PCSK9 is associated with mortality inpatients with septic shock: data from the ALBIOS study
PCSK9與敗血性休克患者的死亡率有關:來自ALBIOS研究的數據。
影響因子 : 6.871
發表時間：2021-02-01
來源期刊：JOURNAL OF INTERNAL MEDICINE
DOI：10.1111/joim.13150
文章類型：雜誌文章
作者列表：Vecchié A, Bonaventura A, Meessen J,Novelli D, Minetti S, Elia E, Ferrara D, Ansaldo AM, Scaravilli V, Villa S,Ferla L, Caironi P, Latini R, Carbone F, Montecucco F, ALBIOS Biomarkers StudyInvestigators.
Background. Pro-protein convertasesubtilisin/kexin 9 (PCSK9) is a proenzyme primarily known to regulatelow-density lipoprotein receptor re-uptake on hepatocytes. Whether PCSK9 canconcurrently trigger inflammation or not remains unclear. Here, we investigatedthe potential association between circulating levels of PCSK9 and mortality inpatients with severe sepsis or septic shock.
背景 : 前蛋白轉化酶枯草桿菌蛋白酶/kexin9 (PCSK9) 是一種已知的主要調節肝細胞低密度脂蛋白受體再攝取的酶原。PCSK9是否可以同時引發炎尚不清楚。在這裡，我們調查了嚴重膿毒症或膿毒休克患者的PCSK9循環水平與死亡率之間的潛在關聯。
Methods. Plasma PCSK9 levels at days 1, 2and 7 were measured in 958 patients with severe sepsis or septic shockpreviously enrolled in the Albumin Italian Outcome Sepsis (ALBIOS) trial.Correlations between levels of PCSK9 and pentraxin 3 (PTX3), a biomarker ofdisease severity, were evaluated with ranked Spearman’s coefficients. Coxproportional hazards models were used to assess the association of PCSK9 levelsat day 1 with 28- and 90-day mortality.
方法 : 在958例先前參加ALBIOS 試驗的嚴重敗血症或敗血性休克患者中，測量了第1、2和7天的血漿PCSK9水平。PCSK9和PTX3 (疾病嚴重程度的生物標誌物) 的水平之間的相關性通過排名Spearman係數進行評估。Cox比例風險模型用於評估第1天PCSK9水平與28天和90天死亡率的關係。
Results. Median plasma PCSK9 levels were278 [182– 452] ng mL1 on day 1. PCSK9 correlated positively with PTX3 at thethree time-points, and patients with septic shock within the first quartile ofPCSK9 showed higher levels of PTX3. Similar mortality rates were observed inpatients with severe sepsis across PCSK9 quartiles. Patients with septic shockwith lower PCSK9 levels on day 1 (within the first quartile) showed the highest28- and 90-day mortality rate as compared to other quartiles.
結果 : 第1天血漿PCSK9的中位數mL-1為278[182-452] ng。PCSK9在三個時間點與PTX3呈正相關，PCSK9第一個四分位數內膿毒休克的患者顯示出更高的PTX3水平。在PCSK9四分位數的嚴重敗血症患者中觀察到相似的死亡率。與其他四分位數相比，第1天PCSK9水平較低的敗血性休克患者 (在第一個四分位數內) 28天和90天的死亡率很高。
Conclusion. In our sub-analysis of theALBIOS trial, we found that patients with septic shock presenting with lowerplasma PCSK9 levels experienced higher mortality rate. Further studies arewarranted to better evaluate the pathophysiological role of PCSK9 in sepsis.
