PCSK9 is a critical regulator of the innate immune response and septic shock outcome

Abstract

A decrease in the activity of proprotein convertase subtilisin/kexin type 9 (PCSK9) increases the amount of low-density lipoprotein (LDL) receptors on liver cells and, therefore, LDL clearance. The clearance of lipids from pathogens is related to endogenous lipid clearance; thus, PCSK9 may also regulate removal of pathogen lipids such as lipopolysaccharide (LPS). Compared to controls, Pcsk9 knockout mice displayed decreases in inflammatory cytokine production and in other physiological responses to LPS. In human liver cells, PCSK9 inhibited LPS uptake, a necessary step in systemic clearance and detoxification. Pharmacological inhibition of PCSK9 improved survival and inflammation in murine polymicrobial peritonitis. Human PCSK9 loss-of-function genetic variants were associated with improved survival in septic shock patients and a decrease in inflammatory cytokine response both in septic shock patients and in healthy volunteers after LPS administration. The PCSK9 effect was abrogated in LDL receptor (LDLR) knockout mice and in humans who are homozygous for an LDLR variant that is resistant to PCSK9. Together, our results show that reduced PCSK9 function is associated with increased pathogen lipid clearance via the LDLR, a decreased inflammatory response, and improved septic shock outcome.

Fig. 1. Response to LPS in wild-type and Pcsk9^∑/− mice

Top: Plasma cytokine concentrations at baseline (top) and after LPS. A multiplex cytokine assay measured TNFα, IL-6, IL-10, JE, and MIP-2 at baseline and 6 hours after intraperitoneal injection of LPS (20 mg/kg). There were no significant differences at baseline shown in the bottom right frame (n = 9), whereas all cytokines were significantly (P < 0.05) lower in Pcsk9^−/− mice (n = 9) than wild-type control mice (n = 9) after exposure to LPS. Bottom: (A) Activity phenotype. All 10 wild-type control mice had an activity index of 0 (no movement) by 6 hours after LPS administration compared to none of 10 Pcsk9^−/− mice. Analysis of variance (ANOVA) demonstrates a difference between Pcsk9^−/− and wild-type mice (P = 2.2 × 10⁻¹⁶) that was statistically significant (*P < 0.05) in hours 3 to 6. Typically, an activity index less than 0.5 represents a terminal state. (B) Body temperature. Wild-type control mice demonstrated a progressive loss of body temperature over the 6 hours after LPS administration such that 6 of the 10 mice had a body temperature <32°C, whereas none of 10 Pcsk9^−/− mice had their temperature drop below 32°C. ANOVA demonstrates a difference between Pcsk9^−/− and wild-type mice (P = 0.00035) that was statistically significant (*P < 0.05) in hours 4 to 6. Typically, a sustained temperature lower than 32°C represents a terminal state. (C) Blood pressure. Mean arterial pressure is preserved and was significantly higher in Pcsk9^−/− mice than in wild-type control mice 6 hours after administration of LPS (P = 0.014 by ANOVA). (D) Left ventricular ejection fraction. Left ventricular function is preserved in Pcsk9^−/− mice compared to wild-type control mice as evidenced by significantly higher left ventricular ejection fraction 6 hours after LPS infusion (P = 0.031 by ANOVA).

Fig. 2. Effect ofPCSK9on LPS uptake by HepG2 cells

(A) Representative histogram of LPS uptake by HepG2 cells. HepG2 cells at 80% confluence were cultured in 80% Dulbecco’s modified Eagle’s medium (DMEM) and 20% human plasma from pooled healthy donors. Three hours before LPS treatment and at 4 and 19 hours after treatment, cells were treated with recombinant human PCSK9 (3 µg/ml) or vehicle control. Cells were then treated with Alexa Fluor 488–conjugated LPS. After 24 hours of LPS treatment, cells were analyzed by flow cytometry for MFI. (B) Group mean data from n = 6 flow cytometry runs. In HepG2 cells known to have only a trace of secreted PCSK9, the MFI of cells exposed to Alexa 488–bound LPS was 1.49 ± 0.06, whereas the addition of PCSK9 (3 µg/ml) reduced the MFI to 0.52 ± 0.03. This effect is highly significant with a P value of 8.20 × 10⁻⁷ by unpaired t test.

Fig. 3. Double-blind, randomized control study of PCSK9 inhibition in a murine model of polymicrobial peritonitis

All study personnel were blinded as to group assignments (PCSK9-blocking antibody or isotype control IgG). C57BL/6 mice [n = 14 (PCSK9-blocking antibody) and n = 15 (isotype control)] were followed for 72 hours after CLP, and the time of death was recorded. Treatment with the antibiotic imipenem and the blinded study drug (PCSK9-blocking antibody of isotype control) began 6 hours after CLP, and both were administered daily thereafter. (A) There was a significant improvement in survival with PCSK9-blocking antibody (71%) versus isotype control treatment (40%) by log-rank test (P = 0.034). (B) Serial plasma inflammatory cytokine measurements beginning the first day (day 1) after CLP are shown. Using ANOVA, PCSK9-blocking antibody increases the clearance rate of TNFα (P = 0.027), IL-6 (P = 0.051), IL-10 (P = 0.068), JE (P = 0.0085), and MIP-2 (P = 0.040). Ab, antibody.

Fig. 4. Effect of PCSK9 genetic variants on human septic shock

(A) LOF and GOF 28-day survival curves in VASST. Patients in VASST having at least one PCSK9 LOF allele (solid black line) had increased survival over 28 days compared to patients without a LOF or GOF allele (dashed black line) and those patients having a GOF allele (gray line) (P = 0.0029 by log-rank test). (B) Survival curve for SPH cohort by LOF and GOF genotype. Similarly, in the SPH replication cohort, patients carrying at least one LOF allele (solid black line) demonstrated increased survival compared to those who did not (dashed black line), and patients carrying a GOF allele (gray line) had decreased survival (P = 0.022 by log-rank test). (C) Plot of individual and pooled hazard ratios by SNP in a combined VASST + SPH sepsis cohort. (D) Plasma cytokines taken upon enrollment (median, 18 hours after the onset of septic shock) in VASST patients. Patients carrying a PCSK9 LOF allele had significantly (using ANOVA), or a trend toward, lower plasma cytokine concentrations compared to patients carrying a GOF allele (Fig. 3C) for TNFα (P = 0.01), IL-6 (P = 0.042), the anti-inflammatory cytokine IL-10 (P = 0.008), the human CC chemokine MCP-1 (P = 0.035), and the human CXC chemokine IL-8 (P = 0.075). (E) Survival in VASST of LOF and No LOF patients by LDLR rs688 genotype. In patients carrying two copies of the LDLR LOF allele of rs699, survival was not altered by PCSK9 LOF (P = 0.67 by log-rank test). In contrast, for patients carrying at least one wild-type allele of LDLR, PCSK9 LOF was significantly associated with improved survival (P = 0.009 by log-rank test).