Depletion of chitinase-3-like protein 1 (CHI3L1) in glioma cells restraints tumor growth and affects neovasculature in intracranial murine gliomas

doi:10.21203/rs.3.rs-3208637/v1

Download PDF

Article

Depletion of chitinase-3-like protein 1 (CHI3L1) in glioma cells restraints tumor growth and affects neovasculature in intracranial murine gliomas

https://doi.org/10.21203/rs.3.rs-3208637/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Chitinase-3-like protein 1 (CHI3L1) is a secreted, non-enzymatic glycoprotein that binds proteins and carbohydrates and interacts with cell-surface and extracellular-matrix proteins, proteoglycans, and polysaccharides. Multiple interacting partners of CHI3L1 make dissection of its functions challenging. While many studies reported an upregulation of CHI3L1 mRNA/protein in various tumors, its exact roles in tumorigenesis remain elusive. We performed a comprehensive analysis of CHI3L1 expression in multiple public datasets including TCGA and single-cell RNAseq datasets to determine the cellular source of CHI3L1 expression in gliomas. The highest CHI3L1 mRNA/protein levels were detected in glioblastoma (GBM), a high-grade diffusive brain tumor. CHI3L1 knockout in human U87-MG glioma cells grossly affected transcriptional profile and in vitro invasiveness of these cells and strongly reduced the growth of intracranial U87-MG tumors in athymic mice. Remarkably, CHI3L1 knockout in glioma cells resulted in normalization of tumor vasculature and diminished infiltration of glioma-associated myeloid cells. Mechanistically, CHI3L1 depleted cells had reduced MMP2 expression/activity, which was associated with reduced invasion; and downregulated SPP1 (osteopontin), a crucial factor driving myeloid cell accumulation in GBM. Altogether, we demonstrate that CHI3L1 is a key player in GBM progression, and its targeting represents a novel strategy to treat GBM patients.

Biological sciences/Cell biology/Mechanisms of disease

Biological sciences/Cancer/Cancer microenvironment

CHI3L1

YKL-40

glioblastoma

genome editing

extracellular matrix

neoangiogenesis

Chitinase-3-like protein 1 (CHI3L1, YKL-40) is a secreted non-enzymatic 40 kDa glycoprotein that binds proteins and carbohydrates, including chitin, heparin, and hyaluronic acid (1). The ability of CHI3L1 to bind proteins and carbohydrates permits interactions with a spectrum of cell-surface and extracellular matrix proteins, proteoglycans, and polysaccharides. Several interacting partners of CHI3L1 play multi-facet roles in physiology and disease, making CHI3L1 signaling network difficult to apprehend. CHI3L1 interacts with IL13Rα2, CD44v3/IL13Rα2 complex, syndecan-1 and integrin α_Vβ₅ and signals through multiple pathways involved in allergic reactions, wound healing, inflammation resolution, epithelial-to-mesenchymal transition (EMT), tumor metastasis and angiogenesis (1–5).

CHI3L1 mRNA/protein levels are elevated in diseases characterized by chronic inflammation and tissue remodeling such as cancer. Its principal source are activated macrophages, chondrocytes, neutrophils, and cancer cells (5). Increased serum CHI3L1 levels correlate with malignancy, poorer prognosis, and shorter overall survival in breast, colon, prostate, ovaries, brain, thyroid, lung, and liver cancer patients (6). CHI3L1 affects the activity of matrix metalloproteinases (MMPs), influencing cell adhesion and migration, tissue remodeling, fibrosis, and tumorigenesis (3). Recent studies implicate CHI3L1 in the regulation of immune checkpoint inhibitors expression and melanoma progression (7).

In the brain, CHI3L1 upregulation occurs during neuroinflammation, mostly in reactive astrocytes and macrophages but not in microglia (8,9). CHI3L1 protein was elevated in 65% cases of glioblastoma (GBM), the most common and lethal primary brain tumor with a median overall survival of 15 months (10,11). CHI3L1 overexpression in the Cancer Genome Atlas (TCGA) glioma dataset was identified as one of markers of the mesenchymal subtype, the most aggressive GBM subtype (12). CHI3L1 serum levels were substantially elevated in numerous GBM patients and correlated positively with higher tumor grade and tumor burden (10). Higher CHI3L1 expression was associated with poorer radiation response, shorter time to progression and shorter overall survival (13).

CHI3L1 affects endothelial tubulogenesis similarly to vascular endothelial growth factor (VEGF) (14) and increases VEGF expression by syndecan-1-integrin α_Vβ₅-pFAK signaling in human U87-MG glioma cells. Elevated CHI3L1 levels were associated with augmented VEGF in subcutaneous U87-MG and SNB-75 gliomas (15). CHI3L1 counteracts the effects of VEGF inhibitor bevacizumab in GBM, likely via vascular mimicry, production of angiogenic factors and modulation of pericyte coverage of blood vessels (16). The exact role of CHI3L1 in GBM vascularization remains poorly understood.

We performed a comprehensive analysis of CHI3L1 expression in TCGA and single-cell glioma datasets and glioma cell lines to pinpoint the cellular source of CHI3L1 and resolve its roles. We evaluated the effects of CHI3L1 knockout in human U87-MG glioma cells on transcriptomic patterns, tumor cell proliferation, migration, and invasion in vitro. We demonstrate reduced growth of intracranial CHI3L1 KO gliomas in nude mice along with normalization of tumor vasculature and decreased infiltration of glioma associated myeloid cells. The results show the important role of CHI3L1 in glioma progression.

CHI3L1 is highly overexpressed in glioblastoma and is associated with the mesenchymal subtype markers

We determined CHI3L1 expression in glioma bulk and single-cell RNA sequencing (scRNAseq) datasets. Comparison of CHI3L1 expression in various tumors in the GEPIA database (17) shows that CHI3L1 expression is highest in GBM among 32 cancer types (Fig. 1A) with 63.5 times higher expression in GBM vs. normal brain (Supp. Figure 1). The quantitative PCR (RT-qPCR) analysis of CHI3L1 expression in the 76 gliomas dataset (18) shows upregulated expression of CHI3L1 in malignant (GBM) and benign gliomas (pilocytic astrocytoma) compared to normal brain tissue (Fig. 1B). Similarly, high CHI3L1 expression was detected in the bulk-RNAseq dataset of 31 gliomas dataset (19) (Fig. 1C). The analysis of the published scRNAseq dataset of 28 IDH1 wild-type GBMs shows that CHI3L1 was expressed predominantly by malignant cells, with some expression in glioma-associated macrophages (Fig. 1D-E). Interestingly, CHI3L1 expression positively correlated with the mesenchymal signature (MESlike1), the aggressive transcriptional subtype of GBM (Fig. 1E-F). We cross-checked the genes listed in the MESlike1 signature with the proteins correlated with CHI3L1 in the Brain Protein Atlas database (20,21). Out of 27 genes assigned to the MESlike1 signature, 25 correlated positively with CHI3L1 at a protein level in GBMs (Fig. 1G).