結論 : 在我們對ALBIOS試驗的分析中，我們發現敗血症性休克患者血漿PCSK9水平較低，死亡率較高。需要進一步的研究以評估PCSK9在敗血症中的病理生理作用。
Introduction
Multiorgan dysfunction during sepsis and itsevolution towards septic shock are critically influenced by dysregulation ofthe host’s inflammatory response [1]. Different biomarkers have been studied tokeep track of these alterations, such as Creactive protein (CRP),procalcitonin, osteopontin and pentraxin 3 (PTX3) [2-4]. Several signalingcascades are involved in the defensive mechanisms against infections with acentral role played by cholesterol and lipids [5]. Indeed, lipids in microbialcell walls bind to toll-like receptors (TLRs) and trigger inflammatoryresponses.（研究了不同的生物標誌物來反應這些變化，如CRP、降鈣素原、骨橋蛋白和pentraxin 3 (PTX3)。一些信號級聯參與了對抗感染的防禦機制，膽固醇和脂質扮演了核心角色。事實上，微生物細胞壁中的脂質與toll樣受體(TLRs)結合併觸發炎症反應。
Pro-protein convertase subtilisin/kexin 9(PCSK9) is a proenzyme which regulates circulating cholesterol lipids. Theactive form of PCSK9 binds to the low-density lipoprotein (LDL) receptor(LDL-R) on hepatocytes, interrupts LDLR’s cell surface recycling pathway andstimulates the endocytosis and degradation of LDLRs in lysosomes [6]. Thisfunction of PCSK9 has also been demonstrated during infections, when freemicrobial lipids are incorporated in LDLs and subsequently cleared from bloodby binding to hepatocyte LDL-Rs [7-9]. Hence, this process of pathogen lipidclearance reduces the amount of microbial lipids available for TLR binding,thus decreasing the inflammatory response [10].（PCSK9的這種功能也在感染過程中得到證實，當游離微生物的脂質進入ldl中，隨後通過與肝細胞LDL-Rs結合從血液中清除。因此，這一清除病原體脂質的過程減少了可與TLR結合的微生物脂質數量，從而降低了炎症反應。）
To date, PCSK9 was mostly studied incardiovascular diseases showing that the reduction of PCSK9 levels isassociated with lower plasma LDL concentrations and better outcomes [11-13].The role of PCSK9 in sepsis was studied mostly in rodents. PCSK9 knockout(Pcsk9-/-) mice showed a reduction of inflammatory signals and an improvementin prognosis [14]. On the other hand, PCSK9 inhibits reverse cholesteroltransport (RCT) as a consequence of increasing LDL-R degradation, which maybenefit the immune response. When RCT is downregulated, there is an enrichmentof cholesterol in cell membranes, including macrophages and other immune cellswith a subsequent improvement in their immunogenic functions [15]. Recently,two single-centre studies evaluating the role of PCSK9 in sepsis have beenpublished. In a cohort of patients with sepsis enrolled at the first day ofemergency department (ED) admission, 28-day mortality was 5%, and subjectsdeveloping organ failure (especially cardiovascular and respiratory) presentedhigher PCSK9 levels [16]. In another study, 469 patients with bacteraemiaenrolled within 24 h from ED admission showed higher PCSK9 levels compared tothe control group, and PCSK9 correlated positively with CRP [17]. The mortalityrates at 28 and 90 days were 15% and 21%, respectively, and lower PCSK9 levelswere recorded in nonsurvivor patients [17].（最近，兩項評估PCSK9在膿毒症中的作用的單中心研究已經發表。在急診(ED)入院第一天入組的膿毒症患者中，28天死亡率為5%，出現器官衰竭(特別是心血管和呼吸系統)的受試者出現較高的PCSK9水平。在另一項研究中，469例在ED住院24小時內入組的菌血症患者PCSK9水平高於對照組，PCSK9與CRP呈正相關。第28天和第90天的死亡率分別為15%和21%，非存活患者的PCSK9水平較低。）
In sum, the role of PCSK9 in sepsis is notcompletely clear. For this reason, we aimed to assess the association betweenplasma PCSK9 levels and outcomes in a large cohort of patients with severesepsis or septic shock.
Materials and methods
Study design
This study is a sub-analysis of the Albumin Italian Outcome Sepsis(ALBIOS) study, an open-label controlled trial in which 1818 patients withsevere sepsis or septic shock were enrolled in 100 intensive care units (ICUs)and randomly assigned to receive crystalloid solution alone or crystalloidsolution plus 20% albumin from day 1 until day 28 or discharge from ICU(whichever came first) [18]. For the current sub-study, blood samples collectedat days 1, 2 and 7 after enrolment were available for 958 patients.
