Contribution des niveaux de PCSK9 à la sévérité phénotypique

La contribution des niveaux de PCSK9 à la sévérité phénotypique de l’hypercholestérolémie familiale est indépendante du génotype du récepteur LDL, tutoriel & guide de travaux pratiques en pdf.

Methods

Population

A total of 292 HeFH (133 men and 159 women) and 56 control subjects, matched as a group for gender and body mass index (BMI), were recruited in the Quebec City area to participate in the study. All eligible subjects were required to discontinue their use of lipid-lowering medications for at least 6 weeks before blood sample collection. Subjects with a previous history of cardiovascular disease, acute liver disease, hepatic dysfunction, persistent elevations of serum transaminases, secondary hyperlipidemia due to any cause, a recent history of alcohol or drug abuse, diabetes mellitus, a history of cancer or any other conditions that may interfere with optimal participation in the study were ineligible.
All the HeFH subjects were carriers of one of the nine previously identified French-Canadian mutations in the LDLR gene.6 Of the 292 HeFH subjects that were included in the study, 183 had the >15 kb deletion at the 5’ end of the gene,3 57 had the W66G mutation in exon 3,18 24 had the C646Y mutation in exon 14,4 17 had the Y468X mutation in exon 10,19 4 had the C347R mutation in exon 8,6 3 had the E207K mutation in exon 4,4 1 had the C152W mutation in exon 4,6 1 had the R329X mutation in exon 7,6 and 2 had the 5 kb deletion in exons 2 and 3.4 The >15 kb deletion and the R329X, Y468X and C646Y point mutations were considered to be NR mutations because the receptor encoded by these alleles exhibits <2% of the normal activity. The 5 kb deletion and the C152W, C347R, E207K and W66G point mutations were considered to be DR mutations because the receptor encoded by these alleles exhibits 2% to 30% of the normal activity.
The study was approved by the Laval University Medical Center ethical review committee, and informed consent was obtained from each patient. This trial was registered at clinicaltrials.gov as NCT02225340.

DNA analysis

All the FH patients were screened for nine mutations in the LDLR gene using genomic DNA; the two deletions were analyzed by Southern blotting,20 and seven point mutations were analyzed by restriction enzyme fragment analysis.6, 19 Genotyping of apoE was performed by PCR amplification of a 244 bp fragment of exon 4 of the apoE gene with oligonucleotides F4 and F6 and digestion of the resulting PCR fragments with the restriction enzyme Hhal.21
Plasma lipids, lipoproteins and apoproteins
Blood samples were collected after a 12-hour fast in tubes containing disodium EDTA (Na2EDTA) and benzamidine (0.03%).22 The samples were then immediately centrifuged at 4°C for 10 min at 3000 rpm to obtain plasma, which was stored at 4°C and processed within 5 days. Cholesterol and triglyceride (TG) levels were determined in the plasma and lipoprotein fractions by enzymatic methods (Randox Co., Crumlin, UK) using an RA-500 analyzer (Bayer Corporation Inc., Tarrytown, NY) as previously described.23 Plasma VLDL (d < 1.006/mL) was isolated by preparative ultracentrifugation, and the HDL fraction was obtained after precipitation of LDL in the infranatant (d > 1.006 g/mL) using heparin and MnCl. The cholesterol and TG content of the infranatant fraction was measured before and after the precipitation step to obtain the LDL-C level. Plasma apoB levels were measured using a Behring Nephelometer BN-100 (Behring Diagnostic, Westwood, MA) with reagents and calibrators (Dade Behring, Mississauga, ON) provided by the manufacturer.

Measurement of plasma PCSK9 and Lp(a) levels

Commercial ELISA kits were used to measure plasma PCSK9 levels (Circulex, CycLex Co, Nagano, Japan)24 and Lp(a) levels (ALPCO Diagnostics, Salem, NH, US). The intra-assay coefficients of variation for the PCSK9 and Lp(a) measurements were estimated at 1.5 and 4%, respectively.

Statistical analysis

Comparisons between different groups were calculated using ANOVA and Student’s t-test. A multiple regression linear analysis was performed to evaluate the independent contribution of PCSK9 to LDL-C variability after adjusting for age, LDLR mutation and BMI. Plasma PCSK9 levels were log-transformed to normalize their distribution. Spearman’s correlation was used to assess the relationship between PCSK9 and Lp(a) levels. Differences with P≤0.05 were considered to be statistically significant. All the statistical analyses were performed using JMP software v10.0.0.

