The CHI3L1 protein plays role as a protecting factor against autophagy in glioblastoma cells
Agnieszka Rusak, Marzena Janiszewska, Maciej Raczkowski, Supatcharee Cael, Mateusz Guźniczak, Igor Buzalewicz, Klaudia Krawczyńska, Samir F. El-Mashtoly, Michał Kulus, Tomasz Górnicki, Piotr Dzięgiel, Marzenna Podhorska-Okołów, Jürgen Popp, Christoph Krafft
TL;DR
This study shows that the CHI3L1 protein protects glioblastoma cells from autophagy, suggesting that inhibiting it could be a new treatment strategy.
Contribution
The novel finding is that CHI3L1 inhibition induces autophagy in glioblastoma cells, offering a potential new therapeutic approach.
Findings
CHI3L1 inhibition leads to increased autophagy in glioblastoma cells.
CHI3L1 expression changes in response to cell starvation and X-ray doses.
Inhibiting CHI3L1 could be a new therapy to upregulate autophagy in glioblastoma.
Abstract
The CHI3L1 protein supports various types of cancer progression and metastasis, where the background players were the upregulation of angiogenesis and microenvironment modulation. In glioblastoma (GB), high vascularisation is a key feature of these tumours, making anti-angiogenic therapy a pivotal treatment option. Autophagy, a dual-faced mechanism, may be useful as a target in GB treatment. This work presents the role of the CHI3L1 protein in autophagy in GB. U-87 MG glioblastoma cells and a GB spheroid model consisting of U-87 MG cells, macrophages and endothelial cells were used in the studies. A new tissue-like phantom was designed for the radiotherapy of spheroids. The role of CHI3L1 in autophagy regulation was analysed after cell starvation and treatment with G721-0282, a small molecule inhibitor of CHI3L1, as well as X-ray doses. The biological responses were evaluated using the…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6- —Leibniz-Institut für Photonische Technologien e.V. (4465)
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsStudies on Chitinases and Chitosanases · Nicotinic Acetylcholine Receptors Study · 14-3-3 protein interactions
Introduction
Glioblastoma (GB; grade IV according to the WHO classification system) is still one of the most aggressive brain tumours, with an average survival time of approximately eight months after diagnosis [1]. The results of the most appropriate treatments available for GB, including temozolomide (TMZ), radiotherapy, and surgical resection, are not effective due to the chemoresistant and radioresistant phenotypes arising and the rapid recurrence after resection. The clinical impasse in GB treatment has led to new targeted therapy propositions which are currently being tested in clinical studies [2]. One of the most promising targets in glioblastoma (GB) is the inhibition of angiogenesis, or the formation of new blood vessels. However, the effects of bevacizumab, an antibody that targets vascular endothelial growth factor A (VEGF-A), are short-lived and limited to a specific GB subtype. Sensitising GB tumour cells to radiotherapy is one of the most interesting possibilities associated with the induction of autophagy—the mechanism responsible for protecting glioblastoma cells against radiation [3]. Autophagy, the mechanism by which cellular components are degraded, is considered a potential target in solid tumours, with both overexpression and inhibition being investigated. In GB tumours, autophagy inhibition is thought to play a protective role and is associated with chemoresistance and radioresistance phenotypes. For this reason, inhibiting autophagy is a recent approach to designing therapy for this type of tumour [3]. On the other hand, the stimulation of autophagy, which leads to apoptotic cell death, has also been indicated as useful in targeting GB. During treatment with temozolomide (TMZ), one of the most potent chemotherapeutic drugs for GB, autophagy occurs as a side effect of the therapy [4], and also as a result of bevacizumab administration [3].
CHI3L1 (YKL-40) is a 40 kDa glycoprotein with multiple functions in relation to the cancerous process, primarily supporting angiogenesis in solid tumours [5–7]. In GB, CHI3L1 overexpression was detected in the cytoplasm of cancer cells, as well as in the serum of patients, and was found to correlate with a poor prognosis and a poor response to X-ray radiation [8, 9]. In vivo models of breast and colon cancer demonstrated the anti-angiogenic potential of CHI3L1 inhibition [10, 11]. Moreover, increased expression of the CHI3L1 protein was observed after X-ray irradiation and was found to be connected with the protective function of this protein in U-87 MG glioblastoma cells. Inhibition of the CHI3L1 protein was also found to decrease angiogenesis and compensate for the inhibition of VEGF-A [11]. Additionally, the inhibition of CHI3L1 changes the expression of membrane proteins, such as N-cadherin and VCAM-1, in glioblastoma spheroids [12]. Currently, the broad anti-cancer effect of CHI3L1 inhibition has been demonstrated following the use of antibodies directed against this protein, as well as selective inhibitors G721-0282 in osteosarcoma and glioblastoma cells, and K284 in lung cancer cells [12–15]. For this reason, targeting CHI3L1 using selective small molecule inhibitors has the potential to improve GB therapy with regard to angiogenesis, autophagy and immune modulation. In this paper, we present CHI3L1 targeting in glioblastoma spheroids following X-ray treatment, alongside analysis based on optical methods: O-PTIR (optical photothermal infrared spectroscopy) and DHT (digital holographic tomography).
In general, infrared spectroscopy is recognised as a promising tool for soft tissue analysis [16–19]. Optical phototermal infrared spectroscopy (O-PTIR) is a relatively new implementation of infrared spectroscopy used in tissue analysis. It enables label-free, semi-quantitative and qualitative analysis of biological components based on an optical detection of a photothermal response that is induced by absorption of infrared radiation [20]. The advantages of O-PTIR compared to conventional infrared techniques include to collect artefact-free IR spectra in reflection mode on inexpensive glass slides and to collect rapidly infrared images at selected absorption bands with submicrometer resolution. These O-PTIR advantages made O-PTIR attractive for analysing protein, nucleic acid, lipid and metabolomic changes in biological samples, especially cancer tissue. Although immunohistochemical reactions are still the gold standard widely used by pathologists, O-PTIR may serve as a valuable complement from a pathological point of view [18, 21, 22]. In this paper, we present for the first the application O-PTIR to analyse biological responses in GB spheroid models treated with the CHI3L1 selective inhibitor G721-0282 and X-rays.