Study design, inclusion and exclusioncriteria, and main results were published elsewhere [18]. The study wasconducted in accordance with the Declaration of Helsinki as revised in 2008 andapproved by the Institutional Review Board of the Fondazione IRCCS Ca』 Granda –Ospedale Maggiore Policlinico, Milan, Italy (coordinating centre) and by theinstitutional review boards of all other participating centres. Written informedconsent or deferred consent was obtained from each patient.
Study end-points
The predictive value of plasma levels of PCSK9on day 1 towards 28-day overall mortality is the primary outcome of thissub-study. Secondary outcomes include the association of plasma levels of PCSK9assessed on day 1 with 90-day mortality and with PTX3 concentrations.
Detection of PCSK9 and PTX3
After blood draw, the samples were centrifugedand ethylenediaminetetraacetic acid (EDTA) plasma was stored at 70°C untilanalysis. PCSK9 was measured using colorimetric enzyme-linked immunosorbentassay (ELISA) according to the manufacturer’s instruction (R&D Systems,Minneapolis, MN, USA). The lowest limit of detection was 0.625 ng mL1 . Meanintra- and interassay coefficients of variation were below 8%, as previouslyreported [11].
After a thorough review of the currentlyavailable literature, we found that median concentrations of PCSK9 in nonsepticsubjects range from 114 to 220 ng mL1 [16, 17, 19-21]. In particular, ametaanalysis from 15 studies including 3556 subjects not on statins or fibratesfound that the normal range of plasma PCSK9 was between 170 and 220 ng mL1[16]. Other two studies reported the following values for a control cohort: 114(87–147) ng mL1 [20] and 180 (65–357) ng mL1 [21], respectively. Recently,Rannikko et al. reported a median PCSK9 level of 188 (139–264) ng mL1 in controlsamples collected 1 year after recovery from bacteraemia [17,19]. In healthyvolunteers who underwent experimental endotoxaemia, baseline (pre-endotoxaemia)PCSK9 levels were 199 57 ng mL1 [22].
Definitions
As per ALBIOS protocol [18], severe sepsis is definedby the presence of a proved or suspected infection in at least one site, two ormore signs of systemic inflammatory response syndrome (SIRS), and the presenceof at least one severe and acute sepsis-related organ dysfunction, as measuredby the modified Sequential Organ Failure Assessment (SOFA) score [18,23].Septic shock is defined by the presence of severe sepsis and a cardiovascularSOFA score of 3 or 4 [18,23].
The estimated glomerular filtration rate(eGFR) was used to evaluate the renal function based on the Chronic KidneyDisease Epidemiology Collaboration (CKD-EPI) equation [24].
Statistical analysis
Categorical values are expressed as absoluteand relative frequencies, and continuous data as median and interquartile range(IQR). Comparisons of continuous variables were performed using Wilcoxonsigned-rank test and Kruskal–Wallis test, as appropriate. Categorical variableswere compared using chi-square test. Ranked Spearman’s correlation coefficientswere used to evaluate correlations between PCSK9 and PTX3. The prognostic
value of PCSK9 quartiles towards 28-day and90- day mortality was assessed by Cox proportional hazards model, expressed ashazard ratio (HR) and 95% confidence interval (CI). Only factors showingstatistical significance in the unadjusted model were included in the adjustedmodel. Variables with non-normal distribution (lactate, glycaemia, white bloodcells [WBC], PTX3) were log-transformed when included as covariates in Coxproportional hazards model. Kaplan–Meier analyses with log-rank test wereperformed to estimate the cumulative survival and to calculate thecorresponding risk difference according to PCSK9 quartiles. Restricted cubicsplines were performed to assess a nonlinear association of PCSK9 assessed on day1 towards 28- and 90-day mortality. For all statistical analyses, a two-sided P< 0.05 value was considered statistically significant. Statistical analyseswere performed with IBM SPSS Statistics for Windows, Version 23.0 (IBM CO.,Armonk, NY), GraphPad Prism version 8.0.2 for Windows (GraphPad Software, SanDiego, California USA, www.graphpad.com) and R statistics (R: A language andenvironment for statistical computing. R Foundation for Statistical Computing,Vienna, Austria, http://www.R-project.org).
