Article Images (0)
×![]()
featured_play_listSave article to folder
X
Create new folder
Full Text
PDF
Table of Contents
Abstract
1. Introduction
2. Materials And Methods
2.1. Ethical Approval And Glioma Tissue Information
2.2. Immunohistochemistry And Immunofluorescent Staining Of Human Glioma Sections
2.3. Human Glioblastoma (GBM) Cell Line Cultivation
2.4. Rat Microglial Cell (MG) Preparation And Human GBM/MG Coculture Cultivation
2.5. RNA Preparation, CDNA Synthesis, And QRT-PCR
2.6. Quantitative Analysis Of Cytokine Secretion Dynamics
2.7. Immunocytochemistry
2.8. In Vitro Migration Assay
2.9. Statistical Analysis
3. Results
4. Discussion
Acknowledgments
Appendix A. Supplementary Data
Abstract
Gliomas are the most common parenchymal tumors of the central nervous system (CNS). With regard to their still unclear etiology, several recent studies have provided evidence of a new category of pathogenic elements called human endogenous retroviruses (HERVs) which seem to contribute to the evolution and progression of many neurological diseases such as amyotrophic lateral sclerosis (ALS), schizophrenia, chronic inflammatory polyneuropathy (CIDP) and, particularly, multiple sclerosis (MS). In these diseases, HERVs exert effects on cellular processes such as inflammation, proliferation, and migration. In previous studies, we demonstrated that in MS, the human endogenous retrovirus type-W envelope protein (HERV-W ENV) interferes with lesion repair through the activation of microglia (MG), the innate myeloid immune cells of the CNS. Here, we now show that HERV-W ENV is also present in the microglial cells (MG) of the tumor microenvironment (TME) in gliomas. It modulates the behavior of glioblastoma (GBM) cell lines in GBM/MG cocultures by altering their gene expression, secreted cytokines, morphology, proliferation, and migration properties and could thereby contribute to key tumor
HERV-W envelope protein is present in microglial cells of the human glioma tumor microenvironment and differentially modulates neoplastic cell behavior Laura Reiche a,1, Benedikt Plaack a,1, Maike Lehmkuhl a, Vivien Weyers a, Joel Gruchot a, Daniel Picard b, Hervé Perron d,e, Marc Remke b, Christiane Knobbe-Thomsen c,f, Guido Reifenberger c, Patrick Küry a,g, David Kremer h,* a Department of Neurology, Medical Faculty and University Hospital Düsseldorf, Heinrich Heine University Düsseldorf, Germany b Department of Pediatric Oncology, Hematology and Clinical Immunology, Medical Faculty and University Hospital Düsseldorf, Heinrich Heine University Düsseldorf, Germany c Institute of Neuropathology, Medical Faculty and University Hospital Düsseldorf, Heinrich Heine University Düsseldorf, Germany d R&D Division, GeNeuro Innovation, Lyon, France e GeNeuro, Plan-les-Ouates, Switzerland f ViraTherapeutics GmbH, Rum, Austria g Department of Neurology, Inselspital, Bern University Hospital and University of Bern, Bern, Switzerland h Department of Neurology and Neurorehabilitation, Hospital Zum Heiligen Geist, Academic Teaching Hospital of the Heinrich-Heine-University Düsseldorf, Kempen, Germany A R T I C L E I N F O Keywords: Glioma Glioblastoma Human endogenous retrovirus type-W envelope protein Microglia Tumor microenvironment A B S T R A C T Gliomas are the most common parenchymal tumors of the central nervous system (CNS). With regard to their still unclear etiology, several recent studies have provided evidence of a new category of pathogenic elements called human endogenous retroviruses (HERVs) which seem to contribute to the evolution and progression of many neurological diseases such as amyotrophic lateral sclerosis (ALS), schizophrenia, chronic inflammatory polyneuropathy (CIDP) and, particularly, multiple sclerosis (MS). In these diseases, HERVs exert effects on cellular processes such as inflammation, proliferation, and migration. In previous studies, we demonstrated that in MS, the human endogenous retrovirus type-W envelope protein (HERV-W ENV) interferes with lesion repair through the activation of microglia (MG), the innate myeloid immune cells of the CNS. Here, we now show that HERV-W ENV is also present in the microglial cells (MG) of the tumor microenvironment (TME) in gliomas. It modulates the behavior of glioblastoma (GBM) cell lines in GBM/MG cocultures by altering their gene expression, secreted cytokines, morphology, proliferation, and migration properties and could thereby contribute to key tumor properties.
