pharmacogenomics Flashcards
Applications and main focus areas of
genomic medicine ?
Precision oncology ( Sequencing is by now routine in clinical
oncology)
Genetic disease diagnostics
Pharmacogenomics(Genetic variation impacts drug response , for example ADME genes , MHC genes, Drug receptor genes )
Complex diseases
how can Variability data guide diagnostic strategies ?
if with the interogation of little variants we knwo alot about pathogenicity then we know that the rest of the variability is not that important for example
what can be the Impact of genetic variants on drug response ?
Altered pharmacokinetics
if the parent compound is active and the metabolite inactive then ultrarapid metabolisers ( incresed functionality of the enzyme) have less of an effect but is the metabolite is active and the parent compound not then we see the opposite .
Altered pharmacodynamics
the binding site of a drug might be affected positively or negatively
there is a spectrume of the posible changes in fanctionality of the protein
Altered immune response
might affect binding to HLA ( HUMAN LEUCOCYTE ANTIGEN) and activation of T cells
Preemptive genetic testing on a national level? iS IT AN OPTION ?
- First-line therapy for all patients
- Management of ADRs
VS - Universal genetic testing
- First-line therapy to negative patients
- Alternative therapy to positive patients
From population genomic data to public health recommendations, WHAT NEEDS TO BE DONE ?
Building a universal model for cost-effectiveness estimation
Preemptive genetic testing on a national level?
* First-line therapy for all patients
* Management of ADRs
VS
* Universal genetic testing
* First-line therapy to negative patients
* Alternative therapy to positive patients
For each country:
* The number of patients needed to test to prevent one ADR case (NNT)
* The number of patients would be unnecessarily denied the first-line drug (NUD)
* The maximum cost difference between first-line and alternative drug that allows preemptive genetic
testing cost-effective (incremental cost threshold, ICT)
Approaches to decode the human genetic
variation?
- PCR-based (candidate SNPs)
- MS-based (candidate SNPs)
- Chip-based (candidate SNPs and limited discovery)
- Sequencing-based (candidate SNPs and discovery)
how can we use PCR and what are some advantages
-for example in Allelic discrimination
*Gene expression analysis
* Cancer research
* Drug research
*Disease diagnosis and
management
Viral quantification
*Food testing
* GMO food
* Animal and plant breeding
* Gene copy number
Advantages of qPCR
1. High throughput: Samples & Targets
2. Efficient
3. Quantification abolute (standard curve) and relative (ΔΔCt)
4. Faster
Pharmacogenetic profiling paradigms –
chips vs sequencing ?
inexpensive, less info, requires less patients vs expensive, more info, requires more patients
Sequencing strategies ?
whole genome sequencing
Whole exome sequencing
Targeted sequencing
Sequencing region :
whole genome
Sequencing Depth:
>30X
Covers everything-
can identify all kinds
of variants including
SNPs, INDELs and SV.
Sequencing region:
whole exome
Sequencing Depth :
>50X ~ 100X
Identify all kinds of
variants including
SNPs, INDELs and SV
in coding region.
Cost effective
Sequencing region:
specific regions
(could be customized)
Sequencing Depth :
>500X
Identify all kinds of
variants including
SNPs, INDELs in
specific regions
Most Cost effective
WES has become very widely used but it has major limitations for the analysis of
common pharmacogenetic variants
Conclusions and take-home messages on sequencing
- There are various technical approaches to study pharmacogenetic variation
- Choice of method depends on throughput, “validation-vs-discovery” & costsensitivity
- Genotyping of candidate SNPs allows the probing of established
associations (however, even for those many are not implemented) - Extensive missing heritability when only common SNPs are considered
- Sequencing is generally more expensive than PCR or chip-based approaches
at the per-sample level - Sequencing allows identification of all variants in the analyzed loci,
including rare and novel variants - WES misses various common pharmacogenetic variants of relevance
-Short-read sequencing is prone to misalignments in complex loci, such as
CYP2B6 and CYP2D6, which is solved by long read NGS
Definition of Pharmacogenetics/Pharmacogenomics (PGx)
“Pharmacogenetics deals with genetically determined variation in how
individuals respond to drugs. ”
“The study of the variability in drug response because of heredity”
“Pharmacogenomics focuses on the identification of genetic variants that
influence drug effects, typically through alterations in pharmacokinetics or
pharmacodynamics.”