Results
Patients』 characteristics
PCSK9 modifications across 1 week
Plasma PCSK9 and PTX3 concentrations. (a)PCSK9 levels at days 1, 2 and 7 are shown for 958 patients. Data are expressedas median and interquartile range. P for Wilcoxon signed-rank test, * P =0.001, § P < 0.001. (b) PTX3 levels on day 1 across quartiles of PCSK9assessed on day 1 by clinical condition (severe sepsis/septic shock). Data areexpressed as median and interquartile range. P for Kruskal–Wallis acrossquartiles.
PCSK9 levels on day 1 are not associatedwith mortality in patients with severe sepsis
Discussion
The main finding of this sub-study of the ALBIOS trial is thatpatients with septic shock presenting with lower plasma PCSK9 levels (1stquartile, 19– 182 ng mL1 ) on day 1 experienced a higher rate of 28- and 90-daymortality. Our results are concordant with those recently published by Rannikkoand colleagues showing that reduced plasma PCSK9 concentrations were associatedwith an increased 7-, 28- and 90-day mortality in a cohort of 481 patients withbacteraemia [17]. Our study, however, presents some novelties. The mostimportant differences as compared to prior studies [16, 17] rely on the numberof patients and setting of care. We considered a larger population (n = 958) ofcritically ill patients, who were enrolled in the multicentric (100 centres)ALBIOS cohort [18]. All ALBIOS patients were admitted to the ICU (as this wasan inclusion criterion), whilst only a small number of patients (n = 44. 9%)were admitted to the ICU in the study by Rannikko et al. [17] and none noted inthe study by Boyd et al. [16], as both of them were conducted in patientsadmitted to the ED. The different setting, i.e. ED vs. ICU, indicates a potentialdifference in the stage or the severity of sepsis, which may account fordifferent outcomes. In the present sub-study, 32% of patients died within 28days compared to 5% and 15% in the study by Boyd et al. and Rannikko et al.,respectively [16, 17]. Finally, we followed the time course of PCSK9 across 1week compared to 4 days as done by Rannikko et al. [17], thereby giving aclearer picture of PCSK9 fluctuations during sepsis. Differently from Rannikkoand colleagues who considered only blood culture-positive cases and excludedcontaminants [17], we included all patients with severe sepsis or septic shockirrespective of blood culture results, which may sometimes result negative inspite of a severe clinical picture (negative cultures were n = 313 [33%] in ourcohort and n = 604 [33%] in the whole ALBIOS cohort [18]).
The role of both endogenous and microbial lipids and PCSK9 in sepsiswas previously investigated in basic research studies with conflicting results.Walley and colleagues studied the role of PCSK9 in two animal models. Pcsk9-/-mice were found to have lower pro-inflammatory cytokine levels and betterhaemodynamic parameters after lipopolysaccharide (LPS) injection as compared towild-type mice [25]. In a second mouse model of polymicrobial sepsis obtainedthrough the caecal ligation and puncture, the administration of a PCSK9blocking antibody reduced circulating endotoxins and increased survival [25].According to the authors, this result was explained by an increased clearanceof pathogen lipids followed by a decreased inflammatory response, resulting inbetter outcomes in mice with reduced PCSK9 function. On the contrary, Bergerand colleagues found that LPS-induced mortality was not reduced neither in micetreated with a PCSK9 inhibitor nor in Pcsk9/ mice [26].
