WO2017218798A1

WO2017218798A1 - Systems and methods for diagnosing familial hypercholesterolemia

Info

Publication number: WO2017218798A1
Application number: PCT/US2017/037715
Authority: WO
Inventors: Robert Hegele
Original assignee: Boston Heart Diagnostics
Priority date: 2016-06-15
Filing date: 2017-06-15
Publication date: 2017-12-21

Abstract

A method is provided for diagnosing familial hypercholesterolemia (FH) based on mutations and copy number variations in certain monogenic contributor genes, low-density lipoprotein levels, and the presence of certain polygenic contributor single nucleotide polymorphisms. Methods account for multiple contributors to FH and therefore capture a larger percentage of diagnoses than monogenic mutations alone.

Description

SYSTEMS AND METHODS FOR DIAGNOSING FAMILIAL HYPERCHOLESTEROLEMIA

Cross-Reference to Related Applications

This application claims priority to U.S. Provisional patent application Ser. No.

62/350,388, filed June 15, 2016, the contents of which are herein incorporated by reference in their entirety.

Field of the Invention

The present invention relates generally to methods for diagnosing familial

hypercholesterolemia using a weighted combination of biomarkers in a patient.

Background

Familial hypercholesterolemia (FH) is a relatively common disorder resulting in high cholesterol levels and, more specifically, high levels of low-density lipoprotein (LDL-C) which can result in cardiovascular disease. While the homozygous state for familial

hypercholesterolemia (resulting from two abnormal copies of the LDLR gene) is readily identifiable with symptoms that manifest early, the more common heterozygous condition is less readily diagnosed. The heterozygous form of FH (HeFH) is characterized by lifelong elevations in plasma low-density lipoprotein (LDL) cholesterol, typically > 5.0 mmol/L (194 mg/dL), sometimes occurring with characteristic physical signs, such as cholesterol-rich fatty deposits. Heterozygous FH is often accompanied by personal or family history of early cardiovascular disease (CVD). See Nordestgaard, et al., 2013, 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; Eur Heart J. 2013;34:3478-3490, incorporated herein by reference. Recent population-based surveys, including screening with DNA sequencing, suggest HeFH has a prevalence of - 1 in 217 individuals in Northern Europe. See Benn, et al., Mutations causative of familial

hypercholesterolaemia: Screening of 98,098 individuals from the Copenhagen general population study estimated a prevalence of 1 in 217; Eur Heart J. 2016;37: 1384-1394, incorporated herein by reference. Familial Hypercholesterolemia was historically thought to be monogenic, but later research has identified cases of FH that are not explained by mutations in single known monogenic contributors. For example, large-scale whole exome sequencing efforts indicate that about 4% of individuals with early coronary heart disease (CFID) have HeFH resulting from one of several loss-of-function mutations in the LDLR gene encoding the LDL receptor. Do, et al., Exome sequencing identifies rare LDLR and APOA5 alleles conferring risk for myocardial infarction. Nature. 2015;518: 102-106, incorporated herein by reference. Other large-scale sequencing efforts indicate that within subgroups of individuals with severe

hypercholesterolemia, defined as untreated LDL-C > 190 mg/dL (5.0 mmol/L), only about 2% had a pathogenic mutation in an autosomal dominant FH gene4. See Khera, et al., Diagnostic yield of sequencing familial hypercholesterolemia genes in patients with severe

hypercholesterolemia. J Am Coll Cardiol. 2016, incorporated herein by reference. Furthermore, high-throughput DNA sequencing has shown that 20 to 40% of individuals with phenotypic HeFH have no mutation in canonical FH genes such as LDLR, APOB or PCSK9. See Talmud, et al., Use of low-density lipoprotein cholesterol gene score to distinguish patients with polygenic and monogenic familial hypercholesterolaemia: A case-control study, Lancet.

2013;381 : 1293-1301, incorporated herein by reference.

Recent studies have hypothesized other potential contributors to FH. A few individuals have rare mutations in minor genes such as APOE, ABCG5, ABGCG8, LIPA or STAP1, underlying a phenotype that resembles FH. See Hegele, et al., Targeted next-generation sequencing in monogenic dyslipidemias. Curr Opin Lipidol. 2015;26: 103-113, incorporated herein by reference. Others carry a disproportionately high burden of multiple small effect common single nucleotide polymorphisms or SNPs each of which incrementally raises plasma LDL cholesterol by a fraction of a mmol/L, but which cumulatively raise LDL cholesterol into the FH range (polygenic contributors). Other individuals with apparent HeFH have none of the above associated causes, suggesting that mutations inaccessible by exome sequencing, such as deep intronic variants or large-scale copy number variations (CNVs), mutations in as yet undefined genes, gene-by-gene interactions, gene-by-environment interactions, non-mendelian mechanisms (e.g. epigenetic imprinting) or purely environmental factors could explain their phenotype. See Wang, et al., Multiplex ligation-dependent probe amplification of LDLR enhances molecular diagnosis of familial hypercholesterolemia, J Lipid Res. 2005;46:366-372, incorporated herein by reference.

Given the number of possible contributors to HeFH described above, specific diagnosis is difficult but can be useful in tailoring treatment methods for high cholesterol and understanding patient response to treatment. Furthermore, such a diagnosis may be important where it is a condition for third party reimbursement of novel LDL-lowering therapies. See Hegele, et al., Nonstatin low-density lipoprotein-lowering therapy and cardiovascular risk reduction-statement from atvb council, Arterioscler Thromb Vase Biol. 2015;35:2269-2280; Hegele, et al., Improving the monitoring and care of patients with familial hypercholesterolemia, J Am Coll Cardiol.

