Introduction
Proprotein convertase subtilisin kexin type 9 (PCSK9) acts as a negative regulator of the low-density lipoprotein (LDL) receptor in the liver and therefore increases serum LDL cholesterol (LDL-C) [
13]. Since hypercholesterolemia is still one of the main risk factors for cardiovascular disease (CVD) [
22] and CVD mortality [
1], PCSK9 evolved as a new target in the treatment of hypercholesterolemia [
6].
Given that subjects with a loss of function mutation of PCSK9 gene show an additional cardiovascular protection independent of LDL-C lowering [
4,
7,
19], a different role of PCSK9 in the cardiovascular system has been suggested. Whereas distinct extrahepatic functions of PCSK9 have already been described [
31], a direct effect of PCSK9 on cardiac function is still a matter of debate [
3].
Previously, we could demonstrate that terminally differentiated ventricular rat cardiomyocytes constitutively express PCSK9 [
28]. Furthermore we found that oxLDL, which also plays a pivotal role in the pathophysiology of cardiovascular disease [
11,
35], impaired cardiomyocyte function in a PCSK9-dependent way. However, the mechanism by which PCSK9 is involved in this process remained elusive. Nevertheless, oxLDL increased the expression of PCSK9 in cardiomyocytes and silencing of PCSK9 in cardiomyocytes attenuated oxLDL-dependent effects on cell shortening [
28]. This might indicate either, that intracellular PCSK9 affects cell function or that cardiomyocytes release PCSK9 which then affects the function of cardiomyocytes.
Therefore, we used isolated cardiomyocytes from rats and mice to clarify whether modification of intracellular or extracellular PCSK9 is involved in oxLDL effects on cardiomyocyte and/or cardiac function.
Material and methods
Materials
OxLDL (human high oxidized low density lipoprotein), Low-oxidized LDL and LDL was produced by KB Kalen Biomedical, Montgomery Village, USA and purchased from Biotrend Chemikalien, Cologne, Germany. OxLDL stock solution was stored at 4 °C and used within six weeks. Human recombinant PCSK9 (hPCSK9) was purchased from Cayman Chemical, Michigan, USA. Malondialdehyde-modified human serum albumin (MDA-HSA) was produced by Cell Biolabs, Inc., California, USA and purchased from Biotrend Chemikalien.
Animals
Adult Wistar rats (175–225 g) and C57BL6/JR (25–35 g) were ordered from Janvier Labs (Le Genest Saint Isles, France). PCSK9 knockout mice (B6;129S6-Pcsk9tm1Jdh(PCSK9−/−), 25–30 g) and appropriate controls (B6129SF2/J (PCSK9+/+), 8–9 weeks old) were supplied from The Jackson Laboratory (Maine, USA) with consent of the University of Texas Southwestern Medical Center. Mice and rats were housed according to the Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85–23, revised 1996). All protocols were approved by the Justus-Liebig-University Giessen (permission number: 666_M and 561_M). In this study only male animals were used, to avoid the effect of sex on experiments.
Isolation and cultivation of cardiomyocytes
Ventricular heart muscle cells were isolated from rats and mice as described previously [
5,
25]. In brief, rats or mice have been sacrificed under deep anesthesia (isoflurane 5%) before the heart was excised and transferred to ice-cold saline. Thereafter the aorta was cannulated and the heart was connected to a Langendorff perfusion system where it was perfused with Powell medium (NaCl 110 mM, KCl 2.5 mM, KH
2PO
4 1.2 mM, MgSO
4 × 7H
2O 1.2 mM, Hepes 25 mM, D( +)-glucose-monohydrate 10 mM, pH adjusted to 7.4), that contained collagenase (25 mg/ml, 265 u/mg, Worthington Biochemicals, Lakewook, USA) and CaCl
2 (25 µM), for 25 min at 37° C. Thereafter, ventricular tissue was minced and incubated in the abovementioned solution for another 5 min at 37° C. The remaining cell solution was filtered through a 200-µm nylon mesh. The filtered material was resuspended in Powell medium, centrifuged and after a stepwise increase in calcium concentration finally transferred to culture medium (Medium 199, supplemented with creatine, carnitine, taurine, and 2% penicillin–streptomycin, Biochrom). Cells were plated onto culture dishes (35 mm) which were precoated with 4% (
v/
v) fetal calf serum (FCS) (PAA, BioPharm). After one hour cell culture medium was refreshed and PCSK9 (200 ng/ml), oxLDL (5–20 µM), Pep 2–8 (10 µM), alirocumab (1.5 mg/ml) or combinations thereof were added (as indicated) to the culture medium. Cells were then cultured under serum-free conditions for 24 h at 37° C in the above mentioned medium. Where indicated, small interfering RNA (siRNA) directed against SURF-4 (final concentration 0.05 µM) was added 6 h before administration of oxLDL (5–20 µM) to the culture medium. SiRNA was purchased from Qiagen, Venlo, Netherland.