CHI3L1 depleted human glioma cells show altered expression of genes involved in extracellular matrix reorganization and cell adhesion

To evaluate the effect of CHI3L1 absence, CRISPR/Cas9 genome editing was applied and human glioma cells depleted of CHI3L1 (CHI3L1 KO) were generated (Fig. 2A). U87-MG cells exhibited high CHI3L1 protein level compared to normal astrocytes and were selected for further experiments (Fig. 2B). While WG9 cells showed highest CHI3L1 mRNA/protein levels, they failed to establish tumors in recipient mice (data not shown). We used the Cancer Dependency Map portal (22,23) and found U87-MG to have the highest gene effect (a negative parameter predicting the effect of gene knock-out on cell viability) for CHI3L1 depletion among the commonly used human glioma cell lines (Fig. 2C).

U87-MG cells were transfected with a commercial Cas9-GFP plasmid and validated guide RNAs specific for two sites in the CHI3L1 gene. Transfected cells (GFP+, Suppl. Figure 2A) were sorted individually and expanded clonally. Multiple independent clones were subjected to RT-qPCR and ELISA to evaluate efficacy of knockout (Suppl. Figure 2B-C). Two clones with the lowest CHI3L1 expression (KO1 and 2) were utilized further (Fig. 2D-E).

We performed RNAseq on wild-type (WT) and two CHI3L1 KO clones. A Venn diagram shows numbers of commonly up- or down-regulated genes among DEGs in KO1 and KO2 clones compared to WT cells (Fig. 2F). Around 50% of DEGs were affected in both KO cell lines. A Volcano plot shows DEGs in KO1 cells (Fig. 2G) and Reactome pathway enrichment analysis reveals that many genes down-regulated in CHI3L1 KO clones belong to ECM organization, migration, and invasion pathways (MMP2, MMP24, BMP1, PCOLCE, PCOLCE2, COL5A2, COL1A1, ADAMTS2, ADAMTS9, EFEMP1, TGFB3) (Fig. 2H).

Genes highly up-regulated in CHI3L1 KO cells included genes associated with formation and re-organization of ECM and cell-to-cell adhesion (FBLN1, PXDN, MMP1, PCDHB8) (Fig. 2G). Genes involved in PD-1 and IL-10 signaling were downregulated in CHI3L1 KO cells. CSF1 encoding macrophage colony stimulating factor or TGFB3 encoding transforming growth factor 3 were consistently downregulated in both clones (Fig. 2H). Some of the genes determined as mesenchymal-associated in single-cell database (24): ANXA1, EFEMP1, SPP1, NPC2, TNFRSF1A were most down-regulated in CHI3L1 KO compared to WT cells (Fig. 2G-H).

We cross-checked DEGs in CHI3L1 KO cells with the Brain Protein Atlas study (20,21) and found a positive correlation between CHI3L1 and proteins encoded by genes down-regulated in CHI3L1 KO cells (EFEMP1, ANXA1, NPC2, SERPINH1 and HLA-DRA) (Fig. 2I).

Altogether, CHI3L1 KO cells downregulated genes involved in ECM reorganization, invasion, and immunosuppression.

Glioblastoma cells depleted of CHI3L1 form smaller tumors in recipient mice

CHI3L1 KO glioma cells implanted intracranially into Foxn1^nu mice developed significantly smaller tumors compared to WT controls as demonstrated by magnetic resonance imaging (MRI) 21 days post-implantation (Fig. 3A-B). Mice in CHI3L1 KO group gained weight during tumor growth, which is an indicator of a lesser tumor burden in this group (Fig. 3C). Tumor volume and body mass of tumor-bearing animals were negatively correlated (Fig. 3D).

Human CHI3L1 (hCHI3L1) was detected in blood serum of mice bearing WT tumors and observed to drop substantially in mice bearing CHI3L1 KO tumors (Fig. 3E, left). hCHI3L1 serum level positively correlated with tumor volume in the WT group (Fig. 3E, right). Mouse CHI3L1 was detectable in the serum of naive mice and increased in mice bearing WT tumors but not CHI3L1 KO tumors (Fig. 3F, left). mCHI3L1 levels did not correlate with tumor volume (Fig. 3F, right).

Importantly, diminished tumor growth in KO group was not due to reduced cell proliferation, as WT and CHI3L1 KO U87-MG cells had comparable proliferation rate in vitro (Fig. 3G-H). Adding the recombinant hCHI3L1 to culture media of CHI3L1 KO glioma cells did not increase CHI3L1 KO cell proliferation (Fig. 3I). Thus, extracellular CHI3L1 does not control glioblastoma proliferation in vitro. Notably, CRISPR/Cas9 manipulation did not affect proliferation of U87-MG cells (Suppl. Figure 2D).

Glioma-derived CHI3L1 enhances invasion and myeloid cell infiltration into the tumor

CHI3L1 KO cells showed lowered expression of MMP2, coding for metalloproteinase 2 (Fig. 4A). We performed gelatin zymography and found a significantly lower gelatinolytic activity in conditioned media from CHI3L1 KO1 glioma cells compared to WT cells (Fig. 4B). We and others had demonstrated that glioma invasion is strongly enhanced by microglial cells (25–27). Invasion of WT and CHI3L1 KO (KO1) glioma cells in the presence/absence of human HM-SV40 microglial cells was determined with a matrigel assay. We found reduced invasion of CHI3L1 KO cells in microglia presence (Fig. 4C).

