If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Molecular and Clinical Medicine, Institute of Medicine, Sahlgrenska Academy, Wallenberg Laboratory, University of Gothenburg, Gothenburg, Sweden
Department of Clinical Chemistry and Transfusion Medicine, Institute of Biomedicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg, SwedenDepartment of Clinical Chemistry, Sahlgrenska University Hospital, Gothenburg, Sweden
Corresponding author. Department of Molecular and Clinical Medicine, The Sahlgrenska Academy, University of Gothenburg, Wallenberg Laboratory, Bruna Stråket 16, SE-413 45 Göteborg, Sweden.
Department of Cardiology, Sahlgrenska University Hospital, Gothenburg, SwedenDepartment of Molecular and Clinical Medicine, Institute of Medicine, Sahlgrenska Academy, Wallenberg Laboratory, University of Gothenburg, Gothenburg, SwedenClinical Nutrition Unit, Department of Medical and Surgical Sciences, University Magna Graecia, Catanzaro, Italy
Corresponding author. Department of Molecular and Clinical Medicine, The Sahlgrenska Academy at the University of Gothenburg, The Wallenberg Laboratory, Bruna Stråket 16, SE-413 45 Göteborg, Sweden.
Department of Molecular and Clinical Medicine, Institute of Medicine, Sahlgrenska Academy, Wallenberg Laboratory, University of Gothenburg, Gothenburg, Sweden
Among individuals with genetic familial hypercholesterolemia (FH), those with previous cardiovascular disease have higher lipoprotein(a) levels.
•
Statins have the same low-density lipoprotein cholesterol–lowering effect irrespective of the type of FH mutation.
•
Antiproprotein convertase subtilisin/kexin type 9 low-density lipoprotein cholesterol–lowering trended slightly better in nonsense FH mutation carriers.
Background
Cardiovascular disease (CVD) is a major cause of mortality and morbidity. Increased low-density lipoprotein cholesterol (LDL-C) level is its major risk factor. Familial hypercholesterolemia (FH) is a genetic disorder characterized by elevated LDL-C since birth and subsequent premature CVD. There is a heterogeneity in the CVD onset in patients with FH. This is potentially due to the presence of other independent risk factors. Lipoprotein(a) [Lp(a)] is an LDL-like particle and represents a strong risk factor for CVD.
Objective
Our objective was to understand the contribution of Lp(a) in the susceptibility to CVD in individuals with genetic diagnosis of FH.
Methods
We measured Lp(a) levels in 2 independent and well-characterized genetic-FH cohorts: the FH-Gothenburg cohort (n = 190) and the FH-CEGP Milan cohort (n = 160). The genetic diagnosis was performed by targeted next-generation sequencing (FH-Gothenburg and part of the FH-CEGP Milan cohort), or by Sanger sequencing.
Results
We show that among individuals with genetic diagnosis of FH, those with previous CVD had higher Lp(a) levels. In addition, analyzing the response to the lipid-lowering therapies, we have also shown that statins had the same LDL-C–lowering effect irrespective of the type of FH-causative mutation. However, when we examined the lipid-lowering effect of proprotein convertase subtilisin/kexin type 9 inhibition by antibodies, we observed a trend in a better reduction of the LDL-C level in carriers of nonsense mutations.
Conclusion
In conclusion, our results suggest that Lp(a) contributes to CVD onset in individuals with genetic diagnosis of FH. Our finding supports the importance to identify an efficacious therapy to lower Lp(a) in patients with FH to prevent CVD onset or recurrence.
Cardiovascular disease (CVD) is a major cause of mortality and morbidity in Western Countries. Increased low-density lipoprotein cholesterol (LDL-C) level is a major risk factor for CVD
Coronary heart disease risk factors ranked by importance for the individual and community. A 21 year follow-up of 12 000 men and women from The Copenhagen City Heart Study.
Nonoptimal lipids commonly present in young adults and coronary calcium later in life: the CARDIA (Coronary Artery Risk Development in Young Adults) study.
Familial hypercholesterolemia (FH) is a genetic inherited disorder mainly caused by mutations in genes encoding for the LDL receptor (LDLR), apolipoprotein B (APOB, the main protein of LDL) or proprotein convertase subtilisin/kexin type 9 (PCSK9).
Familial hypercholesterolaemia is underdiagnosed and undertreated in the general population: guidance for clinicians to prevent coronary heart disease: consensus statement of the European Atherosclerosis Society.
Statins represent the pivotal LDL-C–lowering drug in individuals with FH to prevent CVD. However, the efficacy of statins in FH shows high individuals variability.
Despite the strong penetrance of the disease, there is a heterogeneity in the onset of CVD in patients with FH. This is potentially due to differences in the response to lipid-lowering therapy or to the presence of other independent risk factors including high levels of lipoprotein(a) [Lp(a)].
Lp(a) (or LPA) is an LDL-like particle synthetized by the liver and consists of an APOB-100 covalently linked to a very large glycoprotein known as apolipoprotein(a). Circulating Lp(a) levels are predominantly controlled by genetic variations on the LPA gene
To date, Lp(a) is not effectively lowered by any approved drugs. However, novel therapeutic drugs able to lower Lp(a) are currently under clinical trials.
Mendelian randomization studies with genetic variables primarily increasing Lp(a) have shown that high Lp(a) is an independent risk factor for CVD in the general population independently from the other traditional risk factors including LDL-C.
Association of LPA variants with risk of coronary disease and the implications for lipoprotein(a)-lowering therapies: a Mendelian randomization analysis.
