or si-NC; si-UHRF1 or si-NC; DRD1-OE or DRD1-NC; c-Myc-OE or c-Myc-NC) were seeded at a density of 10, 20, 30, 40, 50 and 100 cells per well in uncoated 96-well plates. Serum-free stem-cell medium was refreshed every week. Spheres were left to grow for 14 days before manual scoring of the 60 inner wells. Extreme limiting dilution analysis was performed with publicly available software at Extreme Limiting Dilution Analysis web [
19].
3D tumor spheroid invasion assay
Tumor sphere invasions were performed in vitro by embedding the spheres in matrigel according to a standardized protocol [
20]. Images were obtained after 96 h of culture. The spheroid area was considered as the starting point for quantification.
GBM brain organoid co-culture invasion ex vivo system
GFP-transfected human GSCs were cultured to form spheroids and then co-cultured with mature brain organoids for 24 h, then treated with SKF83566 100 µM and then co-cultured for 72 h. Images of fluorescent human GSCs were captured by confocal microscopy (Leica TCS SP8; Wetzlar, Germany) and a Z-stack images were generated. The spheroid area was considered as the starting point for quantification. The invasive ratio was determined by dividing the area of tumor cells present outside the initial tumor sphere with its initial area. ImageJ (National Institutes of Health; Bethesda, United States) software was used to analyze the GSCs-related invasion.
Western blotting
Cells were lysed in radioimmunoprecipitation assay buffer (RIPA; P0013C, Beyotime; Haimen, China) supplemented with a protein inhibitor cocktail (20–201, Millipore Sigma; Burlington, MA, USA) for 30 min on ice, and protein concentrations were determined with the BCA assay according to the manufacturer’s instructions (Beyotime). Protein lysates (20 µg) were separated with 10% SDS-polyacrylamide gel electrophoresis (150 V, 60–90 min) and transferred to polyvinylidene difluoride (PVDF) membranes (GVW2932A, 0.22 μm, Millipore Sigma) through wet transfer (220 mA, 100 min). The membranes were then blocked in Tris-buffered saline containing 0.1% Tween-20 with 5% skim milk for 1 h and incubated with human antibody overnight at 4 °C. After rinsing, the blots were incubated with appropriate secondary antibodies (dilution 1:3000; goat anti-rabbit: A0208, goat anti-mouse: A0216, Beyotime; Haimen, Jiangsu, China). Bands were visualized with a chemiluminescent HRP kit (WBKLS0500, Millipore Sigma). Chemiluminescence signals were imaged and quantitated using the ChemiDoc XRS+ (Bio-Rad; Hercules, CA, USA). The human antibodies used were the following: caspase-7 (dilution 1:1000, 12,827, Cell Signaling Technology, Danvers, MA, USA), cyclin dependent kinase 4 (CDK4, (dilution 1:1000, 4060, Cell Signaling Technology), PCNA (dilution 1:1000, 13,110, Cell Signaling Technology), β-actin (dilution 1:1000, 3700, Cell Signaling Technology), N-cadherin (dilution 1:1000, 22018-1-AP, ProteinTech; Rosemont, IL, USA), MMP-2 (dilution 1:1000, 10373-2-AP, ProteinTech), DRD1 (dilution 1:1000, WC3224679, Invitrogen/ThermoFisher Scientific), GAPDH (dilution 1:5000, 5174, Cell Signaling Technology), DRD5 (dilution 1:1000, sc-376,088 Santa Cruz Biotechnology; Dallas, TX, USA), c-Myc (dilution 1:1000, 18,583, Cell Signaling Technology), UHRF1 (dilution 1:1000, sc-373,750, Santa Cruz Biotechnology), and SOX2 (dilution 1:1000, ab79351, Abcam; Cambridge, MA, USA). Horseradish peroxidase-labeled goat anti-rabbit secondary antibodies were provided by Zhongshan Golden Bridge Bio-technology (Beijing, China). All experiments were repeated three times.
RNA isolation and quantitative RT-PCR
Total RNA was extracted from cells with TRIzol reagent (Invitrogen/ThermoFisher Scientific) and reverse-transcribed with the Rever Tra Ace qPCR RT Kit (Toyobo; Osaka, Japan). cDNA was amplified with SYBR Green on the Roche Light Cycler 480 for quantification (Indianapolis, IN, USA). The relative expression levels of mRNA were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Sequences of the primers used were the following: c-Myc: F-CCCCTACCCTCTCAACGACA and R-CTTCTTGTTCCTCCTCAGAGTCG. UHRF1: F-AACTACAACCCCGACAACCC and R-ACCACGTTGGCGTAGAGTTC.
