The Impact of Micrornas and Alternative Splicing in Pharmacogenomics

Home , Clearance (pharmacology), Drug resistance, NOVA2

cogenetic history see2). Since then, the The impact of microRNAs and number of cases described in the literature has increased exponentially alternative splicing in and they were mainly focused on single-nucleotide polymorphism pharmacogenomics (SNP), insertions and deletions (indels) of nucleotides, copy number variation 1 2 3 (CNV) and missense mutations (for F Passetti , CG Ferreira and FF Costa review see Roses3 and Giacomini et al.4). 1 Laboratory of Bioinformatics and Computational Biology, Division of Clinical and The most studied gene family so far Translational Research, Research Coordination (CPQ), Instituto Nacional de Caˆncer, Rio is the cytochrome P450 enzyme5–7 (for de Janeiro, Brazil; 2Division of Clinical and Translational Research, Research Coordination more details see Table 2). The genetic (CPQ), Instituto Nacional de Caˆncer, Rio de Janeiro, Brazil and 3Cancer Biology and variability observed in this gene family Epigenomics Program, Children’s Memorial Research Center, Northwestern University’s has been associated with differences in Feinberg School of Medicine, Chicago, IL, USA drug metabolism in several pathologies, such as cancer, cardiovascular diseases and rheumatoid arthritis.5–8 The Pharmacogenomics Journal (2009) 9, more than half of the human genes, There is one SNP in CYP2C9 that 1–13; doi:10.1038/tpj.2008.14 thereby changing the sequence of key causes an amino acid change that proteins related to drug resistance, modify the ability of the enzyme to activation and metabolism. Further- metabolize the anticoagulant warfar- Keywords: polymorphisms; miRNAs; alter- more, alternative splicing and miRNAs in.9 This genetic variation has impor- native splicing; drug metabolism; drug can work together to differentially tant implications in the dose a patient resistance; complex diseases and therapy control genes. This perspective article will receive of warfarin depending on will highlight recent exciting discov- the genetic background of the indivi- eries in pharmacogenomics and also dual. Another gene that metabolizes Pharmacogenomic studies emphasize discuss how players such as miRNAs warfarin is the vitamin K epoxide the use of genomic information to and alternative splicing may affect reductase complex 1 (VKORC1). The enhance success in finding new med- the way we design and apply future allele variations in VKORC1 strongly icines and also to improve those that therapies. alter warfarin dosage when compared are already used in clinics. Therefore, with CYP2C9 genetic variant: almost 9,10 this field has a special interest in Introduction three times. knowing how patients metabolize The main goal of the emerging drugs depending on their genetic The improvement of patient treatment discipline of pharmacogenomics is to background. Most of the studies so is always a concern for modern med- use personalized therapy based on an far have focused on the impact of icine. One aspect that can influence individual’s genotype. This term has single genetic differences on drug the effectiveness of therapies in pa- evolved from the older pharmacoge- metabolism. However, this may be tients during drug-based treatment is netics, which was focused on studying only the tip of the iceberg in terms of the genetic background. One of the inherited differences in a single gene how interpatient variability can influ- major challenges for proper drug de- responsible for drug metabolism and ence the response to drugs. For exam- velopment by companies is the differ- response. The emergence of the term ple, control of gene expression by ence in response between individuals pharmacogenomics was possible after microRNAs (miRNAs) and alternative and between populations. The idea the availability of the human haplo- 11 splicing are cellular mechanisms that that genetic variability between pa- type map (HapMap) and of high- have an effect on proteome diversity tients might influence response to throughput genotyping platforms that and have already been implicated in drugs was described and termed ‘phar- have been facilitating more systematic complex diseases such as cancer, ar- macogenetics’ by Vogel.1 Pharmacoge- genetic screens for new and clinically thritis and others. Changes in the netics became a field of importance for important drug targets.12 As already sequence of a miRNA and/or varia- scientists and physicians in the early discussed by Bertino et al.,13 our tions in the miRNA target region of a 1950s, when available technologies understanding of an individual’s ge- transcript can have a major impact on were able to measure enzyme activity, netic background will provide knowl- post-transcriptional regulation. Events drug metabolites and drug response edge to predict the host response to of alternative splicing can occur in (for a timeline review of the pharma- specific drugs. microRNAs and alternative splicing in phamacogenomics F Passetti et al 2