However, the role of Lp(a) in CVD prevalence in patients with FH is still a matter of debate. Previous studies described higher Lp(a) levels in patients with FH with previous CVD. However, most of these studies were performed in patients with diagnosis of FH established using clinical evidence and not confirmed by genetic tests.
Relation of serum lipoprotein(a) concentration and apolipoprotein(a) phenotype to coronary heart disease in patients with familial hypercholesterolemia.
Lipoprotein(a) levels in familial hypercholesterolemia: an important predictor of cardiovascular disease independent of the type of LDL receptor mutation.
They have also shown that among the individuals with genetic diagnosis of FH, those with CVD had higher Lp(a) than those without previous CVD events. Furthermore, they showed that individuals with a nonsense mutation in the LDLR had higher Lp(a) and higher CVD incidence than those with missense mutation on the same gene. However, in this very elegant study, the genetic diagnosis of FH was performed using a microarray containing only a selection of mutations in LDLR and of different mutations in APOB. In addition, only selected mutations (namely the most frequent null and defective mutations) on the LDLR were included in the analysis stratified by type of mutation.
Here, to understand the contribution of Lp(a) in the susceptibility to CVD, we performed a comprehensive genetic diagnosis of FH using targeted next-generation sequencing of the main FH-causative genes namely LDL-R, APOB, and PCSK9 in individuals from two European Lipid Clinics. We show that among individuals with definite genetic diagnosis of FH, those with previous CVD had higher circulating levels of Lp(a) in these 2 independent cohorts. Analyzing the response to the lipid-lowering therapies, we have also shown that statins had the same LDL-C–lowering effect irrespective of the type of mutation (nonsense vs missense) in both FH study cohorts. However, when we examined the lipid-lowering effect of PCSK9 inhibition by antibodies, we observed a trend in a better reduction of the LDL-C level in carriers of nonsense mutations.
Materials and methods
Study cohorts
The FH-Gothenburg cohort
In the present study, we included 190 adult individuals with genetic diagnosis of FH from the Lipid Clinic, Department of Cardiology, Sahlgrenska University Hospital in Gothenburg, Sweden. At the time of the first visit, all the individuals underwent a physical and clinical examination. Clinical profile (including lipid profile, liver panel, and other biochemical parameters) was determined in freshly isolated 12-hours fast plasma or serum at the Department of Clinical Chemistry, Sahlgrenska University Hospital, as previously described.
LDL-C after lipid-lowering therapy (namely treated LDL-C) refers to the LDL-C level at the last available assessment. The diagnosis of CVD was performed based on the presence of previous myocardial infarction or stroke (self-reported or present in personal reports). Individuals referred to the lipid clinic for genetic diagnosis have had a myocardial infarction > 5 months ahead of the visit and therefore not in an acute state.
All subjects gave their written informed consent to participate in the study. The study was approved by the regional ethics committee of Gothenburg.
The FH-CEGP Milan cohort
We included 168 adult subjects with genetic diagnosis of FH from the Dyslipidemia Center “E. Grossi Paoletti” of ASST Grande Ospedale Metropolitano Niguarda, Milan, Italy. At the time of the first visit, all the individuals underwent a physical and clinical examination. Biochemical profile was determined after at least 1 month of lipid-lowering treatment suspension in freshly isolated 12-hour fast plasma or serum as previously described.
LDL-C after lipid-lowering therapy (namely treated LDL-C) refers to the LDL-C level at the last available assessment. Cardiovascular events were defined as previous diagnosis of angina, acute myocardial infarction, stroke, and peripheral arterial disease. All subjects gave their written informed consent to use their clinical data for the study, which was approved by the ethics committee of ASST Grande Ospedale Metropolitano Niguarda, Milan, Italy (approval no 19-022016).
Genetic diagnosis of FH
For the FH-Gothenburg cohort, the genetic diagnosis of FH was performed by targeted next-generation sequencing as previously described.
Specifically, DNA was isolated from whole blood. Targeted next-generation sequencing was performed using SEQPRO LIPO RS (Progenika Biopharma, Derio, Spain, http://www.progenika.com/) (for the first 165 samples) or using SEQPRO LIPO IS (Progenika Biopharma, Derio, Spain, http://www.progenika.com/) (the remaining samples). The exons and their flanking regions of LDLR, APOB (exon 26 and 29), PCSK9, and low-density lipoprotein receptor adaptor protein 1 (LDLRAP1), or of LDLR, APOB, PCSK9, LDLRAP1, apolipoprotein E (APOE, exon 4), and signal-transducing adaptor protein 1 (STAP1) were sequenced using SEQPRO LIPO RS or by SEQPRO LIPO IS, respectively.
For the FH-CEGP Milan cohort, the genetic diagnosis of FH was performed by Sanger sequencing (first 90 cases) or by targeted next-generation sequencing (the remaining samples) within the LIpid TransPort Disorders Italian GEnetic Network (LIPIGEN) project.
For the FH-Gothenburg cohort, fasting plasma samples were collected at time of enrollment and stored at −80°C. Lp(a) levels were then measured on all samples simultaneously (on April 2018) by immunoturbidimetric assay (Randox laboratory, United Kingdom) using an ABX Pentra400 (Horiba, Japan) analyzer. The precision of the Randox assay has been determined at 2 levels in an in-house verification: 1) at the 0.19 g/L level, a total coefficient of variance (CV) of 2.4% was determined and 2) at the 0.50 g/L level, a total CV of 1.6% was determined. For the present analysis, Lp(a) measurements below the assay detection limit were considered equal to the lower detection limit (10 mg/dL).
