Immunological and clinicopathological characteristics of C1RL in 2143 glioma patients

Glioma is a deadly and immunosuppressive brain tumor. Complement C1r Subcomponent Like (C1RL), as a prognostic biomarker in several kinds of tumors, has attracted increasing attention from oncologists. However, the role of C1RL in glioma remains unclear. Through 2143 glioma patients from 5 public datasets, the relationships between C1RL expression and clinicopathological characteristics were analyzed. Furthermore, the C1RL associated genes were screened and GO analysis was conducted to investigate their biological process enrichment. In addition, tumor purity, leukocytes infiltration and overall survival were calculated based on C1RL expression. We found that C1RL expression was upregulated in glioblastoma (GBM), especially mesenchymal GBM and primary GBM. Increased C1RL accompanied IDH1 wt both in lower grade glioma (LGG) and GBM. C1RL associated genes were mainly enriched in biological processes which are related to immune response. C1RL expression was also correlated with less tumor purity and more M2 macrophage infiltration. Higher C1RL expression predicted unfavorable survival for patients with glioma and therapeutic resistance in GBM. Our results implied that C1RL is involved in immunological activities and is an independent unfavorable prognostic biomarker for patients with glioma. C1RL is a potential immunotherapeutic target for glioma in future clinical practice.


Introduction
Glioblastoma (GBM, WHO grade IV) and lower grade glioma (LGG, WHO grade II and III) are incurable brain tumor. Existing therapeutic strategies only prolong the survival of glioma patients to a limited extent. Patients with glioma eventually die from tumor recurrence, even with aggressive treatment.
Novel therapies that have been successful in other tumors, such as PD-1 inhibition [1] and Bevacizumab [2,3] , failed to extend the overall survival time of patients with glioma. Tumor treating fields (TTF), a novel therapy which was approved to treat GBM by Food and Drug Administration (FDA) recently, is not widely used in clinical practice because of its high price and difficulty in using [4,5] .
Current poor situation push us to explore the mechanism of glioma development and novel therapy.
Immunosuppressive microenvironment significantly contributes to the progression and therapeutic resistance of glioma. On one hand, glioma cells induce less immune response and more 3 immunosuppression. Compared to other malignancies, glioma accompanies with less mutant burden and then less infiltrating T cells [6] . GBM cells block T cell activation and proliferation in response to T cell receptor stimulation by generating extracellular vesicles which carrying PD-L1 [7] . Glioma cells promote the expression of PD-L1 on macrophages which are derived from healthy donators [8,9] .
Intratumorally immunosuppressive education from glioma also contributes to the rise of systemic immunosuppressive myeloid-derived suppressor cells (MDSCs) [10] . On the other hand, brain provides an immunosuppressive bed for glioma. Compared to the melanoma in flank, melanoma in brain accompany with less CD8 T cells [11] . Moreover, the antigen-specific cytotoxicity is systemically impaired in mice with brain melanoma [11] . Naïve T cells are sequestered in large numbers in the bone marrow in cancer patients. This phenomenon characterizes a variety of tumors only when tumors are placed into the intracranial compartment [12] . Complement C1r Subcomponent Like (C1RL) was found as a prognostic marker in hepatocellular carcinoma [13] and renal cell cancer [14] . A gene based analysis showed a significant association between non-Hodgkin lymphoma as well as with diffuse large B-cell lymphoma and C1RL gene [15] .
C1RL also mediates the progression of Burkitt's lymphoma [16] . C1RL is a protein coding gene and associated with Ovary Adenocarcinoma and Leukorrhea. In terms of molecular function, C1RL protein, homologous to C1r, was identified in active form of serine hydrolase [17] . C1RL protein cleaves prohaptoglobin in the endoplasmic reticulum [18] . In addition, pro-C1s is also proteolytically cleaved into two fragments of sizes identical with those of the two chains of active C1s by C1RL protein [19] .
However, the immunological and clinicopathological characteristics of C1RL in glioma remain unclear.
In the present study, we employed 2143 glioma specimens from 5 datasets to further explore the clinicopathological and biological characteristics of C1RL in glioma. Clinicopathological features include WHO grades, histology, GBM status, IDH mutation, GBM subtypes, overall survival and therapeutic resistance. Biological process of C1RL associated genes was enriched to explore the biological characteristics of C1RL. Moreover, the relationship between C1RL expression and tumor 4 purity and leukocytes infiltration were analyzed.

Statistics of C1RL expressions
Unpaired t test was used in the comparisons of C1RL expressions of two groups. Ordinary one way ANOVA (multiple comparison) was applied to compare C1RL expressions in three or more groups.
P<0.05 was considered significant.