New players in pharmacogenomics alternative splicing of pre-mRNAs. Dur- several questions still to be answered ing the maturation of an mRNA, exons in the biology of miRNAs (for review see The elucidation of the sequence of the can be spliced out, intronic sequences Eulalio et al.32). miRNAs are generally human genome in 200114,15 and the can be retained and cryptic splice sites 18–25 nt long and were first described subsequent release of other versions are can be used to form more than one in the early 1990s in the worm Caenor- allowing a better understanding of mRNA from a single gene. The rate of habditis elegans as regulators of develop- structural variations and how they can alternative splicing of mRNAs has been ment and differentiation.33 It is affect diseases.16 Recently, the identifi- the focus of different studies and it estimated that the human genome has cation and analysis of functional ele- seems that more than 60% of the thousands of miRNA genes, but only ments in 1% of the human genome by human genes produce at least one B700 have been described so far. This the ENCODE Project represented a alternative mRNA.25 The functional class of non-coding genes is predicted major step towards a more comprehen- implication of alternative splicing in to regulate at least 30% of all the sive characterization of all functional normal and pathological states has human protein-coding genes by target- elements in the human genome. The been studied by several groups,26,27 ing their 30-UTR sequences.34 There is ENCODE project — standing for EN- but there are still many questions to also evidence that miRNAs can regulate Cyclopedia Of DNA Elements — has set be answered. Future studies will be the expression of large ncRNAs, indicat- out to identify all the functional ele- focused on the functional relevance of ing that these small genes have a big ments in the human genome.17 It is splice variants in the context of whole impact in transcriptome networks in becoming clear that the definition of genome studies, instead of a single eukaryotic cells.22 ‘gene’ is changing and noncoding tran- gene to understand how they will affect Several groups studying the biology scripts (also referred as to noncoding cellular networks and pathways. of miRNAs have been able to describe RNAs) are an important component of Recently, reports have shown that each step in the processing and me- the information that is being tran- changes in the sequence of miRNAs chanism of action of these regulators. scribed by eukaryotic cells (for review and/or variations of the miRNA target It is already known that they are see18–20). Noncoding RNAs (ncRNAs) region within the transcripts can have initially transcribed as pri-miRNAs are transcripts that do not have pro- major effects on post-transcriptional which can be processed into pre- tein-coding potential but might be still regulation.28,29 More importantly, ge- miRNAs and subsequently into mature functional.18 ncRNAs are a large group netic variations such as SNPs can affect miRNAs. Mature miRNAs have the of transcripts that can differ in size the way miRNAs regulate their targets, ability to affect the translational effi- which is indicative of their mechanism indicating that this could be impor- ciency of various protein-coding genes of action.18 These ncRNAs can vary in tant in drug metabolism and in phe- at the same time. In the past 5 years, size range from B18 to 25 nucleotides notypic variation.30 In the same way there have been several reports impli- for the families of microRNAs (miRNAs) that miRNA can regulate protein trans- cating miRNAs in posttranscriptional and small interfering RNAs (siRNAs), lation, alternative splicing can have regulation of proteins with diverse B20 to 300 nucleotides for small RNAs several implications which can affect roles (for review see Filipowicz et commonly found as transcriptional and the biological activity of proteins (for al.35). Evidence based on computa- translational regulators or up to and example enzymes, antagonist proteins, tional studies has already revealed that beyond 10 000 nucleotides in length for and so on). For example, splice variants there is a ‘seed’ region of B8 nt at the RNAs involved in various other pro- of the human BCL2L1 gene (also known 50-end of miRNAs that is essential to cesses (for review see18–20). miRNAs can as BCL-X) switches an antiapoptotic miRNA function.36 This region is im- block mRNA translation and affect both protein to a proapoptotic one.31 Thus, portant for binding to the mRNA targets the expression of protein-coding alternative splicing differences can have and for mRNA target degradation. One genes21 and long ncRNAs.22 A growing a major impact on drug metabolism and of the concepts in the biology of number of reports have been showing therefore resistance. miRNAs that is of particular therapeutic that miRNAs are master regulators of This article is aimed at describing relevance is that one miRNA can down- important gene networks in eukaryotic new discoveries in the field of phar- regulate multiple target proteins by cells.23 Associations of deregulated ex- macogenomics as well as discuss how interacting with different target mRNAs pression of miRNAs in complex diseases players such as miRNAs and alterna- (‘one hit multiple targets’ concept).37 have been also described (for review see tive splicing may affect the way we The hypothesis of using the ‘one-hit- Gartel and Kandel24). miRNAs have design and apply future therapies into multiple targets’ concept to treat dis- been considered as attractive drug tar- the clinic. eases was previously discussed by Wur- gets in complex diseases such as cancer dinger and Costa.37 as they may be differentially expressed microRNAs in malignant cells compared to normal microRNAs in complex diseases cells altering the regulation of expres- Biology of microRNAs Recently, miRNA-deregulated expression of many important genes. miRNAs are part of a group of ncRNAs sion has been extensively described in Another molecular mechanism that that can block mRNA translation and a variety of diseases, including cancer. produces gene expression diversity is affect mRNA stability but there are Deregulation of miRNAs in other