Study How Genetic Variations Affect Drug Response
Pharmacogenomics is an extension of pharmacogenetics, a science
described in terms of five stages of development:
Dr. Werner Kalow
1. some clinical observations predicted genetic alterations of drug response
2. additional case discoveries led to the term “pharmacogenetics”
3. many systemic case studies, and the realization of its wide applicability
4. the recognition of systematic pharmacogenetic differences between human populations
5. most human drug-response differences were multifactorial, caused by many genetic
alterations plus environmental factors
Why Pharmacogenomics?
One drug does not fit all
* Only 50-75% of patients respond adequately to
medications
* Abnormal response accounts for around 6% of all
hospitalisations
* Adverse drug reactions (ADRs) cost up to $30 billion/year
in US alone
* Around 32% Drugs were affected by a post-market safety
event
PGx explains 20-30% of abnormal drug response
By studying pharmacogenomics, we would like to…
- Understand genetically induced inter-individual variability in drug response
- Improve drug efficacy and minimize adverse drug reactions
- Assist drug development process by
- Guiding the design of candidate drugs for the least variable drug response
- Defining patient groups that can benefit from specific drugs (cancer drugs)
- Identifying alternative drugs for patients experiencing advese drug reactions when
taking first-line treatments - Identifying novel drug targets (PCSK9)
Pharmacogenes ?
- Pharmacokinetic (PK) genes – genes related to drug absorption, distribution,
metabolism and excretion (ADME)
(what body does to the drug) - Pharmacodynamic (PD) genes – genes encode drug targets
(what drug does to the body) - Other genes related to off-target effects
Pharmacokinetic (PK) genes?
- Phase I
E.g., Cytochrome P450 (CYP)
enzymes, others (DPYD, CES1 etc.) - Phase II
E.g., UDP-glucuronosyltransferase
(UGT) family, glutathione Stransferase
(GST), sulfotransferase
(SULT) family - Transporters
E.g., Solute carrier (SLC) family, ATPbinding
cassette (ABC) family
Clinically important PK genes
* CYP gene family: involved in >90% of drug metabolism, highly relevant to adverse drug
reactions
* 57 CYP genes in the human genome, 8 of the encoded enzymes cover the majority of
drug metabolism
conservation of CYP genes ?
EXAMPLE
Duplicated CYP2D6 can increase
capability for detoxification of
alkaloids in plants
Selection of CYP2D6 duplication
carriers in Africa
Different types of genetic variations ?
Single Nucleotide Polymorphism (SNPs)/ Single Nucleotide variations (SNVs)
- A>T, C>G
* Small insertion/deletion (indel)
- A>AT, C>CCTTTTT, AT>A, ATATAT>A
* Structural variations (SVs)
- genomic deletions, duplications, insertions, inversions and other complex
rearrangements that affect >50bp
* Based on genomic location: exonic/intronic variants, regulatory variants,
intergenic variants
* Based on consequence: coding/non-coding variants; missense variants,
synonymous variants, splicing variants
what is Important to know about genetic variations ?
- Reference genome/assembly (DNA sequence assembled from a number of
individual donors) - GRCh38/hg38, released by Genome Reference Consortium in 2017
- GRCh37/hg19, released in 2009
- RS number/RSID (Reference SNP, a unique label to identify a specific SNP)
- rsxxxxxx
Allele frequency
how to calculate?
F1 = 6/10 = 0.6 (60%)
F2 = 4/10 = 0.4 (40%)
1/1
1/2
2/2
1/1
1/2
F1 + F2 = 1
or
1/1
2/3
2/2
3/3
1/2
F1 = 3/10 = 0.3 (30%)
F2 = 4/10 = 0.4 (40%)
F*3 = 3/10 = 0.3 (30%)
F1 + F2 + F*3=1
Will the frequency of CYP2D6*4,
one of the most well-studied
CYP2D6 allele, differ between
Europeans and Asians?