2016;67: 1286-1288.

Summary

The present invention generally provides methods for diagnosing FH in a patient. The invention provides a weighted model for FH diagnosis that accounts for copy number variation and mutations in one or more monogenic contributors to FH, measured LDL-C levels, and the presence of one or more polygenic contributor single nucleotide polymorphisms (SNPs).

Systems and methods of the invention account for the various contributors to FH and assign weights to each of the contributors in order to deliver an accurate diagnostic tool that provides patients and medical professionals a valuable tool in identifying and, thereby, treating cardiovascular disease and high cholesterol levels especially where caused by FH.

One component of the invention includes an optimized set of SNPs, with weights based on homozygosity or heterozygosity of the SNP and the contribution of the SNP to LDL-C levels. Systems of the invention may include a next generation sequencing panel for dyslipidemias including FH, which assesses simultaneously from batches of 24 clinical samples the monogenic and polygenic determinants of severely elevated LDL cholesterol. The panel includes major genes (LDLR, APOB, PCSK9 and LDLRAPl) and minor genes (APOE, ABCG5, ABCG8, LIP A and STAPl) underlying monogenic FH (monogenic contributors) as well as SNP genotypes for LDL cholesterol (polygenic contributors).

Systems and methods may also include detection of copy number variations in monogenic contributors and, specifically the LDLR gene. In certain embodiments, systems and methods may include multiplex ligation primer amplification (MLPA) assays to detect large- scale CNVs in the LDLR gene.

Aspects of the invention include methods for diagnosing familial hypercholesterolemia by identifying in nucleic acid from a patient, a plurality of single nucleotide polymorphisms (SNPs) in one or more polygenic contributors to familial hypercholesterolemia, determining if the patient is homozygous or heterozygous for the SNPs, and creating a genetic risk score for familial hypercholesterolemia score using a weighted, multivariate model based on the identified SNPs and the heterozygosity or homozygosity of the patient for the trait-raising allele. In certain embodiments, methods may further include treating the patient for familial hypercholesterolemia based on the genetic risk score. Each identified SNP may be weighted based on measured effect of the identified SNP on LDL-C in a cohort of individuals with familial hypercholesterolemia. The SNPs may comprise one, two, three, four or more selected from the group consisting of: rsl2740374, rsl 1206510, rs515135, rs6544713, rsl501908, rs3846663, rs2650000, rs6511720, rsl0401969, and rs6102059. In certain embodiments, the SNPs may comprise all of the above SNPs.

The weighted multivariate model may include the following components:

.23*rsl2740374 + .09*rsl 1206510 + .16*rs515135 + .15*rs6544713 + .07*rsl501908 + .07*rs3846663 + .07*rs2650000 + .26*rs6511720 + .05*rsl0401969 + .06*rs6102059 = genetic risk score (GRS), where rsl2740374, rsl 1206510, rs515135, rs6544713, rsl501908, rs3846663, rs2650000, rs6511720, rsl 0401969, and rs6102059 are assigned a value of 0, 1, or 2 based on presence of the corresponding trait-raising alleles of SNPs in nucleic acid from the patient and heterozygosity or homozygosity of the patient with respect to the corresponding SNPs. In certain embodiments, a GRS greater than or equal to 1.96 may indicate an extreme weighted GRS or (wGRS) and an elevated risk of familial hypercholesterolemia.

Certain aspects of the invention relate to methods for diagnosing a patient with familial hypercholesterolemia including the steps of comparing a nucleic acid sequence obtained from a patient and encoding a monogenic contributor to familial hypercholesterolemia to a reference genome to identify a mutation and assaying the monogenic contributor to familial

hypercholesterolemia for copy number variation. Further steps may include testing a polygenic contributor to familial hypercholesterolemia for a single nucleotide polymorphism (SNP) and determining low-density lipoprotein cholesterol (LDL-C) levels in the patient. Steps may include diagnosing familial hypercholesterolemia in the patient using a weighted multivariate model where the model factors in the presence of the mutation in the monogenic contributor to familial hypercholesterolemia, the presence of the copy number variation in the monogenic contributor to familial hypercholesterolemia, the presence of the S P in the polygenic contributor to familial hypercholesterolemia, and the LDL-C level in the patient.

Certain methods include treating the patient with a statin or a monoclonal antibody against PCSK9 based on the above diagnosing step. In various embodiments, the nucleic acid sequence encoding a monogenic contributor to familial hypercholesterolemia may include one, two, three, four, or more of: LDLR, APOB, PCSK9, STAP1, APOE, ABCG5, ABCG8, LDLRA1, and LIP A. The nucleic acid sequence encoding a monogenic contributor to familial

hypercholesterolemia may include nucleic acids encoding each of the above genes.

The SNP may include one, two, three, four, or more selected from the group consisting of: rsl2740374, rsl 1206510, rs515135, rs6544713, rsl501908, rs3846663, rs2650000, rs6511720, rsl0401969, and rs6102059. In certain embodiments, the weighted multivariate model may include all of the above listed SNPs.

In certain embodiments, the LDL-C level being greater than about 5.0 mmol/L may be indicative of an increased likelihood of familial hypercholesterolemia in the patient. In some embodiments, the LDL-C level being greater than about 8.0 mmol/L may be indicative of an increased likelihood of familial hypercholesterolemia in the patient. The reference genome may be selected from the Human Gene Mutation Database or the University College London FH mutation database. The copy number variation can be determined by multiplex ligation- dependent probe amplification (MLPA).