1. Introduction
Gliomas are primary central nervous system (CNS) malignancies that account for approximately one third of all primary brain tumors [1], with glioblastoma, isocitrate dehydrogenase (IDH)-wildtype being the most common and most aggressive type of glioma that is associated with poor outcome despite aggressive multimodal therapy [2]. In this regard, several recent studies have pointed to a new class of pathogenic elements called human endogenous retroviruses (HERVs) which impact the evolution and progression of a variety of neurological diseases. These effects are based on the modulation of cellular processes such as inflammation, proliferation, migration, and other parameters of cellular homeostasis. HERVs are ancient retroviral elements accounting for up to 8 % of the human genome which originate from mammalian germ-line retroviral infections millions of years ago. Usually epigenetically silenced or functionally irrelevant they can be re-activated by environmental factors such as infections with contemporary viruses which trigger their expression [3]. The resulting production of viral particles * Corresponding author. Department of Neurology and Neurorehabilitation Hospital Zum Heiligen Geist, Von-Broichhausen-Allee 1, 47906, Kempen, Germany. E-mail address: David.kremer@artemed.de (D. Kremer). 1 Note: Laura Reiche and Benedikt Plaack contributed equally. Contents lists available at ScienceDirect Microbes and Infection journal homepage: www.elsevier.com/locate/micinf https://doi.org/10.1016/j.micinf.2024.105460 Received 31 March 2024; Received in revised form 28 October 2024; Accepted 19 November 2024 Microbes and Infection xxx (xxxx) xxx Available online 20 November 2024 1286-4579/© 2024 The Authors. Published by Elsevier Masson SAS on behalf of Institut Pasteur. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ). Please cite this article as: Laura Reiche et al., Microbes and Infection, https://doi.org/10.1016/j.micinf.2024.105460 and/or proteins, especially from members of the HERV-W and HERV-K family, is correlated with the onset and progression of diseases such as amyotrophic lateral sclerosis (ALS), type 1 diabetes, schizophrenia, chronic inflammatory polyneuropathy (CIDP), and particularly MS [4]. In previous studies, it was demonstrated that in MS the human endogenous retrovirus type-W envelope protein (HERV-W ENV) promotes a pro-inflammatory anti-regenerative phenotype in microglia, the innate myeloid immune cells of the central nervous system (CNS), and drives these cells to damage axons and contributes to disease progression [5,6]. Regarding a potential contribution to pathogenic processes in malignancies, the exact roles of HERVs are still unclear. However, a large body of evidence points to HERV expression in different types of cancers such as for instance, melanoma [7], prostate cancer [8], and lymphoma [9]. In this context, viruses such as Kaposi’s sarcoma-associated herpesvirus (KSHV), Epstein Barr virus (EBV), and hepatitis B virus (HBV) have been hypothesized to lead to a transactivation of HERVs [10]. However, only a few studies have so far investigated the role of HERVs in brain tumors such as gliomas. Earlier studies demonstrated full-length HERV-K expression in both patient tissue samples and glioma cell lines, while the absence of detectable splice products initially suggested that HERV-K does not contribute to glioma malignancy [11]. However, more recent work by Shah and colleagues demonstrated that, in fact, HERV-K (subtype HML-2) contributes to the glioblastoma stem cell niche [12] and that its differential expression reduces survival in glioblastoma patients probably via epigenetic regulation [13]. In contrast, there have been no studies so far that investigated the potential role of HERV-W in gliomas. Of note, it is known that glioma homeostasis strongly relies on an interaction between tumor cells and the tumor microenvironment (TME; [14,15]). The TME is composed of the extracellular matrix (ECM) surrounding the tumor cells, growth-promoting and -inhibiting factors, nutrients, chemokines, gases such as oxygen (O2) and nitric oxide (NO), as well as other non-malignant cell types in the immediate vicinity, including inflammatory cells like macrophages and activated microglia as well as reactive astrocytes and residual neurons [14,15]. Brain-resident microglia and monocyte-derived macrophages entering brain tumors via a leaky blood-brain-barrier (BBB) account for up to 50 % of GBM-associated cells and provide an environment facilitating cancer progression [16]. Glioma cells, on the other hand, secrete a multitude of chemokines that stimulate microglia, macrophages, and, in a paracrine manner, other glioma cells to produce factors that increase tumor growth and invasiveness [17]. Shedding light on the potential impact of HERV-W ENV on the pathophysiological processes underlying the glioma/TME interaction may, therefore, facilitate the identification of new therapeutic targets in glioma treatment. This would be particularly important for glioblastoma, the most malignant subtype of glioma with the greatest unmet therapeutic need. It is for this reason that we deliberately decided to perform our experiments in cell lines derived from this tumor.
2. Materials and methods
2.1. Ethical approval and glioma tissue information
The storage of primary human tumor samples in the CNS tumor tissue bank at the Institute of Neuropathology and the use of these samples and associated patient data for research purposes have been approved by the Ethics Committee of the Medical Faculty of Heinrich Heine University (study number: #3005). In addition, a project-specific ethics vote was obtained from the Ethics Committee of the Medical Faculty of Heinrich Heine University for the studies reported here (study number 2020-841). Human glioma tissue samples were retrieved from the CNS tumor tissue bank Düsseldorf covering samples from adult-type diffuse gliomas classified according to the 2021 WHO classification of brain tumors [18] and corresponding to the CNS WHO grades 2, 3, or 4. Specifically, the investigated glioma cohort consisted of 5 astrocytomas, IDH-mutant, CNS WHO grade 2, 8 astrocytomas, IDH-mutant, CNS WHO grade 3, 2 oligodendrogliomas, IDH-mutant and 1p/19q-codeleted, CNS WHO grade 2, 5 oligodendrogliomas, IDH-mutant and 1p/19q-codeleted, CNS WHO grade 3, and 13 glioblastomas, IDH-wildtype, CNS WHO grade 4. Non-neoplastic brain RNA samples for real-time RT-PCR analyses were purchased from Ambion and Stratagene. The preparation of neonatal rat primary mixed glial cultures to obtain microglial cells was approved by the Institutional Review Board (IRB) of the Zentrale Einrichtung für Tierforschung und wissenschaftliche Tierschutzaufgaben (ZETT) at the Heinrich-Heine University Düsseldorf with the licenses O69/11 and V54/09.