Brain resident microglia and monocyte-derived macrophages accumulate and are reprogrammed by GBM to support tumor progression, immunosuppression, and resistance to therapy (28,29). One of the reprograming factors in GBM is SPP1 (osteopontin) (25). The expression of SPP1 was significantly down-regulated in CHI3L1 KO glioma cells (Fig. 4D), which was corroborated on protein level (Fig. 4E). We determined the infiltration of myeloid cells in intracranial gliomas by staining for IBA1, a pan-myeloid cell marker. We found the density of IBA1 + cells significantly decreased in CHI3L1 KO tumors compared to WT tumors (Fig. 4F-G).

Depletion of glioma-derived CHI3L1 results in normalization of tumor vasculature

GBM is one of the most highly vascularized solid tumors and it is characterized by an extensive proliferation of microvessels (30,31), which leads to a high microvessel density (MVD). We compared the vasculature of WT and CHI3L1 KO U87-MG tumors by staining for von Willebrand Factor (vWF), an activated endothelium marker (Fig. 5A). The gross vasculature differed considerably in WT and CHI3L1 KO tumors; with more numerous non-capillary vessels (Fig. 5C left) and lesser MVD in the latter (Fig. 5C right). The number of vessels did not correlate with the tumor volume (Suppl. Figure 3A-D), suggesting that the increase in non-capillary vessels in CHI3L1 KO tumors was independent of tumor volume but resulted from CHI3L1 depletion instead.

We examined the structure of the blood vessel walls by staining tumor sections for laminin, an endothelium basement membrane marker (Fig. 5B). We observed that in CHI3L1 KO gliomas blood vessels were continuously lined with laminin in contrast to WT tumors. Laminin immunopositivity significantly increased in CHI3L1 KO gliomas vs. WT (Fig. 5D). The integrity of morphologically abnormal vessels was assessed using endothelium CD31 staining and 3D rendering. CD31 + vessels formed continuous tubes in CHI3L1 KO tumors in comparison to those in WT group (Fig. 5E). These findings show that in the absence of tumor-derived CHI3L1, blood vessels are better structured and likely more functional.

Coverage of blood vessels by astrocytic endfeet contributes to the blood-brain barrier by sealing endothelial cell junctions and controlling the transport of substances from blood (32). Astrocytes clear waste and excess of extracellular fluid from the perivascular space via a water channel aquaporin 4 (AQP4) (33), localized at the astrocytic endfeet. The expression of AQP4 was low in WT U87-MG gliomas but considerably increased in CHI3L1 KO gliomas (Fig. 5F and H). A linear regression analysis for CHI3L1 and AQP4 protein levels in the BPA database shows a negative correlation between the two proteins in GBM (Fig. 5I). Reduced staining for glial fibrillary acidic protein (GFAP) demonstrates that astrocytes were less activated in CHI3L1 KO tumors in comparison to WT controls (Fig. 5G and J). Aquaporin 4 was localized at the GFAP + cell protrusions covering blood vessels as demonstrated with double staining for AQP4 and GFAP (Fig. 5K). We propose that tumor-derived CHI3L1 promotes deregulation of the vascular network in gliomas by increasing the number of endothelial sprouts, altering the structure of vascular walls, and decreasing the coverage of blood vessels by astrocytes (Fig. 6).

In this study, we demonstrated an upregulated CHI3L1 expression in malignant gliomas exploring scRNAseq databases and bulk RNAseq data from two Polish glioma patient cohorts. CHI3L1 expression was the highest in GBM compared to other tumor types in the GEPIA database and malignant cells were determined as the predominant source of CHI3L1 expression in GBM. We found a positive correlation between genes listed in the MESlike1 signature and CHI3L1 protein in the Brain Protein Atlas, which further strengthens the link between CHI3L1 expression and mesenchymal phenotype of GBM.

The CHI3L1 KO cells formed significantly smaller tumors compared to WT cells. The tumor volume correlated with the human CHI3L1 levels in sera of tumor-bearing mice. We verified several potential mechanisms that glioma tumors are known to employ to grow out: 1) upregulation of ECM rearrangement regulators; 2) induction of myeloid cells infiltration/reprograming; 3) stimulation of neoangiogenesis.

MMP-2 is a crucial ECM degrading enzyme that is produced by glioma cells and activated through interactions with activated microglia (27). We found reduced MMP-2 expression and proteolytic activity in CHI3L1 KO cells. The CHI3L1 KO clone 2 (KO2) showed a statistically insignificant reduction of MMP-2 activity vs. WT cells, which might be due to remaining production of CHI3L1 in KO2 cells (Fig. 2E). CHI3L1 KO cells had diminished expression/production of SPP1, a well-established multifunctional protein associated with the regulation of tumor invasion and immunosuppression in GBM (25,34). Accumulation of myeloid cells in CHI3L1 KO tumors was lowered, which we attribute to the lower SPP1 production.

In our glioma cohorts, increased expression of CHI3L1 was found in both GBM and, unexpectedly, in pilocytic astrocytomas (PA) which are non-invasive tumors (35). The high level of CHI3L1 mRNA in PA might not be manifested at protein level as CHI3L1 protein have been reported to be low in PA (36). This might be due to the presence of mRNA-stabilizing protein HuR that is known to be up-regulated in cancer, including high-grade gliomas (37), and drives excessive protein synthesis in high-grade but not low-grade gliomas.

The genes of the ECM reorganization pathway along with genes associated with PD-1 and IL-10 signaling were down-regulated in CHI3L1 KO cells. CSF1, CCL2 and TGFB3 are important factors driving accumulation for myeloid cells in glioblastoma (38) and their downregulation might profoundly change the tumor microenvironment into less permissive for tumor growth (39,40). The genes associated with the mesenchymal signature were strongly down-regulated in CHI3L1 KO cells. We conclude that CHI3L1 might be a potential positive regulator of some of the mesenchymal subtype genes.

Genes associated with “DNA replication” were down-regulated in CHI3L1 KO cells. To make sure CRISPR/Cas9 procedure did not impact cell proliferation, three clones of U87-MG transfected with control gRNA were tested in a proliferation assay and proliferated as WT controls. (Suppl. Figure 2E).