Digital holographic tomography (DHT) is another label-free, non-destructive method that can be used to analyse morphological changes in biological samples based on their 3D refractive index (RI) distribution [20, 27–30]. Based on this, DHT enables the qualitative analysis of intracellular structures such as the cytoplasm, nucleus, nucleoli and lipid droplets, providing 3D morphological and optical information with submicron and subnanometre resolution, which goes beyond the optical diffraction limits of conventional microscopy techniques used in life science, such as confocal or fluorescence microscopy. In this study, the DHT technique was applied to analyse the content of autophagosomes related to U-87 MG treatment with G721-0282 [14, 23–25].
Materials and methods
Cell culture
The commercially available human glioblastoma cell line U-87 MG (ATCC^®^ CRL-1421™, American Type Culture Collection, Old Town, Manassas, VA, USA) was provided in DMEM with 4.5 g/L glucose (Capricorn, Ebsdorfergrund, Germany). The HMEC-1 human microvascular endothelial cell line (ATCC) was cultured in MCDB131 medium supplemented with 10 mM L-glutamine, 10 ng/mL FGF (ThermoFisher Scientific, Wilmington, DE, USA) and 1 µg/mL hydrocortisone (Sigma-Aldrich). The THP-1 human monocyte cell line, derived from acute monocytic leukaemia (ATCC), was cultured in RPMI-1640 (Gibco, ThermoFisher Scientific), supplemented with 50 µM β-mercaptoethanol (Sigma-Aldrich). All media were supplemented with 10% FBS (fetal bovine serum) (Sigma-Aldrich) and 2 mM L-glutamine, as well as a streptomycin and penicillin solution. The cells were passaged with TrypLe (Gibco) when confluence did not exceed 70%, with the medium changed twice weekly. Cell cultures were provided under standard conditions (37 °C, humid atmosphere, 5% CO₂) in a HeraCell 150i incubator (ThermoFisher Scientific) [14].
Monocyte differentiation
THP-1 monocytes, which are a typical suspension culture, were stimulated by the addition of 100 nM PMA (from Sigma-Aldrich) to a complete RPMI-1640 medium (from Gibco). After 24 h, the differentiation of monocytes into adherent macrophages was observed [26].
G721-0282 compound
G721–0282 (IUPAC name 2-({6-butyl-1,3- dimethyl-2,4-dioxo-1 H,2 H,3 H,4 H-pyrido[2,3-d]pyrimidin-5-yl}sulfanyl)-N-(prop-2-en-1-yl)acetamide), the small molecule inhibitor of CHI3L1 protein was obtained from MolPort (compound number: MolPort-003–169–389 Mw = 376.48 g/mol, purity > 90%), synthesised by ChemDiv, Inc (San Diego, CA, USA). The main 100 mM stock was prepared in DMSO (dimethylsulphoxide) (Sigma-Aldrich, St. Louis, MO, USA) [13]. A dilution to 100 µM was prepared in culture medium, as this was found to be the most effective concentration in previous studies [14]. Concentration of DMSO in 100 µM of G721-0282 was 0.1% and has no influence on cells as we previously described [12, 14].
Autophagy protocol
To induce autophagy, a nutrient-free medium of Hank’s Balanced Salt Solution (HBSS) (Gibco) was used [25, 27]. In the U-87 MG cell monolayer culture, the medium was replaced with HBSS or a medium containing 100 µM G721-0282 for 4 h. Cell culture in standard complete medium served as a control. For the DHT method, 4000 U-87 MG cells were seeded in a 35 mm µ-Dish (Ibidi, Gräfelfing, Germany) containing 600 µl of complete DMEM medium. The next day, the medium was discarded and the cells were treated with 100 µM G721–0282, HBSS or complete medium for 4 h.
Label-free digital holographic tomography (DHT)
To compare the dynamics of autophagy in glioblastoma cells line U-87 MG induced by HBSS and medium with 100 µM G721-0282 for 4 h the digital holotomography (DHT) cutting-edge fast, label-free quantitative imaging technique enabling the multiparametric phenotyping of living cells was applied. The series of digital holograms (DHs), formed by interference between the reference plane wave with object waves scattered by samples, were recorded using a commercial off-axis Mach-Zehnder interferometric setup with a rotatable scanning mirror (3D Cell Explorer, Nanolive, Tocholenz, Switzerland). DHs were acquired at 520 nm (sample exposure 0.2 mW/mm²) using a dry objective (60×, NA = 0.8, Nikon) for each scanning mirror position. Phase and amplitude data were extracted and processed to reconstruct 3D refractive index (RI) maps and their 3D visualizations using STEVE software (version 1.6.3496, Nanolive, Tocholenz, Switzerland). The RI is an important biophysical parameter directly related to the density and chemical composition of cell organelles. The technique used the RI that as a contrasting parameter allowing the visualization of cells and their organelles, which are characterized by different values of this physical quantity. Moreover, based on differences in RI values between organelles it is possible to use RI for digital staining and label-free characterization of single-cell and its organelles. In this study, three kinds of samples (control, HBSS, G71-0282) were examined. For each sample group, at least 20 3D-RI tomograms were recorded, each consisting of 96 2D-RI tomograms. Thus, at least 60 3D-RI tomograms and 5760 2D-RI tomograms were analyzed. Based on the recorded data, it was possible to extract quantitative measures characterizing the cells and the autophagy process occurring in them. In addition to morphometric parameters such as volume and maximum height of cell, additional quantitative measures were also determined, namely: the average refractive index of the entire cell, cytoplasm and nucleus with nucleoli, the average gradient of RI value changes, and average dry mass. Dry mass can be determined based on RI data using the following linear calibration model [28, 29]:
\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$DryMass=\frac{V}{\alpha}\left(\frac{{RI}_{S}}{{RI}_{medium}}-1\right)\left[g\right]$$\end{document}where α is the proportionality constant called the RI increment, V is the volume of organelle / cell, RI_s_ is the average value of the RI of the examined organelle / cell, and RI medium is the average value of the medium surrounding the cells. In this study, it was assumed that α equals 0.19 mL/g.