（內源性和微生物脂質以及PCSK9在膿毒症中的作用在之前的基礎研究中被研究過，但結果相互矛盾。Walley和他的同事在兩個動物模型中研究了PCSK9的作用。與野生型小鼠相比，經脂多糖(LPS)注射後，Pcsk9-/-小鼠具有更低的促炎細胞因子水平和更好的血流動力學參數。在第二個通過盲腸結紮和穿刺獲得的多微生物膿毒症小鼠模型中，PCSK9阻斷抗體的使用減少了循環內毒素，增加了存活率。根據作者的說法，這一結果可以解釋為對病原體脂類的清除增加，隨後炎症反應降低，導致PCSK9功能降低的小鼠效果更好。相反，Berger及其同事發現，無論是PCSK9抑制劑治療的小鼠，還是PCSK9 /小鼠[26]，lps誘導的死亡率都沒有降低。）
In a cohort of patients with sepsis or septic shock, subjects withat least one PCSK9 loss-of-function (LOF) allele were found to have anincreased 28- day survival [25]. Another large study conducted in a populationof 10 924 black individuals from the REasons for Geographic and RacialDifferences in Stroke (REGARDS) cohort did not find any association between PCSK9LOF variants and the incidence of serious infections [27]. However, differentlyfrom these two studies, we only measured circulating plasma PCSK9concentrations and this may partially explain the different results concerningthe outcomes.
In the present study, the first quartile of PCSK9 was in the rangebetween 19 and 182 ng mL1 , meaning that at least one-quarter of patients inour cohort presented with circulating PCSK9 levels below the 『normal range』based on the data by Boyd et al. collected in patients admitted to the ED forsepsis [16]. Interestingly, the same authors reported that in theirexperimental model on hepatic cells, they found a reduction in LPS uptake whenPCSK9 exceeded the threshold of 250 ng mL1 . This reduction became linear in theplasma range of PCSK9 of 250–1000 ng mL1 , which approximately corresponds tothe ranges of quartiles other than the first one in our cohort.
In our paper, we showed thatPCSK9 and PTX3 were positively associated. PTX3 belongs to the CRP family(pentraxins) and can be considered as an acute-phase reactant [28]. PTX3 hasbeen evaluated as a biomarker of disease severity and outcomes in infections[29-34]. PCSK9 was found to induce an inflammatory response on macrophages byreleasing pro-inflammatory cytokines, such as tumor necrosis factor (TNF)-a,interleukin (IL)-1 and IL-6 [35]. This has also been confirmed inhypoxia/reoxygenation (H/R) stress, where primary murine cardiomyocytesreleased TNF-a, IL-6 and IL-1b and enhanced the production for these pro-inflammatorycytokines by macrophages undergoing H/R [36]. Therefore, the positiverelationship between PTX3 and PCSK9 should not be surprising and is similar towhat reported by Rannikko et al., who observed a significant positivecorrelation between PCSK9 and CRP levels [17]. Interestingly, in the presentstudy, despite the positive association observed between PTX3 and PCSK9 levelsin the overall cohort, PTX3 concentrations were higher in the first quartile ofPCSK9 compared to the others in patients with septic shock. This suggests thatpatients within the first quartile were characterized by a large inflammatoryburden and experienced a higher death rate. Differently from classicalacute-phase reactants (e.g. CRP and PTX3), PCSK9 may act not only as a triggerto release pro-inflammatory cytokines, but also as a marker of bacterialclearance. This was recently shown in experimental endotoxaemia in healthyvolunteers showing that soon after LPS administration, PCSK9 reduced of 40% at6 h and recovered to baseline levels at 24 h [22]. Based on this experiment,PCSK9 may fall to lower levels when a high burden of bacterial endotoxin ispresent, thus explaining why in our cohort patients with lower PCSK9 levelsexperienced a higher rate of 28- and 90-day mortality, as observed also byRannikko et al. [17].