For the FH-CEGP Milan cohort, Lp(a) levels were measured on freshly isolated fasting plasma samples at the time of enrollment (since 2015) using the immunoturbidimetric assay (Tina-quant Lipoprotein (a) Gen.2, Roche Diagnostics, Switzerland) using a Cobas c311 autoanalyzer. All samples of this cohort were measured using the same assay. The precision of the Roche assay has been determined in a total CV of 1.7% for 40.7 nmol/L level and a total CV of 1.2% for 156 nmol/L level.
To evaluate the assay agreement between the different methods, a subset of plasma samples (N = 27) from the Gothenburg cohort, was aliquoted, refrozen, and analyzed both in Milan using the Roche assay and in Gothenburg using the Randox assay. The samples were chosen to get a representation of the whole measurement range. The result from this method comparison is presented in Supplement Figure 1.
Statistical analysis
For descriptive statistic, continuous traits are shown as mean and standard deviation or as median and interquartile range. Categorical traits are shown as number and proportion. For continuous traits, P-value was calculated by linear regression unadjusted, or adjusted for age, gender, and BMI when appropriate. Non-normally distributed values were log-transformed before entering the model. For categorical traits, P-value was calculated by chi-square or by Fisher-exact test as appropriate. Lp(a) levels in patients with FH for the FH-CEGP Milan cohort stratified by CVD status were compared using one-tailed T-test. For assays agreement evaluation, a Bland–Altman plot was generated. To evaluate correlation between the 2 methods, results from the Gothenburg method were plotted against results from the Milan method, a regression line was fitted, and both the Pearson and the Spearman correlation coefficients were calculated. Statistical analyses were performed using the IBM Statistical Package for Social Sciences (IBM SPSS, version 20.0, Inc. Chicago, IL) and GraphPad Prism7.02. P-values <.05 were considered statistically significant.
Results
Characteristics of the FH-Gothenburg cohort
A total of 190 individuals with genetic diagnosis of FH from the lipid clinic, Department of Cardiology, Sahlgrenska University Hospital in Gothenburg, Sweden (namely FH-Gothenburg cohort) were examined. The characteristics of the participants are shown in Table 1. Briefly, individuals were all adults (mean age 46 ± 15 years) with mean BMI 26 ± 4 kg/m2, mean LDL-C level before treatment 299 ± 68 mg/dL, and median current Lp(a) levels 24 (10–56) mg/dL. Fifty-one percent of them were men, 10% had hypertension, 3% had diabetes, and 7% were active smokers. Most of the cohort was under statin treatment (85%), whereas 7% (n = 13) was under PCSK9 mAbs treatment alone or in combination with statin (10/13 were treated with PCSK9 mAbs plus statins, 3/13 with PCSK9 mAbs only). Among the 190 genetic-FH patients, 61 had personal history of CVD events (43%) defined as previous stroke or myocardial infarction. Most individuals from the FH-Gothenburg cohort were carriers of mutation in LDLR gene (93.7%, n = 178), 4.7% (n = 9) were carriers of mutations in APOB, and only a minority of them were carriers of mutations in PCSK9 (n = 2) or STAP1 (n = 1) (1.1 and 0.5%, respectively). Most of the individuals of the FH-Gothenburg cohort were carriers of mutations previously described as FH-causative (80% were carriers of mutations described as pathogenic, 12% possibly and 2% probably pathogenic). In addition to these, we detected 7 new mutations with unknown pathogenicity (not previously described as FH-causative mutations) in 10 individuals (6%). The totality of these mutations was detected on the LDLR. Specifically, we identified 2 mutations in splicing sites, 1 in-frame deletion, and 4 different amino acid changes (accounting for the 57% on the individuals with mutation of unknown pathogenicity) (Supplementary Table 1).
Table 1Characteristics of patients with FH from Gothenburg
Trait
FH Gothenburg
N
190
Women/Men, n (%)
92/95 (49/51)
Homozygosity, compound heterozygosity, or double heterozygosity, n (%)
Individuals with genetic diagnosis of FH and previous CVD have higher Lp(a) levels in the FH-Gothenburg cohort
To test whether individuals with genetic-FH and previous CVD had higher Lp(a), we examined the differences in the clinical characteristics of the study participants stratified by CVD status. We found that individuals with previous CVD had approximately 2-fold increased Lp(a) levels compared with those without CVD (P = .037, Fig. 1A). Importantly, we did not detect differences in the untreated or in the treated LDL-C comparing these 2 groups (Fig. 1B). These data suggest that Lp(a) is a risk factor for CVD in patients with FH even stronger than LDL-C. No differences in age, gender, hypertension, or smoking were found between the 2 groups (Supplementary Table 2). However, genetic-FH individuals with previous CVD had higher BMI (28 ± 5 kg/m2 vs 26 ± 4 kg/m2, P = .031) and higher prevalence of diabetes (10% vs 0%, P = .003). In addition, the proportion of individuals under statin therapy was higher in genetic-FH individuals with CVD than in those without previous CVD.
Figure 1Patients with FH with previous cardiovascular disease events have higher circulating level of Lp(a) in the FH-Gothenburg cohort. Plasma levels of (A) Lp(a) and (B) untreated and treated LDL-C in patients with FH from the Gothenburg cohort stratified by cardiovascular status. Data expressed as median and quartile range or as mean and SD as appropriate. P value has been calculated by linear regression unadjusted. Non-normally distributed traits were log-transformed before entering the model. Lp(a), lipoprotein(a).
The response to lipid-lowering therapy does not differ between genetic-FH patients with nonsense or missense mutations in the FH-Gothenburg cohort
To test whether the response to lipid-lowering therapy was different in genetic-FH with nonsense or missense mutations, we firstly examined the absolute and relative LDL-C reduction in the overall FH-Gothenburg cohort stratified by type of mutation. Individuals homozygous for FH-mutations were excluded from the analysis. We found that there were no differences between individuals with nonsense or missense mutations in the absolute nor in the relative LDL-C in the overall FH-Gothenburg cohort (Fig. 2A and B).