Spheroids formation
For the glioblastoma model, the U87-MG and HMEC-1 cells were mixed with macrophages and then centrifuged at 300 g for 10 min. The cells were then seeded at a density of 2 × 10³ cells per well (6 × 10³ cells per well), as previously described [14]. To form the spheroids, a low-adhesion 96-well plate (3D PrimeSurface^®^ 96 V, catalogue number MS-9096VZ, Akita Sunitomo Bakelite, Akita, Japan) and complete DMEM media (Capricorn) were used. After 48 h required for the formation, the spheroids were treated with a concentration of 100 µM of the G721-0282 compound for 72 h, after which the process was repeated for a further 72 h. Untreated spheroids (0 µM compound) in culture medium were used as a control.
Phantom for X-ray radiation
Four treatment plans were prepared in order to irradiate the research material with homogeneous doses of 2, 4, 8 and 16 Gy, based on tomographic sections of a tissue-like phantom containing the research samples. In the present study, an IBA RW3 phantom was used, which is typically used to evaluate the conformity of IMRT/VMAT (dose intensity modulation techniques) radiotherapeutic treatment plans with dosimetric reality on the therapeutic device. The material used to construct the phantom has a physical density of 1.045 g/cm³, comparable to the density of water. The proportions of the elements carbon (C), hydrogen (H), oxygen (O) and titanium (Ti) were determined to be 0.9041, 0.0759, 0.008 and 0.012, respectively. In the central part of the phantom (see Fig. 1A–C), cylindrical vials containing the research material were arranged in parallel so that the tissue-like material in the phantom surrounded them, simulating the clinical situation while ensuring homogeneous distribution of the absorbed dose.
Fig. 1. Designed tissue phantom for spheroid exposition to X-rays. Phantom diagrams: A- cross-section of the CT scan with samples; B- frontal CT reconstruction at the half-width of the samples; C- 3D reconstruction of the phantom with visualisation of the samples as the clinical target volume (CTV) and the planning target volume (PTV), marked with a red contour; D- geometry diagram of the radiation beams used; E- 3D reconstruction of the beam plan; F- the implementation of the prepared treatment plans involved the utilisation of electromagnetic radiation beams with a nominal energy, as defined in the plans and generated in the TrueBeam 4 linear accelerator (Varian, Palo Alto, CA, USA). The dose distribution is shown qualitatively (i.e. as an isodose distribution) in F. The dose distribution shown in G covers the area between 95% and 107%; H: the dose-volume histogram for structures. The CTV is indicated by the yellow line and the PTV by the red line
To enhance the spatial resolution of the images obtained and achieve more accurate mapping of the shape of the system’s individual elements, a tomographic examination was performed with a single-layer thickness of 0.15 mm. Following transmission to the Eclipse 18.0 treatment planning system (Varian), the CT images were subjected to a preparatory procedure to determine a uniform dose distribution, based on computational algorithms (ACUROS 18.0) and the electron densities of the tomographic sections.
The initial phase of treatment plan formulation involved delineating the regions necessary for calculating the dose distribution: the external contour that defines the calculation area (body outline), the CTV (Clinical Target Volume) area, which is the sum of the areas of the vials containing the research material, and the PTV (Planning Target Volume) area, which is the irradiation area that encompasses the CTV structure with an adequate margin. A conforming treatment plan was formulated in compliance with ICRU 83 guidelines [30]. The plan covered the PTV region, ensuring a uniform dose distribution by using three orthogonal beams of electromagnetic radiation, each with a nominal energy of 6 MeV. The configuration comprised two lateral beams measuring 6 × 22 cm and employing wedge modulation, as well as a front beam measuring 22 × 22 cm and lacking modifiers (Fig. 1D–E).
Four absorbed dose values consistent with clinical reality: 2 Gy, which is typical for conventional fractionation, doses of 4 Gy and 8 Gy, which are typical for palliative radiotherapy, a dose of 16 Gy, which is used for stereotactic radiotherapy and radiosurgery [31]. All spheroids exposed to X-ray radiation were previously incubated with 100 µM of G721-0282 for 72 h. Forty-eight hours after exposure to X-ray radiation, the spheroids were harvested for further analysis, as described below.
Slides preparation
The spheroids were then counterstained with haematoxylin for 4 min at room temperature (RT) and fixed in 4% paraformaldehyde (PFA) at 4 °C overnight. The next day, they were embedded in 2% agarose as previously described [14]. After this step, the spheroids were embedded in paraffin and all subsequent steps were carried out as routine for pathological tissue slides.
Preparations for transmission electron microscopy (TEM)
The spheroids were fixed in a solution of 2.5% paraformaldehyde in a 0.1 M cacodylate buffer solution (Serva Electrophoresis, Heidelberg, Germany) [14, 32, 33]. Subsequently, individual spheroids were encapsulated within approximately 50 µL of 2% agarose gel. Thereafter, the standard fixation and embedding protocol was implemented, as described before [34]. Briefly, the spheroids were washed in cacodylate buffer, post-fixed in 1% osmium tetroxide (Serva Electrophoresis), washed again, dehydrated in an ascending series of ethanol solutions (Stanlab, Lublin, Poland) and acetone, and then embedded in epoxy resin (Serva Electrophoresis). The spheroids were then sectioned into 600-nm semi-thin sections within the epoxy blocks and stained with toluidine blue to facilitate an initial assessment of the specimens. Approximately one-third to one-half of the way through the thickness of the spheroids, 60-nm ultra-thin sections were cut using the Ultra 45° Diamond Knife (Diatome, Nidau, Switzerland) and collected on rhodium-coated copper grids (Ted Pella, Redding, California, USA). The sections were then counterstained with a solution of uranyl and lead citrate (Delta Microscopies, France). The counterstained sections were analysed using a JEM-1011 transmission electron microscope (Jeol, Tokyo, Japan) at 80 kV. Images were acquired using a Morada camera (Olympus, Münster, Germany).
