Introduction

Radiotherapy (RT) is an integral part of adjuvant glioblastoma multiform (GBM) radiochemotherapy [1, 2]. However, the existing regimen is not without some contradictions [3] and has numerous clinical negative side-effects [4, 5]. Some of these effects are related to RT-induced changes in tumour microenvironment (TME) that may promote GBM aggressiveness and contribute to therapeutic resistance [6, 7].

One of the key TME components is extracellular matrix (ECM) mainly composed of glycosylated proteoglycan (PG) and glycosaminoglycan (GAG) molecules, which exert diverse functions in tumour stroma in a cell-specific and context-specific manner and contribute to the formation of a permissive provisional matrix for tumour growth [8]. Biological role and importance of PGs in cancer development are described in details in these reviews [8, 9].

PGs play critical roles in the regulation of brain cancer cell signalling and migration and drive multiple oncogenic pathways promoting critical tumour-microenvironment interactions [9]. GBM radiochemotherapy results in changes in heparan sulfate PG (HSPG) glypican-1 and HS-modifying enzyme heparanase in GBM tumours [10, 11]. Radiation increases the production of hyaluronic acid (HA) (a signalling ligand for CD44 receptor) by GBM cells and affects biomechanical tension in GBM microenvironment, providing pro-invasive extracellular signalling cue [12]. Irradiation of primary human GBM developed in nude rats with a dose of 50 Gy increases content of CD44-positive cells and activates proliferation, migration and invasion of primary GBM cells [5]. In their turn, radioresistance of GBM cells depends on NG2/CSPG4 expression, which suggested as an important prognostic factor for GBM patient survival [13].

Along with the effect of RT on tumour cells, there are few data on the effect of irradiation on normal brain tissue. It was shown that low dose (1 Gy) X-ray irradiation of experimental rats during initial cortical development causes multiple defects in the formation of cortical afferent and efferent pathways [14]. Irradiation of adult rat spinal cord with a dose of 40 Gy selectively facilitates migration of oligodendrocyte progenitor cells (OPCs) actively expressing NG2/CSPG4 biomarker, but not Schwann cells (SCs) [15]. X-ray irradiation of juvenile rat brain (hippocampus) with a moderate radiation dose (16 Gy) induces active migration and proliferation of olfactory ensheathing cells as compared with SCs in the moderately X-irradiated rat brain [16]. This is practically all data that is known about the effect of radiation on the composition and expression of PGs in tumour and normal brain tissues. Purpose of the study was to investigate whether single X-ray irradiation is able to affect PGs expression levels in normal mouse brain tissue in the animal model in vivo.

Materials and methods

Animals and tissue samples

All experiments were performed on 2-months male C57BL/6 mice (n = 54) obtained from the SPF animal house at the Institute of Cytology and Genetics (Novosibirsk, Russia). Animals were housed in polycarbonate cages with free access to food and water, 12/12 h light/dark cycle, temperature of 25 ± 1 °C, humidity of 50–60%. Animals were killed by cervical dislocation [AVMA Guidelines for the Euthanasia of Animals, 2013]. Freshly excised brain tissue was divided, and one hemisphere was immediately fixed in 10% neutral buffered formalin, then routinely processed and embedded in paraffin using Shandon Citadel 2000 Tissue Processor and HistoStar Embedding Workstation (ThermoFisher Scientific, USA). The other hemisphere was divided on cortex and subcortex structures and was preserved with RNALater solution (ThermoFisher Scientific, USA) according to the manufacturers’ instruction. All procedures with the experimental animals were conducted in accordance with the European Communities Council Directive 2010/63/EU and approved by the local Committee for Biomedical Ethics of Institute of Molecular Biology and Biophysics FRC FTM. All efforts were made to minimize animal suffering and to reduce the number of animals used.

Irradiation and the radiation dose determination

This study was performed using the mouse brain irradiation model described in [17]. Briefly, intraperitoneal injection with domitor (Orion Corporation, Finland; 0.125 mg/kg of mouse body weight) followed after 10 min by zoletil injection (Valdepharm, France; 20 mg/kg) into another part of mouse abdomen was used to anaesthetise mice. The anaesthetised animals were fixed in 2-post immobilization device and irradiated using the research accelerator complex VEPP-4 (Budker Institute of Nuclear Physics SB RAS, Novosibirsk, Russia) or the clinical linear accelerator Elekta Axesse (Elekta Ltd., UK) (Meshalkin National Medical Research Center, Novosibirsk, Russia).

Irradiation zones included the whole mouse head for VEPP-4, or brain zone delineated using computer tomography (CT) scans (for ElektaAxesse). A single dose of 7 Gy was given by either VEPP-4 or ElektaAxesse accelerators, and treatment mode was set up according to the VEPP-4 settings or calculated using treatment planning software for ElektaAxesse, respectively. Irradiation doses were determined using radiochrome film Gafchromic EBT3 (Ashland Specialty Ingredients, USA), scanned by Canon 9000f Mark II (Canon, Japan) and quantification analysis was performed using the ImagePro6 programme.