Aberrant vasculature is a hallmark of GBM pathology. The tumor vessels are collapsed and disorganized, which contributes to hypoxia and immunosuppression, and creates specific TME (16,41–43). The astrocytes are displaced by glioblastoma, which disrupts the blood-brain barrier and results in a damaging ‘permeability and retention effect’ (44). The delivery of drugs to the tumor remains a key challenge in the treatment of GBM (45). Therefore, normalization of vasculature is desirable for an effective anti-tumor activity in GBM. Previous studies reported on the impact of CHI3L1 on vasculature in GBM based on in vitro studies or extracranial mouse models (14,15,46). In our study, intracranial tumors depleted of glioma-derived CHI3L1 presented features of normalized vasculature network, along with AQP4 localized to the astrocytic endfeet surrounding the vessels. Previously, CHI3L1 impact on neoangiogenesis was associated with VEGF signaling (1,15,16). We have found a positive correlation for VEGFA and CHI3L1 in the Ivy GAP database (47) (Suppl. Figure 3E), and for mVEGF and mCHI3L1 concentration in blood of tumor-bearing mice in CHI3L1 KO group (Suppl. Figure 3F). This suggests that CHI3L1 and VEGF might impact neoangiogenesis synergistically, but the exact interplay between them remains elusive.

Altogether, we demonstrate that CHI3L1 is an important therapeutic target in GBM, and the findings presented here, open new avenues to explore the role of CHI3L1 in driving GBM pathogenicity.

Patient samples

We analyzed data from two cohorts: 76 glioma samples described in our previous study (18) and 31 glioma samples collected and analyzed by RNA sequencing in our previous study (19). The collection of samples was approved by the Bioethics Committee of respective hospitals. RNAs from normal brains were purchased from: Agilent 540005 (lot 0006127195, Female 66 years), Biochain R1234035-50 (lot B304105, Male 25 years), Human Brain Reference RNA AM6050 (lot 1207015, mix of RNA, male and female donors).

Human cell lines and cell culture

Human glioblastoma cell lines LN229, U87-MG, U87-MG-RFP (ATCC, Manassas, VA) and WG9 (48) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) with 10% fetal bovine serum (FBS) (Gibco, MD, USA) and antibiotics (100 U/mL penicillin, 100 µg/mL streptomycin). Normal human astrocytes (NHA; Lonza, Cologne, Germany) were cultured in dedicated Clonetics media (Lonza, Walkersville, MD, USA) in CO₂/air (5%/95%) at 37°C (Heraeus, Hanau, Germany). Cells were passaged when confluent up to 10x from the initial seeding.

RNA extraction and quantitative PCR (RT-qPCR)

Cells were washed with PBS and collected using a cell scraper and centrifugation. RNA was isolated using RNeasy Mini Kit (QIAGEN, USA) and reverse transcription was performed using SuperScript III Reverse Transcriptase (Invitrogen) on 500 ng RNA. Quantitative PCR was performed on 12.5 ng of cDNA in duplicates using SYBR® Green detection reagents and primers listed in Table 1.

CRISPR/Cas9 genome editing

U87-MG-RFP human glioma cells (2 × 10⁶) were electroporated with 0.5 µg pCMV-Cas9-GFP plasmid and 2 sgRNAs targeting CHI3L1 (150 pmol each, Sigma, Munich, Germany) in SE electroporation buffer (100 µL) using a Nucleofactor 4DX according to manufacturer’s instructions (Lonza, Cologne, Germany). Cells were then seeded in 20 mL standard culture media and maintained under standard culture conditions for 24 h before sorting.

Fluorescence-activated cell sorting and clonal selection

Cells were detached and sorted to 96-wells flat-bottom culture plates into 100 µL of high glucose GlutaMAX™ DMEM supplemented with 15% FBS (Gibco, MD, USA) using BD FACSAria II cell sorter (BD Pharmingen). Gates for sorting were set for RFP^high and GFP⁺ events (Suppl. Figure 2A). Media were changed every 3 days. When colonies were established, cells were upscaled to a 6-well culture plate. Flow cytometry experiments were performed at the Laboratory of Cytometry, Nencki Institute of Experimental Biology.

Enzyme-linked immunosorbent assay (ELISA)

To select CHI3L1 U87-MG KO clones, cell culture supernatants were analyzed for the presence of human CHI3L1 using Human Chitinase 3-like 1 Quantikine ELISA Kit (R&D Systems, USA). Swere centrifuged for 10 min at 300 × g to remove residual cells. Tests were performed according to manufacturer’s protocol and measurements were acquired with a scanning multiwell spectrophotometer.

RNA sequencing

RNA was isolated using the RNeasy kit (Qiagen) and RNA quality/yield was verified using Bioanalyzer 2100 (Agilent Technologies, Santa Clara, CA). mRNA libraries were prepared using KAPA Stranded mRNAseq Kit (Roche, Basel, Switzerland) according to the manufacturer’s protocol (KR0960-v6.17). mRNAs were enriched from 500 ng of total RNA using poly-T oligo-magnetic beads (Kapa Biosystems, MA, USA), fragmented, cDNA synthesized, and adapters were ligated. Loop structures were cut by USER enzyme (NEB, Ipswich, MA, USA). The obtained dsDNA fragments were amplified with NEB starters (Ipswich, MA, USA). Quality of libraries was determined using Bioanalyzer High Sensitivity dsDNA Kit (Agilent Technologies, Palo Alto, CA, USA) and concentrations measured using Quantus Fluorometer and QuantiFluor ONE Double Stranded DNA System (Promega, Madison, Wisconsin, USA). Libraries were sequenced on HiSeq 1500 (Illumina, San Diego, CA 92122 USA) on the rapid run flow cell with a paired-end settings (2× 76 base pairs). Fastq files were aligned to hg38 human reference genome with STAR program (49), and reads were counted to genes using feature Counts algorithm SUBREAD package (50). Gene counts were normalized, and differential analysis was performed using the DESeq2 (51); genes differentially expressed (DEG) had FDR corrected p-value < 0.05. REACTOME pathway analyses (52) were performed using R package clusterProfiler (53).

Intracranial glioma implantation

The study was approved by the First Local Ethics Committee in Warsaw, Poland (1123/2020). Male mice (Foxn1^nu, Janvier, France) were housed in individually ventilated cages, fed standard chow, and kept under standard day/night conditions. Ten weeks old mice were anesthetized with isoflurane anesthesia (Temsega, Pessac, France). Appropriate analgesics were used for the procedure. WT or CHI3L1 KO U87-MG-RFP glioma cells (5×10⁴ cells in 1 µl of DMEM) were implanted aseptically into the right striatum using a stereotactic apparatus and 1 µL syringe with a 26-gauge needle (Stoelting Co., Wood Dale, USA).