Immunohistochemistry
Immunohistochemical reactions were performed on 4 μm sections of spheroid material. First, PT-Link (Dako, Glostrup, Denmark) and EnVision FLEX Target Retrieval Solution (97 °C, 20 min; pH 9.0) were used for deparaffinization and antigen retrieval. The slides were then blocked in 1% BSA at room temperature (RT) for 30 min and endogenous peroxidase was blocked at RT for 5 min using EnVision FLEX Peroxidase Blocking Reagent (Dako). Incubation of the slides were provided (overnight in 4 °C) with primary anti-CHI3L1 antibody (rabbit polyclonal, 1:100; ab180569, Abcam, Cambridge, UK). For secondary antibodies, EnVision FLEX/HRP (RTU, Dako, 1 h, RT) mouse and rabbit antibodies was used. Primary antibody was diluted in 1%BSA in PBS/0.1%Tween20. Incubation with DAB substrate (Dako) was provided (10 min., RT) and then counterstained in EnVision Flex Hematoxilin solution (Dako) (5 min., RT). Dako Mounting Medium (Dako) was used to cover the slides [35–37].
Western blot
Lysates were prepared from spheroids in RIPA buffer containing 5 mM phenylmethanesulfonyl fluoride (PMSF), EDTA and Heat™ Protease Inhibitor Cocktail × 100 (all from Thermo Scientific, Wilmington, DE, USA). Total protein content was measured using a bicinchoninic acid assay (Pierce BCA Protein Assay Kit) and a NanoDrop 1000 (Thermo Fisher). The samples were denatured in a buffer consisting of 250 mM Tris at pH 6.8, 40% glycerol, 20% β-mercaptoethanol (v/v), 0.33 mg/ml bromophenol blue and 8% sodium dodecyl sulfate (SDS), for 10 min at 95 °C. SDS-PAGE electrophoresis was performed using a Mini Protean 3 instrument (Bio-Rad, Hercules, CA, USA) with a 10% polyacrylamide gel to detect the CHI3L1 (40 kDa) protein level and a 12% gel to detect the LC3B (14 kDa) protein level with 20 µg of protein was applied per lane [14, 38]. Next, wet transfer was applied using a Tris–glycine buffer containing 20% methanol and 0.05% SDS, with a PVDF membrane (Immobilon, Millipore, Bedford, MA, USA) with a pore size of 0.45 μm–0.2 μm for LC3B protein. Transfer was performed for 1 h at 140 V or 0.5 h and 70 V for LC3B. The blocking agent used was 5% skimmed milk powder in 0.05% TBST for all membranes. Primary antibody incubation was conducted overnight at 4 °C at the following concentrations: for CHI3L1 (rabbit polyclonal antibody, ab180569, Abcam), for LC3B (rabbit polyclonal antibody, NB600-1384, Novus Biologicals, Abingdon, UK) (all 1: 1000 in 5% skim milk in 0.05% TBST) and β-tubulin (rabbit polyclonal, 1:1000, ab6046, Abcam; 0.1% BSA in 0.1% TBST). Incubation with donkey anti-rabbit secondary HRP-conjugated antibodies (1:6000, Jackson ImmunoResearch, Suffolk, UK) was carried out for 1 h at room temperature (RT) in 5% milk in 0.05% TBST. Luminata Forte Immobilon Western HRP Substrate (Thermo Fisher Scientific) was used for the chemiluminescence reaction, followed by the ChemiDoc™ MP visualisation system (Bio-Rad) and ImageLab software (Bio-Rad) (exposure time from 1 s to 3 min). In the Western blot method for evaluating autophagy, the intensity of the LC3B II band was considered a signal for autophagosome accumulation when compared to the intensity level of the LC3B I band [39].
Optical photothermal infrared (O‑PTIR) spectroscopy
Each spheriod sample was analysed using the O-PTIR system mIRage-R (Phototermal Spectroscopy Corp., Santa Barbara, CA, USA) which was equipped with a 785 nm probe laser and was purged with dried air from an air purifier. Before O-PTIR analysis of spheroids on glass slides, xylene washing (15 min., RT) was used to remove paraffin. A 16 × 16 array of O-PTIR spectra were collected with a step size of 2 μm from a 30 × 30 μm field of view. The QCL was tuned with a speed of 1000 cm^− 1^/sec from the range 934 to 1800 cm^− 1^ and from 2700 to 2992 cm^− 1^ with auto focus option at 1660 cm^− 1^. The QCL power was set to 47% (0.76 mW at 1660 cm^− 1^), and the probe laser power to 37% (5.4 mW at 785 nm), and 16 scans were averaged for each spectrum. Before sample measurements, an automated background procedure was performed to optimize beam alignment and normalize QCL emission which also included residual water band suppression. Initial spectral analysis and export to the csv format were done with the PTIR Studio software ver. 4.6 (Phototermal Spectroscopy Corp.). RamanMetrix (Biophotonics Diagnostics, Germany) was used to analyze five data set with a total of 1280 spectra. After a quality test removed low quality spectra, second derivatives were calculated from the remaining 1158 spectra, vector normalized and subjected to principal component analysis in the spectral range 1800 –934 cm^− 1^.
Statistical analysis
The normality of the distribution was analysed using the Shapiro–Wilk test. Statistical analysis of spectral intensity was performed using One-Way ANOVA and the post hoc Duncan test, with a significance level of p < 0.05 (SPSS software version 30.0.0, IBM Corp., Armonk, NY, USA). Spectra analysis and processing were performed using Quasar 1.11.1 software [38]. Origin PRO 24 software (OriginLab Corp., Northampton, MA, USA) was used for the analysis and visualisation of RI-data based quantities.
Results
CHI3L1 inhibition induces moderate autophagy in glioblastoma cells compared to starvation
Autophagy was analysed by DHT and Western blot methods in U-87 MG glioblastoma cells treated with G721-0282, HBSS or complete medium. The presence of autophagy in U-87 MG cells treated with G721-0282 was confirmed by DHT examination, which indicated the existence of autophagosomes (see Fig. 2).