In certain aspects, systems of the invention may include a composition for detecting coding region mutations present in a human DNA sample and associated with a low-density lipoprotein cholesterol (LDL-C) polygenic trait score (PTS) calculated from the allele status of single nucleotide polymorphisms (SNPs) comprising at least one oligonucleotide capture probe corresponding to each SNP of the set consisting of: rsl 1206510, rsl2740374, rs515135, rs6544713, rs3846663, rsl501908, rs2650000, rs6511720, rsl0401969 and rs6102059. The capture probes may be designed to coordinate with the human GRCh37/hgl9 build.

Aspects of the invention may include methods for determining a polygenic risk susceptibility profile based on the ascertainment of allele status of SNP probes listed in 1.1 in each subject and creating an aggregate score, i.e. PTS. The score for any given allele may be determined by the identification of the LDL-C trait raising allele. Where an allele is present in a test subject, the score may be given as a 1. Where two alleles are measured at each SNP target, the score may be a 0 or 1 or 2. The resultant score may then be multiplied by the weighting factor for the SNP. In various embodiments, the cutoff for the 90th percentile of the PTS may be 1.96 and the cutoff for the 95th percentile may be 2.02.

In certain embodiments diagnostic systems and methods may also include familial history of FH and CVD as a component of the diagnosis.

Brief Description of the Drawings

FIG. 1 diagrams steps of polygenic diagnostic methods of the invention.

FIG. 2 diagrams steps of diagnostic methods of the invention including polygenic, monogenic, LDL-C, and copy number variation factors.

FIG. 3 is a graph of LDL cholesterol level based on various monogenic and polygenic contributors to FH.

FIG. 4 shows percentage of FH individuals captured by methods of the invention among various ranges of measured LDL cholesterol.

FIG. 5 shows percentages of individuals from various samples and subgroups with an extreme weighted GRS > 1.96.

FIG. 6 shows distribution of weighted GRSs for elevated LDL cholesterol.

FIG. 7 illustrates polygenic risk scores in FH with no mutation.

FIG. 8 illustrates polygenic risk scores in FH.

FIG. 9 shows a schematic of a computing device that may appear in the methods of the invention.

FIG. 10 illustrates steps of the LipidSeq workflow according to certain embodiments.

Detailed Description

The present invention relates to diagnosing FH in a patient while accounting for multiple FH contribution factors. The invention provides a weighted model including factors such as copy number variation and mutations in monogenic contributors to FH, measured LDL-C levels, and polygenic contributors including an optimized panel of single nucleotide polymorphisms (SNPs) identified in a patient. Systems and methods of the invention comprise an accurate diagnostic tool that provides patients and medical professionals a valuable tool in identifying and, thereby, treating cardiovascular disease and high cholesterol levels especially where caused by FH.

FIG. 1 illustrates steps according to certain polygenic FH diagnostic methods of the invention. SNPs are identified in nucleic acid samples from a patient 281 and the patient is determined to be heterozygous or homozygous at the SNP 283. The SNPs are then weighted based on presence and homo or heterozygosity to create a genetic risk score 287 from which a FH diagnosis may be determined 289.

SNPs for polygenic risk analysis may include one, all, or some combination of the following SNPs: rsl 1206510, rsl2740374, rs515135, rs6544713, rs3846663, rsl501908, rs2650000, rs6511720, rsl0401969 and rs6102059. The SNPs may be weighted in the model based on the presence on one or both of alleles of the SNP (assigned a 0, 1, or 2 as shown in Table 1) and then further weighted depending on the specific SNPs contribution to increased or lowered LDL-C levels (weighting factors described in Table 1). The presence, in one or both alleles, of an SNP may be determined using any of several known methods including allele-specific amplification primers or allele-specific probes capable of determining whether the genotype of the individual is heterozygous or homozygous for the one or more polymorphisms described in table 1.

According to certain methods of the invention, FH diagnosis may include analyzing certain SNPs to establish a genetic risk score (GRS). In certain embodiments the GRS may be determined by the following weighted model:

23*rsl2740374 + .09*rsl 1206510 + .16*rs515135 + .15*rs6544713 + .07*rsl501908 +

.07*rs3846663 + .07*rs2650000 + .26*rs6511720 + .05*rsl0401969 + .06*rs6102059 = GRS where rsl2740374, rsl 1206510, rs515135, rs6544713, rsl501908, rs3846663, rs2650000, rs6511720, rsl0401969, and rs6102059 are assigned a value of 0, 1, or 2 based on presence of the corresponding SNPs in nucleic acid from the patient and heterozygosity or homozygosity of the patient with respect to the corresponding SNPs. Table 1 shows the 10 S Ps discussed for assessing polygenic genetic risk scores according to systems and methods of the invention along with the chromosome they may be found on and the rsID. Ref refers to the reference nucleobase anticipated at the location, Alt refers to the SNP variant nucleobase and risk refers to which of the Ref or Alt nucleobase is associated with an increased LDL-C level in patients. The Ref score, Het score, and Horn score provides the value to be assigned in the above GRS calculation for each respective SNP based on the presence in the patient's nucleic acid sample of the Reg geno, Het geno, or Horn geno, respectively. For example, if a patient's sequence for SNP rs515135 was C_C, then rs515135 in the above equation would be assigned the value 2.