Adenoviral infection of cardiomyocytes and HepG2 cells
Isolated cardiomyocytes were infected with hPCSK9. Therefore an adenoviral plasmid with the insert of hPCSK9 gain of function mutation D374Y (pAd/CMV/V5 hPCSK9, Addgene, Massachusettes, USA) was generated by gateway cloning [
15,
17] (pENTR™ Directional TOPO®Cloning Kit,Thermo Fisher Scientific, Massachusettes, USA). An adenoviral plasmid with the insert of LacZ was generated for control (pAd/CMV/V5 LacZ, Addgene, Massachusettes, USA). HEK 293 A cells have then been transfected with these adenoviral plasmids (hPCSK9, LacZ). The adenovirus produced by HEK 293 A cells was collected and used to infect adult ventricular cardiomyocytes as well as HepG2 cells with hPCSK9 and LacZ, respectively. The supernatants of HepG2 cells that contained hPCSK9 or LacZ were collected 72 h after infection, centrifuged (6000×
g) and then stored at − 80 °C. Cardiomyocytes have been infected with hPCSK9/LacZ or incubated with supernatants of HepG2 cells over 24 h before load-free cell shortening was determined.
Cell shortening
Cells were stimulated via two AgCl electrodes with biphasic electrical stimuli composed of two equal but opposite rectangular 50 V stimuli of 5 ms duration as described before [
18]. Cells were stimulated at 2 Hz frequency. Four signals were registered from each cell. The mean of these four measurements was used to define the contractile responsiveness of a given cell. Cell lengths were measured at a rate of 500 Hz via a line camera. Cells were used in M199 with an extracellular calcium concentration of 1.25 mM. Data are expressed as ΔL/L (%) in which the shortening amplitude (ΔL) is expressed as percent of the diastolic cell length (L). Furthermore, maximal contraction and relaxation velocity (µm/s) were analyzed.
ELISA
Release of PCSK9 into the supernatant from cardiomyocytes as well as from hPCSK9 overexpressing HepG2 cells was quantified by the use of a commercial available kit purchased by Cusabio Biotech Co. China (Rat PCSK9 ELISA kit) and Cloud-Clone Corp. Texas, USA (Human PCSK9 ELISA).
Langendorff perfusion of hearts isolated from rat our mouse
The left ventricular function of isolated rat hearts was evaluated as described before [
30]. Therefore hearts were excised as mentioned above and attached to a 16-gauge needle. The cannulated heart was then connected to a Langendorff perfusion system where it was perfused retrograde with an oxygenated saline buffer (NaCl 140 mM, NaHCO
3 24 mM, KCl 2.7 mM, NaH
2PO
4 0.4 mM, MgCl
2 1 mM, CaCl
2 1.8 mM and glucose 5.0 mM, pH 7.4). A polyvinylchloride balloon, which was connected to a pressure transducer, was carefully inserted into the left ventricle to assess left ventricular pressure (LVDP), dp/dt max as well as dp/dt min.
To validate a possible function of PCSK9 on left ventricular function isolated hearts were perfused with a monoclonal antibody against PCSK9 (alirocumab used in a concentration of 750 µg).