Fig. 2. Autophagy in U-87 MG glioblastoma cells. Comparison of different stages of autophagy in U-87 MG glioblastoma cells after 4 h of starvation (HBSS), as well as upon inhibition of CHI3L1 by 100µM of G721-0282, with control (cells in complete medium): A- representative 2D-RI maps and digitally stained 3D-RI maps of cells (yellow arrows indicate autophagosomes), B- example of an extracellular autophagosome observed in samples after starvation, C- changes in protein levels by Western blot
The DHT results obtained indicate that the samples contained cells at various stages of autophagy, and the dynamics of this process were significantly greater in starved cells. In these cells, areas with higher RI (RI > 1.37 for HBSS, RI > 1.35 for G721-0282 - violet regions on Fig. 2A) were observed, often in the perinuclear zone of the cytoplasm, which are likely to be specific structures occurring exclusively during autophagy [40]. This can be explained by early stage of autophagy, which involves the reorganization of intracellular compartments leading to localised increases in density (e.g. at ER-mitochondria contact sites, where phagophores are initiated) [41] generating local increase of RI gradient. In the case of cells treated with G721-0282, the RI maps obtained suggest an early stage of autophagy, in which autophagosomes are still forming inside the cells, near the cell nucleus, while in starved cells after 4 h, advanced autophagy is accompanied by morphological changes characteristic of early apoptosis, such as cell membrane wrinkling, and large autophagosomes (≈ 5 μm) are already present in the extracellular environment in this type of sample (Fig. 2B). These observations were also confirmed by the Western blot method (see Fig. 2C).
Western blot analysis based on the LC3B II/I ratio after four hours of incubation showed increased autophagy in U-87 MG cells treated with 100 µM of G721-0282 compared to control cells cultured in a complete medium. The highest active CHI3L1 level also occurred in control cells (Fig. 2C). Furthermore, the expression level of CHI3L1 corresponded with the level of p-STAT3. In U-87 MG cells treated with HBSS buffer, starvation conditions led to the inhibition of CHI3L1 protein expression. A low LC3B II/LC3B I ratio and low LC3B intensity may indicate the advanced, late phase of autophagy, where autophagosome degradation occurs [39]. This same pattern of a weak signal from LC3B II but high autophagy up-regulation was observed in Western blot analysis of spheroids trated with X-ray radiation. Analysis of these spheroids in TEM confirmed the low LC3B II/I ratio observed at higher dose of radiation with autophagocytosis.
The significant differences in the dynamics of autophagy induced by starving and G721-0282 were also confirmed by analysis of the quantitative measures extracted from the RI data (see Fig. 3), which are consistent with the results presented in [25]. With advancing autophagy, an increase in the average RI value of the entire cell, cytoplasm and nucleus can be observed (Fig. 3A), which is associated with a decrease in the spatial size of cells during phagocytosis and an increase in the dry mass of their organelles (Fig. 3B). This also leads to an increase in the average RI gradient of whole cells (Fig. 3C), directly related to an increase in their density in relation to the medium in which they are located. Furthermore, analysis of morphometric parameters showed that progressive autophagy leads to a decrease in cell volume (Fig. 3D), while increasing the maximum width of cells (Fig. 3E), which explains the increase in cell density. The quantitative analysis of the obtained parameters shows that the dynamic of autophagy is significantly greater in starved cells than in those treated with G721-0282, confirming previous results.
Fig. 3. Autophagy in U-87 MG glioblastoma cells as quantitative analysis of: A- refractive index, B- dry mass, C-average RI gradient, D-volume and E- maximum height of examined cells/organelles. The U-87 MG glioblastoma cells after 4 h of starvation (HBSS), as well as upon inhibition of CHI3L1 by 100µM of G721-0282, with control (cells in complete medium). Analysis was done in Origin PRO 24 software
X -radiation influence CHI3L1 expression
Exposure of glioblastoma spheroids to X-rays after pre-treatment with 100 µM G721-0282 resulted in changes in CHI3L1 expression, as observed using the Western blot method, compared to the untreated control. However, this effect was not linear (Fig. 4A). The highest level of CHI3L1 was observed in spheroids treated with 2 Gy; no CHI3L1 expression was observed in spheroids treated with 4 Gy. In the control group, which was neither exposed to X-rays nor treated with G721-0282, the level of CHI3L1 protein was detectable, albeit at a lower level compared to the spheroids treated with 2, 8 and 16 Gy (Fig. 4A). Changes in CHI3L1 expression levels in the spheroids were also observed in immunohistochemical reactions. Moreover, cavities in the spheroids were mostly observed when a 2 Gy dose was applied; however, cavities also occurred after a 16 Gy dose (Fig. 4B).
Fig. 4. Changes in the expression of the CHI3L1, LC3B and pSTAT3 proteins in spheroids treated with of G721-0282 and X-ray radiation; A- Western blot method: control was untreated with G721-0282 and X; B - Immunohistochemical analysis of CHI3L1 expression in spheroids treated; spheroids were treated with 100 µM of G721-0282 for 72 h, followed by various doses of X-ray radiation (magnification 200x).
X–ray radiation modulate autophagy when CHI3L1 is inhibited
In spheroids treated with X-ray radiation and pre-treated with 100 µM of G721-0282, the level of autophagy, as assessed by the intensity of the LC3B II band relative to the LC3 I band and detected by Western blot analysis, was highest when a dose of 2 Gy was applied (Fig. 4A). In this group of spheroids, CHI3L1 expression levels were also the highest. At higher doses of 4 and 8 Gy, the autophagy process also occurred, but at a lower rate than at the 2 Gy dose. At the highest dose of 16 Gy, fast autophagosome degradation was probably observed at the latest stage of autophagy, as evidenced by the decreased difference between LC3B I and LC3B II (Fig. 4A). In control, untreated spheroids, autophagy occurred at a lower physiological level. Activation of STAT-3 was observed in spheroids treated with X-ray radiation at 2, 8 and 16 Gy, whereas at 4 Gy it was observed only at a very low level and was not observed in the untreated control. The expression level of CHI3L1 corresponded with the level of p-STAT3, and no expression of CHI3L1 or p-STAT3 was observed after treatment with 4 Gy or in the untreated control spheroids. At the higher doses of 8 and 16 Gy, a weak signal from LC3B II indicates the late phase of autophagy with autophagosome degradation (Fig. 4A). Conducted immunofluorescence detection of apoptosis on spheroids slides with Annexin V and propidium jodide assay revealed that there was no apoptosis in any of the groups of spheroids. (Fig. S1.)