SiRNA transfections and lentiviral transduction
SiRNAs to knockdown of DRD1, c-Myc and UHRF1 (GenePharma; Shanghai, China) was performed on P3, BG5 and BG7 for 48 h using Lipofectamine 2000 (11668-027, Invitrogen/ThermoFisher Scientific) according to the manufacturer’s protocol. Lentiviral vectors expressing human shRNA targeting DRD1 (shDRD1, GenePharma) or scrambled-control (shNC, GenePharma) were used to generate stable cell clones expressing shDRD1 or a nonspecific shRNA as control. Clones were selected in 1 mg/mL of puromycin (Selleckchem; Houston, TX, USA). Western blot analysis was used to evaluate siRNA and shRNA knockdown efficacy. The siRNA sequences used were as follows: DRD1#1, 5’-GGGUCCUUCUGUAACAUCU-3’; DRD1#2, 5’-CCAUCAUUUAUGCCUUUAA-3’; DRD1#3, 5’-CCAGCCCUAUCAGUCAUAU-3’; c-Myc#1, 5’-GAGGAUAUCUGGAAGAAAU-3’; c-Myc#2, 5’-CAAGGUAGUUAUCCUUAAA-3’;c-Myc#3, 5’-GACGAGAACAGUUGAAACA-3’;UHRF1#1, 5’-GGACGAAGUCUUCAAGAUU-3’, UHRF1#2, 5’-GCAAUGUCAAGGGUGGCAA-3’; and UHRF1#3, 5’-CCAGUUGUUCCUGAGUAAA-3’;si-NC: 5’-UUUUCCGAACGUGUCACGUTT-3’. The UHRF1 overexpression plasmid was purchased from WZ Biosciences Inc (Jinan, Shandong, China) The transcript variant was pEnter, into which the coding region of UHRF1 (NM_001048201) was inserted in the MCS area.
Establishment of intracranial GBM xenografts and SKF83566 treatment
Four-week-old male thymus-free nude mice (Foxn1nu mut/mut; SLAC Laboratory Animal Center; Shanghai, China) were housed under specific pathogen-free conditions at 24 °C with a 12-hour diurnal cycle. For orthotopic transplantation, mice were randomly grouped (
n = 5 per group) and injected intracranially with 5 × 10
5 P3 or BG5 human GSCs resuspended in 10 µL PBS as previously described [
21]. On day 10 or day 30 after cell implantation, mice were administered drug or vehicle control by intraperitoneal injection (SKF83566 at a dose of 20 mg/kg/day or DMSO). The animals were sacrificed following apparent neurological signs and weight loss. Growth of P3 and BG5-expressing luciferase xenografts was monitored with an IVIS Spectrum bioluminescence imaging system (Perkin-Elmer; Waltham, MA, USA). Brains were collected and fixed in 4% formaldehyde for hematoxylin and eosin (H&E) staining and IHC analysis.
Chromatin immunoprecipitation (ChIP) assays
The EZ-ChIP Immunoprecipitation Kit (Cell Signaling Technology) was used to perform ChIP assays. In brief, human GSCs were cross-linked with 1% formaldehyde for 10 min and quenched with 0.125 M glycine. Cells were collected by centrifugation, washed, resuspended, lysed and sonicated. Chromatin extracts were pre-cleared with agarose beads from the ChIP kit and incubated overnight with c-Myc antibody or normal rabbit IgG as a control. After washing, elution and reverse cross-linking, qPCR was performed on the DNA obtained. The sequences of the primers used for c-Myc and UHRF1 binding sites in the UHRF1 promoter were as follows: Primer 1, F-GGACTTAAGAGTTCAGGGGGTC and R-CCTAGTGGGCCACGGCT; and Primer 2, F-CTGTCCAGGCTGACCAAGG and R-AAAGTATCGGCTGGTGGCTG. The following antibodies were used: anti-c-Myc (1:100, #18,583, Cell Signaling Technology) and normal rabbit IgG (1:100, Cell Signaling Technology).
Dual-luciferase reporter gene assays
The dual luciferase assay was performed according to the manufacturer’s protocol (Promega). Briefly, cells were seeded into 96-well plates, incubated for 24 h, and transfected with the following plasmids as indicated: c-Myc-NC + pGL3-UHRF1-WT + TK, c-Myc-OE + pGL3-UHRF1-WT + TK, and/or c-Myc-OE + pGL3-UHRF1-mut + TK expression vectors. At 48 h post-transfection, cells were lysed with passive lysis buffer. The dual luciferase assay was then performed on lysates using a Mithras LB 940 microplate reader (Berthold).
