Keep an Eye on CXCL-13: A Promising Biomarker to Help IL-10 and IL-6 for the Diagnosis of Intraocular Lymphomas

Intraocular lymphomas (i-OL) are rare and aggressive subsets of central nervous system (CNS) lymphomas. They include primary vitreoretinal lymphomas (PVRL) but also concomitant localizations of primary CNS lymphomas (PCNSL) and systemic lymphomas (SCNSL). In most cases, i-OL are difficult to detect and often misdiagnosed as their clinical features can mimic other ocular conditions. To date, no independent biological tool can firmly diagnose i-OL; the combination of cytologic examination, immunochemistry, flow cytometry, and molecular analysis is required. Hence i-OL diagnosis remains challenging and diagnostic delays are still frequent. Recent data have shown that the diagnosis is achieved earlier, in part thanks to improved diagnostic tools, especially soluble biomarkers. Indeed, Interleukin (IL)-10 and 6 play an important role in the diagnostic work-up. Nowadays, the IL-10:IL-6 ratio and the recently published ISOLD diagnostic score based on these two biomarkers (Costopoulos, 2016) are of great help for i-OL diagnosis. However, some difficult cases cannot be classified in the ISOLD certainty clusters and remain in the grey zone. Interestingly, CXC chemokine ligand (CXCL) 13, the ligand of CXCR5, a chemokine that partly regulates B-cell trafficking, has been shown to increase in cerebrospinal fluids of CNS lymphoma patients (Rubenstein, 2013). Furthermore, CXCR5 overexpression has also been reported in PVRL cells (Chan, 2003). For the first time, we measured the concentration of CXCL-13, in addition to IL-10 and IL-6, in vitreous samples. Our aim was to determine the benefit of CXCL-13 for prompt i-OL diagnosis.

Our retrospective study includes vitreous samples from 47 patients at the time of diagnosis: 28 i-OL patients (16 PVRL including 6 with concomitant cerebral involvement, 7 PCNSL with intraocular relapse, and 5 SCNSL with intraocular impairment) and 19 non lymphoma patients (uveitis). Vitreous samples were frozen within 6 hours and assays were performed on freshly thawed samples. IL-10 and IL-6 were measured using the quantitative Cytometric Bead Array® technique (human IL-10 and IL-6 CBA kits; BD Biosciences^TM, Le Pont de Claix, France) on FACSCanto II cytometer (BD Biosciences^TM) following the manufacturer's instruction, with a detection limit of 2.5 pg/ml. As commercialized CBA kit for CXCL-13 is not available, we custom-built a CBA test detecting CXCL-13 within a range of 19.5 to 5,000 pg/ml. The CBA technique allows multiplex dosages on small sample volumes (50µl). Statistical analyses were performed with Prism 5 software (GraphPad Software, San Diego, CA, USA). Intergroup comparisons were assessed with the non parametric Mann-Whitney U -test. P -values <0.05 were considered statistically significant.

We confirm that IL-10 is significantly increased in i-OL (median IL-10: 439 [28-5,000pg/ml]) compared to uveitis (median IL-10: 2.5 [2.5-18pg/ml]; P <0.0001). Conversely, IL-6 is increased in uveitis (median IL-6: 324 [17-5,000pg/ml]) compared to i-OL (median IL-6: 54 [7-5,000pg/ml]); P <0.0005). As expected, the previously described IL-10:IL-6 ratio is strictly below 1 for all uveitis cases. However, 3 i-OL also have a ratio <1 which can mislead the diagnosis towards uveitis. Of them, one was a secondary localization of a systemic DLBCL (patient #1) and 2 were intraocular relapses of a preexisting PCNSL. Of note, in the latter, intraocular impairment were diagnosed early, with low IL-10 levels. That explains why the IL-10:IL-6 ratio approach failed. The use of the ISOLD scoring system correctly reclassifies patient #1; the other two remain in the ISOLD grey zone. Concerning CXCL-13, the two cohorts are significantly different with higher CXCL-13 levels in i-OL (median CXCL-13: 4,544 [94-5,000pg/ml] with CXCL-13 >5,000 pg/ml in 13/28 samples) than in uveitis (median CXCL-13: 185 [11-2,970]); P <0.0001). Interestingly, the 2 uncertain ISOLD classified cases display high CXCL-13 level (>5,000 and 2,637 pg/ml), which is an additional argument for i-OL diagnosis.

CXCL-13 was tested for the first time using a self-made CBA approach in vitreous samples of patients suffering from i-OL and in a case-control cohort. CXCL-13 is significantly higher in i-OL compared to uveitis; thus it could add a major value for the ISOLD grey-zone patients. More patients are needed to confirm these promising preliminary results and to define new diagnostic algorithms for all i-OL diagnosis.