2.2. Immunohistochemistry and immunofluorescent staining of human glioma sections
For tissue staining, sections from formalin-fixed and paraffinembedded glioma specimens were cut on a microtome (4 μm thick). Sections were mounted onto glass slides and dried prior to deparaffinization. Deparaffinization was performed in 100 % xylene for 15 min at room temperature, followed by 1 min each in 100 % xylene, 100 % ethanol, 96 % ethanol, 90 % ethanol, 70 % ethanol, 50 % ethanol, followed by deionized water. Antigen retrieval was performed in citrate buffer (Target Retrieval Solution, Citrate pH 6, Agilent Dako) for 20 min in a steam cooker. After antigen retrieval, slides were left to cool down for 15 min and then transferred into deionized water. Subsequently, slides were washed 3 times for 5 min in TBS-T at room temperature on a shaker. For immunohistochemical staining, slides were treated with Peroxidase-Blocking Solution (Agilent Dako) for 10 min at room temperature in an incubation chamber. Subsequently, slides were washed 3 times for 5 min in TBS-T at room temperature on a shaker. Blocking was performed in the incubation chamber for 10 min using serum-free protein block (Agilent Dako). Blocking solution was removed without subsequent washing and primary antibody HERV-W ENV (GN-mAB_03 [3B2H4]; provided by GeNeuro SA), diluted 1:1000 in Antibody Diluent, Dako REAL (Agilent Dako) was applied and incubated in the incubation chamber overnight at 4 ◦C. Slides were then washed 3 times in TBS-T for 5 min each at room temperature. Anti-mouse secondary antibody (Vector Labs) was diluted 1:1000 in Antibody Diluent, Dako REAL (Agilent Dako) and slides were incubated for 30 min at room temperature. Slides were washed 3 times for 5 min in TBS-T and then incubated with an ABC HRP complex for 1 h at room temperature (Vector Labs) (Dako, VectorLabs, Ultravision). Slides were washed 3 times for 5 min and then incubated with DAB Quanto Substrate and Chromogen mixture (Thermo Fisher Scientific) for 3 min. Slides were washed in deionized water, counterstained in hematoxylin for 2 min and then washed in tap water. Slides were transferred into deionized water and then moved through ascending concentrations of ethanol (50 %, 70 %, 90 %, 96 %, 100 %) into 100 % xylene and mounted using Eukitt (Sigma). Slides were scanned on a Leica Aperio slide scanner. For immunofluorescent staining, slides were blocked for 10 min using serum-free protein block (Agilent Dako). Blocking solution was removed without subsequent washing and primary antibody was applied and incubated overnight at 4 ◦C. The following primary antibodies were used after dilution in Antibody Diluent, Dako REAL (Agilent Dako): mouse anti-HERV ENV (1:500), rabbit anti-GFAP (1:1000; DAKO Agilent, Cat# Z0334), goat anti-IBA1 (1:500, Abcam ab5076), and rabbit anti-CD3 (1:200, Dako Agilent, Cat# A0452). Following washing steps with TBS-T, secondary antibodies (anti-mouse, anti-rabbit, anti-goat) conjugated with either Alexa Fluor488 or Alexa Fluor594 (1:1000; Thermo Fisher Scientific, Darmstadt, Germany) diluted in Antibody Diluent, Dako REAL (Agilent Dako) were applied for 30 min at RT. After three short washing steps in TBS-T and one washing step in PBS, nuclei are stained with 4′,6-diamidino-2-phenylindole (DAPI, 1 mg/ml stock; 1:2500; Thermo Fisher Scientific, cat #62248) in PBS for 3 min at RT. Sections were mounted with ProLong™ Gold Antifade Mountant (Thermo Fisher Scientific, cat #P10144) and imaged using a confocal laser scanning microscope 510 (CLSM 510, Zeiss, Jena, Germany).
2.3. Human glioblastoma (GBM) cell line cultivation
Human glioblastoma (GBM) cell lines A172, LN229, and T98G (obtained from American Type Culture Collection; ATCC) were cultured in T-75 flasks with GBM culture medium consisting of Dulbecco’s Modified Eagle’s Medium (DMEM; Thermo Fisher Scientific, Waltham, UK) supplemented with 10 % fetal bovine serum (FBS; Capricorn Scientific, Palo Alto, CA, USA) and 50 U/ml penicillin/streptomycin (Invitrogen, Carlsbad, USA) maintained in a 5 % CO2 atmosphere at 37 ◦C, following ATCC’s guidelines. Medium was exchanged tertian until cells were grown confluently. For passaging and seeding for experiments, cell detachment was performed using trypsin-EDTA (Gibco) for 3 min, and after stopping trypsinization with GBM culture medium, spun down for 5 min at 300×g at 4 ◦C. GBM cells were prepared for analysis by either plating cells directly into 24-well plates at a density of 10000 cells per well for quantitative reverse transcription-polymerase chain reaction (qRT-PCR) and cytokine arrays or by being seeded onto 13 mm glass coverslips placed within 24-well plates at a density of 5000 cells per well for subsequent immunocytochemical staining. In both cases, GBM cells were cultured with GBM medium for 24 h in an incubator (37 ◦C, 5 % CO2, and 90 % humidity) before cells were either further cocultured with primary rat microglia (MG; to generate GBM/MG cocultures) or stimulated with 1000 ng/ml HERV-W ENV protein (Protein’eXpert, Grenoble, France) and the appropriate buffer solution (ENV buffer: 20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1.5 % SDS, 10 mM DTT) as previously described in Refs. [5,6,19,20]. To mitigate any potential side effects from the recombinant production of HERV-W ENV protein, endotoxin levels were assessed using the limulus amebocyte lysate (LAL) test and were found to be below the detection threshold (<5EU/ml), as detailed previously [5]. For extended experiments and due to the known high proliferation rates of GBM cells, cell cultures were re-stimulated every other day (48 h).
2.4. Rat microglial cell (MG) preparation and human GBM/MG coculture cultivation
For GBM/MG cocultures, microglial cells were isolated using magnetic-activated cell sorting (MACS) from primary mixed glial cell cultures (as previously described in Refs. [5,6]). Briefly, mixed glial cell cultures were generated by isolating and subsequent dissociating cortices from P0/P1 Wistar rats. Primary cells were then cultured in T-75 flasks using DMEM supplemented with 10 % FBS, 2 mM L-glutamine (Invitrogen, Carlsbad, USA), and 50 U/ml penicillin/streptomycin. The culture medium was exchanged every second day until the cells reached confluence. After 10 days, the flasks were placed on an orbital shaker and agitated at 180 rpm at 37 ◦C for 2 h initially and subsequently for an additional 22 h. Supernatants containing microglia were harvested at both intervals, seeded onto bacterial dishes, and incubated at 37 ◦C, 5 % CO2, and 90 % humidity to allow cell adherence. Microglial cells were then detached using L-accutase (Gibco), and the cell suspension was centrifuged at 300×g for 5 min at 4 ◦C, followed by MACS with CD11b/c microbeads as per the manufacturer’s instructions (Miltenyi Biotec). The purity of microglial cell cultures, verified regularly by Iba1/GFAP staining, consistently showed a 99 % purity rate for microglia. To establish cocultures of human GBM cells with primary rat microglia (GBM/MG cocultures), microglia (100000 cells/well) were either directly co-seeded with the GBM cells (+MG) or placed in a hanging Millicell® cell culture insert (*MG; pore size 0.3 μM; Merck, PCSP24H48) preventing direct cell-cell contact of GBM and MGs. After 24 h, GBM/MG cocultures were treated with 1000 ng/ml HERV-W ENV protein or the appropriate buffer solution as described above.