Yes. Europeans (20%) and Asians
(2%).
* Genetic isolation
Will the frequency of CYP2D6*4,
one of the most well-studied
CYP2D6 allele, differ between East
Asians and South Asians?
Yes. East Asians (0.3%) and South
Asians (10.4%).
* Genetic isolation?
Will the frequency of CYP2D6*4,
one of the most well-studied
CYP2D6 allele, differ between
Swedish and Bulgarian?
No . Swedish (19.2%) and Bulgarian
(19.3%).
* Same ethnicity
Will the frequency of CYP2D6*4,
one of the most well-studied
CYP2D6 allele, differ between
Swedish and Finnish?
Yes. Swedish (19.2%) and Finnish
(10%).
* Genetic isolation
For some countries, calculating the
national frequency could be
complicated. What are these
countries?
The United States, Brazil,
Singapore
* Population admixture
Which ethnicity has the most
complex genetic composition (high
numbers and high frequency of
genetic variants)?
- African
- Human origin?
Which ethnicity/population has the
highest numbers of novel genetic
variants that are not observed in
other ethnicities/populations ?
- Ashkenazi Jews
- Culture isolation, population
bottleneck (~ 350 individuals
what do you know about CYP genetic variants ?
- Pharmacokinetic (PK) genes, particularly CYPs, are polymorphic and harbor
many genetic variants.
Around 8 CYP genes are clinically most important with CYP2D6 being most
polymorphic.
Star (*) alleles are used to name well-characterized (mostly common) CYP
variants.
Allele frequency of CYP alleles differ substantially across ethnicities and
populations.
Why are CYP star alleles important?
- Common
- Have impact on activity of encoding enzymes
- Decrease enzyme function (deceased function allele)
- Abolish enzyme function (Loss-of-function allele)
-Increase enzyme function (Gain-of-function allele)
enzyme Function alleles and drug exposure, what is it dependent on ?
it is dependent on where the enzyme acts, the produg to metabilise it into the active compound or to the drug deactivating it
general dogma though is :
Changes in allele function - Changes in enzyme activity - Changes in drug
exposure
PharmVar database
-Data source for the categorical function of star alleles
Star allele functions are well-characterized and can be looked up in PharmVar
PharmVar, short for Pharmacogene Variation Consortium, is a global initiative focused on cataloging and standardizing genetic variations within genes that encode proteins involved in drug metabolism and response. These variations can influence an individual’s response to pharmaceutical drugs, affecting factors such as drug efficacy and potential adverse reactions.
The PharmVar database provides curated information on pharmacogenetic variations, including their frequencies in different populations, clinical significance, and recommendations for drug dosing or selection based on an individual’s genetic profile. This information is valuable for healthcare professionals in personalized medicine, allowing them to tailor drug treatments to individual patients based on their genetic makeup, ultimately leading to better therapeutic outcomes and reduced risks of adverse drug reactions.
what is The Clinical Pharmacogenetics Implementation Consortium
(CPIC) ?
CPIC (and others) publishes guidelines to translate genotypes into metabolizer
phenotypes and provide therapeutic recommendations.
*A group of international pharmacogeneticists who are
interested in facilitating use of pharmacogenetic tests
for patient care.
* Curates pharmacogenetic data from the literature.
- Allele definition
-Allele function
- Allele frequency
-Evidence linking Genotype to Phenotype
* Publishes detailed gene/drug clinical practice
guidelines
- From variant function to allele function to protein
function to patient response to dose/drug
recommendation
Allele function and activity score,
Translate an individual’s genotype into phenotype,
From metabolizer phenotype to dose recommendations
what are the possible activity scores for a metabolier of one drug ?