Magnetic resonance imaging

The animals were anesthetized with 1.5-2% isoflurane (Baxter, Deerfield, IL, USA) in oxygen. Animal heads were scanned with 7T BioSpec 70/30 MR system (Bruker, Ettlingen, Germany) equipped with Avance III console and actively shielded gradient system B-GA 20S (amplitude 200 mT/m). Structural transverse MR images covering the whole brain were acquired with T2-weighted TurboRARE (TR/TEeff = 7000/30 ms, RARE factor 4, spatial resolution = 86µm×86µm×350µm, 42 slices, no gaps, number of averages 4, scan time ~ 23 min).

Collecting brain tissues and blood samples

The animals were anesthetized 21 days post-implantation with injections of ketamine (160 mg/kg) and xylazine (20 mg/kg) and perfused transcardially with an ice-cold PBS and 4% paraformaldehyde (PFA). Beforehand, the right atrium was perforated and 500 µL of blood was collected. Brains were removed and fixed in 4% PFA for 48 at 4°C. Then, brains were placed in 30% sucrose in PBS at 4°C. Finally, brains were frozen in Tissue Freezing Medium (Leica, Germany) and cut into 12 or 50 µm thick coronal sections and stored. Collected blood was kept for 2 h at RT until coagulated and centrifuged for 10 min at 3,000 x g. Blood serum was supplemented with proteinase inhibitor PMSF (phenylmethylsulfonyl fluoride, 1 µM, Sigma, Munich, Germany) and stored at -80°C. Blood sera were analyzed using Human/Mouse Chitinase 3-like 1 Quantikine ELISA Kits (R&D Systems, USA).

Fluorescence microscopy

For vWF, IBA1, AQP4 immunostaining, brain sections were thawed, dried for 2 h in RT, rehydrated in PBS for 15 min and washed 3x in 0.05 M TBS (Tris-buffered saline, Sigma, Munich, Germany). Then, sections were incubated with 10% donkey serum (DS) with 0.1% Triton X-100 (in TBS) for 2 h at RT and stained overnight at 4°C with primary antibodies listed in Table 2. Secondary antibodies were incubated for 2 h at RT, sections were washed and mounted with Vectashield Vibrance® Antifade Mounting Medium with DAPI (Vector Labs, US). For GFAP, CD31 and laminin staining, brain sections were thawed and dried for 5 min at 37°C temperature and then kept in chilled methanol in -20°C for 30 min. Next, sections were washed 3x for 5 min in PBST (0.1% Triton X-100 in PBS), blocked with 3% DS in 0.4% Triton X-100 in PBS for 2 h at RT, stained overnight with primary antibodies at 4°C followed by incubation with secondary antibody for 2 h at RT. Sections were washed 3x in PBST for 5 min, once in PBS for 5 min, followed by washing in ultrapure water for 5 min and mounted with Vectashield Vibrance® Antifade Mounting Medium with DAPI (Vector Labs, US). In negative controls primary antibody was omitted. Images were captured using Zeiss Axio Imager M2 microscope. Image processing and quantification was performed using ImageJ software (NIH). Confocal images were taken using Zeiss LSM 780 or Zeiss LSM800 Airyscan microscope at the Laboratory of Imaging Tissue Structure and Function, Nencki Institute. The images shown in the Fig. 5E are 3D renderings of z stacks: 639×639 µm and at least 40 µm of z stack (z steps 3.5 µm) using Bitplane Imaris Image Analysis software.

Matrigel invasion assay

Cell culture inserts (12 µm pore size, Millpore, Tullagreen, Ireland) were coated with 1 mg/mL Growth Factor Reduced Matrigel Matrix (BD Biosciences, CA, USA) in DMEM and dried at 37°C for 5 h. U87-MG glioma cells were seeded at 4×10⁴ on matrigel-covered membrane. Inserts were transferred to a 24-well plate with or without human SV40 microglial cells (HM SV40). After 18 h, cells invading through the matrigel were fixed with 95% methanol and stained with DAPI (4′,6- Diamidino-2-Phenylindole; 0.01 mg/ml, Sigma, Munich, Germany). The membranes were cut out and cells were visualized using fluorescence microscope (Leica DM4000B, ×10 objective). Cell nuclei were counted from five independent fields per slide using ImageJ software.

Gelatin zymography

U87-MG cells (5×10⁵ cells per well) were seeded onto 24-wells plate in 1 mL DMEM with 10% FBS (Gibco, MD, USA) and antibiotics. After 24 h, media were replaced with DMEM, and cells were incubated for 24 h. The culture media were collected, centrifuged for 10 min at 300 × g and subjected to gelatin zymography. Samples were diluted in 5× Laemmli buffer without dithiothreitol and separated electrophoretically on 8% acrylamide–bisacrylamide (29:1, Sigma, Munich, Germany) gels containing 2 mg/mL porcine gelatin (Sigma, cat. 9000-70-8). Then, gels were rinsed twice with 2.5% Triton X-100 before incubation in a renaturation buffer containing for 24 h (37°C, 100 rpm). Gels were stained with Coomasie brilliant blue, washed with PBS, and quantified by densitometry. Culture media with FBS were used as a positive control. Full composition of ingredients used for gelatin zymography is listed in Table 3.

Public data

Analysis of public data was performed using R Statistical Software (version 4.2.2; R Foundation for Statistical Computing, Vienna, Austria). RNAseq data (FPKMs) were downloaded from the Ivy Glioblastoma Atlas Project (47) and normalized RNAseq and protein data from The Brain Protein Atlas (BPA) (20,21). Correlation of CHI31L1 mRNA and protein expression level was computed in the BPA dataset, using Kendall correlation and von Waerden test followed by post-hoc U-Mann-Whitney tests to determine differences between tumor regions. Exact Fisher tests was performed for overlap of genes, whose expression correlated with CHI3L1, and signature by Neftel et al. (24).

Data availability statement

RNAseq data are available in the NIH GEO database with the accession number GSE231816. All other data supporting the findings of this study are available within the manuscript or the supplementary information or from the corresponding author upon a reasonable request.

Acknowledgements

We would like to thank Beata Kaza, Katarzyna Poleszak, Maria Pasierbińska, Paulina Pilanc, Aleksandra Ellert-Miklaszewska, Katarzyna Jaślan, Małgorzata Całka, Artur Wolny for sharing their expertise. Special thanks to Brian Lam for sharing normalized data from the Brain Protein Atlas.