2.5. RNA preparation, cDNA synthesis, and qRT-PCR
RNA preparation, cDNA synthesis, and qRT-PCR analyses were performed following protocols described in Ref. [21]. Briefly, lysates from GBM monocultures, GBM/+MG cocultures, and GBM cells from GBM/*MG cocultures were prepared using 350 μl of β-mercaptoethanol (1:100, Sigma-Aldrich) supplemented RLT lysis buffer (Qiagen, Hilden, Germany) and were immediately frozen on dry ice. The isolation of total RNA was carried out with the RNeasy Mini Kit (Qiagen), incorporating DNase digestion as per the manufacturer’s guidance. The assessment of RNA quality and concentration was conducted using a Nanodrop 1000 spectrophotometer (Thermo Fisher Scientific), with samples subsequently stored at − 80 ◦C for further analysis. For RNA extraction from human glioma specimens, cryosections were preserved and snap-frozen prior to RNA purification, which was performed using the Maxwell(R) RSC simply RNA Tissue Kit on a Maxwell(R) Benchtop Instrument (Promega). RNA was quantified using a QuantiFluor RNA Dye on a QuantusTM Fluorometer (Promega). Total RNA from cells and tissues was reverse transcribed using the High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific). The levels of gene transcripts were measured quantitatively using either a 7900HT Sequence Detection System (Thermo Fisher Scientific) for cell cultures or a StepOne™ Real-Time PCR System (Thermo Fisher Scientific) for human glioma samples applying SybrGreen Universal Master Mix (Thermo Fisher Scientific). Primer sequences, designed with PrimerExpress 2.0 (Thermo Fisher Scientific), are listed in the supplementary data table (SI Appendix, Table S1). ARF1 was identified as the most accurate and stable normalization gene when compared to ODC or GAPDH. Each sample was measured in duplicates and relative transcript levels were calculated following the ΔΔCt method after normalizing to -MG buffer within each data group. Results are presented as z-scores. To establish statistical significance, the raw data were tested on a group-by-group basis across GBM/-MG, GBM/+MG, and GBM/*MG categories, focusing on comparing the impacts of stimulation with HERV-W ENV protein against the buffer control.
2.6. Quantitative analysis of cytokine secretion dynamics
GBM/+MG cell culture medium was collected after 3 days of stimulation with either buffer or HERV-W ENV protein, as outlined above. Note that the presented results reflect the cytokine dynamics between days 2 and 3, supernatants were collected 24 h post medium change/restimulation on day 2. Approximately 1.5 ml of supernatant from three technical replicates was pooled, then centrifuged at 300×g for 5 min at 4 ◦C to eliminate any cell debris or floating cells. After harvesting, pure supernatants were immediately frozen on dry ice and stored at − 80 ◦C until analysis. For cytokine assessment, supernatants were gently thawed on ice. Secreted cytokine levels were determined using the Human Cytokine Antibody Assay (ab133997, Abcam) according to the manufacturer’s instructions by following overnight incubations for any binding steps. Images were captured using a Fusion FX imaging system (Vilber Lourmat). Densitometric quantification of cytokine levels was performed using ImageJ BioVoxxel software [22]. After background subtraction (uncultured GBM medium used as blank), the mean pixel densities of the 42 cytokines were normalized to the mean pixel densities of the positive controls (6 in total) on a buffer-treated array membrane for each specific GBM cell line group (e.g. A172/+MG, buffer vs. ENV). Mean pixel densities below 1 post background subtraction were considered irrelevant and omitted from analysis. Data are shown as means ± SEM based on two independent experiments.
2.7. Immunocytochemistry
Immunocytochemical staining was conducted following cell fixation with 4 % paraformaldehyde (PFA) at room temperature (RT) for 10 min. To block non-specific antibody binding, cells were incubated in a blocking solution composed of 10 % normal goat or donkey serum (NGS/NDS) mixed in PBS with 0.1 % Triton X-100 at RT for 45 min. This was followed by incubation in primary antibody solution containing 10 % NGS/NDS in PBS with 0.03 % Triton X-100. The primary antibodies were applied at the following concentrations overnight at 4 ◦C: rabbit anti-Ki67 (1:250; Abcam, ab16667), rat anti-Ki67 (1:250; Thermo Fisher, cat#14-5698-82), goat anti-Iba1 (1:500; Abcam, ab5076), and rabbit anti-CC3 (1:500; Cell Signaling Technology, #9661). After three PBS washing steps, secondary antibodies labeled with Alexa Fluor488 or Alexa Fluor594 (anti-rabbit, anti-rat, anti-goat; 1:500; Thermo Fisher Scientific) and supplemented with DAPI (0.02 μg/ml; Roche Diagnostic GmbH, Mannheim, Germany) were applied for 90 min at RT. For visualization of GBM cell line clustering and morphologies in GBM/+MG cocultures, F-actin was stained by additionally adding Phalloidin CruzFluor™ 594 Conjugate (1:1000; Santa Cruz, sc-363795) to the secondary antibody solution. Cells were mounted using Citifluor (Citifluor, London, United Kingdom). For image acquisition, a Zeiss Axionplan2 microscope (Zeiss, Jena, Germany) was used. Ki67 expression analysis was conducted using ImageJ BioVoxxel software [22]. Nine images per coverslip (with two coverslips per condition) were captured at 20 × magnification, maintaining consistent exposure times across all experiments. For quantification, the ratio of marker-positive cells to the total number of DAPI-stained (nuclear) cells was calculated, with results presented as the mean ± SEM of the relative change in Ki67 expression in response to ENV protein stimulation versus buffer across each culture model (GBM/-MG;/+MG;/*MG).