0-2 but it can be more than 2 for ultrarapid metabolisers
Ultrarapid metabolizers (UM),Rapid metabolizers (RM) =Duplicated allele or increase function allele
Normal metabolizers (NM)
Intermediate metabolizers (IM)= Decreased function allele or LOF allele
Poor metabolizers (PM)=Two LOF alleles
Other resources for genotype-phenotype translation
- The Dutch Pharmacogenetics Working Group (DPWG)
- The Canadian Pharmacogenomics Network for Drug Safety (CPNDS)
- The French National Network of Pharmacogenetics (RNPGx)
- Drug labels
- FDA (Table of Pharmacogenetic Associations; Table of Pharmacogenomic
Biomarkers in Drug Labeling) - EMA
- Pharmaceuticals and Medical Devices Agency, Japan (PMDA)
- Swiss Agency of Therapeutic Products (Swissmedic)
- Health Canada (Santé Canada) (HCSC)
FDA_Table of Pharmacogenetic Associations ?
Section 1: Pharmacogenetic Associations for
which the Data Support Therapeutic Management
Recommendations
e.g abacavir ( maybe the only one )
Section 2: Pharmacogenetic Associations for
which the Data Indicate a Potential Impact on
Safety or Response
e.g codein
Section 3: Pharmacogenetic Associations for
which the Data Demonstrate a Potential Impact on
Pharmacokinetic Properties Only
e.g. amitriptyine
make a Comparison between drug labels and guideline
recommendations
Why discrepent?
one might say that knowing the PGx an action should be taken while the other might only be recomending or not saying anything
CPIC and DPWG
- >20% differences in therapeutic recommendations were identified (Bank PCD. et al.
Clin Pharmacol Ther. 2018)
* DPWG, CPIC, CPNDS, and RNPGx
- Discordances in fluoropyrimidines, irinotecan, clopidogrel, warfarin and statins
(Abdullah-Koolmees H. et al. Front Pharmacol. 2021)
* FDA and EMA
- Only 54% overlap in drug-gene interactions (Ingelman-Sundberg M. Pharmacol Res.
2020)
* FDA, EMA, CPIC and DPWG
- Only 18% concordance Source of information considered
- Clinical study/Preclinical study/Case report
* Time point when the evidence was collected
* Therapeutic recommendations vary between countries
* Update of the gene-drug information
- English version of DPWG recommendations hasn’t been updated since 2011
* Activity score assignment
PharmGKB what is it ?
PharmGKB is a comprehensive resource for pharmacogenomic information.
- PharmGKB (The Pharmacogenomics Knowledgebase) collects, curates and
disseminates knowledge about clinically actionable gene-drug associations and
genotype-phenotype relationships.
it give you info form all databases , as it may vary
it can also show you the pathways
you can Browse Drugs with Variant Annotations
Well-established pharmacognomic biomarkers can be related to …
Associated with Adverse Drug Reactions
Modulate Drug Efficacy
Summarize the allele frequency-related databases
- Allele frequency
- gnomAD
- dbSNP
- The literature
- SNP information (genomic coordinate, variant type, nucleotide substitution)
- dbSNP
User
gnomAD ns dbSNP
In summary, while both gnomAD and dbSNP contain information about genetic variants in humans, gnomAD focuses on allele frequency data from large-scale sequencing studies, particularly rare variants, while dbSNP is a comprehensive database that catalogs various types of genetic variants and provides unique identifiers for each variant.
how do I go From allele frequency to diplotype frequency ?
- Allele: *1, *2, *3…
- Diplotype: 1/1, 1/2, 2/3…
- Assume in a cohort people are either wild-type or CYP2C19*2 carriers
- Allele frequency: CYP2C19*2 = 0.2
- Diplotype frequency
CYP2C192/2 = ?
CYP2C191/1 = ?
CYP2C191/2 = ?
for hoozygous we just do: 0,2^2 for example but for heterozygus we have to do: 2* 0,2*0,8
how do I go From allele frequency to metabolizer phenotype frequency?
If the numbers dont add up to 1 then the *1 is missing
add all the same variant functions together
multiply and add together all the possible allele combinations that give you every functional group
where does the Substantial pharmacogenetic variability come from ?
