Identification of immune-related candidate biomarkers in plasma of patients with sporadic vestibular schwannoma

Vestibular schwannoma (VS) is an intracranial tumor arising from neoplastic Schwann cells, and typically presenting with hearing loss. The traditional belief that hearing deficit is caused by physical expansion of the VS, compressing the auditory nerve, does not explain the common clinical finding that patients with small tumors can have profound hearing loss, suggesting that tumor-secreted factors could influence hearing ability in VS patients. We conducted profiling of patients' plasma for 66 immune-related factors in patients with sporadic VS (N>170) and identified and validated candidate biomarkers associated with tumor size (S100B) and hearing (MCP-3). We further identified a 9-biomarker panel (TNR-R2, MIF, CD30, MCP-3, IL-2R, BLC, TWEAK, eotaxin, S100B) with outstanding discriminatory ability for VS. These findings revealed possible therapeutic targets for VS, providing a unique diagnostic tool that may predict hearing change and tumor growth in VS patients, and may inform the timing of tumor resection to preserve hearing.

Methods
Study population and specimen collection
From 07/2015–04/2021, blood was prospectively collected from patients undergoing VS resection at Massachusetts Eye and Ear (MEE) in Boston, MA on the day of surgery, typically within 30 minutes of inducing general anesthesia and ≥1 hour before tumor microdissection (discovery cohort). Blood was similarly prospectively collected from 08/2021–06/2023 at Stanford Hospital in Palo Alto, CA (validation cohort). Blood from controls was collected at the Massachusetts General Hospital (MGH) Blood Donor Center in Boston, MA, and the Research Blood Components in Watertown, MA for the discovery cohort, and from the Stanford Hospital Blood Donor Center in Palo Alto, CA for the validation cohort. At collection, fresh blood was stored in EDTA vacutainer tubes (Becton Dickinson, NY, US) and kept at 4˚C without freezing. The whole blood samples were centrifuged at 2000 g for 10 min at 4°C. Plasma was separated and spun at 2000 g for 5 min at 4°C. Centrifuged plasma was filtered through 0.8 μm filter units (MF-Millipore MCE membrane, SLAA033SB; Millipore, Burlington, MA, US) and stored at -80°C until further use.

Eligible patients had unilateral, sporadic VS that had not been previously resected or irradiated. Of 186 and 50 enrolled VS patients in the discovery and validation cohorts, 163 and 50 met inclusion criteria, respectively, and were included in the analyses for comparison with 70 and 43 controls, respectively.

Clinical data
Clinical and demographic data were collected from patient charts, operative reports, pathology reports, and pre-operative radiographic imaging. Patient variables included age at tissue collection; pre-surgical tumor volume measured via high-resolution axial contrast-enhanced T1-weighted brain MRI; internal auditory canal protocol; and pre-surgical pure-tone audiometric threshold and WR measurements. MRI and hearing tests were those nearest to resection, typically ≤3 months. WR was defined as the percentage of spoken monosyllabic words discernable from a list typically read at 70 dB or the level at which a patient's speech intelligibility curve plateaus. Pure-tone audiometric thresholds at 0.5, 1, 2, and 3 kHz were used to calculate the PTA. Hearing groups were defined according to the American Academy of Otolaryngology–Head and Neck Surgery (AAO-HNS) Hearing Classification Guidelines. GH was defined as WR>70% and PTA<30 dB (AAO-HNS class A hearing). Otherwise, patients were classified as having PH (AAO-HNS class B, C and D hearing). A deaf ear was assigned a PTA of 125 dB and WR score of 0%. Serviceable hearing (SH) was defined as WR>50% and PTA<50 dB (AAO-HNS class A and B hearing). Otherwise, patients were classified as having non-serviceable hearing (NSH) (AAO-HNS class C and D hearing). Tumor resection was defined as GTR if there was complete tumor removal or a small tumor remnant no greater than 5x5x2 mm was left behind to preserve nerve integrity. Otherwise, it was defined as STR.

Biomarker measurements
Luminex (65 cytokines, chemokines, growth factors, and soluble receptors); electrochemoluminescence (IL-18, TNF-α, and FGF-2); and ELISA (S100B) assays were conducted as described in detail below. Potential biomarkers were defined as those detectable in the plasma of >75% of patients

Luminex assay
Simultaneous multiplex profiling of 65 immune-related factors composed of cytokines, chemokines, and growth factors was performed using a customized multiplex bead-based immunoassay - Immune Monitoring 65-Plex Human ProcartaPlex™ Panel (#EPX650-10065-901; ThermoFisher Scientific, Waltham, MA, US) according to manufacturer's instructions. The fluorescence-based signal was acquired on the Magpix instrument (Luminex, Austin, TX, US), and the values of analytes were calculated using ProcartaPlex Analyst 1.0 Software (ThermoFisher Scientific). The analytes with values below the lower limit of quantification in more than 75% of all samples were excluded from the analysis. For the accepted analytes, the values between 0 pg/mL and the lower limit of quantification, and those exceeding the upper limit of quantification, were approximated with the lowest and highest concentration representing these limits, respectively.

Electrochemiluminescence assay
The following candidate biomarkers were measured using electrochemiluminescence-based human assays from Meso Scale Diagnostics (MSD, Rockville, MD, US): IL-18 (U-PLEX Human IL-18), and TNF-α (U-PLEX Human TNF-α). All assays were conducted according to the manufacturer's protocols and signal detection was performed on the QuickPlex SQ 120 device (MSD). Pre-analytical data processing was done using MSD Discovery Workbench software (v4.0.12).

ELISA assay
The S100B ELISA Kit (#EZHS100B- 33K; EMD Millipore, Billerica, MA, US) was used to measure S100B protein levels in the plasma of VS patients and controls. ELISA assays were performed by adhering to the manufacturer's protocol. Absorbance was measured using Spectra MAX 190 plate Reader (Molecular Devices, Sunnyvale, CA, US), and the standard curve was plotted using SoftMax Pro software (v5.2; Molecular Devices, San Jose, CA, US).