Temporal Profiling of the Unfolded Protein Response Identifies PERK Activation as a Marker of Complicated Sepsis Outcomes

Authors:
Surmai Shukla, Timothy R. Billiar, Andrew C. Sayce

Body of Abstract:
Background:

Sepsis is a life-threatening condition caused by a dysregulated host response to infection, leading to organ dysfunction. The cellular stress response during sepsis extends beyond inflammation to involve dysregulated proteostasis and Endoplasmic Reticulum (ER) stress. We aimed to characterize unfolded protein response (UPR)-related molecular signatures in sepsis to uncover potential mechanistic drivers of organ failure in sepsis and its relationship to the Endotypes in MARS consortium.

Methods:

Whole-blood human bulk RNA transcriptomic data (n=802) was analyzed. UPR arm- specific (ATF6, IRE1 α, PERK) gene signatures were developed based on previously published Perturb-Seq gene classification. Briefly, z-score based gene activation, a composite representative of pathway activation, was compared across clinical recovery classes and four clinical endotypes of the MARS sepsis registry, MARS1= immunoparalysed high mortality group; MARS2 = hyperinflammatory innate activation endotype; MARS3 = increased adaptive immunity and low risk endotype, MARS4 = interferon pathway driven hyperinflammatory group. Associations were evaluated using principal componentanalysis, group comparison tests, and linear regression.

Results:

UPR-related genes alone can differentiate clinical endotypes and reveal marked heterogeneity. Global UPR score differed between controls and patients with sepsis (p<0.05). MARS1 (highest mortality) showed decreased global UPR activity as compared with MARS2 (p<0.001) and MARS3 (p<0.01). PERK scores were elevated in non-survivors (p=<0.001) and PERK was significantly increased in patients who died from sepsis (p<0.001). This effect was strongest in MARS1 (p<0.01) but also noted in MARS2 (p<0.05). MARS4 had higher activation of PERK but no significant difference was detected for mortality. Greater activation of ATF6 was seen in MARS2 and MARS3 as compared with MARS1 endotype. IRE1α scores were increased in sepsis (p<0.001) and endotype differences mirrored those findings for ATF6a. Conclusion: ER-stress biology mirrored MARS endotype heterogeneity, implying that regulation of proteostasis correlates with clinical outcomes such as immune status and mortality. Dysregulated UPR signaling differentiates clinical sepsis endotypes;sepsis triggers UPR with predominant PERK activation in the patients who die, with suppression of ATF6/IRE1 pathway. We present the first quantitative approach to measure severity of ER stress in human transcriptomics. Our results indicate that manipulation of UPR sub-pathways may offer a precision therapeutic window in sepsis patients directed by endotype and associated clinical features.