Introduction: Glioblastoma is a common primary malignant brain tumour in adults with poor survival. Standard of care treatment with surgery, temozolomide (TMZ) and radiation therapy (RT) fails to improve patient outcomes. CXCL16 is a unique chemokine existing as a transmembrane form and a soluble form. Each form of CXCL16 has a different function; from chemotaxis of cells bearing CXCR6 to the accumulation of immune cells at an inflamed site. Here we evaluate CXCL16 as a biomarker of tumour aggression in primary and matched recurrent glioblastoma tissue, in patient-derived cell lines and in plasma.
Material and Methods: Sections of formalin-fixed, paraffin-embedded tumour tissue from 72 glioblastoma patients were labelled for CXCL16, CD68, IBA1, CD34 and GFAP by immunohistochemistry. Each slide was digitally imaged and a H-Score calculated for each marker in each tumour tissue using the HALO image analysis platform. Kaplan-Meier analysis with log rank tests were used to examine overall survival (OS) and progression free survival (PFS) of glioblastoma patients with respect to median CXCL16, CD68 and IBA1 H score. Twelve patient-derived glioblastoma cell lines were treated with patient appropriate doses of TMZ (35mm) and RT (2Gg). RNASeq (Taqman Assay MTO) was performed post treatment. The soluble CXCL16 concentration in patient plasma at both primary and recurrent surgery was analysed by ELISA.
Results: Using the Paired Sample t-Test, the difference between mean CXCL16 H-scores of primary and recurrent tumours was significant, t(72)=-3.54, p<0.001, d=-0.40. Upon examining the % of CXCL16 cell labelling as either weak, moderate or strong, it is the weak and moderate CXCL16 positive cells which were significantly different between primary and recurrent tumours (t(72)=-3.70, p<0.001, d=-0.42), (t(72)=-2.33, p<0.05, d=-0.26) respectively. There was decreased expression of CD68 in recurrent tissue. Although there were significant differences in CD68 and CXCL16 expression, there was no difference in OS between patients with primary and recurrent GBM tumours. There was no significant difference between the level of CXCL16 mRNA before and after treatment of glioblastoma cell lines with temozolomide and radiation (p=0.434).
Conclusion: The medium expression of CXCL16 and CD68 was decreased in recurrent compared to primary tumours, but these changes did not correlate with a change in patient survival. Whilst the expression of CXCL16 in patient-derived glioblastoma cells lines did not change after standard treatment, this suggests the changes in the tissue at recurrence may not due to treatment effects. The role CXCL16 plays in tumour recurrence needs further investigation.