Figure 2The response to lipid-lowering therapy does not differ between patients with FH with nonsense or missense mutations in the FH Gothenburg cohort. (A) Absolute and (B) relative LDL-C reduction according to type of mutations in patients with FH from the Gothenburg cohort. Data are expressed as mean ± SD. P value calculated by linear regression unadjusted. A total of 10 FH patients were simultaneously under statin and PCSK9 mAbs treatment. Individuals homozygous for FH mutations have been excluded from the analysis. LDL-C, low-density lipoprotein cholesterol; FH, familial hypercholesterolemia; PCSK9, proprotein convertase subtilisin/kexin type 9.
Then we examined the absolute and relative LDL-C reduction in individuals after statin or PCSK9 mAbs therapy separately. We found that there were no differences in the absolute or in the relative LDL-C reduction after statin therapy based on the type of mutation. When we examined the response to the PCSK9 mAbs therapy (alone or in combination with statin), we found a slightly higher absolute and relative LDL-C reduction in individuals with nonsense compared to those with missense mutations suggesting a better response to PCSK9 mAbs therapy in individuals with nonsense mutation. However, this slight difference was not significant. No differences in the anthropometric traits or in CVD, active smoking, hypertension, or diabetes prevalence were found between the 2 groups. However, individuals with nonsense mutation had a more severe phenotype with higher prevalence of tendon xanthomas, and higher total cholesterol and untreated LDL-C (Supplementary Table 3). In addition, the proportion of individuals under statin therapy was higher in carriers of nonsense than in carriers of missense mutation.
Lp(a) is a risk factor for CVD in individuals with genetic diagnosis of FH and the type of FH mutation does not affect the response to statin therapy in an independent genetic-FH cohort
To confirm our finding in the FH Gothenburg cohort, we used an independent genetic-FH cohort from Milan, Italy (namely FH CEGP Milan cohort). A total of 168 individuals with genetic diagnosis of FH from Dyslipidemia Center “E. Grossi Paoletti” ASST Grande Ospedale Metropolitano Niguarda were included in the present study. Anthropometric, clinical, and biochemical characteristics of the FH CEGP Milan cohort are shown in Table 2. Briefly, 60% of them were men, 20% had hypertension, and 4% had diabetes. When we compared the 2 independent cohorts, we found that individuals from the FH-CEGP Milan cohort were younger (39 ± 17 years vs 46 ± 15, P < .0001), with lower BMI (24 ± 4 kg/m2 vs 26 ± 4 kg/m2, P = .002), and with a better liver and glucose profile (Supplementary Table 4). However, individuals from the FH-CEGP Milan cohort had a more severe phenotype with higher prevalence of tendon xanthomas and arcus cornealis, and with higher levels of total cholesterol, LDL-C (both untreated and treated LDL-C), and Lp(a). The proportion of individuals under statin therapy and the prevalence of CVD were lower in the FH-CEGP Milan cohort, whereas the number of individuals under PCSK9 mAbs therapy was higher than the FH-Gothenburg cohort. For Lp(a) measurements, we have evaluated the assay agreement between the 2 methods and, lack of uniformity was demonstrated (Supplementary Fig. 1). Nevertheless, both assays performed sufficiently well to demonstrate the association of Lp(a) with CVD in the respective cohorts.
Table 2Characteristics of patients with FH from FH-CEGP Milan cohort
Trait
FH CEGP Milan
N
168
Women/men, n (%)
68/100 (40/60)
Homozygosity, compound heterozygosity, or double Heterozygosity, n (%)
The totality of mutation classified as of unknown pathogenicity were detected on LDLR.
4 (2)
BMI, body mass index; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; AST, aspartate aminotransferase; ALT, alanine aminotransferase; γ-GT, gamma-glutamyltransferase; HOMA-IR, homeostatic model assessment; PCSK9 mAbs, proprotein convertase subtilisin kexin type 9 monoclonal antibodies.
Data expressed as mean (±SD), median and quartile range or as number (proportion) as appropriate. Lipid profile was measured before starting any treatment.
∗ Data on xanthomas available for N = 162.
† Data on arcus cornealis available for N = 162.
‡ Treated LDL-C at last available assessment.
§ Data on statin use available for N = 140.
‖ The totality of mutation classified as of unknown pathogenicity were detected on LDLR.
To confirm our finding on the increased Lp(a) levels in genetic-FH with previous CVD, we examined the differences in the clinical characteristics of the FH-CEGP Milan cohort stratified by CVD status. Consistently with our results in the FH-Gothenburg cohort, we found that individuals with previous CVD had approximately 2-fold increased Lp(a) levels compared with those without CVD (one-tailed P = .045, Fig. 3A) in the FH-CEGP Milan cohort. Importantly, we did not detect differences in the untreated nor in the treated LDL-C comparing these 2 groups (Fig. 3B).In the FH-CEGP Milan cohort, genetic-FH individuals with previous CVD were older, with a higher BMI, and higher prevalence of hypertension than those without previous CVD (Supplementary Table 5). In addition, HDL level and the proportion of active smokers were lower in individuals with previous CVD than in those without. No other differences in biochemical traits were detected between the 2 groups.
Figure 3Patients with FH with previous cardiovascular disease events have higher circulating level of Lp(a) in the FH CEGP Milan cohort. Plasma levels of (A) Lp(a) and (B) untreated and treated LDL-C in FH patients from the CEGP Milan cohort stratified by cardiovascular status. Data expressed as median and quartile range or as mean and SD as appropriate. P value has been calculated by one-tailed T-test. Non-normally distributed traits were log-transformed before entering the model. Lp(a), lipoprotein(a); LDL-C, low-density lipoprotein cholesterol; FH, familial hypercholesterolemia.