In addition hearts from PCSK9 knockout and control mice were prepared and attached to an Aortic Cannula (Ø 1 mm, Hugo Sachs Elektronik-Harvard Apparatus, March, Germany). The cannulated heart was then connected to a Langendorff perfusion system (Hugo Sachs Elektronik-Harvard Apparatus, March, Germany) where it was perfused retrograde with a 37 °C warm modified Krebs Henseleit buffer (NaCl 118 mM, KCl 4.7 mM, MgSO4 0.8 mM, KH2PO4 1.2 mM, G/glucose 5 mM, CaCl2 2.5 mM, NaHCO3 25 mM, pyruvate 1.9 mM; pH 7.4; sterile-filtered with an 0.2 µm filter; saturated with 95% oxygen and 5% carbogen). The perfusion pressure was adjusted to 70 mmHg (transduced by a Replacement Transducer Head for APT300 Pressure Transducer, Hugo Sachs Elektronik-Harvard Apparatus) and kept constant during the whole experiment. A small ballon (built up from a cling-film), which was connected to a pressure transducer (Combitrans 1-fach Set Mod.II University Giessen, B. Braun, Melsungen, Germany), was carefully inserted into the left ventricle and inflated up to 12–14 mmHg (end diastolic pressure) to assess left ventricular function. Additionally the hearts were paced at 600 bpm to ensure a stable heart rate. Left Ventricular Pressure (LVPD = systolic pressure-diastolic pressure) was determined as well as dLVP/dtmax and dLVP/dtmin. For the ischemia/reperfusion protocol: After stabilization phase (5 min) the perfusion and pacing was switched off for 45 min to generate a no flow ischemia. Thereafter the heart was reperfused for another 120 min.
RNA Isolation and qRT-PCR
RNA was isolated from liver samples and isolated cardiomyocytes using peqGold Trifast (peqlab, Biotechnologie GmbH, Erlangen, Germany) according to manufacturer’s protocol. To avoid genomic DNA contamination, samples were treated with DNase (1U/ µg RNA; Invitrogen, Karslruhe, Germany) for 15 min at 37 °C. The Reverse Transcription of RNA (1 µg/10 µl) into cDNA was performed with Superscript RNase reverse transcriptase (200 U/µg RNA; Invitrogen) and oligo dTs at 37 °C for 60 min. Quantitative Real-Time PCR was performed using the Icycler MyiQ® detection systems (Bio-Rad, Munich, Germany) in combination with IQ SYBR green real-time supermix (Bio-Rad, Munich, Germany). Primer for surfeit locus protein-4 (SURF-4) and beta-2-microglobulin (B2M) (Table
1) were used and quantification was performed applying the ΔΔ Ct method [
20].
Table 1
List of primer used in this study
B2M | GCCGTCGTGCTTGCCATTC | CTGAGGTGGGTGGAACTGAGAC |
SURF-4 | ATTTCGCCGACCAGTTCCTT | CAGGTAACCACAGCTCCAGG |
PCSK9 (human) | CACCATGGGCACCGTCAGCTCCAG | AAACTGGAGCTCCTGGGAGGCC |
PCSK9 (mouse) | TTGAACAAACTGCCCATCGC | CCCAACAGGTCACTGCTCAT |
Western Blot
Isolated cardiomyocytes were incubated with cell lysis buffer (Cell Signaling, Technology, Frankfurt, Germany) as described before [
29]. Samples were treated 1:1 with Leammli buffer (Sigma-Aldrich, Taufkirchen, Germany) and loaded (~ 100 µg protein) on a 10% SDS-PAGE (Life Technology, Darmstadt, Germany). Following electrophoresis proteins were subsequently transferred to a nitrocellulose membrane. Anti-PCSK9 polyclonal antibody produced in goat (#AF3985, R&D Systems, Minnesota, USA) was used to detect the expression of PCSK9 protein. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) served as loading control. Detection of GAPDH was performed using anti-GAPDH monoclonal antibody produced in mouse (#CB1001, Calbiochem®, Schwalbach, Germany). Secondary antibodies were purchased from Dako (now Agilent Technologies, Santa Clara, USA).