Once a GRS is calculated, the value can be compared to a threshold value to aid in diagnosis of FH. In certain embodiments, a GRS greater than or equal to about 1.8, 1.85, 1.9, 1.95, 1.96, 2.0, 2.02, 2.05, or other values may indicate an extreme GRS and indicate a likely FH diagnosis.

In certain embodiments, the probes to SNP regions described in table 1 may be used to yield raw data files in the format of two fastq files per test subject. The fastq files, forming reads from one forward and one reverse read from the probes in table 1, may be combined and aligned to the GRCh37/hgl9 human genome build. The file can be further processed to perform local realignments of the reads and then PCR duplicate sequence regions may be removed to yield a consensus sequence file. This consensus sequence can then be compared to the GRCh37/hgl9 reference sequence and variants to the reference sequence in the form of single nucleotide variants (SNVs) can called. The variant sequence data may be converted into a variant call format (vcf) file. The vcf file may then be queried with a script that searches for the scaffold coordinates outlined in Table 1. If the scaffold location is not present in the vcf file, the script returns the reference allele value as a homozygote, otherwise, the script returns the value from the vcf file - either heterozygous or homozygous for the non-reference allele. The score could be a 0 or 1 or 2. The resultant score can then be multiplied by the weighting factor. For the 10 SNP targets, the LDL-C PTS could range from 0 to 2.42 based on the weights given in table 1, where 2.42 is the highest possible PTS. In certain embodiments, the cutoff for the 90th percentile of polygenic FH risk is 1.96 and the cutoff for the 95th percentile is 2.02. The weighted LDL-C PTS value may then be output to a text file or other output or written report as described elsewhere herein that represents a clinical finding or diagnosis, which is ready for review for by a clinician or medical professional. The clinician can then relate the genomic finding and evaluate the PTS in the context of physical, demographic, clinical and biochemistry data that exists for the patient to create a genomic diagnosis of polygenic susceptibility for elevated LDL- C cholesterol to be included in a clinical report; a diagnosis of FH; and/or a treatment regimen for the patient for lowering LDL-C levels.

FIG. 2 illustrates steps for multifactor methods of diagnosing FH according to certain embodiments. The patient's nucleic acid sequences encoding for certain monogenic contributor genes to FH is determined and compared to a reference to identify mutations 371. These monogenic contributor genes may include one, all, or some combination of LDLR, APOB, PCSK9, STAP1, APOE, ABCG5, ABCG8, LDLRAl, and LIP A. The reference genome may be any known standard such as GRCh37/hgl9. Mutations may be verified according to the Human Gene Mutation Database or the University College London FH mutation database.

The patient's nucleic acid may then be assayed to identify CNV in a monogenic contributor gene 373, especially LDLR. Polygenic contributor SNPs may then be identified 375 in the sequencing data or using targeted primers or probes as described above. LDL-C levels are also determined 377 using any known method including standard lipid panels using a blood sample from the patient. The monogenic mutations, identified CNVs, polygenic SNPs, and LDL-C levels are then input into a weighted model to determine an FH diagnosis 279 and potential treatments.

In certain embodiments, where LDL-C levels are measured to be greater than or equal to 5.0 mmol/L, the presence of one or more monogenic mutations may be assigned a weight of about .473 and the presence of a CNV in the monogenic contributor may be assigned a weight of about .064. Polygenic SNPs may be fed into a GRS calculation as described above where a GRS above a certain threshold (e.g., 1.96) indicates an extreme GRS and is therefore added to the model with a weight of about .134. For example, if a patient was determined to have an extreme GRS, an APOB mutation, and a large CNV in LDLR, the weighted model would output a FH diagnostic risk percentage of 67.1% (.671).

Weights for monogenic and polygenic contributors may be adjusted based on the LDL-C cholesterol level according to the percentages given in FIG. 4.

Nucleic acid may be isolated from a patient according to any known method. In certain embodiments, circulating cell-free nucleic acid is obtained from an individual. Circulating cell- free nucleic acid may be any fragments of DNA or ribonucleic acid (RNA) that are present in the blood of an individual. Cell-free nucleic acid may be from sub-cellular sources such as mitochondria or other organelles or cell fragments from any cell type in the human body. In a preferred embodiment, the circulating cell-free nucleic acid is one or more fragments of DNA obtained from the plasma or serum of the individual.

In certain embodiments, monogenic mutations may be identified using the LipidSeq method as described in Hegele, et al., Targeted next-generation sequencing in monogenic dyslipidemias. Curr Opin Lipidol. 2015;26: 103-113 and Johansen, et al., Lipidseq: A next- generation clinical resequencing panel for monogenic dyslipidemias. J Lipid Res. 2014;55:765- 772; each of which is incorporated herein in its entirety. The LipidSeq workflow is illustrated in FIG. 10.

In certain embodiments, the LipidSeq NGS method and pipeline may be used for identifying and reporting both small-scale variants and polygenic risk scores, while CNVs may require an independent MLPA method be run in parallel. In certain embodiments, bioinformatic annotation tools may be used to predict CNVs from NGS data. In such embodiments, a single NGS plus bioinformatics platform may be used to identify and report small-scale sequence variants, large-scale CNVs and genetic risk scores useful in models of the invention. MLPA may be reserved for confirming predicted CNVs from NGS results.

In certain embodiments, one or more steps of the methods of the invention, such as running the multivariate model, may be performed by a computing device 511 comprising a processor 309 and a tangible, non-transient memory 307. Computing devices may generate a written diagnostic report with results of the model. Written reports may be an electronic document and may be sent, electronically (e.g., through email) to a recipient. The written report may be sent to an output device such as a display monitor or a printer.