These results support our findings in FH-Gothenburg cohort and confirm Lp(a) as risk factor for CVD in individuals with genetic diagnosis of FH.
To confirm that the response to lipid-lowering therapy does not differ between genetic-FH patients with nonsense or missense mutations, we examined the absolute and relative LDL-C reduction in the FH-CEGP Milan cohort stratified by type of mutation. Individuals homozygous for FH mutations were excluded from the analysis. We found that there were no differences between individuals with nonsense or missense mutations in the absolute nor in the relative LDL-C in the overall FH-CEGP Milan cohort (Fig. 4A and B). The same result was obtained when only individuals under statin therapy were analyzed. When we examined the response to the PCSK9 mAbs therapy, we found a slightly higher and borderline significant relative LDL-C reduction in individuals with nonsense compared with those with missense mutations (Fig. 4B, P = .044). The same trend was observed in the absolute LDL-C reduction although not significant. No other differences between the 2 groups were detected except for higher prevalence of active smokers and higher Lp(a) levels in individuals with nonsense than in those with missense mutation (Supplementary Table 6). These results support our finding in the FH-Gothenburg cohort and suggest that the type of FH mutation does not affect the response to statin therapy. However, it seems that individuals with nonsense mutations have a better response to PCSK9 mAbs therapy resulting in a higher relative LDL-C reduction.
Figure 4The response to lipid-lowering therapy does not differ between patients with FH with nonsense or missense mutations (with the exception for the PCSK9 mAbs) in the FH CEGP Milan cohort. (A) Absolute and (B) relative LDL-C reduction according to type of mutations in patients with FH from the Gothenburg cohort. Data are expressed as mean ± SD. P value calculated by linear regression unadjusted. Individuals homozygous for FH mutations have been excluded from the analysis. PCSK9, proprotein convertase subtilisin/kexin type 9; LDL-C, low-density lipoprotein cholesterol; FH, familial hypercholesterolemia.
In this work, we show that individuals with genetic diagnosis of FH and previous cardiovascular events have higher Lp(a) levels than those without CVD in 2 independent cohorts. Lp(a) is an independent risk factor for CVD
and high levels of this lipoprotein may contribute to the variability of the onset of CVD observed in individuals with FH. Our result is in line with previous finding describing higher Lp(a) in individuals with genetic diagnosis of FH and previous CVD events.
Lipoprotein(a) levels in familial hypercholesterolemia: an important predictor of cardiovascular disease independent of the type of LDL receptor mutation.
Clinical characteristics and evaluation of LDL-cholesterol treatment of the Spanish Familial Hypercholesterolemia Longitudinal Cohort Study (SAFEHEART).
In the present study, we examined Lp(a) levels in 2 independent cohorts in which the genetic diagnosis of FH has been performed by targeted next generation sequencing or Sanger sequencing. Specifically, in the FH-Gothenburg cohort, we performed genetic diagnosis of FH by screening comprehensively the main FH-causative genes (namely, LDLR, APOB, PCSK9, and LDLRAP1, or LDLR, APOB, PCSK9, LDLRAP1, APOE, and STAP1). In this cohort, 93.7% of the individuals with genetic diagnosis of FH were carriers of mutation on LDLR; 4.7% were carriers of mutations on APOB; and only a minority of them were carriers of mutations on PCSK9 or STAP1 (1.1 and 0.5%, respectively). We identified 2 individuals homozygote for LDLR mutation (LDLR p.Gly505Asp, or LDLR p.Glu408Lys), one compound heterozygote (carrying the LDLR p.Ser637Cys and LDLR p.Asn665Lys), and 4 double heterozygote (LDLR p.Cys296*plus PCSK9 p.Arg46Leu; LDLR p.Cys261Phe plus PCSK9 c.524-4A>G, LDLR c.2390-2A>G plus APOB p.Arg3527Gln, and APOB p.Arg3527Gln plus PCSK9 p.Arg46Leu). Most of the individuals were carriers of mutations previously described as pathogenic, possibly pathogenic or probably pathogenic (80, 12, and 2%, respectively). In addition, our unbiased approach allowed us to detect 7 novel mutations (not previously described in patients with FH) in 10 (6%) individuals from the FH-Gothenburg cohort. Of the 7 novel mutations, 5 where resulting in an amino acidic change, 1 in an in-frame deletion, and 1 occurred in a splicing site. Pathogenicity of these mutations should be confirmed.
In the FH-CEGP Milan, the genetic diagnosis of FH was performed using Sanger sequencing of LDLR gene (the first 90 patients), or by targeted next generation sequencing of LDLR, APOB, PCSK9, and LDLRAP1. This difference in sequencing methods in the FH-CEGP Milan cohort does not affect the FH diagnosis reliability. However, the comprehensive screening of the main FH-causative genes increases the diagnosis rate and allows the identification of new possible causative mutations. For these reasons, we switched from Sanger sequencing to next generation sequencing to perform the FH genetic diagnosis also in the FH-CEGP Milan cohort. In this cohort, all the individuals were carriers of mutation on the LDLR. Of these individuals, 2 were double heterozygous carrying additional mutations on APOB or on PCSK9 (LDLR p.Arg350* plus PCSK9 p.Asp50Asp; LDLR p.Gly343Ser plus APOB p.Thr3826Met plus PCSK9 p.Ala30Gly; LDLR Cys222_Asp227dup plus APOB Thr3754Ile + Thr2749Ala). In addition, 2 novel mutations in 2 subjects have been detected in the FH-CEGP cohort. One is a duplication of the nucleotide 1125 that results in a frameshift mutation (p.Lys376Glnfs*5), and one is an intronic deletion that affects a splicing site (c.1587-2delA). Pathogenicity of these mutations should be confirmed as well.