Immunohistochemistry
Tumor samples were fixed in 4% formaldehyde overnight at 4 °C, paraffin embedded, sectioned (5 μm), and mounted on slides. Heat-induced epitope retrieval was performed by heating samples immersed in 1 mmol/L citric acid buffer, pH 7.2, in a microwave. Samples were blocked in 200 µL of blocking buffer (950 µL of TBST/5% BSA + 50 µL of serum from the species of the secondary antibody) for 30 min at room temperature and incubated with the primary antibody at 4 °C overnight. Sections were rinsed with PBS, and detection was performed through standard procedures with horse‐radish peroxidase‐linked secondary antibody (goat anti‐rabbit or anti‐mouse) and diaminobenzidine as a substrate (CTS002‐NOV, HRP‐DAB IHC Detection Kit, Novus Biologicals; Littleton, CO, USA). Slides were counterstained with Mayer’s hematoxylin and evaluated under light microscopy (MC21‐N, Sony ics412 CCD, Sony; Tokyo, Japan). Primary antibodies used were the following: Ki67, (1:500, Abcam, #ab15580); c-Myc, (1:200, #ab32072, Abcam); UHRF1, (1:200, #sc-373,750, Santa Cruz Biotechnology); MMP-2(1:200, #10373-2-AP, ProteinTech) and N-cadherin (1:200, #22018-1-AP, ProteinTech). Positively stained cells were counted under a 40x light microscope in 10 randomly selected non-overlapping fields of view, with 3 different tissue sections in each group, followed by intergroup comparisons. All experiments were repeated three times.
Statistical analysis
All statistical analyses and experimental graphs were performed with GraphPad Prism 8 software (La Jolla, CA, USA). Paired or unpaired Student’s t-tests were performed for two-group comparisons and one-way analysis of variance (ANOVA) for multi-group comparisons. Two-way ANOVA was performed for cell viability multi-group comparisons. Three independent experiments were performed, and results were expressed as the mean ± the standard error of the mean (SEM). P-values determined from different comparisons are indicated as follows: *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. P-values < 0.05 were considered to be statistically significant.
Discussion
Dopamine receptors have been linked to multiple physiological and pathological functions [
29,
30]. More recently, studies have shown that dopamine receptors are important in glioma initiation and progression and therefore represent potential therapeutic targets [
31‐
33]. Our team has previously shown that thioridazine, a DRD antagonist, inhibits late autophagy by impairing the fusion between autophagosomes and lysosomes. Moreover, the combination of thioridazine and TMZ significantly reduced the growth of brain tumors in tumor-bearing mice [
34]. In this study, using a 3D co-culture model, we performed an in-depth analysis of ultra-low input RNA sequencing data to compare invasive and non-invasive tumor cells. We specifically focused on finding key differentially expressed genes between invasive and non-invasive patient derived GSCs. By integrating the gene expression data with the CMap database, we identified SKF83566, known for its inhibitory effect on dopamine receptors, as a potential therapeutic agent.
We further show that SKF83566 inhibits tumor cell invasion and malignant progression by specifically targeting DRD1 but not DRD5. Additionally, our results indicate a significantly higher expression of DRD1 compared to DRD5 in GSCs and also that DRD1 regulates c-Myc activity. Furthermore, we show that c-Myc actively binds to the promoter region of UHRF1and stimulates its transcription.