Autophagosomes are accumulated in spheroids during CHI3L1 inhibition and X-ray radiation
Representative images obtained during TEM observations are shown in Fig. 5. The spheroids consisted of three cell lines, some of which could be distinguished in TEM because they exhibited characteristics similar to those of cells cultured separately [32, 42–44]. This type of spheroids we previously described in TEM analysis [14]. The most pronounced differences visible between the control group and the group with highest dose of X-ray radiation included the extent of endocytosis (mostly macropinocytosis) and the number of autophagic vesicles. Figure 5AB shows intense pinocytosis with numerous macropinosomes, particularly in HMEC-1 cells. Notably, unequivocal signs of autophagy are not visible. As the X-ray dose increases, some autophagosomes become visible (Fig. 5C–F). The highest dose of X-ray radiation show notably decreased pinocytosis and the most visible autophagosomes in all cell types (Fig. 5G–H).
Fig. 5TEM analysis of cell morphology in spheroids. Cells without additional treatment show signs of extensive macropinocytosis (arrowheads) A - and B- particularly in HMEC-1 cells (letter H); C and D- Some autophagosomes (arrows) appear in HMEC-1 and U87-MG cells (letter U), and their number increases with the exposition to X-ray radiation (E and F). In these cells, both autophagy and macropinocytosis occur simultaneously. For spheroids with the highest dose of X-ray radiation (G and H), macropinocytosis is significantly reduced, and the switch to autophagy is more pronounced. Spheroids were treated with 100 µM of G721-0282 for 72 h, followed by various doses of X-ray radiation. Letter H—HMEC-1; letter U—U87-MG; letter M—T macrophages. Arrowheads—macropinocytosis at different stages; arrows—autophagosomes.
Biochemical component changed mostly at 2 Gy dose and CHI3L1 inhibition
Main variations in normalized O-PTIR spectra were evident as background slopes and broad spectral contributions of the glass substrate between 934 and 1200 cm^− 1^. These variations are suppressed in second derivative spectra in which minima represent maxima of bands in original spectra. Beside suppression of broad variations, overlapping subbands are better resolved in second derivative spectra which are assigned to protein amide I at 1695, 1658 and 1630 cm^− 1^, protein amide II at 1540 and 1515 cm^− 1^, hydrocarbons at 1456 and COO at 1388 cm^− 1^ of amino acids side chains in proteins, and protein amide III at 1250 cm^− 1^ [45]. Spectral contributions of lipids are neglible after to deparaffinization by xylol washing. Spectral contributions of nucleic acids, carbohydrates and other biomolecules are weak below 1250 cm^− 1^. Negative PC1 scores separate the A100 µM 2 Gy data from the other data. This is consistent with positive PC1 loading bands at 1655, 1630 and 1540 cm^− 1^ that are more negative in second derivative of A100 µm 2 Gy than in the other derivatives. More loading bands occur at 1470/1460 and 1235/1206 that are inconclusive because they agree with low intensities at the QCL chip transitions. PC2 and PC3 loadings represent smaller variations that did not contribute significantly to the separation in PC2 and PC3 scores. Figure 6.
Fig. 6O-PTIR analysis of spheroids treated with G721-0282 and X-ray radiation. A- O-PTIR spectra normalized to amide I; B- second derivatives used as input for principal component analysis (PCA); C- PCA-scatter plots of PC1 to 3; D- loading plots of PC1 to 3. Analysis provided with RamanMetrix software
Discussion
The role of CHI3L1 in the regulation of autophagy in GB is currently unknown. Autophagic upregulation in this tumour is observed during TMZ treatment and leads to a chemoresistant GB phenotype. For this reason, autophagy may be considered a therapeutic target in GB; however, the pro-tumour and also anti-tumour aspects of this process must be considered, as autophagy has two faces in GB [2, 3]. Inhibition of autophagy is connected with sensitisation to TMZ treatment. However, induction of autophagy may stimulate immunisation and glioblastoma cell death, thereby leading to sensitisation to chemotherapy [4, 46]. The CHI3L1 protein plays a role in inducing autophagy in lung cancer, as demonstrated by Hong and collaques in the studies performed on in vivo and in vitro lung cancer models as well as on lung cancer tissue [47]. The contribution of the CHI3L1 protein to autophagy was recently demonstrated by Li and colleagues in an in vivo model of enterocolitis, intestinal cell cultures and samples derived from neonatal patients with necrotising enterocolitis [48]. These results showed that inhibiting CHI3L1 leads to the downregulation of autophagy in intestinal epithelial cells [48]. Overexpression of the CHI3L1 protein is observed in glioblastoma patients and is associated with a poor prognosis. Recent studies by Zhou et al. also showed CHI3L1’s involvement in modulating oxidative stress-related genes, supporting its role as a biomarker and target for targeted therapy [49].
In contrast, our findings suggest that the CHI3L1 protein may have different role in this process. Evaluation of autophagy in U-87 MG cells using Western blot and DHT methods showed that inhibition of the CHI3L1 protein induces up-regulation of autophagy. Compared to cells cultured in a complete medium, inhibiting CHI3L1 for four hours using the G721-0282 compound at a concentration of 100 µM increased autophagy. In the standard autophagy induction model, in which the cells are starved in HBSS buffer [25, 27], the advanced late phase of autophagy occurred at this same time point. Autophagy up-regulation was indicated in Western blot method as a increasing LC3B II/I ratio, which is one of the main autophagy markers in immunoblotting analysis [39, 50].