C1RL associated genes siftings and GO analyses
Pearson correlation coefficients between C1RL and all other genes were calculated in RStudio 1.1.453 with the algorithm of cor.test. C1RL-associated genes were defined as genes with r value > 0.4 in GBM dataset (TCGAmic) and r > 0.5 in glioma datasets (TCGAseq, CGGAmic, and CGGAseq). All the C1RL associated genes were introduced into DAVID (https://david.ncifcrf.gov/) for further GO analyses. The top 10 terms of biological processes of GO analyses results were listed in Figure 2.

Tumor purity and leukocytes infiltration
The ESTIMATE algorithm package was used to analyze the tumor purity. The CIBERSORT tool (https://cibersort.stanford.edu/) was used to evaluate the leukocytes infiltration. The heatmap was produced in MORPHEUS (https://software.broadinstitute.org/morpheus/) online. The color showed the Z score (subtract mean, divide by standard deviation) of all the expression data. The samples were ordered according to the expression of C1RL.

Survival analyses
The log-rank tests and Kaplan Meier survival curves were used to describe the survival differences between two groups. The survival analysis of GSE16011 dataset was conducted in R2 (https://hgserver1.amc.nl/cgi-bin/r2/main.cgi).

Results
C1RL expression was upregulated in GBM, especially mesenchymal GBM, primary GBM and IDH1 wt GBM. accompanied with lower C1RL than that in primary GBM (Fig. 1.M-N). IDH mutation status is a wellaccepted marker for glioma classification. C1RL is higher in IDH1 wt GBM than that in IDH1 mt GBM ( Fig. 1.O-Q). In addition, C1RL is higher in IDH1 wt LGG than that in IDH1 mt LGG ( Fig. 1.P and Q).
These results suggested that higher C1RL expression accompanies higher malignancy in glioma, especially in GBM.
C1RL associated genes enriched in the biological processes of immune response.
The biological function of C1RL, especially in tumor, has not been clarified thoroughly. Therefore, we aimed to speculate the possible biological function of C1RL through the biological function of C1RL associated genes. C1RL associated genes were the genes which have the similar expression trends as C1RL in glioma samples. All the C1RL genes from each dataset were introduced to the GO analysis.
The biological processes were listed in reverse order of their p values. The GO analyses showed that C1RL associated genes mainly enriched in the biological processes of immune response, inflammatory 6 response, IFN-γ mediated signaling pathway, and innate immune response (Fig. 2.A-D).
In order to determine whether C1RL plays a positive role in anti-glioma immune response or not, the expression relationship between C1RL and existing biomarkers were analyzed. CD86 protein is the receptor of CTLA4 and mainly expresses on dendritic cells and monocytes. Galectin-9, encoded by LGALS9, was identified as the ligand of Tim-3 and play a key role the apoptosis of T cells. TGFB1 encodes a secreted ligand of the TGF-beta (transforming growth factor-beta) superfamily of proteins.

C1RL expression was correlated with less tumor purity and more M2 macrophage infiltration
Immune response is based on the migration of immune cells. Both tumor purity and the infiltration of 22 leukocytes were assessed for each sample from TCGA datasets. The samples were displayed in order of their C1RL expression levels. Both immune score and stromal score exhibited a positive correlation with C1RL expression trends (Fig. 4.A and B, top panels). In addition, the tumor purity showed an inverse correlation with C1RL expression trends (Fig. 4.A and B, middle panels). Moreover, C1RL expression is mostly related to M2 macrophage infiltration among 22 leukocytes (Fig. 4.A and B, bottom panels).