Microenvironmental interactions are particularly complex in human brain tumors, as the central nervous system contains a variety of organ-specific molecules, growth factors, and cell types. Among these molecules, neurotransmitters, especially the monoamine class, which includes dopamine, have been shown to have a strong influence on cell proliferation and differentiation, particularly in neural stem and progenitor cells [
35]. Dopamine, upon release, binds to its G-protein-coupled receptors (GPCRs), which are divided into two families, D1-like and D2-like. The D1-like family includes DRD1 and DRD5, while the D2-like family includes dopamine receptor 2 (DRD2), dopamine receptor 3 (DRD3), and dopamine receptor 4 (DRD4) [
36,
37]. In general, D1-like receptors interact with Gs alpha subunits that increase intracellular, second messenger, cyclic AMP (cAMP) levels [
38]. This leads, among others, to an activation of Protein Kinase A (PKA) that phosphorylates many target proteins. In contrast, D2-like receptors stimulate the Gi alpha subunit, which leads to a decrease in intracellular cAMP levels [
39]. Thus, D1 and D2-like receptors should exhibit opposite biological activities. In the brain, D1-like receptors are known to regulate neuronal growth and development by modulating synaptic plasticity that influences cognitive processes. In contrast, D2-like receptors are known to regulate neurotransmitter release, motor control, motivation, and reward [
40,
41]. In DRD expressing cancers, activation of D1 receptors should in theory lead to an inhibition of tumor growth, whereas activation of D2 receptors should lead to an increased intracellular signaling and an increased tumor progression. However, this view is oversimplified since it is at present acknowledged that the activation of dopamine receptor pathways is dynamic and exhibits significant variability across different cancer types [
42]. Considering the inherent heterogeneous nature of many cancers, characterized by diverse GPCR cell signaling events [
43], it is conceivable that inhibiting dopamine receptors could paradoxically result in both a suppression and promotion of tumor growth. In this context, it was recently shown that DRD1 agonist treatment led to an inhibition of auto-lysosomal degradation in GBM cells and that this process was calcium overload dependent and related to an inhibition of the mammalian target of rapamycin (mTOR) [
44]. These results contradict to a certain extent our current findings. It should, however, be emphasized that the tumor models used were different. In the study above, the authors based their conclusions on experiments performed on standardized GBM cell lines (U87 and U251). We show in Fig.
3B that DRD1 expression is considerably lower in three standardized GBM cell lines (A172, LN229 and U251) compared to the GSC cell lines used in the present work (P3, BG5 and BG7). It is therefore conceivable that different cell signalling pathways are activated/inhibited between standardized GBM cell lines and GSCs.
Our results suggest that SKF83566, an inhibitor of D1-like receptors, suppress GBM growth, and that this effect is regulated through c-Myc, a key molecule involved in cancer progression. As mentioned above, it is important to note that GBM growth is regulated by a complex interplay between various molecular pathways and cellular interactions. Therefore, the role of DRD1 needs to be understood in a broader context of these pathways and cellular interactions. However, we here show that an inhibitor of D1-like receptors, not an agonist, suppress GBM through inhibition of c-Myc.
Unlike previous studies, we show that DRD1, in contrast to DRD5, is highly expressed in GSCs and promotes GBM invasion and stem cell self-renewal. DRD1 upregulates the mRNA and protein levels of the c-Myc oncogene. We show that c-Myc binds directly to the promoter region of UHRF1 and upregulates its transcription, which induces the expression of other genes associated with invasion. Notably, the expression of c-Myc and UHRF1 in both normal brain tissue and NHA was much lower than in GBM. This difference in expression of c-Myc and UHRF1 in tumor and normal cells might suggest that SKF83566 has a therapeutic window by selectively inhibiting GBM invasion and cell stemness. A previous study has shown that UHRF1 also has functions in the establishment and maintenance of DNA methylation patterns in mammalian cells [
45]. Here, we show that high expression of UHRF1 promotes GSC proliferation and invasion into the surrounding brain environment. Thus, a potential UHRF1-regulated methylation may modulate molecular switches critical for GBM invasion, thereby promoting the GBM malignant behavior.
While emphasizing the need for further studies into the specific mechanisms by which DRD1 regulates c-Myc, our research shows some interesting findings. For instance, in pancreatic cancer, LINC00261 has been shown to suppress c-Myc transcription by obstructing p300/CBP (CREB binding protein, cAMP response element-binding protein) recruitment to the c-Myc promoter region and diminishing H3K27Ac levels through direct interaction with p300/CBP’s brominated structural domain [
46]. Our studies, employing a cAMP ELISA kit, revealed that DRD1 knockdown causes a reduction in cAMP levels. Furthermore, analysis of our RNA sequencing data indicates that SKF83566 increases CREB3 regulatory factor (CREBRF) expression in the human GSCs. Notably, CREBRF has been observed to inhibit CREB3 activity [
47], suggesting that DRD1’s role extends beyond merely elevating cAMP levels since it seems to modulate CREB3 expression by restraining CREBRF. Future research should explore the possibility that DRD1 augments CBP’s expression and binding through the cAMP second messenger in GBM, thereby enhancing c-Myc transcription.
In conclusion, we show that SKF83566, an inhibitor of DRD1, suppresses GBM development by targeting the DRD1-c-Myc-UHRF1 axis. These results present new insight into the role of the dopamine receptor family in the development of GBM and its treatment.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.