Results

The demographic, anthropometric and genotypic characteristics of the 292 HeFH (133 men and 159 women) and the 56 control subjects (28 men and 28 women) are presented in Table 1. The mean age was significantly higher in the control group than in the HeFH group (38.8 ± 13.0 y vs. 34.9 ± 13.2 y; P=0.04). The two groups had a similar gender distribution and a similar BMI (24.1 ± 2.1 kg/m2 vs. 24.6 ± 4.2 kg/m2; P=0.37). In the HeFH group, the >15 kb deletion at the 5’ end of the gene and the W66G missense mutation in exon 3 were the most common mutations, representing 82.5% of all mutated alleles. Two hundred and twenty-six HeFH were carriers of a NR mutation, and 66 HeFH were carriers of a DR mutation. The distribution of apoE genotypes was similar among the NR and DR groups (data not shown).
The biochemical characteristics of the controls and HeFH are presented in Table 2. Compared with controls, HeFH had higher total cholesterol levels (+74%; P<0.0001), TG levels (+50%; P=0.001), LDL-C levels (+120%; P<0.0001), plasma apoB levels (+104%; P<0.0001), and PCSK9 levels (+56%; P<0.0001) as well as lower HDL-C levels (-24%; P<0.0001). The Lp(a) concentrations were not significantly different between the two groups. As illustrated in Figure 1, the LDL-C concentration was significantly correlated with PCSK9 levels in both HeFH (r=0.31; P<0.0001) and controls (r=0.37; P=0.005).
The potential independent association of PCSK9 levels with the variability in LDL-C concentration among HeFH was assessed using multiple linear regression analysis. As shown in Table 3, age and PCSK9 levels were the most important predictors of LDL-C variance in HeFH, representing 13.6 and 8.1% of the total variance, respectively; furthermore, the type of LDLR mutation (DR vs. NR) and BMI accounted for 3.9 and 1.8% of the LDL-C variance, respectively. Interestingly, PCSK9 levels also significantly predicted LDL-C levels in the whole cohort (P<0.0001) independent of age, BMI, and the FH/control status and accounted for 5.2% of the LDL-C variance.
PCSK9 levels were significantly higher in HeFH than in controls (317.9 ± 107.1 vs. 203.3 ± 59.8 ng/mL; +56.4%; P<0.0001) (Figure 2A). NR carriers had significantly higher LDL-C levels than DR carriers (6.48 ± 1.59 mmol/L vs. 5.97 ± 1.44 mmol/L; +8.5%; P=0.02), but PCSK9 levels were not significantly different between these two groups (314.5 ± 107.7 ng/mL and 329.7 ± 105.0 ng/mL; P=0.31) (Figure 2B). The concentrations of Lp(a) and total apoB were positively correlated with PCSK9 levels in HeFH (Lp(a): r=0.14, P=0.01; apoB: r=0.35, P<0.0001) (Figure 3). Total apoB concentrations were positively correlated with PCSK9 levels in controls (r=0.40; P=0.002).