- Pharmacogenes are polymorphic
- Star alleles
- Inter-population differences in star allele frequencies
- Consequent metabolizer phenotypes
Approaches identifying genotype-phenotype relationship
Forward genetics Phenotype to Genotype
Reverse genetics the opposite
First generation sequencing (Sanger sequencing) details and uses ?
- Read length: 400 to 1200 bp
- High accuracy
- Easy to perform in a common genetic laboratory
- Low-throughput
- Time-consuming
- Trained staff required
- Still used today in small scale project or validate result from more advance
sequencing technologies
Second generation sequencing (Massive parallel sequencing,
next-generation sequencing, short-read sequencing)
- Defined by methods in which millions of DNA templates are sequenced
spontaneously in a single reaction - Emerged around 1995 and commercially available since 2005
- Principle (Video)
-DNA fragmentation
-Amplification of DNA fragment by PCR - Short-read sequencing
- Short reads assembly
- Read length: 50 to 600 bp
- High-throughput, cost-effective
- Automation
- Can be very accurate
- Difficult to analyze complex genomic loci (repetitive regions, structure variants)
- Widely-used in industry and academia
Third generation sequencing (long-read sequencing)
- Defined by methods that can produce substantially longer reads than second
generation sequencing - Developed after 2011
- Two long-read sequencing platforms
- Nanopore sequencing by Oxford Nanopore Technology
- Single molecule real time (SMRT) sequencing by Pacific Biosciences (PacBio)
- Read length: 5,000 to 30,000 bp
- No amplification bias from PCR
- Suitable to analyze complex genomic loci
- Expensive
- Much less downstream bioinformatic tools coupled
- Getting more used in detecting genetic variations in disease genes and complex
pharmacogenes, e.g. CYP2D6
Sequencing a human genome cost
?
Before: 3 billion USD, 13 years
Now: 500 USD, half a day
Sequencing projects across the globe,Sequencing projects across the globe
what is true about those
multiple around the world , in many different palces
Sequencing reveals large number of rare genetic variants , what do we know on this ?
- 2,504 human genomes (1KGP, 2015)
88M variants, 76M variants with frequency < 5%, 64M variants with frequency < 0.5% - 76,156 human genomes (gnomAD v3, 2023)
> 644M short nuclear variants, >390M variants with frequency < 0.1% - Pharmacogenomic variations
rare variants are also important
CYP2D6 has the most variants
The majority of sequencing identified variants have low frequencies. gnomAD is
a good resource to analyze these variants.
From genotype to phenotype
* Are star alleles sufficient to inform metabolizer phenotype?
No cause we see that the phenotypic spectrum is overlapping between different diplotypes
* Are star alleles sufficient to inform metabolizer phenotype?
* No. Something is missing in genetic factors that can explain the variability in
drug response (missing heritability).
* What is missing?
20-30%
genetic factors
~50% known variants
~50% missing
heritability :
rare variants
log read sequencing of complex loci
haplotypes of specific CYP genes
Two variants in the same allele – we call that haplotype
specific polimorphic gene regulation
SIFT (Sorts Intolerant From Tolerant substitutions) how does it work ?
- Searches for similar amino acid sequences
- Chooses closely related sequences that may share similar function
- Obtains the multiple alignment of these chosen sequences
- Calculates normalized probabilities for all possible substitutions at each position
from the alignment
Normalized probabilities < cutoff : Deleterious
Normalized probabilities >= cutoff : Tolerated
Here’s how SIFT works:
Sequence Alignment: SIFT starts by aligning protein sequences from different species that are evolutionarily related to the protein of interest. This alignment helps identify conserved regions and residues within the protein.
Calculation of Conservation Scores: SIFT calculates a conservation score for each amino acid residue in the protein based on the degree of conservation observed in the sequence alignment. Highly conserved residues are assigned higher conservation scores, indicating that changes to these residues are less likely to be tolerated without affecting protein function.