In both FH-Gothenburg and Milan cohorts, the diagnosis of CVD was performed based on the presence of previous myocardial infarction or stroke. Lp(a) levels were measured by turbidimetric assays performed using frozen plasma or freshly isolated plasma for FH-Gothenburg and Milan, respectively. We have evaluated the assay agreement between the 2 methods and demonstrated a lack of uniformity. Despite the differences in Lp(a) measurement methods between the 2 cohorts, we detected 2-fold higher Lp(a) levels in FH individuals with CVD than in those without CVD in both cohorts. This finding was consistent in the 2 independent cohorts, suggesting that the final finding is comparable between the 2 cohorts and not affected by the type (frozen or fresh) of samples used for the measurements. However, variation due to assay methodology could affect risk assessment in individual patients, accentuating the need to improve calibration and standardization of Lp(a) assays.
Use of a reference material proposed by the International Federation of Clinical Chemistry and Laboratory Medicine to evaluate analytical methods for the determination of plasma lipoprotein(a).
The same 2-fold increase in Lp(a) that we detected in individuals with previous CVD, was detected by Alonso et al in individuals with genetic diagnosis of FH and CVD.
Lipoprotein(a) levels in familial hypercholesterolemia: an important predictor of cardiovascular disease independent of the type of LDL receptor mutation.
Nenseter et al, detected a 3-fold increased Lp(a) levels in individuals with genetic diagnosis of FH susceptible to CVD (with premature CVD) compared with those CVD resistant (later or no CVD).
Very recently, Ellis et al showed that the combined presence of high Lp(a) and FH confers 5.3-fold increased risk of premature CAD, compared with a 1.9-fold and 3.2-fold increase due to the presence of the 2 factors separately.
All together these findings strongly suggest that Lp(a) is a risk factor for CVD in individuals with genetic diagnosis of FH.
When we looked at other clinical characteristics of the individuals with genetic diagnosis of FH stratified based on the CVD status, we observed that those with previous CVD event had higher classical CVD risk factors, namely diabetes, hypertension, and BMI, but we did not detect any difference in smoking status. This is consistent with a previous report on 1690 individuals with clinical diagnosis of FH in which those with previous CVD events had higher classical CVD risk factors with the exception of smoking.
When we stratified the FH cohorts based on type of mutation (nonsense vs missense mutation), we found that individuals with nonsense mutation had an overall more severe phenotype (namely, higher total cholesterol, and LDL-C) in FH Gothenburg with similar trend in FH Milano. This is consistent with the possible presence of residual LDLR activity in missense mutations, which may mitigate the FH phenotype. In other words, nonsense mutations induce a complete loss of function of LDLR and may result in a more severe phenotype; missense mutations alter LDLR activity without inducing a complete loss of function. This results in a possible presence of residual LDLR activity in missense mutations, which may mitigate the FH phenotype.
In addition, we observed that individuals carrying nonsense mutations had higher Lp(a) levels than those carrying missense mutations in the FH CEGP Milan cohort (with a similar but not significant trend in the FH Gothenburg). It has been proposed that LDLR mutations result in high Lp(a) levels with a gene-dosage effect.
Lipoprotein(a) levels in familial hypercholesterolemia: an important predictor of cardiovascular disease independent of the type of LDL receptor mutation.
In this study, only the most prevalent null and defective mutations on the LDLR were included in the analysis. Here we extend these findings in 2 independent cohorts (one from Sweden and one from Italy) by a comprehensive analysis of all the nonsense or missense mutations detected on the sequenced genes. Considering that most individuals from our 2 cohorts were carriers of mutations on the LDLR, our results together with previous data support the role of LDLR in Lp(a) circulating levels although the mechanisms remain still known.
When we looked at the response to lipid-lowering therapies, we found the same lowering effect of statins irrespective of the type of mutation in both FH cohorts. The efficacy of statins in FH shows high interindividual variability.
It has been proposed that this variability may be a function of the type of FH-causative mutation. Previous studies investigated this hypothesis with conflicting results. Most of these studies, consistently with our results, detected similar response to statins irrespective of the type of mutation.
Influence of genotype at the low density lipoprotein (LDL) receptor gene locus on the clinical phenotype and response to lipid-lowering drug therapy in heterozygous familial hypercholesterolaemia. The Familial Hypercholesterolaemia Regression Study Group.
Heterozygous familial hypercholesterolaemia: the influence of the mutation type of the low-density-lipoprotein receptor gene and PvuII polymorphism of the normal allele on serum lipid levels and response to lovastatin treatment.
Is responsiveness to lovastatin in familial hypercholesterolaemia heterozygotes influenced by the specific mutation in the low-density lipoprotein receptor gene?.
However, other studies described a better response to statins in individuals with more deleterious mutations (null allele) than individuals with milder mutations (defective allele),
Genetic determinants of responsiveness to the HMG-CoA reductase inhibitor fluvastatin in patients with molecularly defined heterozygous familial hypercholesterolemia.
Influence of LDL receptor gene mutation and apo E polymorphism on lipoprotein response to simvastatin treatment among adolescents with heterozygous familial hypercholesterolemia.
Association of specific LDL receptor gene mutations with differential plasma lipoprotein response to simvastatin in young French Canadians with heterozygous familial hypercholesterolemia.