These results were confirmed quantitatively in DHT method, where increasing the average RI value of the entire cell is observed with advancing autophagy [40, 41]. In DHT, cellular starvation in HBSS buffer was the strongest autophagy inducer, but G721-0282 also induced autophagy at an early stage. U-87 MG cells morphology reflected differences between cells treated with G721-0282, HBSS buffer and complete medium, where more round cell shape was associated with autophagy up-regulation.
Moreover, starvation of U-87 MG cells for only four hours was also associated with a complete decrease in CHI3L1 levels. These findings were observed as increasing LC3B II/I ratio in Western blot method and suggest that CHI3L1 expression may be precisely, dynamic regulated by environmental conditions such as nutrient depletion. Decreased CHI3L1 expression or activity as a result of G721-0282 treatment was also associated with upregulation of autophagy, also observed in Western blot method. This findings suggest that CHI3L1 may be an independent autophagy protective factor in glioblastoma cells. The function of CHI3L1 in glioblastoma cells has not yet been established.
Furthermore, the highest CHI3L1 levels in glioblastoma U-87 MG cells was associated with the highest levels of p-STAT3, showed in Western blot method, which suggests a role for CHI3L1 in activating the STAT3 pathway. We observed this effect in our previous study on the anticancer role of CHI3L1 inhibition by G721-0282 [14]. Guetta-Terrier and colleagues demonstrated the participation of CHI3L1 in STAT-3 phosphorylation in glioma stem cells (GSCs). The authors also suggested that CD44 is a receptor for CHI3L1 and demonstrated the impact of the CHI3L1 protein on AKT and β-catenin phosphorylation and nuclear translocation [51]. STAT-3 is involved in the transformation of glioblastoma cells into the mesenchymal subtype, and has potential to improve GBM treatment [52–54].
Recently, CHI3L1 inhibitors have been extensively investigated using the glioblastoma spheroid model. One of these inhibitors, the K284 compound and its chemical modification, was presented by Kaur and colleagues [55]. In this paper, the authors revealed that the 11G compound decreased the migration of glioblastoma cells and the viability, mass and size of spheroids, and therefore may be a new compound that could potentially be used in targeted therapy for GBM [55].
In our previous studies, we present the broad anti-cancer effects of CHI3L1 inhibition using the G721-0282 compound [12, 14]. Shortly, we demonstrated that this inhibition led to decreased angiogenesis and motility of glioblastoma cells in in vitro assays, as well as changes in the cytokine background and membrane protein expression, such as VCAM-1, which is involved in the EMT (epithelial-mesenchymal transition) pathway. Building on this, the dependence of autophagy on CHI3L1 in response to treatment with the CHI3L1 inhibitor G721-0282 in glioblastoma cells appears to offer a new approach to modulating anti-cancer treatment.
In our spheroid model, which served as an in vitro model of a GB tumour in this study, three cell types were present: U-87 MG glioblastoma cells, HMEC-1 endothelial cells, and macrophages [12, 14], the effect of autophagy in relation to the CHI3L1 protein was also observed in Western blot method. The highest level of autophagy based on increasing LC3B II/I ratio was detected when the highest level of CHI3L1 protein was observed following the application of a 2 Gy dose of X-ray radiation. The highest level of CHI3L1 in these conditions may have occurred as a compensatory effect caused by the biological inactivation of the CHI3L1 protein, as the same effect was previously observed [14] or may also be induced by X-ray radiation. Surprisingly, CHI3L1 ptotein expression was not detectable in spheroids treated with a 4 Gy dose, but up-regulation of autophagy observed as an increasing LC3B II/I ratio occured compared to untreated spheroids (no inhibitor, no X-ray radiation). In Western blot method we observed up-regulation of CHI3L1 level as a results of X-ray radiation treatment, except dose of 4 Gy. Changes in CHI3L1 level was observed also in immunohistochemical reactions. The absence of CHI3L1 in the spheroids treated with a 4 Gy dose showed that the expression of this protein is not constitutive in the glioblastoma cells, but rather is regulated by the external environment. This effect of blocked CHI3L1 expression was consistent with our observation during starvation in U-87 MG glioblastoma cells. These findings suggest that glioblastoma cells in spheroids may also switch off CHI3L1 protein expression under unfavourable conditions. Under these conditions, such as starvation or X-ray radiation in our experiments, down-regulation of CHI3L1 protein expression coexists with up-regulation of autophagy. For the other side, we observed that CHI3L1 inhibition also leads to autophagy upregulation. All this together suggesting the role of this protein in protecting glioblastoma cells from autophagy.
According to this findings we found in TEM analyses the upregulation of autophagy in spheroids increased with X-ray radiation dose. This indicates that autophagy was up-regulated by X-ray doses independently of CHI3L1 levels. It is well known that radiotherapy strongly induces autophagy [3]. Ultrastructural analysis alo showed that apoptosis was not detected in any of the spheroid groups, and this findings were also confirmed by directly apoptosis detection with immunofluorescence assay, hovewer TEM is the most classic and reliable method for apoptosis and allow to distinguish between apoptotic and necrotic cells, as well as auophagosome structures [56, 57]. These results show that the application of X-ray doses and the inhibition of CHI3L1 lead to autophagy, but not apoptosis. They also reveal the mechanism of CHI3L1 involved in this process.