Histological study

Routine staining of 3.5-μm tissue sections with Haematoxylin and Eosin (H&E) was performed on Microm HMS740 (ThermoFisher Scientific, USA) and documented by light microscopy (Axiostar Plus, Carl Zeiss). The staining was analysed by two qualified pathologists independently; ten microscopic fields were estimated for each specimen.

Real-time RT-PCR analysis

Total RNA was extracted from the tissue samples using the QIAzol Lysis Reagent (Qiagen, USA) according to the manufacturer’s instructions. The integrity and quality of the isolated RNA were checked by agarose gel electrophoresis, and total RNA concentration was measured with Qubit–iT RNA Assays Kit (ThermoFisher Scientific, USA) according to the manufacturer’s instructions. cDNA was synthesized from 0.5 μg of total RNA using a First Strand cDNA Synthesis kit (Fermentas, Hanover, MD, USA) and one-tenth of the product was subjected to PCR analysis. Real-time PCR for mouse PGs (syndecan-1, glypican-1, perlecan, versican, brevican, CSPG4/NG2, CD44, decorin, biglycan, neurocan) was performed using the CFX96 Real-Time PCR Detection System (Bio-Rad, USA) and the Taq-polymerase (IMCB, Russia), SYBR Green (ThermoFisher Scientific, USA). PCR conditions were: 95 °C for 3 min, then 95 °C for 20 s, 59 °C for 15 s, and 72 °C for 30 s for 40 cycles; the total reaction volume was 25 μl; PCR primers were.

  • Dcn-F CCCCTGATATCTATGTGCCC; Dcn-R GTTGTGTCGGGTGGAAAATC;

  • Bgn-F GCCTGACAACCTAGTCCACC, Bgn-R CAGCAAGGTGAGTAGCCACA;

  • Ncan-F CCAGCGACATGGGAGTAGAT, Ncan-R GGGACACTGGGTGAGATCAA;

  • Bcan-F GTGGAGTGGCTGTGGCTC, Bcan-R AACATAGGCAGCGGAAACC;

  • Vcan-F GGAGGTCTACTTGGGGTGAG, Vcan-R GGGTGATGAAGTTTCTGCGAG;

  • NG2-F TCTTACCTTGGCCCTGTTGG, NG2-R ACTCTGGTCAGAGCTGAGGG;

  • CD44-F CAAGTTTTGGTGGCACACAG, CD44-R AGCGGCAGGTTACATTCAAA;

  • Sdc1-F GGTCTGGGCAGCATGAGAC, Sdc1-R GGAGGAACATTTACAGCCACA;

  • Gpc1-F CTTTAGCCTGAGCGATGTGC, Gpc1-R GGCCAAATTCTCCTCCATCT;

  • Hspg2-F CCGTGCTATGGACTTCAACG, Hspg2-R TGAGCTGTGGAGGGTGTATG.

Gapdh was used as the reference gene: Gapdh-F CGTCCCGTAGACAAAATGGT, Gapdh-R TTGATGGCAACAATCTCCAC. The fold change for each mRNA was calculated by the 2−ΔCt method.

Statistical analysis

ANOVA analysis with Fisher’s Post-Hoc test was performed to determine statistical significance between groups using OriginPro 8.5 software. Value of p < 0.05 was considered to indicate a statistically significant difference. Data are expressed as means ± SD.

Results

Irradiation effects on brain tissue morphology

To perform the study, novel mouse model with a specific combination of anaesthesia, irradiation dose and X-ray sources was developed [17]. Short-term effects of 7 Gy single-dose irradiation on mouse brain morphology (cerebral cortex and hippocampus) were analysed 24, 48 and 72 h after irradiation (6 animals/time-point). To consider the variability of the observed effects from X-ray source characteristics, two different irradiation facilities were used: synchrotron research accelerator VEPP-4 (X-ray spectrum ranged from 100 to 250 keV) (n = 18) or linear clinical ElektaAxesse accelerator with an energy of 6 MeV (n = 18). Control group was composed of anaesthetised mice (n = 18).

Routine histological staining with H&E did not reveal significant pathomorphological changes in the irradiated tissues compared with the control ones (Fig. 1). The observed moderate polymorphism of neurons, mild oedema of molecular cortex layer and very weak oedema of the ECM of other layers seem to be non-specific and quite minimal. Both accelerators demonstrated similar short-term effects on brain tissue morphology, suggesting that single irradiation of 7 Gy dose in wide energy diapason (100 keV–6 MeV) does not affect the morphology of brain tissue in experimental animals in vivo.

Fig. 1
figure 1

Morphology of cerebral cortex and hippocampus in the irradiated mouse brain tissues after 24, 48 or 72 h after irradiation at VEPP-4M or ElektaAxesse X-ray accelerators. The normal tissue panels demonstrate normal structure of cerebral cortex and hippocampus and serve as control to the corresponding samples irradiated at both VEPP-4M and ElektaAxesse. H&E staining. Magnification × 400, scale bar size 50 µm

Full size image

Irradiation effects on PGs expression in mouse brain tissue

Compared to morphology, molecular characteristics of brain tissue are more dynamic parameters which might quickly react to external stimuli like irradiation. We analysed the expression of key glycosylated components of brain tissue as PGs (syndecan-1, glypican-1, HSPG2/perlecan, versican, brevican, CSPG4/NG2, CD44, decorin, biglycan, neurocan) in control and irradiated brain tissues after 24, 48 and 72 h after anaesthesia or irradiation by real-time RT-PCR (Figs. 2, 3).