2.8. In vitro migration assay
To assess cell migration of GBM cell lines, predefined gaps of 500 μm width were created using culture-inserts 2 well for self-insertion (ibidi, Cat 80209). Each side of the insert was seeded with 50000 GBM cells in 35 μl of GBM medium. Depending on the experimental set-up, additional 50000 MGs were co-seeded to each side 24 h later. Following an overnight incubation to allow cell adherence, the inserts were carefully removed with sterile tweezers, and the wells were washed with PBS to discard any non-adherent cells. Subsequently, the medium was replaced with either HERV-W ENV protein- or buffer-supplemented GBM medium. The progression of gap closure was monitored by capturing images at three different but defined positions per well using a Zeiss CLSM 510 microscope system (CLSM 510, Zeiss) in 10× magnification via live cell imaging. Images of the wound gap were taken every 4 h. Analysis of the gap width was performed using a specific ImageJ plugin [23], with each gap being manually identified and marked. The percentage of wound closure was determined by comparing the initial wound area to the area remaining after x hours, with results expressed as a percentage of the initial gap closed.
2.9. Statistical analysis
Data are presented as mean values ± standard error of the mean (SEM) deriving from at least 3 independent experiments. Graphs and statistical analysis were performed using Excel and the GraphPad Prism 8.0.2 software (GraphPad Prism, San Diego, CA, USA; RRID: SCR_002798). The Shapiro-Wilk normality test was used to evaluate the datasets for non-Gaussian distributions. To determine statistical significance for normally distributed datasets, the unpaired two-sided Student’s t-test was applied for comparing two groups with homogenous variances, and Welch’s one-way analysis of variance (ANOVA) with Dunnett’s T3 multiple comparisons test was applied for analyzing more than two groups with heterogeneous variances. When evaluating normally distributed data relative to control conditions normalized to 1, the one-sample t-test was applied. Conversely, for datasets with nonGaussian distribution, the Wilcoxon signed-rank test was used. Statistical significance thresholds were set as follows: *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001.
3. Results
HERV-W ENV can be found in glioma cells but is also present in the tumor microenvironment (TME). As a first step, we investigated whether HERV-W ENV can be found in human gliomas and whether it is located in particular cell populations. To this end, we studied HERV-W ENV protein localization in control brain tissue (Fig. 1A) and different glioma specimens including oligodendroglioma, IDH-mutant and 1p/ 19q-codeleted, CNS grade 2 (Fig. 1B), oligodendroglioma, IDH-mutant and 1p/19q-codeleted, CNS grade 3 (Fig. 1D) and glioblastoma, IDHwildtype, CNS grade 4 (Fig. 1C). We found that control brain tissue featured no staining (Fig. 1A) while the oligodendroglioma, IDH-mutant and 1p/19q-codeleted, CNS grade 2 depicted in Fig. 1B was characterized by a strong staining of infiltrating cells as opposed to tumor cells. A different distribution emerged when we analyzed a glioblastoma, IDHwildtype, CNS WHO grade 4 (Fig. 1C) and another oligodendroglioma, IDH-mutant and 1p/19q-codeleted, CNS WHO grade 3 (Fig. 1D) in which we found a strong HERV-W ENV staining of tumor cells. In a second step, we analyzed the HERV-W ENV mRNA expression profiles of 32 diffuse human gliomas of CNS WHO grades 2 to 4 (II, III, and IV) and compared them to non-neoplastic normal control brain tissue samples (NB) (Fig. 1E). In doing so, we found that, cumulatively, gliomas exhibited significantly higher levels of HERV-W ENV mRNA expression than NB. This was also the case when the CNS WHO grades 3 (III) and 4 (IV) were separately compared to NB. Tumors of the CNS WHO grade 2 (II) did not reach statistical significance probably due to fewer samples. Finally, in order to better understand the composition of HERV-W ENVpositive TME cells, we performed (Fig. 1F–H) immunofluorescent double staining of a glioblastoma, IDH-wildtype demonstrating that HERVW ENV protein signals (green) could only be weakly detected in CD3positive T cells (Fig. 1F, red) and GFAP-positive tumor cells (Fig. 1G, red), while the strongest signal was found in IBA1-positive myeloid cells (Fig. 1H, red). HERV-W ENV stimulation results in a microglia-dependent modulation of gene expression profiles of human glioblastoma cell lines T98, A172, and LN229. In order to better understand how HERV-W ENV and myeloid cells may change the gene expression repertoire of glioma, we stimulated human glioblastoma cell lines T98 (Fig. 2A), A172 (Fig. 2B), and LN229 (Fig. 2C) with either 1000 ng/ml of recombinant HERV-W ENV or buffer in the absence (-MG) or presence (+MG and *MG) of rat microglia for 1, 3 and 5 days, respectively. Of note, for the validity of this scheme, it was important to rule out endogenous HERV-W ENV expression by GBM cell lines, which we had previously done using qRT-PCR (data not shown). Microglia were either directly cocultured on glioblastoma cells (+MG) or indirectly cultivated in cell culture inserts (*MG) preventing direct cell-cell interaction. Of note, general microglial viability was decreased in the *MG condition so that cell culture beyond one day was not feasible. In general, only HERVW ENV stimulation of glioblastoma cells in the direct or indirect presence of microglia led to significant gene regulation in comparison to controls, resulting in the increased expression of genes associated with tumor progression, aggressiveness and survival (TNFα, IL-6, IL-1β, iNOS), genes associated with tumor invasiveness and TME modulation (MCP-1, MCP-3, MMP9) as well as genes associated with tumor growth (CSF-1, CX3CL1). Notably, there were different patterns of spatio-kinetic responsiveness of glioblastoma cell lines to HERV-W ENV stimulation. In A172 cells, there was a highly significant and comprehensive regulation of genes on days 1 and 3 for the +MG condition, which substantially weakened on day 5. In T98 cells, there was a more scattered but still significant gene induction on day 1 for the +MG condition, while on day 3, significant regulation was scarce and weakened further on day 5. In LN229 gene regulation was even more scattered than in T98 but reached significance for several genes on day 1 for the +MG and the *MG condition. In the *MG condition, on day 1, A172 cells showed the strongest overall gene regulation, followed by LN229 and T98 cells, respectively. Regarding specific genes, on average overall upregulation of MCP1 was by far the strongest throughout all cell lines and timepoints. This was followed by IL-6, which, in turn, was followed by CSF-1, MMP9, and IL1β which were then followed by MCP-3, TNFα, iNOS, and lastly VEGF. HERV-W ENV stimulation modulates the protein expression of human glioblastoma cell lines T98, A172, and LN229. Following our observations regarding gene expression levels, we sought to determine whether such changes were also reflected by protein levels. Accordingly, we stimulated T98 (Fig. 3A and A’), A172 (Fig. 3B and B′), and LN229 (Fig. 3C and C′), respectively, with either 1000 ng/ml of recombinant HERV-W ENV protein or buffer in the presence (+MG) of rat microglia for 3 days. Cytokine secretion patterns were assessed using array membranes capturing the presence of 42 distinct human cytokines. This allowed for a comprehensive evaluation of cytokine secretion dynamics in response to HERV-W ENV stimulation. In general, we observed that glioblastoma cell lines exhibited trends similar to gene expression, but overall showed varied response patterns to HERV-W ENV stimulation. In general, MCP-1 protein levels were the most abundant in all cell lines. However, HERV-W ENV protein stimulation elevated its intensities even further in T98 and A172 in comparison to controls. This mirrored the gene regulation trends depicted in Fig. 2. Other strongly secreted proteins such as IL-8, IL-6 (with the exception of the A172 cell line), the GRO α-γ mix (equivalent to CXCL1-3), GRO α alone, and IL-10, exhibited increased signal intensities in response to stimulation with HERV-W ENV across all cell lines, with the most pronounced effects being observed in LN229. Although on the gene expression level we found that M-CSF (commonly known as CSF-1) was significantly downregulated in T98/+MG cocultures on day 3 of HERV-W ENV stimulation (Fig. 2A), secreted protein levels were more than doubled compared to buffer controls. Despite its low secretion levels, MCP-3 exhibited an overall HERV-W ENV protein-induced expression increase, also reflecting gene regulation data. In addition, HERV-W ENV stimulation led to a secretion increase of TNFα in T98 cells, a decrease in LN229, and unaltered levels of this protein in A172. HERV-W ENV stimulation did not modulate the Fig. 1. HERV-W ENV is present in human glioma tissue. (A–D) Anti-HERV-W ENV DAB immunostainings of non-neoplastic control brain tissue (A), oligodendroglioma, IDH-mutant and 1p/19q-codeleted, CNS WHO grade 2 (B), glioblastoma, IDH-wildtype, CNS WHO grade 4 (C) and oligodendroglioma, IDH-mutant and 1p/19q-codeleted, CNS WHO grade 3 (D). See insets for more tissue detail. Scale bars: 100 μm. (E) Real-time RT PCR of 32 diffuse gliomas (cumulated and according to CNS WHO grades 2 (II), 3 (III), and 4 (IV)) and normal brain (NB). (F–H) Immunofluorescent double staining of an IDH-wildtype glioblastoma (HERV-W ENV: green; CD3/GFAP/IBA1: red). Scale bars: 100 μm (A–D), 50 μM, and 20 μM for related insets (F–H). Data are shown as means (±SEM). Statistical significance was calculated using Welch’s ANOVA with Dunnett’s T3 multiple comparisons test. Data were considered significantly different at **p ≤ 0.01, ***p ≤ 0.001. Fig. 2. HERV-W ENV stimulation results in a microglia-dependent modulation of human glioblastoma cell line gene expression. (A–C) Human glioblastoma cell lines T98 (A), A172 (B), and LN229 (C) were stimulated with either HERV-W ENV or buffer in the absence (-MG) or presence (+MG and *MG) of rat microglia for 1, 3 or 5 days, respectively. Data are presented as z-scores and were considered significantly different at *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001 assessed with the one sample t-test. protein levels of VEGF-A, SDF-1 (commonly known as CXCL12), and IL1β in any of the cell lines. Interestingly, GM-CSF secretion was upregulated in T98 cells in contrast to A172 and LN229 cells, which showed no measurable secretion of this cytokine at all. HERV-W ENV stimulation leads to a microglia-dependent clustering of human glioblastoma cell lines T98 and A172. Following our observations regarding gene and protein expression, we sought to investigate if HERV-W ENV also exerts effects on observable glioblastoma cell behavior. To this end, we stimulated human glioblastoma cell lines T98 (Fig. 4A), A172 (Fig. 4B), and LN229 (Fig. 4C) with either 1000 ng/ml of recombinant HERV-W ENV or buffer in the absence (-MG) or presence (+MG) of rat microglia for 3 or 5 days, respectively. For improved analysis, cocultures were stained with the cytoskeletal marker F-actin (red) and the microglial marker anti-Iba1 (green). In doing so, we found that HERV-W ENV or buffer stimulation of T98 and A172 without microglia (-MG) did not change the spatial distribution of tumor cells after 3 or 5 days (Fig. 4A and B). In contrast, when microglia were present (+MG), HERV-W ENV stimulation resulted in progressive clustering of tumor cells after 3 and 5 days, respectively, in comparison to buffer controls. In LN229 no such effects were observed. HERV-W ENV stimulation modulates the proliferation of glioblastoma cell line A172 but does not affect apoptosis. Since cell proliferation is a key aspect of tumor malignancy, we also investigated the impact of HERV-W ENV on this process. To this end, we stimulated human glioblastoma cell lines T98 (Fig. 5A and B), A172 (Fig. 5C and D), and LN229 (Fig. 5E and F) with either 1000 ng/ml of recombinant Fig. 3. HERV-W