In the case of spheroids treated with X-rays, the CHI3L1 protein activity was inhibited by G721-0282, and even when the expression level of this protein was higher, the biological function of CHI3L1 was blocked. Therefore, in all spheroids treated with various doses of X-ray radiation, the upregulation of autophagy was higher compared to untreated spheroids and may be improved by X-ray radiation, which is known to induce autophagy [3] and CHI3L1 inhibition. It is noteworthy that our experiments showed that X-radiation changes the expression of the CHI3L1 protein depending on the radiation dose. This clearly indicates a lack of this protein when a 4 Gy dose is applied, and an increase in CHI3L1 levels after 2, 8 and 16 Gy doses, observed in Western blot method. In spheroids treated with 4 Gy and lacking CHI3L1 expression and treated with 2 Gy with CHI3L1 blocked with inhibitor, autophagy was the highest compared to the other groups. These results are consistent with observations from U-87 MG cells and suggest that the CHI3L1 protein may serve as an autophagy protector and an independent anti-autophagic factor in GB spheroids. Similarly to the results obtained in experiments conducted on U-87 MG cells, the downregulated level of CHI3L1 in GB spheroids observed in Western blot methods and immunochistochemical reactions was also associated with the decreased level of p-STAT3 and may suggest a role for CHI3L1 in activating STAT-3 in the autophagy pathway in GB. Hong and colleagues observed that CHI3L1 induces autophagy in lung cancer by activating the MAPK/JNK pathway, which is known to be involved in the upregulation of autophagy [47, 58]. In the enterocolitis studies conducted by Li and colleagues, the inhibition of CHI3L1 was found to be associated with the downregulation of autophagy by reducing the activation of the PI3K/AKT pathway [48]. The CHI3L1 protein may activate different signalling pathways depending on the type of cancer. In GB, CHI3L1 is mostly connected with the activation of the STAT3 pathway [59].
It is known that autophagy occurring during radiotherapy or chemotherapy leads to a GB resistance phenotype [3]. Huang and colleagues found that bevacizumab induces autophagy and leads to chemoresistance in GB tumours by downregulating the AKT-mTOR signalling pathway. Furthermore, the use of chloroquine to inhibit autophagy made GB cells more susceptible to Bevacizumab [60]. Clinical trials currently focus on inhibiting autophagy using hydroxychloroquine (HCQ) or chloroquine (CQ) alongside standard therapy for patients with newly diagnosed or recurrent glioblastoma. A phase I/II clinical trial evaluating the use of hydroxychloroquine (HCQ) in combination with standard therapy in patients with GB showed that a daily dose of 600 mg of HCQ did not consistently inhibit autophagy, nor did it result in a significant improvement in overall survival. Furthermore, a daily dose of 800 mg of HCQ caused significant adverse effects in many patients, suggesting that this compound has limited applicability in therapy due to its toxicity and the difficulty of achieving effective autophagy inhibition [61]. In the case of the study involving CQ, a dose of 200 mg per day was established as the maximum tolerated dose (MTD) when combined with radiotherapy and TMZ in patients with newly diagnosed GB [62]. While the results suggest potential clinical benefits, further studies are needed to confirm therapeutic efficacy. The studies highlighted the need to develop new, more selective and safer autophagy inhibitors due to limitations related to toxicity and inconsistent autophagy inhibition. These inhibitors could be effectively used in the treatment of GB.
Autophagy inducers, on the other hand, have also shown promising effects in preclinical studies. ABTL0812, an oral anticancer compound which is currently in the second phase of clinical trials, has demonstrated potential in the treatment of glioblastoma multiforme (GB) by inducing autophagy-mediated cell death and inhibiting cancer cell proliferation [63]. The results of this study demonstrated the effectiveness of this compound in models of GB, both as a monotherapy and in combination with standard therapy. This supports further investigation of the compound as a potential novel therapy for this aggressive tumour type.
Based on these findings, it is crucial to determine whether autophagy primarily plays a protective role (in which case the use of inhibitors would be beneficial) or a pro-apoptotic role (in which case the use of inducers would be more effective). In the context of our findings, CHI3L1 inhibition leading to autophagy upregulation is in accordance with these new therapeutic possibilities for GB.
In our studies, we observed cavities in the spheroids that may be the result of autolysis or apoptosis caused by X-ray radiation, whereby dead cells are washed away during slide processing. Notably, the most visible cavities occurred after 2 Gy of X-ray radiation. This is an important observation that confirms the relevance of our X-ray radiation and spheroids model. For GB radiotherapy, the typical dose for adult patients is 60 Gy in fractions of 2 Gy [31, 64].
The evaluation of the biochemical components in the spheroids provided by the O-PTIR method showed that the 2 Gy dose caused the most significant biological effect compared to the higher doses. These findings are consistent with the results obtained using the Western blot method, in which the highest level of autophagy was observed following treatment of the spheroids with 2 Gy. Analysis using the O-PTIR method served as a label-free screening method for the most biologically effective treatment in our studies and allowed differences between spheroid groups to be distinguished in a relatively short time. This method can be used to analyse general biochemical changes in a spheroid model in response to targeted therapy.
Taken together, our results suggest that combining CHI3L1 inhibition with G721-0282 and X-ray radiation at a dose of 2 Gy may be a new approach to designing GB treatment, which could be explored further in the future, especially in the context of using the 2 Gy dose that is typically used for conventional fractionation in radiotherapy.
Limitation of the study
In experiments involving X-ray radiation (2–16 Gy), the addition of an inhibitor blocked the activity of the CHI3L1 protein. However, no assay was performed to directly investigate the activity of the protein under these conditions due to the fact that the CHI3L1 protein has no enzymatic activity. Undoubtedly, investigating the kinetics of CHI3L1 and G721-0282 may provide valuable insights into CHI3L1 inhibition and interactions with X-ray radiation. In silico analysis of CHI3L1 and G721-0282 interactions in a molecular dynamics and docking model has been presented in our previous work [12]. In addition, the radiosensitivity assay demonstrated that the ratio of live glioblastoma cells following X-ray radiation treatment may provide a rationale for the impact of glioblastoma cell sensitisation subsequent to CHI3L1 targeting alone or together with TMZ treatment. This research direction will be continued by our team.
Supplementary Information
Below is the link to the electronic supplementary material.
Supplementary Material 1
Supplementary Material 2
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Machałowski T, Rusak A, Wiatrak B, Haczkiewicz-Leśniak K, Popiel A, Jaroszewicz J, Żak A, Podhorska-Okołów M, Jesionowski T (2021) Naturally formed chitinous skeleton isolated from the marine demosponge Aplysina Fistularis as a 3d scaffold for tissue engineering. Materials 14. 10.3390/ma 14112992