Funding

The study was supported by National Science Centre Poland grant 2020/39/B/NZ4/02683. M.G. is supported by PACIFIC Call 1 (PAN.BFB. S.BDN.619.022.2021, the EU Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant 847639). J.S. is supported by EMBO Postdoctoral Fellowship ALTF 663-2022.

Author contributions

B.K. and S.C. conceptualized the study. S.C. performed most of the in vitro and all the in vivo mouse experiments. B.G. performed the library preparation and RNAseq. B.W. performed the bioinformatics analysis of RNAseq data. B.W. and S.B. performed the bioinformatics analysis of publicly available datasets. J.S. performed the cytometric experiment for cell proliferation. M.G. and M.Z. performed the wet lab work for immunofluorescence stainings. K.W. prepared cDNA from patients’ samples and cell lines. S.C. and B.K. wrote the manuscript. All the authors contributed to the editing of the manuscript and have approved the submitted manuscript.

Ethics declaration

Mouse experiments presented in the study were approved by the 1st Local Bioethical Committee in Warsaw, Poland. Application approval reference number: 1123/2020.

Competing interests

The authors declare no competing interests.

Zhao T, Su Z, Li Y, Zhang X, You Q. Chitinase-3 like-protein-1 function and its role in diseases. Signal Transduct Target Ther. 2020 Sep 14;5(1):201.
Pinteac R, Montalban X, Comabella M. Chitinases and chitinase-like proteins as biomarkers in neurologic disorders. Neurol - Neuroimmunol Neuroinflammation. 2021 Jan;8(1):e921.
Iwata T, Kuwajima M, Sukeno A, Ishimaru N, Hayashi Y, Wabitsch M, et al. YKL-40 secreted from adipose tissue inhibits degradation of type I collagen. Biochem Biophys Res Commun. 2009 Oct;388(3):511–6.
Bigg HF, Wait R, Rowan AD, Cawston TE. The Mammalian Chitinase-like Lectin, YKL-40, Binds Specifically to Type I Collagen and Modulates the Rate of Type I Collagen Fibril Formation. J Biol Chem. 2006 Jul;281(30):21082–95.
Libreros S, Garcia-Areas R, Iragavarapu-Charyulu V. CHI3L1 plays a role in cancer through enhanced production of pro-inflammatory/pro-tumorigenic and angiogenic factors. Immunol Res. 2013 Dec;57(1–3):99–105.
Coffman FD. Chitinase 3-Like-1 (CHI3L1): A Putative Disease Marker at the Interface of Proteomics and Glycomics. Crit Rev Clin Lab Sci. 2008 Jan;45(6):531–62.
Ma B, Akosman B, Kamle S, Lee CM, He CH, Koo JS, et al. CHI3L1 regulates PD-L1 and anti–CHI3L1–PD-1 antibody elicits synergistic antitumor responses. J Clin Invest. 2021 Nov 1;131(21):e137750.
Bonneh-Barkay D, Bissel SJ, Kofler J, Starkey A, Wang G, Wiley CA. Astrocyte and Macrophage Regulation of YKL-40 Expression and Cellular Response in Neuroinflammation: YKL-40 Expression in Macrophages and Astrocytes. Brain Pathol. 2012 Jul;22(4):530–46.
Bonneh-Barkay D, Wang G, Starkey A, Hamilton RL, Wiley CA. In vivo CHI3L1 (YKL-40) expression in astrocytes in acute and chronic neurological diseases. J Neuroinflammation. 2010;7(1):34.
Tanwar MK, Gilbert MR, Holland EC. Gene expression microarray analysis reveals YKL-40 to be a potential serum marker for malignant character in human glioma. Cancer Res. 2002 Aug 1;62(15):4364–8.
Steponaitis G, Skiriutė D, Kazlauskas A, Golubickaitė I, Stakaitis R, Tamašauskas A, et al. High CHI3L1 expression is associated with glioma patient survival. Diagn Pathol. 2016 Dec;11(1):42.
Verhaak RGW, Hoadley KA, Purdom E, Wang V, Qi Y, Wilkerson MD, et al. Integrated Genomic Analysis Identifies Clinically Relevant Subtypes of Glioblastoma Characterized by Abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell. 2010 Jan;17(1):98–110.
Pelloski CE, Mahajan A, Maor M, Chang EL, Woo S, Gilbert M, et al. YKL-40 Expression is Associated with Poorer Response to Radiation and Shorter Overall Survival in Glioblastoma. Clin Cancer Res. 2005 May 1;11(9):3326–34.
Shao R, Hamel K, Petersen L, Cao QJ, Arenas RB, Bigelow C, et al. YKL-40, a secreted glycoprotein, promotes tumor angiogenesis. Oncogene. 2009 Dec 17;28(50):4456–68.
Francescone RA, Scully S, Faibish M, Taylor SL, Oh D, Moral L, et al. Role of YKL-40 in the Angiogenesis, Radioresistance, and Progression of Glioblastoma. J Biol Chem. 2011 Apr;286(17):15332–43.
Holst CB, Pedersen H, Obara EAA, Vitting-Seerup K, Jensen KE, Skjøth-Rasmussen J, et al. Perspective: targeting VEGF-A and YKL-40 in glioblastoma – matter matters. Cell Cycle. 2021 Apr 3;20(7):702–15.
Tang Z, Li C, Kang B, Gao G, Li C, Zhang Z. GEPIA: a web server for cancer and normal gene expression profiling and interactive analyses. Nucleic Acids Res. 2017 Jul 3;45(W1):W98–102.
Król SK, Kaczmarczyk A, Wojnicki K, Wojtas B, Gielniewski B, Grajkowska W, et al. Aberrantly Expressed RECQL4 Helicase Supports Proliferation and Drug Resistance of Human Glioma Cells and Glioma Stem Cells. Cancers. 2020 Oct 11;12(10):2919.
Stępniak K, Machnicka MA, Mieczkowski J, Macioszek A, Wojtaś B, Gielniewski B, et al. Mapping chromatin accessibility and active regulatory elements reveals pathological mechanisms in human gliomas. Nat Commun. 2021 Jun 15;12(1):3621.
Lam KHB, Leon AJ, Hui W, Lee SCE, Batruch I, Faust K, et al. Topographic mapping of the glioblastoma proteome reveals a triple-axis model of intra-tumoral heterogeneity. Nat Commun. 2022 Jan 10;13(1):116.
Lam KHB, Faust K, Yin R, Fiala C, Diamandis P. The Brain Protein Atlas: A conglomerate of proteomics datasets of human neural tissue. PROTEOMICS. 2022 Dec;22(23–24):2200127.
Pacini C, Dempster JM, Boyle I, Gonçalves E, Najgebauer H, Karakoc E, et al. Integrated cross-study datasets of genetic dependencies in cancer. Nat Commun. 2021 Mar 12;12(1):1661.
Tsherniak A, Vazquez F, Montgomery PG, Weir BA, Kryukov G, Cowley GS, et al. Defining a Cancer Dependency Map. Cell. 2017 Jul;170(3):564–576.e16.