High expression of C1RL predicted unfavorable survival and therapeutic resistance in glioma
The median expression value of C1RL was used to separate samples into two groups. We evaluated the prognostic value of C1RL in four glioma datasets. Patients with glioma containing higher C1RL expression had a significantly shorter survival times than their counterparts in the GSE16011mic, TCGAseq, CGGAmic, and CGGAseq datasets (Fig. 5.A-D). However, histopathology may contribute to these significant differences. Next, we compared the survival times among GBM patients with different C1RL expressions in five datasets. All the survival curves exhibited significant differences ( Fig. 5.E-I). Moreover, the effectiveness of well-accepted treatment was evaluated in different C1RL level groups. The primary GBM patients with less C1RL expression showed better response to resection, radio chemotherapy (temozolomide), and standard therapy (Fig. 5.J-L). These results indicate that C1RL may contribute to therapeutic resistance. 7 Discussion C1RL is a negative biomarker for glioma prognosis. C1RL hadn't been mentioned in cancer until the report of a significant association with non-Hodgkin lymphoma overall as well as with diffuse large Bcell lymphoma for the C1RL gene in 2012 [15] . In recent years, C1RL was reported as a prognostic biomarker in hepatocellular carcinoma [13] and renal cell cancer [14] . Our results showed evidences that C1RL highly express on the glioma samples accompanying with bad prognosis. GBM is the WHO grade IV glioma and has the worst prognosis of median overall survival of 14.6 months [23] . The C1RL expression was always the highest in GBM when it was compared with LGG ( Fig. 1.A-H). Unsupervised transcriptome analysis reveals that mesenchymal GBM is related to the worst survival [24] . The C1RL expression in mesenchymal GBM was higher than that in other GBM subgroups ( Fig. 1.I-L). Secondary GBM progresses from LGG within 5-10 years of diagnosis and accompanies with better prognosis than that in primary GBM [25] . Secondary GBM expressed less C1RL than primary GBM (Fig. 1.M and N).
Patients with IDH1 mutant glioma have a better outcome than those with wild-type IDH1 gene [26] .
Less C1RL expressions happened in both IDH1 mutant LGG and IDH1 mutant GBM (Fig. 1.O-Q). C1RL not only predicts higher malignancy but also worse overall survival in glioma. Due to the distinct outcome of the glioma subgroups, the differences of C1RL on different histopathological subgroups ( Fig. 1.A-H) may contribute to the differences of survival outcome. We further investigated the survival difference between the high expression group and the low expression group of C1RL in GBM and even primary GBM. The results confirmed that the expression of C1RL was a survival indicator in primary GBM (Fig. 5.F-I). Resection following with chemoradiation was the well accepted strategy for primary GBM patients. Considering the effect from variant therapies, C1RL may play a role in therapeutic resistance (Fig. 5.J-L). Overall, our results indicated that C1RL is a biomarker of poor outcome for glioma patients.
C1RL probably play an important role in glioma immunosuppression. C1RL protein, is confirmed as an active form of serine hydrolase [17] and cleaves prohaptoglobin and pro-C1s into active forms [18] [19] .
On one hand, due to the suppression of lymphocyte function by haptoglobin [27] , C1RL may modulate 8 the immunosuppression of glioma by releasing active haptoglobin. On the other hand, the association of C1s with C1r and C1q, following ligand recognition, triggers the activation of the classical complement pathway [28] . C1q plays a fundamental role in the pathogenesis of glioma [29] . Thus, C1RL may trigger the classical complement pathway by activating C1s and then contribute to the pathogenesis of glioma. In addition, accumulated evidences show that C1RL is upregulated in inflammation [30] [31] . GO analyses for C1RL associated genes revealed that they mainly enriched in the biological processes of immune response, inflammatory response, IFN-γ mediated signaling pathway, and innate immune response (Fig. 2.A-D). Furthermore, C1RL has a positive correlation with immunosuppressive markers (Fig. 3.A-L). Less tumor purity and more M2 macrophage predict worse outcome of the patients with glioma [32,33] . More immune cells, especially M2 macrophages, migrated into glioma which expressing higher C1RL (Fig. 4.A and B). All these results are consisted to the hypothesis that C1RL paly an immunosuppressive role in glioma.
In conclusion, we analyzed the immunological and clinicopathological characteristics of C1RL in 2143 glioma patients from five datasets. The results indicated that C1RL is a negative biomarker for the patients with glioma. Furthermore, C1RL probably play an immunosuppressive role in the pathogenesis of glioma by triggering the activations of haptoglobin and C1s.

Author Contributions Statement
Xuejun Yang and Junyou Wang participated the design of this study. Luqing Tong performed the statistical analysis and drafted the manuscript. Gaojun Lin and Hui Wang carried out data acquisition.
All authors have read and approved the final manuscript.  The correlation of C1RL with tumor purity and leukocyte infiltration. The correlation of C1RL with nontumor cell infiltration in TCGAmic (A, GBM) and TCGAseq (B, glioma). Samples were listed from left to right according to their C1RL expression levels from low to high. Both immune score and stromal score exhibited in top panels. The tumor purity was showed in middle panels. The leukocytes infiltration levels were listed in bottom panels. High C1RL expression predicted shorter overall survival for GBM patients (E) and primary GBM patients (F-I). The primary GBM with high C1RL showed more resistance to radiochemotherapy (J), resection (K), and standard therapy (L).