A computing device 511 according to methods of the invention generally includes at least one processor 309 coupled to a memory 307 via a bus and input or output devices 305 as shown in FIG. 9.

As one skilled in the art would recognize as necessary or best-suited for the systems and methods of the invention, systems and methods of the invention include one or more servers 511 and/or computing devices 101 that may include one or more of processor 309 (e.g., a central processing unit (CPU), a graphics processing unit (GPU), etc.), computer-readable storage device 307 (e.g., main memory, static memory, etc.), or combinations thereof which

communicate with each other via a bus.

A processor 309 may include any suitable processor known in the art, such as the processor sold under the trademark XEON E7 by Intel (Santa Clara, CA) or the processor sold under the trademark OPTERON 6200 by AMD (Sunnyvale, CA).

Memory 307 preferably includes at least one tangible, non-transitory medium capable of storing: one or more sets of instructions executable to cause the system to perform functions described herein (e.g., software embodying any methodology or function found herein); data (e.g., portions of the tangible medium newly re-arranged to represent real world physical objects of interest accessible as, for example, a picture of an object like a motorcycle); or both. While the computer-readable storage device can in an exemplary embodiment be a single medium, the term "computer-readable storage device" should be taken to include a single medium or multiple media (e.g., a centralized or distributed database, and/or associated caches and servers) that store the instructions or data. The term "computer-readable storage device" shall accordingly be taken to include, without limit, solid-state memories (e.g., subscriber identity module (SIM) card, secure digital card (SD card), micro SD card, or solid-state drive (SSD)), optical and magnetic media, hard drives, disk drives, and any other tangible storage media.

Input/output devices 305 according to the invention may include one or more of a video display unit (e.g., a liquid crystal display (LCD) or a cathode ray tube (CRT) monitor), an anumeric input device (e.g., a keyboard), a cursor control device (e.g., a mouse or trackpad), a disk drive unit, a signal generation device (e.g., a speaker), a touchscreen, a button, an accelerometer, a microphone, a cellular radio frequency antenna, a network interface device, which can be, for example, a network interface card (NIC), Wi-Fi card, or cellular modem, or any combination thereof.

One of skill in the art will recognize that any suitable development environment or programming language may be employed to allow the operability described herein for various systems and methods of the invention. For example, systems and methods herein can be implemented using R, MATLAB, Perl, Python, C++, C#, Java, JavaScript, Visual Basic, Ruby on Rails, Groovy and Grails, or any other suitable tool. For a computing device 101, it may be preferred to use native xCode or Android Java.

Example 1: Polygenic and monogenic etiologies of clinically ascertained hypercholesterolemia

Three hundred and thirteen consecutive unrelated adults aged > 18 years from Ontario, Canada referred with possible HeFH were included in a study of polygenic and monogenic etiologies; patients known to have homozygous FH were excluded. Untreated fasting lipid profiles were recorded. All subjects had untreated Friedewald-determined plasma LDL cholesterol > 5.0 mmol/L (194 mg/dL), plus either a personal or family history of early CVD, plus family history of hyperlipidemia. Physical findings were not uniformly recorded.

Genomic DNA was isolated from whole blood. Target enriched genomic libraries of indexed and pooled samples were generated for target candidate genes in lipid metabolism, including the known causative genes for FH, namely LDLR, APOB, PCSK9, STAPl, APOE, ABCG5, ABCG8, LDLRA1 and LIPA on the LipidSeq Panel. See Hegele, et al., Targeted next- generation sequencing in monogenic dyslipidemias. Curr Opin Lipidol. 2015;26: 103-113;

Johansen, et al., Lipidseq: A next-generation clinical resequencing panel for monogenic dyslipidemias. J Lipid Res. 2014;55:765-772; each of which is incorporated herein in its entirety. The reagents also genotype 10 SNPs statistically shown in table 1 above that are associated with LDL cholesterol levels in the general population that when tallied create a polygenic trait score. Prepared sample libraries were assayed in the MiSeq personal sequencer (Illumina, San Diego CA). The method has average > 300-fold coverage for each base; > 150 base pairs of intronic DNA are sequenced at each intron-exon boundary, in addition to > 1000 base pairs of promoter and 3' untranslated DNA regions. Samples were also run using MLPA for coding regions of the LDLR gene. Sanger sequencing was used to confirm variants detected by NGS.

FASTQ files derived from the MiSeq output were processed individually using custom automated workflow in CLC Genomics Workbench version 8.5.1 (CLCbio, Aarhus, Denmark) for sequence mapping, variant calling and target region coverage statistics. Variant annotation was performed using ANNOVAR (http://www.biobase-international.com/product/annovar) with customized scripts producing a variant call file (vcf).

Variants detected in FH genes have had a long history of archiving and annotation, as well as abundant publications of functional consequences. For instance, > 1700 individual variants previously reported as being causative in FH are reported in the Human Gene Mutation Database (HGMD; http://www.biobase-international.com/product/hgmd) and the University College London (UCL) FH mutation database (http://www.ucl.ac.uk/ldlr/Current/); these reference databases were used for all variants detected by our procedure.

