Lymphotoxin alpha (LTA) secretion correlates with gene expression and myeloid malignancy disease state

Abstract

Abstract Background: Myelodysplastic Syndromes (MDS) are hematopoietic stem cells malignancies whose pathogenesis is, in part, driven by chronic inflammation of the bone marrow. This inflammation fosters malignant clone expansion while suppressing normal hematopoiesis resulting in the characteristic marrow dysplasia and peripheral cytopenias observed in these patients. Although gene expression immune signatures have been suggested as prognostic indicators in MDS, to our knowledge, correlations between protein secretion and gene expression have not been extensively explored. Here, we performed RNA sequencing of bone marrow cells and matched multiplexed cytokine analysis of bone marrow plasma from participants enrolled on the Comprehensive Molecular and Clinical Evaluation of Pediatric and Adult MDS (NCT05350748) study at the National Cancer Institute. Methods: Bone marrow aspirates were collected from participants and total RNA was isolated from cells after RBC lysis. RNAseq libraries were prepared using the TruSeq RNA Exome kit and sequenced on either NextSeq 550DX or NovaSeq 6000. Reads were aligned to the human genome using STAR, and read quantification was performed with RSEM. Raw counts were normalized to log2-transformed transcripts per million (TPM). Low expressing genes were filtered out and only those with log2(TPM) > 0.5 in at least two patients were retained. Bone marrow plasma separated from cells by centrifugation was used for cytokine analysis using the Olink® Target 48 Immune Surveillance Reagent and Target 48 Cytokine Reagent multiplex kits. Normalized protein expression (NPX) levels are reported. Spearman correlation analysis was performed to assess correlations between RNAseq expression, protein abundance, and disease states. P-values less than 0.05 were considered statistically significant. Results: We analyzed the plasma levels of 88 cytokines in 27 cases (4 healthy donors, 2 cases found to have no malignancy, 4 with CHIP or CCUS, 13 with MDS, and 4 AML cases). Unsupervised hierarchal clustering grouped healthy donors with non-malignancy cases and those with CHIP or CCUS together; while MDS and AML cases clustered together. Notably, CHIP or CCUS patients grouped either with non-malignant individuals, or those who had hematologic malignancy depending on the individual cytokine being tested. We then correlated RNAseq data with plasma cytokine levels in 18 participants (9 cases with MDS or MDS/MPN, 3 cases with AML, 4 with CHIP or CCUS, and 2 cases without malignancy). Correlation analysis included 57 cytokines that had paired RNAseq data. Principal component analysis demonstrated that RNAseq better discriminated disease states compared to plasma cytokine levels. Expression at both the gene and protein level of only OLR1, LIF, FLT3LG, LTA, and PDCD1 were positively and significantly correlated (rho=0.70, 0.68, 0.60, 0.58, and 0.55, respectively). Gene expression levels of TNFSF14, TNFSF10, OLR1, IL16, AREG, IFNG, IL1RN, CSF3, KRT18, CD28, IL32, and LTA were significantly correlated with disease state (comparing no malignancy, CHIP or CCUS, MDS/MDS-MPN, and AML) (rho = -0.69, -0.64, -0.63, -0.62, 0.56, -0.56, -0,56, 0.52, 0.51, -0.51, -0.50, and -0.49, respectively.) Plasma protein levels of LTA, HGF, CD80, TNFSF10, FLT3LG, CCL3, IL15 and IL6 were significantly correlated with disease state (rho=-0.72, 0.68, 0.63, -0.59, -0.55, 0.52, 0.49, 0.48, respectively). Interestingly, LTA was the only analyte for which both gene and protein expressions levels associated with disease state. Importantly, both lower LTA gene and protein levels were associated with the more advanced disease states (MDS and AML). LTA gene expression (log2TPM) was 1.24, 1.23, 1.19, and 0.34 in no malignancy, CHIP/CCUS, MDS/MDS-MPN, and AML, respectively while normalized protein values were 5.61, 5.96, 5.29, and 4.84, respectively. Conclusions: LTA (lymphotoxin alpha), also known as TNFbeta, is a member of the tumor necrosis factor family that heterodimerizes with lymphotoxin beta to modulate a variety of inflammatory processes. Our data suggests that LTA may have a protective role in hematologic malignancy and disease progression warranting further investigation. Given the lack of gene expression signatures that have been useful for patient stratification, our results underscore the value of paired secreted protein levels and gene expression patterns to help inform stratification and mechanistic understanding in myeloid malignancies.