Substitution Prediction: When presented with a query amino acid substitution (mutation), SIFT evaluates the conservation score of the wild-type (original) amino acid residue and predicts the impact of the substitution based on this score. If the substitution involves a highly conserved residue (high conservation score), SIFT predicts that it is likely to be deleterious or damaging to protein function. Conversely, if the substitution involves a less conserved residue (low conservation score), SIFT predicts that it is more likely to be tolerated without significant effects on protein function.
Threshold Determination: SIFT uses a predetermined threshold (typically 0.05 or 0.1) to classify substitutions as tolerated or intolerant. Substitutions with a SIFT score below the threshold are predicted to be deleterious (intolerant), while substitutions with a score above the threshold are predicted to be tolerated (tolerant).
Output: The output of SIFT typically includes the predicted impact of the substitution (tolerated or deleterious) along with a numerical SIFT score indicating the degree of conservation at the affected residue.
what is Polyphen-2 (Polymorphism Phenotyping V2)?
A method and server for predicting damaging missense mutations
Based on a machine learning method (Naive Bayes
classifier)
Two training datasets:
* HumDiv, 3,155 damaging variants, 6,321 nondamaging
variants
* HumVar, 13,032 damaging variants, 8,946 nondamaging
variants
what is CADD (Combined Annotation–Dependent Depletion)?
A general framework for estimating the relative
pathogenicity of human genetic variants
- Machine-learning method to train 14.7 million highfrequency
human-derived variants and 14.7 million
simulated variants - Integrating many diverse annotations into a single
score for each variant (higher score à more
deleterious)
what is ANNOVAR?
Generate missense variant predictions from a collection of
tools
VEP
(Ensembl Variant Effect Predictor)
Generate predictions from a collection of tools
Are General variant predictors suitable for assessing functions
of pharmacogenomic variants ?
NO
Sequence conservation
Allele frequency
Disease-related features
are not enough
to find
Disease causing
variants
around 75 % prediction accuracy for CADD
Computational tools to predict functional consequences of
pharmacogenomic variants
better results for ADME optimised
- DPYD gene
- DPYD-Varifier (variant function
prediction) - CYP2D6 gene
-Enzyme activity prediction based
on sequences (long-read) - Hubble.2D6 (haplotype function
prediction) - Few other tools
Challenges for computational prediction
- Overfitting: overlap between training variants and testing variants
- Lack of training data (for predictors designed for specific type of variants, e.g.,
PGx variants) - Method selection
Genotype to phenotype using computational prediction ?
through Aggregated
frequency and then calculating and adding all the probabilities of every diffferent diplotype that is in this cagegory e.g. intermidiate metaboliser
what are the Pharmacodynamic (PD) genes ?
- PD genes encode drug targets that can trigger therapeutic response upon drug
binding - The number of drug targets are not well-defined.
Example of clinically important variations in PD genes:
- Vitamin K epoxide reductase
(VKOR, encoded by VKORC1)
is the target of warfarin - VKORC1 -1639G>A (rs9923231, G3673A)
- Variant carriers have decreased activity of VKOR, thus are more
sensitive to warfarin and require lower dose - MAF = 32.6%, dominant in East Asians
CFTR
* CFTR (cystic fibrosis transmembrane
conductance regulator) encodes a chloride
channel and play important role in ion and water
secretion and absorption in epithelial tissues.
- Abnormal CFTR function is the cause of
cystic fibrosis (CF), a genetic disorder
that induce sticky and thick mucus
secretion and cause difficulty in
breathing and frequent inflammation. - CF occurs in 1 in 2,500 to 3,500 white
newborns - Five classes of CFTR
mutation and
corresponding therapies
different functionalities of the protein
e.g. missfolded , shortened etc
Variants in PD genes can guide drug development
* The PCSK9 case , what happened ?
Two French families with an autosomal
dominant form of familial
hypercholesterolemia (FH)
“gain-of-function”
mutations identified in PCSK9 in
these and other FH patients
Development of PCSK9 inhibitor
what else can genetic variants affect ?