Fig. 2
figure 2

Short-term effects of single-dose irradiation onto the overall transcriptional activity of PG-coding genes in mouse brain tissue. PGs profiling was performed at 24, 48 and 72 h after irradiation. Real-time RT-PCR analysis, stacked column compares the contribution of each value to a total across categories. VEPP-4M or ElektaAxesse–X-ray accelerators

Full size image
Fig. 3
figure 3

Expression levels of individual PGs in the irradiated mouse brain tissue. PGs profiling was performed at 24, 48 and 72 h after irradiation. Real-time RT-PCR analysis, expression normalised to Gapdh, bars represent the mean ± SD from triplicate experiments. ANOVA analysis with Fisher’s Post-Hoc testing was carried out with significance indicated as *p < 0.05 (OriginPro 8.5)

Full size image

Overall transcriptional activity of PGs in the irradiated brain tissues demonstrated variability within 40–50% from that in the control brain tissues, without a clear tendency to its’ activation or inhibition (Fig. 2), different X-ray sources had slightly different effects, especially for subcortex brain zone.

Expression levels of HSPG-coding genes (syndecan-1, glypican-1, perlecan) were significantly lower than those for chondroitin sulfate PG (CSPG)-coding genes, and decorin and neurocan were predominantly-expressed PGs in mouse brain tissue. Interestingly, the transcriptional activity of these highly-expressed genes almost was not affected by irradiation (except the only time-point for decorin), although significant down-regulation of brevican and NG2/CSPG4 expression was observed (3–6-fold and 10–15-fold, respectively, p < 0.05) both in the cerebral cortex and subcortex structures (Fig. 3). In most cases, the revealed effects were similar both by research and clinical accelerators, nevertheless, individual differences were observed for some PGs: irradiation by VEPP-4 demonstrated down-regulation of decorin and brevican in the cerebral cortex, while irradiation by ElektaAxesse had a less pronounced effect or even developed opposite tendency (Fig. 3). Despite these changes not being statistically significant, they point an important issue that X-ray spectrum and/or energy of X-ray source might be of importance for some research application.

In total, the demonstrated changes in brevican and NG2/CSPG4 expression levels did not contribute significantly to the overall transcriptional activity of PGs, although it cannot be excluded that this may reflect a tendency which will be developed and functionally meaningful after multiple brain irradiations.

Discussion

Radiation effects on normal brain ECM remain poorly investigated, and the presented data for the first time demonstrate that a single irradiation of mouse brain with a moderate X-ray dose (7 Gy) can affect the expression levels of some PGs. There are no direct data to compare with, but the obtained data stay in line with the results by Willey et al. on the reduced PG synthesis in human and pig cartilage undergone a single X-ray irradiation with 2 or 10 Gy dose [18].

An important observation is related to the selective down-regulation of some CSPGs brevican and NG2/CSPG4, while HSPGs remained not affected. It is known that glycosylated CSPG molecules act as one of the central brain TME organisers, and non-invasive GBM lesions are associated with a rich ECM containing substantial amounts of CSPGs, whereas they are essentially absent from diffusely infiltrating GBM tumours [19]. The CSPG-rich microenvironment is associated with non-invasive tumour lesions through LAR-CSGAG binding, while the absence of glycosylated CSPGs induces the critical glioma invasion [20]. Our results on the short-term inhibitory effect of X-ray irradiation (even in a single dose mode) on CSPGs expression in the irradiated brain tissues support the mentioned data and suppose CSPG depletion from brain ECM as a potential molecular mechanism of RT-induced GBM relapse.

Interestingly, one of the most irradiation-affected CSPG is neuro-glial antigen 2 (NG2/CSPG4), which is the main biomarker of OPCs. Previously, a profound irradiation-induced decrease in the number of OPCs in the adult rat spinal cord after X-ray exposure was reported [21]. Time-lapse of these two events demonstrates surprising coincidence (fast down-regulation of NG2 expression at 1–3 days after irradiation and decrease in OPCs content at 4–10 days after exposure). In GBM, the up-regulation of NG2/CSPG4 predicts poor survival of patients and was suggested as a prognostic biomarker or potential therapeutic target for the treatment of GBM [22, 23]. It was shown that both elevated expression level of NG2/CSPG4 and number of OPC cells are associated with resistance of the GBM tumours to ionising radiation [13, 24]. All these data are in agreement with the up-to-date ideas that radiation-induced changes in brain TME may promote GBM aggressiveness amplified in the previously irradiated microenvironment [6]. Possibly, NG2/CSPG4 possesses an ambiguous function in brain physiology and gliomagenesis, when an elevated expression of NG2/CSPG4 in GBM cells and its’ downregulation in the TME and surrounding normal brain tissue (for example, upon irradiation/chemotherapy) might cooperate in providing a pro-tumourigenic niche for GBM progression and relapse development.