Neftel C, Laffy J, Filbin MG, Hara T, Shore ME, Rahme GJ, et al. An Integrative Model of Cellular States, Plasticity, and Genetics for Glioblastoma. Cell. 2019 Aug;178(4):835–849.e21.
Ellert-Miklaszewska A, Wisniewski P, Kijewska M, Gajdanowicz P, Pszczolkowska D, Przanowski P, et al. Tumour-processed osteopontin and lactadherin drive the protumorigenic reprogramming of microglia and glioma progression. Oncogene. 2016 Dec 15;35(50):6366–77.
Sliwa M, Markovic D, Gabrusiewicz K, Synowitz M, Glass R, Zawadzka M, et al. The invasion promoting effect of microglia on glioblastoma cells is inhibited by cyclosporin A. Brain. 2007 Feb 1;130(2):476–89.
Markovic DS, Vinnakota K, Chirasani S, Synowitz M, Raguet H, Stock K, et al. Gliomas induce and exploit microglial MT1-MMP expression for tumor expansion. Proc Natl Acad Sci. 2009 Jul 28;106(30):12530–5.
Kaminska B, Ochocka N, Segit P. Single-Cell Omics in Dissecting Immune Microenvironment of Malignant Gliomas—Challenges and Perspectives. Cells. 2021 Aug 31;10(9):2264.
Ochocka N, Segit P, Walentynowicz KA, Wojnicki K, Cyranowski S, Swatler J, et al. Single-cell RNA sequencing reveals functional heterogeneity of glioma-associated brain macrophages. Nat Commun. 2021 Feb 19;12(1):1151.
Quail DF, Joyce JA. The Microenvironmental Landscape of Brain Tumors. Cancer Cell. 2017 Mar;31(3):326–41.
Ghochani Y, Muthukrishnan SD, Sohrabi A, Kawaguchi R, Condro MC, Bastola S, et al. A molecular interactome of the glioblastoma perivascular niche reveals integrin binding sialoprotein as a mediator of tumor cell migration. Cell Rep. 2022 Oct;41(3):111511.
Cabezas R, Ã-vila M, Gonzalez J, El-BachÃ¡ RS, BÃ¡ez E, GarcÃa-Segura LM, et al. Astrocytic modulation of blood brain barrier: perspectives on Parkinsonâ€^TMs disease. Front Cell Neurosci [Internet]. 2014 Aug 4 [cited 2023 May 1];8. Available from: http://journal.frontiersin.org/article/10.3389/fncel.2014.00211/abstract
Troili F, Cipollini V, Moci M, Morena E, Palotai M, Rinaldi V, et al. Perivascular Unit: This Must Be the Place. The Anatomical Crossroad Between the Immune, Vascular and Nervous System. Front Neuroanat. 2020 Apr 16;14:17.
Kijewska M, Kocyk M, Kloss M, Stepniak K, Korwek Z, Polakowska R, et al. The embryonic type of SPP1 transcriptional regulation is re-activated in glioblastoma. Oncotarget. 2017 Mar 7;8(10):16340–55.
Collins VP, Jones DTW, Giannini C. Pilocytic astrocytoma: pathology, molecular mechanisms and markers. Acta Neuropathol (Berl). 2015 Jun;129(6):775–88.
Zhang W, Kawanishi M, Miyake K, Kagawa M, Kawai N, Murao K, et al. Association between YKL-40 and adult primary astrocytoma. Cancer. 2010;NA-NA.
Filippova N, Yang X, Ananthan S, Sorochinsky A, Hackney JR, Gentry Z, et al. Hu antigen R (HuR) multimerization contributes to glioma disease progression. J Biol Chem. 2017 Oct;292(41):16999–7010.
Pyonteck SM, Akkari L, Schuhmacher AJ, Bowman RL, Sevenich L, Quail DF, et al. CSF-1R inhibition alters macrophage polarization and blocks glioma progression. Nat Med. 2013 Oct;19(10):1264–72.
Chang AL, Miska J, Wainwright DA, Dey M, Rivetta CV, Yu D, et al. CCL2 Produced by the Glioma Microenvironment Is Essential for the Recruitment of Regulatory T Cells and Myeloid-Derived Suppressor Cells. Cancer Res. 2016 Oct 1;76(19):5671–82.
Kaminska B, Cyranowski S. Recent Advances in Understanding Mechanisms of TGF Beta Signaling and Its Role in Glioma Pathogenesis. In: Barańska J, editor. Glioma Signaling [Internet]. Cham: Springer International Publishing; 2020 [cited 2023 May 1]. p. 179–201. (Advances in Experimental Medicine and Biology; vol. 1202). Available from: http://link.springer.com/10.1007/978-3-030-30651-9_9
Ghosh M, Lenkiewicz AM, Kaminska B. The Interplay of Tumor Vessels and Immune Cells Affects Immunotherapy of Glioblastoma. Biomedicines. 2022 Sep 15;10(9):2292.
Maddison K, Bowden NA, Graves MC, Tooney PA. Characteristics of vasculogenic mimicry and tumour to endothelial transdifferentiation in human glioblastoma: a systematic review. BMC Cancer. 2023 Feb 23;23(1):185.
Kane JR. The Role of Brain Vasculature in Glioblastoma. Mol Neurobiol. 2019 Sep;56(9):6645–53.
Hempel C, Johnsen KB, Kostrikov S, Hamerlik P, Andresen TL. Brain tumor vessels—a barrier for drug delivery. Cancer Metastasis Rev. 2020 Sep;39(3):959–68.
Mitusova K, Peltek OO, Karpov TE, Muslimov AR, Zyuzin MV, Timin AS. Overcoming the blood–brain barrier for the therapy of malignant brain tumor: current status and prospects of drug delivery approaches. J Nanobiotechnology. 2022 Sep 15;20(1):412.
Faibish M, Francescone R, Bentley B, Yan W, Shao R. A YKL-40–Neutralizing Antibody Blocks Tumor Angiogenesis and Progression: A Potential Therapeutic Agent in Cancers. Mol Cancer Ther. 2011 May;10(5):742–51.
Puchalski RB, Shah N, Miller J, Dalley R, Nomura SR, Yoon JG, et al. An anatomic transcriptional atlas of human glioblastoma. Science. 2018 May 11;360(6389):660–3.
Ciechomska IA, Wojnicki K, Wojtas B, Szadkowska P, Poleszak K, Kaza B, et al. Exploring Novel Therapeutic Opportunities for Glioblastoma Using Patient-Derived Cell Cultures. Cancers. 2023 Mar 2;15(5):1562.
Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics. 2013 Jan 1;29(1):15–21.
Liao Y, Smyth GK, Shi W. The R package Rsubread is easier, faster, cheaper and better for alignment and quantification of RNA sequencing reads. Nucleic Acids Res. 2019 May 7;47(8):e47–e47.
Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014 Dec;15(12):550.
Jassal B, Matthews L, Viteri G, Gong C, Lorente P, Fabregat A, et al. The reactome pathway knowledgebase. Nucleic Acids Res. 2019 Nov 6;gkz1031.
Yu G, Wang LG, Han Y, He QY. clusterProfiler: an R Package for Comparing Biological Themes Among Gene Clusters. OMICS J Integr Biol. 2012 May;16(5):284–7.