Annotated coding and noncoding (± 10 base pair from adjacent exon) variants in vcfs were first filtered to select the rare variants according to minor allele frequencies (MAF) < 1% in 1000 Genomes Project (GIK; http://www.1000genomes.org/), Exome Variant Server (EVS; http://evs.gs.washington.edu/EVS/) or Exome Aggregation Consortium (ExAC;

http://exac.broadinstitute.org/) databases. Polymorphism Phenotype Version 2 (PolyPhen-2) Sorting Intolerant from Tolerant (SIFT) and Combined Annotation Dependent Depletion (CADD) scores were used to evaluate the deleteriousness of the filtered coding variants. Splicing Based Analysis of Variants (SPANR) and Automated Splice Site and Exon Definition Analyses (ASSEDA; www.http://splice.uwo.ca) were used to identify rare deleterious splicing variants. Novel variants found in this study were determined to be likely causative when: 1) they had no listed allele frequencies in GIK, ESV or ExAC databases, no rsID in the dbSNP database, and/or were not reported in HGMD or UCL FH databases; 2) for coding variants, a deleterious score from > two in silico algorithms; and 3) for non-coding variants, a deleterious score for > one in silico algorithm. CNVs detected by MLPA were similarly searched for in HGMD and UCL FH databases. Hereafter, we will use the term "mutation" interchangeably with "rare definite or very likely causative variant" for the sake of brevity. As controls for our annotation pipeline, we used sequence data from the GIK database. A set of 10 genetic markers (reported above in table 1) associated with raising plasma LDL cholesterol were selected from genome-wide association studies. These were the top 10 S Ps according to effect size on LDL cholesterol per allele. Both weighted and unweighted genetic risk scores (wGRS and uwGRS, respectively) were calculated; for the former, the weighting factors were the published beta-coefficients for per-allele change in LDL cholesterol. We chose the 90th percentile for wGRS - i.e. > 1.96 - as the definition for an extreme score from the subjects in the G1K database (G1K; http://www.1000genomes.org/), which reportedly represents a sample of normal healthy control subjects, although lipid profiles are not available. The 90th percentile for the unweighted GRS from G1K was 16/20.

All statistical comparisons were conducted using SAS version 9.2 (SAS Institute, Cary NC). Between-group differences in quantitative traits means were evaluated using unpaired Student's t-test assuming unequal variances. For discrete traits, χ2 analysis was used and odds ratios (ORs) were calculated using the case-control method with the FREQ procedure. Statistical significance for all comparisons was defined as a two-tailed P-value <0.05.

Baseline demographic features of the Ontario hypercholesterolemia sample, overall and subdivided by gender, are shown in Table 2. 91.1% of individuals were self-reported as

Caucasian, with the remainder being of South Asian, Chinese, African or unspecified ethnic background. The mean + standard deviation (SD) age was 51.0 + 15.1 years (range 18.1 to 88.8 years) and mean + SD untreated LDL cholesterol level was 8.91 + 1.87 mmol/L (344 + 72 mg/dL), with range 5.01 to 13.3 mmol/L (194 to 514 mg/dL).

Likely or definite causal variants in LDLR, APOB and PCSK9 genes detected in this study are listed in Table 3. The key findings summarized by mutation type are shown in Table 2. Overall, 148/313 (47.3%) of individuals had at least one probable or definite causal mutation detected by NGS. All these variants were confirmed with Sanger sequencing. A further 20/313 (6.4%)) had a heterozygous large scale pathogenic CNV detected by MLPA, increasing the proportion with a mutation to 168/313 (53.7%). In the G1K database (N = 1092), our pipeline identified no individuals with possible disease-causing FH mutations.

Most FH mutation-positive Ontario subjects had a heterozygous LDLR gene mutation (141/168 or 83.9%), while 15 (8.9%) and 2 (1.2%) of subjects had mutations in APOB and PCSK9 genes, respectively. No potential causative variants in STAP1, APOE, ABCG5, ABCG8, LDLRAP1 or LIPA were found. Twenty (11.9%) of all FH mutation-positive subjects had a large-scale CNV in the LDLR gene. Although no patient had homozygous FH on clinical grounds, 10 FH mutation-positive individuals (6.0%) were found to have mutations in two genes; of the 20 variant alleles in this pool, 14, 4 and 2 were in LDLR, APOB and PCSK9 genes, respectively (Table 3). The mean LDL cholesterol in carriers of two mutant alleles was non- significantly higher than in the rest of the study sample (Table 3 and FIG. 3). Simple

heterozygotes for APOB or PCSK9 mutations had significantly lower mean LDL cholesterol than simple heterozygotes for LDLR mutations (FIG. 3).

Among the 168 FH mutation-positive individuals were 106 were unique mutations, 90 of which (84.9%)) were within the LDLR gene. Among all mutations, 17 were novel to this study, of which 12 were within the LDLR gene: two splicing, five frameshift, one nonsense and four missense (Table 4). When considering only LDLR gene mutations, there were no differences in mean LDL cholesterol levels between subgroups of individuals with different mutation types (FIG. 3).

There was a stepwise increase in the proportion of individuals with monogenic mutations according to plasma LDL cholesterol stratum (FIG. 4). For individuals with LDL cholesterol 5.00 to 5.99 mmol/L (194 to 231 mg/dL), 6.00 to 6.99 mmol/L (232 to 270 mg/dL), 7.00 to 7.99 mmol/L (271 to 309 mg/dL), and > 8.00 mmol/L (> 310 mg/dL), respectively, 42.1, 40.4, 69.8 and 88.0% were positive for a rare mutation.

