Hemolysis Score Is Associated with Increased TNF Cytokines, Elevated BNP, and Hospitalizations in Individuals with Sickle Cell Disease (SCD)

Background: Hemolysis is a prominent feature of sickle cell disease (SCD). Hemolysis triggers inflammation, which leads to endothelial damage, vaso-occlusive events, and ultimately organ injury. However, there is still uncertainty which inflammatory systems are triggered by hemolysis and what the clinical consequences are. If the causative pathways were known, better targeted therapies could be developed. One inflammatory pathway of interest involves tumor necrosis factor (TNF). In vitro and SCD mouse-studies have shown that hemolysis triggers TNF cytokines. In patients with SCD, TNF is elevated and associated with vaso-occlusive crises, acute chest, and chronic kidney disease. The relationship between hemolysis, TNF, and clinical outcomes has not been well studied concurrently. In a cohort of patients with SCD followed prospectively for one year, we investigated the hypothesis that hemolysis causes organ injury through activation of TNF inflammation.

Methods: We used data from the IMPROVE 2 trial, a randomized, placebo-controlled study investigating the impact of inhaled mometasone in non-asthmatic patients with SCD (HbSS, HbSβ⁰-thalassemia) and episodic cough or wheeze. Clinical, laboratory, and O-link proteomic data were collected every 8 weeks for 1 year. The degree of hemolysis was determined by the hemolysis score, a previously validated index, derived from the first component of a principal components analysis of log-standardized values of lactate dehydrogenase (LDH), aspartate aminotransferase (AST), total bilirubin, and reticulocyte percentage. Our outcomes of interest were hospitalization, brain natriuretic peptide (BNP), and urine albumin/creatinine ratio (UACR). Protein concentration of TNF cytokines (TNF, TNFB, TNFRSF9, TNFSF14, TRAIL, TWEAK, TGF-alpha, and TRANCE) were normalized and expressed on a log2 scale in the O-link data set. Pearson correlation analysis was performed to compare continuous variables. The t-test was used to compare the distribution of TNF cytokines for dichotomized outcomes.

Results: Hemolysis score was associated with BNP (Pearson R=0.298, p<0.001) but not with UACR (R=0.01, p=0.840). There was a total of 47 hospitalizations among 28 patients. A high hemolysis score (defined as ≥50^th percentile in the cohort) was an independent predictor for hospitalization (OR=2.14, 95% CI 1.11-4.113, p=0.02).

Regarding the relationship between hemolysis and TNF cytokines, we found that the hemolysis score was associated with TNF (R=0.26, p<0.001), TNFB (R=0.29, p<0.001), TNFRSF9 (R=0.13, p=0.017), TRAIL (R=0.18, p<0.001), and TWEAK (R=0.26, p<0.001), but not with TNFSF14 (R=0.073, p=0.170), TGF-alpha (R=0.04, p=0.511) or TRANCE (R=-0.07, p=0.202). Of note, the use of inhaled corticosteroid was found to have no impact on these markers.

For TNF cytokines and clinical outcomes, we found that those with hospitalizations had significantly higher TNFSF14 (t-test, mean difference 0.47, p<0.001) and TGF-alpha (mean difference 0.19, p=0.003). Patients with elevated BNP (>100 pg/mL) had higher TNF (mean difference 0.44, p=0.002) and TNFB (mean difference 0.36, p<0.001). Elevated UACR (>100) was associated with higher TNF (mean difference 0.19, p=0.007), TNFRSF9 (mean difference 0.185, p=0.011), and TGF-alpha (mean difference 0.16, p<0.001).

Conclusion: A high hemolysis score is an independent predictor for hospitalizations and is associated with elevated BNP in patients with SCD. Hemolysis and elevated BNP were also associated with elevated TNF and TNFB. This suggests that TNF may be involved in the pathogenesis of hemolysis-related cardiac injury and vaso-occlusive events. Anti-TNF agents should be examined to see if they may be effective targeted agents to prevent vaso-occlusive events or cardiac damage in SCD.