1)Genetic variants in drug-target binding sites
* GPCRs
* 108 GPCRs targeted by 475 (34%)
FDA-approved drugs
- Natural occurring variants in GPCR binding site impact drug response and bias
signaling
1 in 6 individuals carries at least one missense variant in the binding sites
- Binding site variants can impact drug response and be used for drug
development
2) Drug off-target effect: HLA genes related hypersensitivity
reactions
- HLA (Human Leukocyte Antigen)
encodes major histocompatibility
complex (MHC) - MHC molecules have peptide
binding grooves
binding sites affected:
HLA-B*57:01 and abacavir induced abacavir hypersensitivity
syndrome
HLA-B*15:02 and carbamazepine induced Stevens-Johnson
syndrome/toxic epidermal necrolysis (SJS/TEN)
- Genetic variations in drug targets are important determinants of drug response
- Genetic variations in HLA genes can trigger drug-induced hypersensitivity
reaction - HLA-B*57:01 is a strong risk allele for abacavir hypersensitivity syndrome
How much impact does pharmacogenomic
testing currently have in clinical practice?
Overall implementation of pharmacogenomic
labels however remains poor particularly for
CYP genes
In total there are 25 guidelines with actionable recommendation affecting 140
medications.
Of these, only 2-3 are implemented in clinical routine
These are testing of:
HLA-B57:01 and abacavir
DPYD and fluoropyrimidines
HLA-B15:02 and carbamazepine in patients of Asian ethnicity
Abacavir hypersensitivity
Manifests as
Morbilliform/maculopapular rash
Fever (often precedes rash)
Myalgias, fatigue
Mucosal ulceration
More rarely (<5%) also as hepatitis and nephritis
Fluoropyrimidine toxicity
Up to 30% of patients have severe treatment-related toxicity, including diarrhea, mucositis,
myelosuppression and hand foot syndrome.
Anti-cancer drugs, such as fluorouracil and capecitabine
Prospective trial showed that genotype-guided dose reduction was beneficial
HLA-B*15:02 and carbamazepine-induced
Steven-Johnson syndrome (SJS)
SJS when <30% of skin surface area is affected
Toxic epidermal necrolysis (TEN) when >30% are affected
Very severe skin reaction
Zhou et al., CPT 2020
Mortality rates: 2-15% for SJS and 25% for TEN
Why is implementation of pharmacogenomic
biomarkers not more widespread?
Overall implementation of pharmacogenomic
labels however remains poor particularly for
CYP genes
Contributing factors:
* Lack of reproducibility (underpowered studies dilute the literature)
* Large variability within genotype groups; effects are stochastic rather than deterministic
Many outliers in the studies make the results less deterministic
- TDM instead of PGx
- Strong effects on pharmacokinetics, weak effects on clinical outcomes/ADRs
- Unclear added value (lack of clinical trials, alternative medications; e.g. DOACs)
Pharmacogenetic expert groups are dominated by pharmacists and pharmacogeneticists
but lack involvement of physicians
Poor congruence between “expert”
recommendations and regulatory guidance
(SmPCs)
- (Lack of pharmacogenetic education)
- Practical limitations in clinical routine
- Lack of cost and reimbursement considerations
UPGx – PREPARE trial
Multi-center, multi-gene, multidrug,
multi-ethnic, and multihealthcare
system trial.
The PREPARE trial involves recruiting patients from primary care practices and providing pharmacogenomic testing to guide medication selection and dosing for selected drugs.
The PREPARE trial involves recruiting patients from primary care practices and providing pharmacogenomic testing to guide medication selection and dosing for selected drugs.
Overall implementation of pharmacogenomic
labels however remains poor particularly for
CYP genes , possible mitigation ?
Incorporation of rare variability to increase the predictive power
- Clinical trials to evaluate the added value
- Fostering of inter-disciplinarity
- Increased cost consideration and feasibility trials in primary care
For CYPs, the most promising PHARMACOGENETIC associations are:
- CYP2C19 for clopidogrel and escitalopram
- CYP2D6 for risperidone and tamoxifen
- CYP2C9 for warfarin
