AUTHORS
Elizabeth A Linton, PhD, Jung Nam Joo, PhD2, Xunde Wang, PhD3, Britney Kruah4, Nancy Asomaning, MS, CCRC5, Swee Lay Thein, MD, DSc6
AFFILIATIONS
1 National Heart, Lung, and Blood Institute, Bethesda, Maryland, United States
2 National Heart, Lung, and Blood Institute, Bethesda, Maryland, United States
3 National Heart, Lung, and Blood Institute, Bethesda, Maryland, United States
4 National Heart, Lung, and Blood Institute, Bethesda, Maryland, United States
5 National Heart, Lung, and Blood Institute, Bethesda, Maryland, United States
6 National Heart, Lung, and Blood Institute, Bethesda, Maryland, United States
ABSTRACT
Introduction
Untargeted proteomics offers a powerful approach to uncover disease mechanisms and identify novel biomarkers and potential therapeutic targets. In diseases with clinical presentations as heterogenous as sickle cell disease (SCD), it is important to conduct such studies among participants with different profiles of disease-modifying and transformative therapy exposure. Here, we present a longitudinal plasma proteomics study examining correlates of mortality among adults with SCD enrolled in a natural history study.
Methods
We used Olink® Explore HT to measure the Normalized Protein eXpression (NPX) of 5,416 plasma proteins in participants with SCD genotypes SS and Sβ0 thalassemia. Samples were collected at a routine baseline outpatient study visit outside of vaso-occlusive crisis. The primary outcome – all-cause mortality as of January 22, 2026 – was ascertained via proxy interview with participants’ family, medical records, and the National and Social Security Death Indices. To identify proteins associated with all-cause mortality, we fit a multivariable Cox proportional hazards model for time to death, with each protein’s scaled NPX included as the predictor. Follow-up continued until death or the end of the study, with participants censored at the date of first stem cell transplant or the end of follow-up. We estimated each protein’s adjusted hazard ratio (adjHR) for mortality, controlling for sex, age in years, and estimated glomerular filtration rate (eGFR 2021 CKD-EPI creatinine equation) from the nearest steady-state labs within +/- 1 month of plasma collection. eGFR was modeled as a continuous variable using penalized smoothing splines with 3 degrees of freedom to account for a nonlinear association with the outcome. Proteins were flagged as statistically associated with mortality if the False Discovery Rate (FDR) was < 0.05.
Results
The median age of the 285 participants analyzed was 31 years at baseline [IQR: 24 – 41]; 48% were female. The median length of follow-up was 14 years, with a 10-year survival rate of 77%. At baseline, 41% were on hydroxyurea; 35 participants later underwent stem cell transplant, of whom 9 died by the end of follow-up.
In the full cohort, after controlling for age, sex, and eGFR, 72 proteins were significantly associated with mortality; two were associated with lower risk of mortality (BPIFB1 [adjHR = 0.64, 95% CI: 0.53 , 0.77] and HMOX1 [adjHR = 0.68, 95% CI: 0.55 , 0.84]), while 70 proteins were associated with higher risk of mortality. The proteins with the strongest associations included: ASGR1 [adjHR = 1.77, 95% CI = 1.40 , 2.23], OCLN [adjHR = 1.74, 95% CI = 1.36 , 2.23], TREM2 [adjHR = 1.73, 95% CI = 1.38 , 2.16], CDCP1 [adjHR = 1.68, 95% CI = 1.38 , 2.06], and CRELD1 [adjHR = 1.67, 95% CI = 1.26 , 2.21]. Forty seven of the 72 proteins statistically associated with mortality had HRs ≤0.67 or ≥1.5.The top Gene Ontology defined biological processes associated with these proteins involved glial cell apoptotic processes, symbiont entry into host / host cell, leukocyte proliferation, myeloid leukocyte activation, and regulation of mononuclear cell proliferation.
When examined separately by hydroxyurea exposure at baseline, no proteins were statistically associated with mortality in participants who were on hydroxyurea. Among those not on hydroxyurea, however, 22 proteins were associated with mortality. BPIFB1 was associated with lower risk of death [adjHR = 0.58, 95% CI: 0.44 , 0.76] while the 5 proteins with the strongest statistical associations with higher mortality were: ASGR1 [adjHR = 2.52, 95% CI = 1.66 , 3.81], SOWAHC [adjHR = 2.51, 95% CI = 1.64 , 3.84], CRELD1 [adjHR = 2.44, 95% CI = 1.63 , 3.66], SCARB2 [adjHR = 2.25, 95% CI = 1.48 , 3.43], and CD300A [adjHR = 2.25, 95% CI = 1.49 , 3.39].
Discussion
High-throughput untargeted proteomics studies in SCD can illuminate mechanisms underlying mortality risk. Here, several plasma proteins were associated with time-to-death, particularly among those not on hydroxyurea at baseline. This suggests that proteomic profiling may help reveal biological pathways involved in mortality and treatment exposure. Future directions for this work include assaying larger cohorts to confirm these findings and comparing proteins associated with time-to-death across other SCD genotypes.
DISCLOSURES
Elizabeth A Linton, PhD: No relevant conflicts of interest to declare.