Presence and type of low density lipoprotein receptor (LDLR) mutation influences the lipid profile and response to lipid-lowering therapy in Brazilian patients with heterozygous familial hypercholesterolemia.
and few studies, on the contrary, reported a better response to statins in individuals with mild mutations compared with individuals with deleterious mutations.
Genetic diagnosis of familial hypercholesterolemia in a South European outbreed population: influence of low-density lipoprotein (LDL) receptor gene mutations on treatment response to simvastatin in total, LDL, and high-density lipoprotein cholesterol.
The type of mutation in the low density lipoprotein receptor gene influences the cholesterol-lowering response of the HMG-CoA reductase inhibitor simvastatin in patients with heterozygous familial hypercholesterolaemia.
All these previous studies examined the response to weak statins (mostly fluvastatin and simvastatin) stratified by type of mutation of only the LDLR gene. In the present study, we examined the response to stronger statins (atorvastatin and rosuvastatin). In addition, we extended the analysis not only at the LDLR mutations but also to all the nonsense or missense mutations detected in LDLR, APOB, PCSK9, and STAP1 gene.
Moreover, for the first time to our knowledge we examined the lipid-lowering efficiency of PCSK9 inhibition stratified by type of mutation. We observed a trend in a better reduction in the LDL-C in carriers of nonsense mutations under anti-PCSK9 treatment alone (FH-Milan cohort) or in combination with statins (FH-Gothenburg cohort). These results need confirmation on larger numbers of patients treated with PCSK9 inhibition.
In conclusion, our results suggest that Lp(a) contributes to CVD onset in individuals with genetic diagnosis of FH. Our finding supports the importance to identify an efficacious therapy to lower Lp(a) in patients with FH to prevent CVD onset or recurrence.
Acknowledgments
The authors thank Lina Håkansson and Carola Gustavsson for their support to the project. The LIPIGEN study is an initiative of the SISA Foundation supported by an unconditional research grant from Sanofi, United States.
Authors' contributions: CPa and RMM performed the analysis and drafted the manuscript; CPi, JS, and SR recruited patients from the FH-Gothenburg cohort; JS contributed to genetic diagnosis and Lp(a) measurement for FH-Gothenburg cohort; KB contributed to data collection for the FH-Gothenburg cohort; CT and LM contributed to genetic diagnosis for the FH-Milan cohort; CPa and LC contributed to data collection for the FH-Milan cohort; SR, LC, and RMM contributed to the study concept and design, contributed to the analysis and interpretation of data. All authors contributed to drafting or revising the article, gave final approval of the version to be published, and agree to be accountable for all aspects of the work.
Financial disclosures
This work was supported by project grant by Amgen, United States and Sanofi-Aventis, United States, the Swedish Research Council, Sweden [Vetenskapsrådet (VR), 2016-01527], the Swedish state under the Agreement between the Swedish government and the county councils (the ALF agreement) [SU 2018-04276], the Novo Nordisk Foundation Grant for Excellence in Endocrinology, Denmark [Excellence Project, 9321-430], the Swedish Diabetes Foundation, Sweden [DIA 2017-205], the Swedish Heart Lung Foundation [20120533], the Wallenberg Academy Fellows from the Knut and Alice Wallenberg Foundation [KAW 2017.0203] (SR). CPa was supported by XXXI cycle of the Doctorate in Pharmacological, Experimental and Clinical Sciences of the University of Milano, Italy.
Method comparison for Lp(a) quantification in FH Gothenburg and FH CEGP Milan cohorts. A total of 27 samples from the Gothenburg cohort representative of the whole measurement range were selected for this method comparison. (A) Results from the Gothenburg method (Randox) were plotted against results from the Milan method (Tina-quant). A regression line was fitted and a Pearson r correlation coefficient was calculated. Using Spearman correlation, r = 0.88. P < .0001 for all the 3 models. (B) Bland-Altman plot visualization of the difference between measurement results for the Milan and Gothenburg methods. The dashed line represents the mean difference between the methods, indicating a minimal bias between the 2 methods.
Appendix
Supplementary Table 1Description of the mutations of unknown pathogenicity in the LDLR gene from the FH-Gothenburg cohort (Ref Seq NM_000527.4)
BMI, body mass index; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; AST, aspartate aminotransferase; ALT, alanine aminotransferase; γ-GT, gamma-glutamyltransferase; HOMA-IR, homeostatic model assessment; PCSK9 mAbs, proprotein convertase subtilisin kexin type 9 monoclonal antibodies.
Data expressed as mean (±SD), median, and quartile range or as number (proportion) as appropriate. Lipid profile was measured before starting any treatment. For continuous traits, P-value calculated by linear regression adjusted for age, gender, and BMI when appropriate. Non-normally distributed values were log-transformed before entering the model. For categorical traits, P-value calculated by chi-square or by Fisher exact test as appropriate.
BMI, body mass index; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; AST, aspartate aminotransferase; ALT, alanine aminotransferase; γ-GT, gamma-glutamyltransferase; HOMA-IR, homeostatic model assessment; PCSK9 mAbs, proprotein convertase subtilisin kexin type 9 monoclonal antibodies.
Data expressed as mean (±SD), median, and quartile range or as number (proportion) as appropriate. Lipid profile was measured before starting any treatment. For continuous traits, P-value calculated by linear regression adjusted for age, gender, and BMI when appropriate. Non-normally distributed values were log-transformed before entering the model. For categorical traits, P-value calculated by chi-square or by Fisher exact test as appropriate.
∗ Data on xanthomas available for N = 93.
† Data on arcus cornealis available for N = 88.
‡ Treated LDL-C at last available assessment.
§ Data on statin use available for N = 109.