Statistics
Data are expressed as means ± SD, indicated in the legend of the figures. The statistical comparison of two groups was performed by two-sided-t-test, if normal distribution of samples could be verified, otherwise Mann–Whitney-U test was applied. Comparison of more than two groups was performed with one-way following a post hoc analysis with the Student–Newman–Keuls test. p levels ≤ 0.05 were regarded as significant and indicated as an asterisk or as indicated in the figure’s legend. Figures were generated with Graph Pad Prism 8.
Discussion
From our previous study it was already known that oxLDL induced the expression of PCSK9 in cardiomyocytes and reduced cardiomyocyte’ function in a PCSK9-dependent way [
28]. To further study the mechanism behind this effect, the current study adds five new findings: (1) oxLDL reduces cell shortening in mice and rat cardiomyocytes. (2) OxLDL induces a paracrine effect of PCSK9 that than reduces cardiomyocyte cell shortening. (3) OxLDL leads to an increase in PCSK9 secretion that requires SURF-4 as in HEK293 cells [
10]. (4) Antagonism of PCSK9 by monoclonal antibody (alirocumab) leads to an immediate increase in left ventricular function of isolated perfused hearts. (5) Hearts isolated from PCSK9 knockout mice show increased postischemic function. Therefore, cardiomyocyte-derived PCSK9 seems to directly reduce cardiomyocyte and cardiac function, respectively. The data suggest that the aforementioned superior effect of PCSK9 inhibition over cholesterol reduction performed by statins is supported by direct effects on cardiomyocyte-derived PCSK9.
Since its discovery in 2003 [
2,
33] PCSK9 has emerged as an important target in the treatment of hyperlipidemia due to its role in hepatic LDL-clearance [
6]. Apart from that distinct extrahepatic functions of PCSK9 have been described during the past years [
12]. PCSK9 for example regulates the abundance of the epithelial sodium channel in the kidney [
34] and is also expressed in pancreatic islets β cells where it contributes to glucose tolerance [
8,
23]. Moreover studies report that patients with a loss of function mutation in PCSK9 show a cardiovascular protection that is higher than predicted by the associated reduction of LDL-C alone [
4,
7,
19]. Therefore a possible direct negative effect of PCSK9 on the cardiovascular system, that is independent of its role in LDL-clearance, could be suggested.
Recently it was reported that PCSK9 plasma levels are associated with a reduced ejection fraction in patients suffering from ST-segment elevation myocardial infarction [
24]. Furthermore a distinct function of cardiomyocyte-derived PCSK9 has only been reported for neonatal mouse cardiomyocytes, where under hypoxic conditions the release of PCSK9 determines autophagy in these cells [
9]. Unfortunately this cell type is not directly comparable to terminally differentiated cardiomyocytes and autophagy not directly linked to mechanical function.
In our previous study we demonstrated that terminally differentiated cardiomyocytes isolated from rats constitutively express (mRNA, protein) and release PCSK9 [
28]. Meanwhile the expression of PCSK9 in neonatal mouse cardiomyocytes as well as in ischemic mice hearts has been described [
9]. In addition we reported that cardiomyocyte-derived PCSK9 is involved in the oxLDL-induced reduction of cardiomyocyte cell shortening [
28]. In the current study we could reveal that the oxLDL-induced reduction of cardiomyocyte function requires PCSK9 because this effect was absent in PCSK9
−/−- but present in PCSK9
+/+-cardiomyocytes. With these experiments we could demonstrate the PCSK9 mediated oxLDL effect already in two different species (mouse, rat). In addition, we found no difference in basal function of cardiomyocytes from both strains. This supports the view that intracellular PCSK9 is not directly involved in cell function. In addition, only when PCSK9 expression is increased by oxLDL [
28] and release is increased by oxLDL (present study) the threshold of extracellular PCSK9 is sufficient to reduce cardiomyocyte shortening. In addition the effect of oxidized LDL seems to depend on the oxidation-level of LDL. Furthermore, an oxidatively modified non-lipid plasma protein (MDA-HSA) had no effect on cardiomyocyte cell shortening.