ENV stimulation modulates the protein expression of human glioblastoma cell lines. (A-C′) Human glioblastoma cell lines T98 (A, A′), A172 (B, B′), and LN229 (C,C′) were stimulated with either HERV-W ENV or buffer in the presence (+MG) of rat microglia for 3 days. Data are shown as mean (±SEM) pixel densities of two biological replicates. HERV-W ENV or buffer in the absence (-MG), presence (+MG and *MG) of rat microglia, respectively. Ki-67-based analysis was performed on day 1 but stimulation was continued for a total duration of 5 days in order to further study the cell clusters described in Fig. 4. (Fig. 5C-C″). In A172, one day of HERV-W ENV stimulation in the absence of microglia (-MG) resulted in a significant decrease in the number of Ki-67-positive cells, while in the indirect presence of microglia (*MG) their number significantly increased in comparison to buffer controls. In T98 (Fig. 5A- Fig. 4. HERV-W ENV stimulation leads to microglia-dependent clustering of human glioblastoma cell lines. (A–C) Human glioblastoma cell lines T98 (A), A172 (B), and LN229 (C) were stimulated with either HERV-W ENV or buffer in the absence (-MG) or presence (+MG) of rat microglia for 3 or 5 days, respectively (Factin: red; Iba1: green). All scale bars: 100 μm. A″) and LN229 (Fig. 5E-E″) no such effects were observed. After 5 days of stimulation (B,D,F) all cell lines still proliferated as demonstrated by Ki67 staining (green), but only T98 (Fig. 5B) and A172 (Fig. 5D) showed clustering when stimulated with HERV-W ENV in the presence of MG (+MG). In none of the cell lines, HERV-W ENV or microglia induced apoptosis as elucidated by CC3-staining (red) confirming that the clustering of T98 and A172 cells does not result from programmed cell death. This was corroborated by gene expression analysis of BAX and TRAIL (apoptosis) as well as RIPK3 and MLKL (necrosis) where no sustained effects of HERV-W ENV were found (data not shown). HERV-W ENV stimulation results in a microglia-dependent increase of migration velocity of the human glioblastoma cell line T98. As migration is a key factor of tumor invasiveness, we also wanted to study the effect of HERV-W ENV on this cellular process. To this end, we stimulated human glioblastoma cell lines T98 (Fig. 6A-A‴), A172 (Fig. 6B-B″), and LN229 (Fig. 6C-C’‘) with either 1000 ng/ml of recombinant HERV-W ENV or buffer in the absence (-MG) or presence (+MG) of rat microglia, respectively. Cell culture inserts were used to create a predefined gap of 500 μm and after 4, 8, 12, and 16h, we measured gap closure. In T98 (Fig. 6A-A‴), HERV-W ENV or buffer stimulation without microglia (-MG) did not significantly change migration velocity. In contrast, when microglia were present (+MG), migration velocity was significantly increased, leading to a gap closure of approximately 80 % after 16h as compared to buffer controls, which reached around 50 % (see yellow lines in 6A‴). In A172 (Fig. 6B-B″) and LN229 (Fig. 6C-C″) no such effects were observed.
4. Discussion
In this study we demonstrate that HERV-W ENV is present in both glioma cells and infiltrating myeloid/microglial cells of the surrounding TME that may modulate neoplastic cell behavior. We found that in response to HERV-W ENV stimulation in the presence of microglia, GBM cells of 3 different GBM cell lines significantly increased their expression of MCP-1, IL6, CSF-1, MMP9, IL-1β, MCP3, TNFα, iNOS, and lastly VEGF, albeit with different kinetics. These findings were reflected further downstream on the protein level. According to the current literature, many of these factors are interconnected in gliomas in a complex network of partly mutual dependencies. IL-6 plays a pivotal role in glioma growth [24], malignancy [25], as well as invasiveness [26], and elevated levels are associated with reduced patient survival [27]. MCP-1 and MCP-3, whose secretion levels were boosted by HERV-W ENV stimulation in the presence of microglia across all three cell lines, are associated with tumor development, tumor invasion, metastasis, angiogenesis and, particularly, immune cell chemotaxis leading, inter alia, to the recruitment of microglial cells into the TME [14,15,28,29]. In Fig. 5. HERV-W ENV stimulation modulates the proliferation of glioblastoma cells but has no effect on apoptosis. (A–F) Human glioblastoma cell lines T98 (A–B), A172 (C–D), and LN229 (E–F) were stimulated with either HERV-W ENV or buffer in the absence (-MG; A,B,C) or presence (direct contact: +MG; A′,B′,C′ and transwell-separated: *MG; A″,B″,C″) of rat microglia for 1 or 5 days, respectively (Ki67: green; CC3: red). Statistical significance was calculated using the one sample ttest: *p ≤ 0.05, **p ≤ 0.01. Scale bars: 100 μm and 50 μm for insets (A–F). addition, IL-6 can increase the expression of matrix metalloproteases (MMPs) such as MMP9 [30] which play a pivotal role in the degradation of the extracellular matrix (ECM). This, in turn, facilitates the migration of tumor cells and thereby contributes to tumor invasion [31] and progression [32]. CSF-1, on the other hand, was found to be involved in high-grade glioma formation [33] as well as glioma progression and microglial recruitment [34]. IL-1β is associated with tumor progression, survival, and invasiveness [35–37]. The pro-inflammatory factors TNFα and nitric oxide (NO) produced by the inducible NO synthase (iNOS) are relevant for the breakdown of the blood-brain-barrier (BBB) during glioma progression [37]. In addition, TNF-α was also found to be overexpressed in gliomas, and to be associated with glioma tumor grade, tumor development, and angiogenesis [38,39]. Lastly, VEGF is the most important stimulator of angiogenesis and is thereby key for the high metabolic demands of GBM [37]. Via its receptors VEGFR1 and VEGFR2, it plays a pivotal role in tumor survival, invasiveness, and progression [37,40]. Beyond the induction of the above-described factors, we also found that via microglia HERV-W ENV increases tumor cell proliferation, enhances tumor cell migration velocity, and