Discussion

The present study examined the extent to which PCSK9 levels correlated with LDL-C levels and LDLR genotype in a large cohort of genetically defined FH heterozygotes. Compared with non-FH controls, HeFH had higher plasma levels of LDL-C, TG, and apoB, whereas HDL-C levels were significantly lower in HeFH. No significant differences were observed in the plasma Lp(a) concentration between the two groups. Moreover, HeFH had an elevated plasma PCSK9 concentration compared with controls, and significant positive correlations were identified between the plasma levels of LDL-C, apoB and PCSK9 in the HeFH group. Plasma PCSK9 levels were also significantly higher in HeFH than in controls, but there were no differences in PCSK9 levels between NR and DR carriers in our cohort. PCSK9 levels among HeFH were associated with variations in LDL-C concentration independent of LDLR genotype, age, and BMI. Finally, we found a positive correlation between Lp(a) and PCSK9 levels in HeFH, suggesting that PCSK9 plays an important role in Lp(a) metabolism.
Previous studies have shown that plasma PCSK9 levels are higher in untreated FH subjects than in non-FH subjects.13-15 Plasma PCSK9 levels are also positively correlated with LDL-C concentration in untreated heterozygous and homozygous FH subjects and in non-FH subjects,12-16 a finding that is consistent with the results of the present study. In fact, PCSK9 has been shown to interact with the LDLR, thereby decreasing the LDLR density on the cell surface and ultimately reducing LDL-C clearance; this leads to an accumulation of LDL particles in the plasma.11, 25, 26 In agreement with previous studies,14, 15 our results showed that because of its stimulatory effect on LDLR catabolism, high PCSK9 levels represent an independent contributing factor to the elevated LDL-C levels in FH heterozygotes, and this may contribute to the phenotypic severity of this disorder.
The impact of the LDLR genotype on plasma PCSK9 levels was also assessed in the present study. Previous studies have shown that the EGF(A) of the LDLR is critical for PCSK9 binding at the cell surface.27 Therefore, any LDLR mutation that disrupts the EGF(A) could result in diminished binding and decreased clearance of PCSK9 via LDLR, leading to elevated plasma PCSK9 levels. Our results showed that plasma PCSK9 levels were significantly higher in HeFH compared with non-FH controls, a finding that supports the concept that LDLR plays a crucial role in PCSK9 clearance. However, there were no significant differences in PCSK9 levels between carriers of a NR mutation and carriers of a DR mutation, in whom EGF(A) integrity is preserved. Our data are consistent with those reported in previous studies that showed that HeFH patients carrying negative or defective LDLR defects had similar PCSK9 levels.15 These results suggest that NR and DR carriers maintain a similar capacity to remove PCSK9 from the circulation either via the fully functional LDLR or via alternative LDLR-independent pathways, as has been suggested in previous reports.28 The presence of a normal LDLR allele in HeFH could also offer some level of compensation, which may differ depending on the function of the mutated allele. Supporting in vitro studies are required to assess the variability in the uptake of LDL particles and PCSK9 based on LDLR functionality.
Our study demonstrated that PCSK9 levels had a greater impact on LDL-C variability than did LDLR genotype and BMI, with independent contributions of 8.1%, 3.9%, and 1.8%, respectively. The only factor with a greater impact on LDL-C concentration than PCSK9 levels was age. Our results are consistent with the analysis of the large multiethnic Dallas Heart study cohort,29 in which PCSK9 levels were correlated with LDL-C concentration but accounted for less than 8% of the variance in LDL-C levels. These findings also support the concept that the impact of PCSK9 levels on plasma LDL-C levels is independent of the functionality of the LDLR mutation.
In the present study, a large variability in PCSK9 concentration, ranging from 119 to 757 ng/mL, was observed among HeFH. In the Dallas Heart Study cohort, 22% of the PCSK9 variability was explained by gender, LDL-C and TG levels, statin use, BMI and fasting glucose.29 Mutations in the PCSK9 gene affect functionality and the plasma levels of its cognate protein. Subjects with gain-of-function mutations have higher plasma PCSK9 levels and higher plasma LDL-C levels,26 whereas mutations causing a defective PCSK9 decrease the LDL-C concentration and CVD risk in non-FH subjects.30, 31 Among the PCSK9 loss-of-function variants, R46L is the most prevalent and affects nearly 1/50 individuals.31-33 This mutation is associated with decreased PCSK9 levels, LDL-C concentrations and CAD risk29, 31 in non-FH subjects. A recent study by Saavedra et al.34 showed that HeFH carrying the R46L had lower LDL-C levels and a reduced CVD risk compared with non-carriers. In the Dallas Heart Study cohort, the addition of the variant R46L in the linear model predicting PCSK9 levels only slightly increased by 1% the predictive value of the model.29 These findings suggest that PCSK9 genetic variants are likely to have a limited impact on variability of PCSK9 and LDL-C levels in our cohort.
Finally, significant positive correlations were found between the levels of Lp(a), apoB and PCSK9. A recent study by Alonso et al. suggested that Lp(a) levels are a significant predictor of CVD independent of the functionality of the LDLR mutation in HeFH.35 It has been shown that PCSK9 impacts the synthesis of Lp(a) via apoB-100 synthesis. Experiments in mice showed that PCSK9 is a determinant of hepatic apoB synthesis by inhibiting the intracellular degradation of newly synthesized apoB and therefore increasing plasma apoB levels.17, 36 In support of these findings, lomitapide, which inhibits microsomal triglyceride transfer protein and therefore blocks the assembly of apoB-containing particles, reduced Lp(a) levels by 15%.37 AMG-145, an antibody against PCSK9, also causes a significant decrease in apoB and Lp(a) levels.38 Therefore, the presence of a significant positive correlation between Lp(a) and PCSK9 levels in HeFH supports the hypothesis that Lp(a) production may be modulated by plasma PCSK9 levels.
This study encompasses several strengths and limitations. The large number of genetically defined FH subjects in the present study is one of the major strengths and increases our statistical power. In addition, the case-control design allowed us to compare the associations between PCSK9, Lp(a) and LDL-C between FH and non-FH control subjects. Data on circulating PCSK9 and lipid levels are presented on only a single blood sample, however, given that lipid profiles naturally experience day-to-day fluctuations, a better estimation would have been achieved via repeated sampling and calculations of means. Finally, we also stress that the association between PCSK9 and LDLR genotype is limited by the relatively small sample size of DR subjects and these will need to be replicated in studies with more subjects.

Conclusion

In summary, these findings confirmed the importance of PCSK9 as a determinant of LDL-C and Lp(a) levels in HeFH. PCSK9 levels had a greater impact on the variability in LDL-C concentration than did the functionality of the LDLR mutation in HeFH. Our results also suggested that the contribution of PCSK9 levels to the phenotypic severity in FH heterozygotes is independent of LDLR genotype.

Author contributions

All of the authors read and approved the final manuscript. PC, TCO and BL designed the research; AJT and JCH conducted the research; JPDC, PC, TCO, BL, and AJT analyzed the data; JPDC, PC, TCO, AJT, and BL wrote the paper; and PC had primary responsibility for the final content. The authors report no conflicts of interest related to this study.

Funding

This study was supported by the Institute of Nutrition and Functional Foods, and the CHUQ Research Center, Laval University, Quebec City, Canada.

Acknowledgments

The authors are grateful for the excellent collaboration of the subjects and for the dedicated staff of the Institute of Nutrition and Functional Food and the Lipid Research Centre. Jean-Philippe Drouin-Chartier is the recipient of a training award from the Food Advancement through Science and Training program funded by the Natural Sciences and Engineering Research Council of Canada. Benoît Lamarche leads the Chair on Nutrition at Laval University.