‖ Data available for N = 65 (Nonsense) and N = 76 (Missense).
Data on statin use available for N = 140 and N = 123 for Milan and Gothenburg cohort, respectively.
105 (85)
92 (67)
<.0001
PCSK9 mAbs use, n (%)
13 (7)
26 (15)
.001
BMI, body mass index; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; AST, aspartate aminotransferase; ALT, alanine aminotransferase; γ-GT, gamma-glutamyltransferase; HOMA-IR, homeostatic model assessment; PCSK9 mAbs, proprotein convertase subtilisin kexin type 9 monoclonal antibodies.
Data are expressed as mean (±SD) or as number (proportion) as appropriate. Lipid profile was measured before starting any treatment. For continuous traits, P-value calculated by linear regression adjusted for age, gender, and BMI when appropriate. Non-normally distributed values were log-transformed before entering the model. For categorical traits, P-value calculated by chi-square or by Fisher exact test as appropriate.
∗ Data on xanthomas available for N = 162 and N = 101 for Milan and Gothenburg cohort, respectively.
† Data on arcus cornealis available for N = 162 and N = 95 for Milan and Gothenburg cohort, respectively.
‡ Data on cardiovascular disease available for N = 141 for Gothenburg cohort.
§ Treated LDL-C at last available assessment.
‖ Data on statin use available for N = 140 and N = 123 for Milan and Gothenburg cohort, respectively.
BMI, body mass index; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; AST, aspartate aminotransferase; ALT, alanine aminotransferase; γ-GT, gamma-glutamyltransferase; HOMA-IR, homeostatic model assessment; PCSK9 mAbs, proprotein convertase subtilisin kexin type 9 monoclonal antibodies.
Data are expressed as mean (±SD), median, and quartile range or as number (proportion) as appropriate. Lipid profile was measured before starting any treatment.
BMI, body mass index; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; AST, aspartate aminotransferase; ALT, alanine aminotransferase; γ-GT, gamma-glutamyltransferase; HOMA-IR, homeostatic model assessment; PCSK9 mAbs, proprotein convertase subtilisin kexin type 9 monoclonal antibodies.
Data are expressed as mean (±SD), median, and quartile range or as number (proportion) as appropriate. Lipid profile was measured before starting any treatment.
Coronary heart disease risk factors ranked by importance for the individual and community. A 21 year follow-up of 12 000 men and women from The Copenhagen City Heart Study.
Nonoptimal lipids commonly present in young adults and coronary calcium later in life: the CARDIA (Coronary Artery Risk Development in Young Adults) study.
Familial hypercholesterolaemia is underdiagnosed and undertreated in the general population: guidance for clinicians to prevent coronary heart disease: consensus statement of the European Atherosclerosis Society.
Association of LPA variants with risk of coronary disease and the implications for lipoprotein(a)-lowering therapies: a Mendelian randomization analysis.
Relation of serum lipoprotein(a) concentration and apolipoprotein(a) phenotype to coronary heart disease in patients with familial hypercholesterolemia.
Lipoprotein(a) levels in familial hypercholesterolemia: an important predictor of cardiovascular disease independent of the type of LDL receptor mutation.
Clinical characteristics and evaluation of LDL-cholesterol treatment of the Spanish Familial Hypercholesterolemia Longitudinal Cohort Study (SAFEHEART).
Use of a reference material proposed by the International Federation of Clinical Chemistry and Laboratory Medicine to evaluate analytical methods for the determination of plasma lipoprotein(a).
Influence of genotype at the low density lipoprotein (LDL) receptor gene locus on the clinical phenotype and response to lipid-lowering drug therapy in heterozygous familial hypercholesterolaemia. The Familial Hypercholesterolaemia Regression Study Group.
Heterozygous familial hypercholesterolaemia: the influence of the mutation type of the low-density-lipoprotein receptor gene and PvuII polymorphism of the normal allele on serum lipid levels and response to lovastatin treatment.
Is responsiveness to lovastatin in familial hypercholesterolaemia heterozygotes influenced by the specific mutation in the low-density lipoprotein receptor gene?.
Genetic determinants of responsiveness to the HMG-CoA reductase inhibitor fluvastatin in patients with molecularly defined heterozygous familial hypercholesterolemia.
Influence of LDL receptor gene mutation and apo E polymorphism on lipoprotein response to simvastatin treatment among adolescents with heterozygous familial hypercholesterolemia.
Association of specific LDL receptor gene mutations with differential plasma lipoprotein response to simvastatin in young French Canadians with heterozygous familial hypercholesterolemia.
Presence and type of low density lipoprotein receptor (LDLR) mutation influences the lipid profile and response to lipid-lowering therapy in Brazilian patients with heterozygous familial hypercholesterolemia.
Genetic diagnosis of familial hypercholesterolemia in a South European outbreed population: influence of low-density lipoprotein (LDL) receptor gene mutations on treatment response to simvastatin in total, LDL, and high-density lipoprotein cholesterol.
The type of mutation in the low density lipoprotein receptor gene influences the cholesterol-lowering response of the HMG-CoA reductase inhibitor simvastatin in patients with heterozygous familial hypercholesterolaemia.
Conflict of interest: SR has been consulting for Amgen, Sanofi-Aventis, Novo Nordisk, Akcea therapeutics, Genzyme, AstraZeneca, Chiesi Farmaceutici Group in the last 5 years. SR has received research grants from Amgen, United States, Sanofi, United States, and AstraZeneca, United Kingdom.
30231-4&id=gr1.jpg){kind=link}
30231-4&id=gr2.jpg){kind=link}
30231-4&id=gr3.jpg){kind=link}
30231-4&id=gr4.jpg){kind=link}