The incubation of cardiomyocytes with supernatants collected from hPCSK9 overexpressing HepG2 cells, led to a significant decline of cardiac function that was absent under control conditions. Furthermore these findings were reproducible by incubation of cardiomyocytes with recombinant PCSK9. Therefore any unspecific effect of HepG2 supernatants on cell shortening can be excluded. The concentration of PCSK9 (200 ng/ml) used in this study to reduce cardiomyocyte function is comparable with the mean serum PCSK9 concentration found in rats [
36] and humans [
27].
Under oxLDL incubation PCSK9 was increased in a concentration-dependent manner. In addition we found that SURF-4, that has shown to be important for PCSK9 secretion [
10,
32], is expressed in cardiomyocytes and that silencing of SURF-4 improves cardiomyocyte function when incubated with oxLDL. Furthermore we could block the oxLDL-induced reduction in cell shortening by alirocumab as well as Pep 2–8 while the inhibitors had no effect on basal cardiac function. This leads to the suggestion that oxLDL induces PCSK9 release and PCSK9 acts in an autocrine way to reduce cardiomyocyte function.
As exogenously added PCSK9 but also endogenous PCSK9 (rat PCSK9) has direct negative effects on cardiomyocyte function, the domain of PCSK9 that induces the reduction of cardiomyocyte cell shortening must lie in the conserved region between human and rat PCSK9. Overall, the total sequence identity of PCSK9 between these two species is about 77% [
14]. Moreover the PCSK9 induced decline of cardiomyocyte function can be antagonized by Pep 2–8 which is known to inhibit the catalytic domain of PCSK9 that is responsible for LDL-R binding [
16,
21,
38]. Therefore the catalytic domain might also be important for the abovementioned PCSK9 effect on cardiomyocyte function.
On the organ level we found that hearts isolated from PCSK9 knockout mice show increased basal cardiac function. In addition a cardiac release of PCSK9 was demonstrated under basal conditions in isolated rat hearts. Furthermore the antagonism of PCSK9 with alirocumab led to an immediate increase in left ventricular function. This is the first report about a direct effect of PCSK9 antagonism on basal cardiac function. Moreover, I/R of isolated mice hearts resulted in an increased expression of PCSK9 in left ventricles. In addition, hearts isolated from PCSK9 knockout mice showed increased postischemic left venctricular function and therefore an improved recovery after suffering from myocardial ischemia. So far an increased expression of PCSK9 in mice hearts, which have been subjected to acute ischemia/reperfusion (present study), myocardial ischemia for 1 week, as well as in human hearts with recent infarcts, was found [
9]. Furthermore, it was shown that increased PCSK9 expression in hypoxic neonatal mouse cardiomyocytes is mediated by Hypoxia inducible factor-1 alpha (HIF-1α) [
9]. Moreover a correlation between the amount of PCSK9 in the zone bordering the infarcted area and left ventricular function (1 week of left coronary artery occlusion) was observed [
9]. In addition, PCSK9 inhibitors have been shown to improve cardiac function in rats with acute I/R injury when applied prior to ischemia [
26]. Finally a clinical study reports a correlation between high PCSK9 plasma levels and a reduction of left ventricular ejection fraction in patients that suffered from ST-segment elevation myocardial infarction [
24]. These studies demonstrate, similar to our findings, a positive effect of PCSK9 inhibition or genetic deletion on cardiac function after ischemia or I/R injury in vivo [
9,
26]. Although in vivo models enable to study the effect of PCSK9 inhibition in a more complex way than ex vivo models they do not allow to distinguish between systemic and cardiac derived PCSK9.
Based on the experiments performed on isolated cardiomyocytes it could be hypothesized that also on the organ level, PCSK9 that is released from, e.g., cardiomyocytes acts in an autocrine way to reduce cardiac function. An increased PCSK9 secretion, that requires SURF-4, can be triggered by oxLDL. The exact mechanisms of PCSK9 mediated impaired cardiomyocyte/myocardial function and in this context binding/interaction partner of PCSK9 in cardiomyocytes have to be identified. Here a modulation of autophagy of cardiomyocytes mediated by PCSK9 could play a role [
9,
37].
In conclusion, this study improves our current understanding about the relationship between high plasma levels of PCSK9 and cardiac function as well as our understanding about the mechanism by which drugs increasing cardiac expression of PCSK9 might affect cardiac function.