Of 145 FH mutation-negative Ontario individuals, 42 (29.0%) had an extreme LDL wGRS > 1.96. This was significantly higher than the proportion of FH mutation-positive individuals (11.9%) who had such an extreme wGRS (OR 3.02, 95% confidence interval 1.61 to 5.68, P<0.0001). Similarly, only 11.8% of 1092 GIK individuals had a wGRS this extreme. We examined the distribution of wGRS in various cohorts (Figure 3) and found no difference between GIK and Ontario FH mutation-positive individuals (mean + SD scores 1.66 + 0.27 and 1.68 + 0.23, respectively, NS). In contrast, the distribution of wGRS in Ontario FH mutation- negative individuals was markedly shifted to the right, with mean + SD score 1.89 + 0.29.

Although the absolute differences in mean scores were modest, as is typical for polygenic effects, the differences in overall distribution of scores were highly significantly between FH mutation-positive and both FH mutation-negative (P = 4.3 X 10-10) and GIK control individuals (P = 2.9 X 10-18).

We then evaluated the mean LDL cholesterol levels in mutation-negative individuals according to extreme wGRS and found no difference (Table 3). We also found no difference in mean LDL cholesterol levels in FH mutation-positive individuals according to extreme wGRS (FIG. 3).

Overall, the percentage of Ontario individuals with severe hypercholesterolemia who had an identifiable probable genetic cause with NGS was 148/313 (47.2%), which increased to 168/313 (53.7%) with MLP A results. This increased further to 210/313 (67.1%) when individuals with an extreme wGRS were included. As LDL cholesterol levels increased, so did the percentage of individuals with an identifiable genetic cause. Specifically, for individuals with LDL cholesterol 5.00 to 5.99 mmol/L (194 to 231 mg/dL), 6.00 to 6.99 mmol/L (232 to 270 mg/dL), 7.00 to 7.99 mmol/L (271 to 309 mg/dL) and > 8.00 mmol/L (> 310 mg/dL), respectively, 57.0%, 59.5%, 79.2 and 92.0% had a genetic basis for their elevated LDL cholesterol (Figure 4). Furthermore, polygenic determinants were less common among individuals with the highest LDL cholesterol levels (FIG. 4).

High throughput NGS has transformed our understanding of HeFH. Here we applied targeted NGS with custom annotation, coupled with MLPA evaluation of large-scale CNV and polygenic GRS assessment in a cohort of 313 individuals with severe hypercholesterolemia, in whom FH was the likely clinical diagnosis.

We found that: 1) monogenic FH-causing mutations detected by targeted NGS were present in 47.3% of individuals; 2) the percentage of individuals with monogenic mutations increased to 53.7% when heterozygous CNVs were included; 3) -85% of monogenic mutations were within the LDLR gene; 4) the percentage further increased to 67.1% when individuals with extreme wGRS were included; 5) the percentage of individuals with an identified genetic component increased from 57.0% to 92.0% as LDL cholesterol level increased from 5.0 to > 8.0 mmol/L (194 to > 310 mg/dL); and 6) individuals with LDLR gene mutations had higher mean LDL cholesterol levels than individuals either with APOB or PSCK9 mutations or an extreme wGRS, with no significant differences between other genetic subgroups.

Within this clinical cohort, NGS alone underestimated the number of individuals with a genetic basis for severe hypercholesterolemia. Less than half of individuals were FH mutation- positive solely based on NGS results, but this increased to more than two-thirds of individuals when CNVs and extreme polygenic wGRS were considered. We suggest that these additional genetic determinants should be considered in routine molecular assessment of patients with severe hypercholesterolemia. The actual proportion of patients with each type of genetic determinant will vary between cohorts and populations.

FIG. 3 shows low density lipoprotein (LDL) cholesterol levels (mean + standard deviations) according to monogenic variant genotype in the Ontario severe hypercholesterolemia sample. Individuals are classified according to presence of 0, 1 or 2 variants (mutations) detected by next-generation sequencing (NGS) in LDLR, APOB, or PSCK9 genes and multiplex ligation primer amplification (MLPA) in the LDLR gene. Familial hypercholesterolemia (FH) mutation- positive individuals (mut.+) are further subgrouped according to extreme weighted genetic risk score (GRS) < 1.96 and > 1.96, respectively. Comparisons of mean LDL cholesterol levels between selected genotype classes are shown, with nominal significant P < 0.05.

FIG. 4 shows percentages of individuals with severe hypercholesterolemia within different LDL cholesterol ranges (numbers of individuals shown), classified as having a monogenic variant detected by NGS and MLPA, or a polygenic basis defined as an extreme weighted GRS > 1.96 (> 90th percentile for elevated LDL cholesterol). FIG. 5 shows percentages of individuals from various samples and subgroups with an extreme weighted GRS > 1.96 (> 90th percentile for elevated LDL cholesterol), including the 1000 Genomes control cohort (G1K; http://www.1000genomes.org/) and individuals with and without a monogenic cause for familial hypercholesterolemia (FH) detected by next-generation sequencing or MLPA in the Ontario cohort. Total numbers of individuals and percentages with extreme weight GRSs are shown. FH mutation-negative with severe hypercholesterolemia had significantly higher odds of an extreme weighted GRS than mutation-positive individuals (OR 3.02, 95% confidence interval 1.61 to 5.68, PO.0001) and normal controls from the 1000 Genomes cohort.

FIG. 6 shows distribution of weighted GRSs for elevated LDL cholesterol in individuals from 1000 Genomes cohort, and Ontario individuals with severe hypercholesterolemia who are FH mutation-positive or negative. There is no significant difference in the distribution of weighted GRSs between 1000 Genomes and FH mutation-positive individuals, while the distribution of weighted GRSs in FH mutation-negative individuals differs significantly from both of these other groups (P = 2.9 X 10-18 and P = 4.3 X 10-10, respectively).