The Pharmacogenomics Journal microRNAs and alternative splicing in phamacogenomics F Passetti et al 3

diseases such as obesity,38 diabetes39 affecting health, disease and responses Mishra et al.64 were able to show that a and schizophrenia40 has also been to drugs and environmental factors. It SNP located near the miR-24 binding described in the literature and the list is becoming clear that understanding site in the dihydrofolate reductase of examples is growing fast. Some lines multigenic diseases will require com- (DHFR) gene 30UTR results in a loss- of evidence have already shown that plex association studies to evaluate of-function mutation. Furthermore, up or downregulation of miRNAs patient risk to specific pathologies. As the SNP affects DHFR expression by correlate with numerous human can- drug metabolism can involve a group interfering with miR-24 post-transla- cers indicating that miRNAs can func- of genes, we believe that the determi- tional regulation, resulting in DHFR tion as oncogenes and/or tumor nation of how patients react to and over-expression and methotrexate re- suppressors (for review see Garzonet metabolize drugs might be focused on sistance in cancer cells.64 MicroRNA et al.41). In cancer, miRNAs are asso- master regulators of gene expression, miR-24 was already implicated in the ciated with the tumorigenesis pro- such as miRNAs. In that regard, reports regulation of important proteins such cess42 and with important cancer have recently shown that changes in as the p16 tumor suppressor gene,65 gene networks such as the p53 path- the sequence of a miRNA and/or ALK4 gene that is involved in erytro- way.43 In addition, they have also been variations in the target region of a poiesis66 and TGFbeta gene that implicated in the metastatic process in transcript that is regulated by a miRNA has a function in skeletal muscle breast cancer.44,45 miRNA expression can have major effects in posttran- differentiation.67 profiles were also evaluated by several scriptional regulation of proteins.28 Sethupathy et al.61 have shown that groups46,47 and they have been used More importantly, variations of se- SNPs in miR-155 target sites located in for prognosis and early diagnosis quence such as SNPs can affect the the 30UTR of the human angiotensin showing how important these players way miRNAs regulate their targets, type 1 receptor (AGTR1) gene down- are in different aspects of complex pointing to a function in drug meta- regulates the expression of the allele diseases. More recently, some reports bolism and in phenotypic variation.30 that has been associated with hyper- have been showing that miRNA de- For example, two groups have evalu- tension. In another study, Tan et al.62 regulated expression in diseases can be ated the presence of SNPs located in were able to identify a SNP in the due to epigenetic changes such as DNA miRNA-binding sites of the 30UTR of 30UTR of HLA-G that influences the methylation and histone modifica- several genes and SNPs in the micro- targeting of three miRNAs (miR-148a, tions.48,49 RNA seed regions by genome-wide miR-148b and miR-152) to this gene. analyses.54,55 Yu et al.54 were able to The authors suggest that allele-specific identify twelve miRNA binding SNPs targeting of these miRNAs can account microRNA variations and the that display an aberrant allele fre- at least in part for the observations impact in pharmacogenomics quency in human cancers. Moreover, that HLA-G is associated with asth- Saunders et al.55 were able to identify ma.62 In cardiovascular disease, Martin SNPs are frequent variations in the approximately 250 SNPs that create et al.63 reported an association of the human genome and estimates suggest novel target sites for miRNAs in hu- human angiotensin II type 1 receptor that they can have a frequency of one mans and may result in phenotypic polymorphism and miR-155. Finally, in every thousand base pairs.11 SNPs differences. Similar studies have eval- Hansen et al.40 identified important have been extensively studied in order uated genetic variants in miRNAs and associations between brain-expressed to understand the susceptibility to the risk of cancer. Wu et al.56 studied miRNAs and schizophrenia for two specific diseases in the general popula- 100 human tumor tissues and 20 SNPs located in mir-206 and mir-198 tion. Initially, SNP association studies cancer cell lines and have identified a sequences. Noteworthy is the fact that were focused in protein-coding genes. mutation in miRNA let-7e that causes these miRNAs have a surprisingly large Several polymorphisms in coding re- a significant reduction of its expres- number of targets in common, eight of gions, or the so-called nonsynon- sion in vivo contributing to tumorigen- which are connected by the same ymous SNPs of genes associated to esis. A similar study by Yang et al.57 has transcription factors.40 There is also a complex diseases such as cancer, were genotyped several SNPs from miRNA study showing that variations within previously identified.50 Recently, there genes in 746 matched normal and the mir-433 target site of the gene has been an explosion of whole-gen- bladder cancer tissues and discovered FGF20 increases the risk of Parkinson’s ome association studies aiming at that SNPs in GEMIN3 gene can affect disease by increasing the levels of the identifying association markers that miRNA binding, thereby significantly protein a-synuclein.28 Finally, poly- can predict diseases based on the increasing bladder cancer risk. morphisms in the precursor microRNA HapMap.51–53 One of the aims of Polymorphisms in miRNA target can affect the biogenesis and proces- the HapMap Project is to generate a sites of protein-coding genes that are sing which will then result in altered haplotype map of the human genome, associated to cancer,58–60 hyperten- control of gene expression.68 All the which describes the common patterns sion,61 asthma,62 cardiovascular dis- examples described here are listed in of human genetic variation. HapMap ease63 and polymorphisms in Table 1. is expected to be a key resource for microRNAs that are associated with The identification of all the genetic researchers to find genetic variants schizophrenia40 were also described. and epigenetic differences that are the

