Salivary Gene Expression Changes after Total Body Irradiation in Leukemia Patients: Toward Non-invasive Biodosimetry.

Blood is the standard matrix for gene expression (GE)-based biodosimetry, but less invasive sampling methods are needed in emergency settings and for longitudinal monitoring. Saliva, which contains leukocytes and an ultrafiltrate of blood, is a promising alternative. In this human in vivo study, we analyzed radiation-induced shifts in salivary gene expression for biodosimetric applications and evaluated their similarity to gene expression patterns in blood. Beyond dose estimation, such saliva-based gene expression signatures may support risk stratification and early prediction of clinical outcomes, particularly when repeated sampling is required or blood collection is not feasible. Matched blood and saliva samples were collected from leukemia patients (n = 31) undergoing fractionated total-body irradiation (TBI) for their myeloablative treatment (1.5-4 Gy total dose, 1.5-2 Gy per fraction). Samples were taken before and 24 h after the first day of radiation treatment. The expression of radiation-responsive genes (GADD45A, CCNG1, CDKN1A, PHPT1, SESN1, FDXR, DDB2, POU2AF1, and WNT3) was analyzed for 28 patients (three patients excluded) using quantitative real-time PCR (RT-qPCR). Significant upregulation was observed for GADD45A (median fold change = 1.52, P = 0.003), CCNG1 (median fold change = 1.73; P = 0.003), and DDB2 (median fold change = 1.60; P = 0.05), as well as WNT3 (median fold change = 1.84; P = 0.038), demonstrating that saliva shows molecular responses to radiation exposure. Notably, 54% of the patients exhibited radiation-responsive upregulation of GADD45A and CCNG1, indicating that saliva can detect radiation-responsive gene expression despite substantial inter-individual variability. Corresponding blood samples were obtained from 16 patients. Statistically significant upregulation of multiple radiation-responsive genes was observed: GADD45A (median fold change = 2.25; P < 0.001), CCNG1 (median fold change = 1.64; P = 0.010), DDB2 (median fold change = 1.91; P < 0.001), CDKN1A (median fold change = 2.45; P = 0.001), SESN1 (median fold change = 1.75; P < 0.001), and FDXR (median fold change = 2.53; P = 0.001). A significant downregulation was observed for POU2AF1 (median fold change = 0.49; P = 0.013). In conclusion, detection of radiation-induced gene expression changes in saliva may serve as a minimally invasive tool for biodosimetry. However, given the greater inter-individual variability observed in saliva compared to blood, further optimization and validation is essential before clinical implementation.