Introduction:Morphological sub-classification of DLBCL into germinal center B-cell-like (GCB) and activated B-cell like (ABC) lacks sufficient reproducibility for risk stratification and predicting survival based on International Prognostic Index (IPI). The importance of stromal cells and their cross-talk with lymphoma cells is important in lymphomagenesis and progression. We postulate that mutations in stromal genes may regulate the rate of tumor progression in DLBCL and understanding the stromal-cellular interaction in DLBCL may help evolve better markers of prognostication and identify novel therapeutic targets.Methods:The study was approved by the Institute ethics committee and patients were enrolled after an informed consent. Frozen samples from 6 cases which included 4 cases of de novo nodal DLBCL (2 cases each of GCB and ABC subtypes as per Han's algorithm) and two control cases of non-neoplastic reactive lymphoid hyperplasia were subjected to AmpliSeqExome - Whole Exome Sequencing(WES) using Ion TorrentTM and compared against human reference genome (hg19). Torrent suiteTM 4.4 version (Life Technologies) and Ion ReporterTM software was used to perform tumor vs normal analysis by choosing variants for missense, frameshift insertions/ deletions, stoploss and single nucleotide variations (SNP). Filtering for common SNPs were done by UCSC common SNPs followed by inclusion in dbSNP and COSMIC and further filtered by predictive functional scores of SIFT < 0.05 and Polyphen (0.15 - 1.0). Variants were visualized using Integrated Genome Viewer (IGV) software. Bioinformatics analysis was done using PANTHER, STRING and PCViz and data mined for mutations in extracellular matrix genes. Selected genes were validated in DLBCL cases using Sanger sequencing BigDye terminator v3.1 cycle sequencing kit (Life Technologies) and ABI PRISM 3130 analyzer (Life Technologies). Further, in-vitro experiments were carried out to evaluate the effect of the conditioned media of CD19- stromal cells on the CD19+ lymphoma cells. Fresh tissue from 5 cases of nodal DLBCL were collected in RPMI 1640, 20% FBS and 1% penicillin and streptomycin and subjected to primary culture at 37°C and 5% CO2 humidified atmosphere. After 24 hours, the cultured cells were sorted into CD19+ and CD19- cells (BD FACSAriaTM) which were re-cultured for 24 hours separately. After this, the conditioned media of CD19- cells (comprising stromal cells predominantly) was transferred onto CD19+ cells and incubated for further 48 hours. Thereafter, cytokine (IL-6, IL-10, TNF-α, IFN-γ, IL-2, IL-4, IL-17) analysis of conditioned media in comparison to mock controls was performed using BD THI/THII kit (560484).Results:Analysis of filtered mutated genes using Panther revealed thirty four genes coding for extracellular matrix mutated specifically in GCB subtype and 26 mutated genes in ABC DLBCL. These genes were then analyzed for association in STRING. STRING identified LAMC3, AGRN, LAMA2, LAMB2, COL5A2, COL13A1,FN1,TGFB3, LTBP1 andADAMTS16 in GCB phenotype whereas LAMB1,LAMA3,COL4A2, COL28A1,COL5A3, IBSP, FGA, MUC6, MUC2, MUC5B,SPARCL1, VWF, GAS6 and USH2A were mutated in ABC DLBCL (Figure 1A and B). Mutations of COL5A2, COL13A1, LAMB2, MUC2, MUC6, and MUC5B were further validated using Sanger sequencing in individual DLBCL cases. This advocated the role of collagen scaffolding and laminin cross-linking in DLBCL progression. In vitro experiments for effect of stromal cells on the secreted cytokine profile revealed increased IL-6, IL-10, TNF-α, IFN-γ, IL-4, IL-17 levels respectively in the media of CD19+ sorted lymphoma cells upon addition of conditioned media of CD19-stromal cells in comparison to controls.Conclusion:Mutational profile of COL5A2, COL13A1, LAMB2, MUC2, MUC6, and MUC5B stromal genes in DLBCL may help in refinement of risk stratification based on morphology. The cross-talk between neoplastic cells and interacting tumor microenvironment cells has implications for therapy in DLBCL.GCB Mutation ProfileABC Mutation Profile [Display omitted] [Display omitted] [Display omitted] DisclosuresNo relevant conflicts of interest to declare.