Incorporation by Reference

References and citations to other documents, such as patents, patent applications, patent publications, journals, books, papers, web contents, have been made throughout this disclosure. All such documents are hereby incorporated herein by reference in their entirety for all purposes.

Equivalents

Various modifications of the invention and many further embodiments thereof, in addition to those shown and described herein, will become apparent to those skilled in the art from the full contents of this document, including references to the scientific and patent literature cited herein. The subject matter herein contains important information, exemplification and guidance that can be adapted to the practice of this invention in its various embodiments and equivalents thereof.

Claims

What is claimed is:

1. A method for diagnosing familial hypercholesterolemia, the method comprising:

identifying in nucleic acid from a patient, a plurality of single nucleotide polymorphisms (SNPs) in one or more polygenic contributors to familial hypercholesterolemia;

determining if the patient is homozygous or heterozygous for the SNPs;

creating a genetic risk score for familial hypercholesterolemia score using a weighted, multivariate model based on the identified SNPs and the heterozygosity or homozygosity of the patient; and

diagnosing the patient with familial hypercholesterolemia where the genetic risk score in greater than a threshold value.

2. The method of claim 1, further comprising treating the patient for familial

hypercholesterolemia based on the diagnosing step.

3. The method of claim 1, wherein each identified SNP is weighted based on measured effect of the identified SNP on LDL-C in a cohort of individuals with familial hypercholesterolemia.

4. The method of claim 1, wherein the SNPs comprise two or more selected from the group consisting of: rsl2740374, rsl 1206510, rs515135, rs6544713, rsl501908, rs3846663, rs2650000, rs6511720, rsl0401969, and rs6102059.

5. The method of claim 1, wherein the SNPs comprise three or more selected from the group consisting of: rsl2740374, rsl 1206510, rs515135, rs6544713, rsl501908, rs3846663, rs2650000, rs6511720, rsl0401969, and rs6102059.

6. The method of claim 1, wherein the SNPs comprise rsl2740374, rsl 1206510, rs515135, rs6544713, rsl501908, rs3846663, rs2650000, rs6511720, rsl0401969, and rs6102059.

7. The method of claim 6, wherein the weighted multivariate model comprises .23*rsl2740374 + .09*rsl 1206510 + .16*rs515135 + .15*rs6544713 + .07*rsl501908 + .07*rs3846663 +

.07*rs2650000 + .26*rs6511720 + .05*rsl0401969 + .06*rs6102059 = genetic risk score (GRS), wherein rsl2740374, rsl 1206510, rs515135, rs6544713, rsl501908, rs3846663, rs2650000, rs6511720, rsl 0401969, and rs6102059 are assigned a value of 0, 1, or 2 based on presence of the corresponding SNPs in nucleic acid from the patient and heterozygosity or homozygosity of the patient with respect to the corresponding SNPs.

8. The method of claim 7, wherein the threshold value is about 1.96.

9. A method for diagnosing a patient with familial hypercholesterolemia, the method

comprising:

comparing a nucleic acid sequence obtained from a patient and encoding a monogenic contributor to familial hypercholesterolemia to a reference genome to identify a mutation;

assaying the monogenic contributor to familial hypercholesterolemia for copy number variation;

testing a polygenic contributor to familial hypercholesterolemia for a single nucleotide polymorphism (SNP);

determining a low-density lipoprotein cholesterol (LDL-C) level in the patient; and diagnosing familial hypercholesterolemia in the patient using a weighted multivariate model comprising:

a presence of the mutation in the monogenic contributor to familial hypercholesterolemia;

a presence of the copy number variation in the monogenic contributor to familial hypercholesterolemia;

a presence of the SNP in the polygenic contributor to familial

hypercholesterolemia; and

the LDL-C level in the patient.

10. The method of claim 9, further comprising treating the patient with a statin or a monoclonal antibody against PCSK9 based on the diagnosing step.

11. The method of claim 9, wherein the nucleic acid sequence encoding a monogenic contributor to familial hypercholesterolemia comprises one or more genes selected from the group consisting of: LDLR, APOB, PCSK9, STAP1, APOE, ABCG5, ABCG8, LDLRA1, and LIP A.

12. The method of claim 11, wherein the nucleic acid sequence encoding a monogenic contributor to familial hypercholesterolemia comprises two or more genes selected from the group consisting of: LDLR, APOB, PCSK9, STAP1, APOE, ABCG5, ABCG8, LDLRA1, and LIP A.

13. The method of claim 9, wherein the SNP comprises one or more selected from the group consisting of: rsl2740374, rsl 1206510, rs515135, rs6544713, rsl501908, rs3846663, rs2650000, rs6511720, rsl0401969, and rs6102059.

14. The method of claim 13, wherein the SNP comprises two or more selected from the group consisting of: rsl2740374, rsl 1206510, rs515135, rs6544713, rsl501908, rs3846663, rs2650000, rs6511720, rsl0401969, and rs6102059.

15. The method of claim 9, wherein the LDL-C level being greater than about 5.0 mmol/L indicates an increased likelihood of familial hypercholesterolemia in the patient.

16. The method of claim 15, wherein the LDL-C level being greater than about 8.0 mmol/L indicates an increased likelihood of familial hypercholesterolemia in the patient

17. The method of claim 9, wherein the reference genome is selected from the Human Gene Mutation Database or the University College London FH mutation database.

18. The method of claim 9, wherein the copy number variation is determined by multiplex ligation-dependent probe amplification (MLPA).