exerts no effect on apoptosis and necrosis. However, we found that HERV-W ENV induces the formation of tumor cell clusters. Even though in gliomas, metastasis only plays a minor role, similar clustering is also observed in circulating tumor cells (CTCs; [41]). It is therefore tempting to speculate that HERV-W ENV-induced glioblastoma tumor clusters undergo changes in their integrin expression profiles and in the expression of other proteins associated with tumor cell-to-cell contact. Future studies analyzing these tumor clusters will have to address this issue. In the past years, the TME has also been identified as a key factor in glioma treatment resistance [42]. In this regard, we also studied whether HERV-W ENV in the presence or absence of microglia exerts an effect on tumor cell response to chemotherapy agents. However, when the glioblastoma cell lines used in our study were exposed to increasing concentrations of temozolomide (TMZ), the clinical standard of care of GBM chemotherapy, we did not find consistent effects due to technical limitations (data not shown). We aim to address this issue in future studies. In addition, using crystal violet-based assays we also investigated whether HERV-W ENV exerts effects on the invasion properties of glioblastoma cells but found no significant differences compared to controls (data not shown). These findings point to the complexity of glioma biology that can only be incompletely simulated in vitro and warrant further studies. Even though Fig. 6. HERV-W ENV stimulation results in a microglia-dependent increase of migration velocity of human glioblastoma cells. (A-C″) Human glioblastoma cell lines T98 (A-A‴), A172 (B–B″), and LN229 (C–C″) were stimulated with either 1000 ng/ml of recombinant HERV-W ENV or buffer in the absence (-MG) or presence (+MG) of rat microglia, respectively. Culture inserts were used to create a predefined gap of 500 μm and after 4, 8, 12, and 16h gap closure was measured. Data are shown as means (±SEM). Statistical significance was calculated using the Student’s two-sided, unpaired t-test: *p ≤ 0.05. Scale bars: 100 μm. it is a scientific consensus that glioma is able to reprogram the surrounding TME and to “hijack” cells such as microglia to facilitate their own proliferation, invasion, migration, and survival [14,43] the relationship between non-neoplastic and neoplastic cells in GBM is reciprocal [29]. However, in our study we deliberately focused on the behavior of glioblastoma cells as we are still clarifying the changes taking place in microglia exposed to HERV-W ENV in conjunction with GBM cells. We felt that including related data would have gone beyond the scope of our current study, and will therefore be the subject of future publications. Interestingly, while we found that HERV-W ENV expression per se is significantly higher in glioma compared to healthy brain, levels are apparently not correlated with increasing tumor malignancy based on the 2021 WHO grading system. On the other hand, we found in vitro evidence that via microglia HERV-W ENV modulates the drivers of glioma progression. These seemingly contradictory observations may, in part, be based on a HERV-W ENV “ceiling effect” or on the different HERV-W ENV-positive cell populations we found in different tumors. In an oligodendroglioma, IDH-mutant and 1p/19q-codeleted (CNS WHO grade 3) and a recurrent glioblastoma, IDH-wildtype (CNS WHO grade 4) we found a strong HERV-W ENV staining of tumor cells while in another oligodendroglioma, IDH-mutant and 1p/19q-codeleted (CNS WHO grade 2) there was a strong staining of infiltrating myeloid/microglia cells. While we have not yet studied this aspect systematically, it is conceivable that in lower grade gliomas, HERV-W ENV is predominantly present in microglia and may thereby support tumor growth. On the other hand, this is not mirrored in our experiments where cell lines originating from highly malignant GBMs reacted strongly to HERV-W ENV. Of note, this aspect is further complicated by the well-described heterogeneity of available GBM cell lines differing greatly in their characteristics [44] which is also reflected in our study. As Kiseleva and colleagues pointed out A172 (isolated from a 53-year-old male) and T98 (isolated from a 61-year-old male) vary morphologically and in surface marker expression but showed the same response to serum [45]. Kiseleva and colleagues concluded that both A172 and T98 sustain the main features of glioblastoma and are suitable tools to study this neoplasm. Complementarily, Demircan and colleagues compared the LN229 (isolated from a 60-year-old female) and U87 – a cell line not used in this study and originally isolated from a male of unknown age – cell lines to healthy human astroglia SVGp12 cells [44]. Accordingly, they found differing proliferation/apoptosis rates (LN229 > U87), migration kinetics (LN229
Acknowledgments
We thank Birgit Blomenkamp-Radermacher, Brigida Ziegler, and Julia Jadasz for their technical assistance. Furthermore, we would like to thank Geneuro SA for providing recombinant HERV-W ENV protein.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi. org/10.1016/j.micinf.2024.105460.
Article Images (0)
×![]()
Bioz is the world’s most comprehensive AI search engine for scientific experimentation. The Bioz search engine offers researchers billions of data-driven product, technique, and protocol recommendations. Bioz taps into the latest advances in Artificial Intelligence (AI) – including Natural Language Processing (NLP), Machine Learning (ML) and Generative AI to mine and structure hundreds of millions of pages of complex and unstructured scientific papers. By harnessing the power of AI, Bioz provides researchers with an unprecedented amount of summarized scientific experimentation knowledge right at their fingertips, ultimately helping to speed up drug discovery and the rate of success in finding cures for diseases. Bioz is used by over 15 Million researchers from over 17,000 different universities and companies in 196 countries.