Table 1. Primers for qPCR.

Target mRNA	Sequence or reference
CHI3L1 human	ref. ID: qHsaCED0044484, BioRad
GAPDH human	forward 5′- AGGGCTGCTTTTAACTCTGGT-3′ reverse 5′- CCCCACTTGATTTTGGAGGGA-3′
18S human	forward 5′-CGGACATCTAAGGGCATCAACA-3′ reverse 5′-AACGAACGAGACTCTGGCATG-3′

Table 2. Antibodies and their source.

Antibody	Working dilution	Company
goat anti-Iba1 antibody	1:500	ab5076, Abcam, UK
rabbit anti-von Willebrand factor	1:1000	A008229-2, Agilent, USA
rabbit anti-aquaporin 4	1:200	ab125049, Abcam, UK
rabbit anti-GFAP antibody	1:200	Z033429-2, Agilent USA
rat anti-CD31 antibody	1:100	#NB600-1475, Novus Biologicals, USA
rabbit anti-laminin	1:50	ab11575, Abcam, UK
donkey anti-goat Alexa Fluor 488	1:1000	A-11055, Invitrogen, USA
donkey anti-rabbit Alexa Fluor 647	1:1000	A-31573, Invitrogen, USA
donkey anti-rabbit Alexa Fluor 488	1:1000	A-21206, Invitrogen, USA
donkey anti-rat Alexa Fluor 488	1:500	A-21208, Invitrogen, USA

Table 3. Zymography gels and buffers.

Ingredient	Composition
Separation gel	380 mM TRIS–HCl (pH 8.8), 0.1% SDS, 0.1% ammonium persulfate, 0.6 μl/ml of TEMED (Sigma, Munich, Germany)
Renaturation buffer	50 mM TRIS–HCl pH 7.6, 10 mM CaCl2, 1 μM ZnCl2, 1% Triton X-100 and 0.02% sodium azide

There is NO conflict of interest to disclose.

Suppl.Figure1.jpg
Supplementary Figure 1. Ratio of CHI3L1 expression in tumor versus normal tissue. Data from the GEPIA database.
Suppl.Figure2.jpg
Supplementary Figure 2. A. Gating strategy for the flow cytometric sort of U87-MG-RFP successfully transfected with pCMV-Cas9-GFP plasmid. B-C.Validation of CHI3L1 knock-out in selected candidate clones. B.RT-qPCR analysis of CHI3L1 expression using 2 sets of primers. CHI3L1 expression given as % of WT control. C. ELISA of cell culture supernatants of sgCHI3L1 clones. CHI3L1 concentration given as % of WT control. Red lines indicate the level in WT reference. D. BrdU incorporation assay for U87-MG-RFP cells transfected with control non-targeting gRNA. CRISPR/Cas9 engineering pipeline does not change the proliferation rate of U87-MG-RFP clones. E. Analysis of CHI3L1 expression in NHA, LN229, U87-MG and WG9 by RT-qPCR.
Suppl.Figure3.jpg
Supplementary Figure 3. A-D.Correlation analysis between tumor volume and (A) no. of non-capillary vessels, (B) microvessel density, (C) laminin signal and (D) AQP4 signal. No correlations were found. E. Correlation analysis for VEGFA and CHI3L1 expression in Ivy GAP RNAseq database. F. Correlation analysis for mVEGF and mCHI3L1 in the sera of tumor-bearing mice.

Download PDF

Version 1

posted

You are reading this latest preprint version

Depletion of chitinase-3-like protein 1 (CHI3L1) in glioma cells restraints tumor growth and affects neovasculature in intracranial murine gliomas

Status:

Version 1

Abstract

Figures

Introduction

Results

Glioblastoma cells depleted of CHI3L1 form smaller tumors in recipient mice

Glioma-derived CHI3L1 enhances invasion and myeloid cell infiltration into the tumor

Depletion of glioma-derived CHI3L1 results in normalization of tumor vasculature

Discussion

Materials and Methods

Patient samples

Human cell lines and cell culture

RNA extraction and quantitative PCR (RT-qPCR)

CRISPR/Cas9 genome editing

Fluorescence-activated cell sorting and clonal selection

Enzyme-linked immunosorbent assay (ELISA)

RNA sequencing

Intracranial glioma implantation

Magnetic resonance imaging

Collecting brain tissues and blood samples

Fluorescence microscopy

Matrigel invasion assay

Gelatin zymography

Public data

Declarations

References

Tables

Additional Declarations

Supplementary Files

Status:

Version 1