（在我們的文章中，我們證明了PCSK9和pctx3呈正相關。tx3屬於CRP家族(pentraxins)，可以認為是急性期的反應物。PTX3已被評估為與疾病嚴重程度和感染結局相關的生物標誌物。發現PCSK9通過釋放促炎細胞因子，如腫瘤壞死因子(TNF)-a、白細胞介素(IL)-1和IL-6，誘導巨噬細胞產生炎症反應。這在缺氧/復氧(H/R)應激中也得到了證實，小鼠心肌細胞釋放TNF-a、IL-6和IL-1b，並通過經歷H/R的巨噬細胞增強這些促炎細胞因子的產生。因此，pctx3和PCSK9之間的正相關關係並不奇怪，這與ranikko等人報導的PCSK9和CRP水平顯著正相關類似。有趣的是，在本研究中，儘管在整個隊列中觀察到PCSK9和pctx3水平之間存在正相關，但在感染性休克患者中，PCSK9的前四分位數中pctx3濃度高於其他四分位數。這表明，在前四分位數內的患者具有較嚴重的炎症和較高的死亡率。與經典的急性期反應物(如CRP和PTX3)不同，PCSK9不僅可以作為促炎細胞因子釋放的觸發器，還可以作為細菌清除的標記物。這最近在健康志願者的實驗性內毒素血症中得到證實，在給藥後不久，PCSK9在6小時下降40%，並在24小時恢復到基線水平。基於本實驗，當細菌內毒素負擔較高時，PCSK9可能下降到較低水平，這也解釋了為什麼在我們的隊列中，PCSK9水平較低的患者28天和90天死亡率較高，ranikko等人[17]也觀察到這一點。）
Taken together, these results seem to suggest that both extremelyhigh and low levels of PCSK9 may be detrimental and associated with adverseoutcomes, as highlighted by the J-shape of mortality curves in our cohort.PCSK9 expression is stimulated by inflammatory mediators [37, 38], andtherefore, higher PCSK9 levels are associated with remarkable systemicinflammation [17], a key aspect of worse outcome in sepsis, as confirmed by thepositive correlation between PCSK9 and PTX3 in our study. On the contrary,severe and persistent bacteraemia may cause a reduction in PCSK9 levels, thatmight be particularly relevant in the early stage of sepsis [22]. Accordingly,we found that patients with lower PCSK9 levels at day 1 had a higher mortality,whilst no differences were observed for PCSK9 at days 2 and 7. Similar resultswere also reported by Rannikko et al. [17]. However, given the observationalnature of our paper, it is difficult to unravel further mechanisms explainingthe role of PCSK9 in sepsis. This represents a knowledge gap that is worthbeing investigated in the future through adequately designed animal models andhuman studies to resolve current controversial findings on PCSK9 in sepsis.（炎症介質刺激PCSK9表達，因此，PCSK9水平升高與顯著的系統性炎症相關，這是膿毒症惡化結局的一個關鍵方面，我們的研究證實了PCSK9和pctx3之間的正相關關係。相反，嚴重且持續的菌血症可能導致PCSK9水平下降，這可能與膿毒症的早期特別相關。因此，我們發現PCSK9水平較低的患者在第1天死亡率較高，而PCSK9在第2天和第7天沒有觀察到差異。ranikko等人也報導了類似的結果。然而，鑑於我們論文的觀察性質，很難進一步解釋PCSK9在膿毒症中的作用。）
Some limitations of our study should be acknowledged. The firstblood sample was collected the day after the enrolment, and so PCSK9 levels maybe partially affected by treatments already performed. Secondly, lipid levelsand records on statin therapy were not available and we could not perform anyanalyses on PCSK9, lipid profile and statin interaction. Moreover, a nonsepticcontrol cohort for a better interpretation of our results is missing. Theoriginal design of the ALBIOS study did not include the collection of patients』blood samples for subsequent analyses other than those enrolled. This biomarkersub-study should, therefore, be considered as exploratory and for this reason,we reported PCSK9 measurements from other studies in the Methods sections tohelp the reader better understand the content of our manuscript. Additionally,in the ALBIOS trial no genotyping for Pcsk9 LOF was performed. Finally, thiswas a clinical study focused on the role of PCSK9 in sepsis and it was notplanned to assess any pathophysiologic mechanisms to further explain theresults shown. Specific studies adequately designed are therefore needed toclarify this aspect, and our explanations remain at a speculative level by now.
應該承認我們研究的一些局限性。第一個血液樣本是在登記的第二天採集的，因此PCSK9的水平可能部分受到已經進行的治療的影響。其次，沒有脂質水平和他汀治療記錄，我們無法對PCSK9、脂質譜和他汀相互作用進行任何分析。此外，我們還缺少一個對我們的結果有更好解釋的非感染性控制隊列。