The Pharmacogenomics Journal microRNAs and alternative splicing in phamacogenomics F Passetti et al 4

Table 1 Examples of polymorphisms that can attenuate the binding of miRNAs to their targets in several diseases

Polymorphism miRNA (s) Target (s) Disease References

DHFR gene 30UTR (829C4T) miR-24 DHFR Cancer drug resistance 64 AGTR1 30UTR (1166C) miR-155 AGTR1 Hypertension 61 HLA-G 30UTR (3142 C4G) miR-148a, miR-148b and miR-152 HLA-G Asthma 62 Angiotensin II type 1 receptor miR-155 Angiotensin type Cardiovascular disease 63 30-UTR (1166 A4C) 1 receptor rs17578796 and rs1700 miR-206 and miR-198 JUN, ATF2 and TAF1 Schizophrenia 40 G–U polymorphism was identified miR-125a NA NA 68 at the mature miR-125a rs17281995 in CD86 and miR-337, miR-582, miR-200a, CD86 and INSR Colorectal cancer 59 rs1051690 in INSR miR-184, miR-212 for CD86 and miR-612, miR-618 for INSR rs12720208 in the 30UTR of FGF20 miR-433 FGF20 Parkinson’s disease 28 (G–4A) mutation at 19 nt miR let-7e NA Human cancers 56 downstream of miRNA let-7e GEMIN3 gene (rs197414) NA GEMIN3 and GEMIN4 Bladder cancer 57 rs11614913 miR-196a2 Several predicted targets Lung cancer 60 ITGB4 30UTR (rs743554) miR-34a Integrin genes Breast cancer 58

Abbreviations: NA, not analyzed; DHFR, dihydrofolate reductase; UTR, untranslated region; AGTR1, angiotensin type 1 receptor. cause of phenotypic variations in single gene.70 The chemical reaction able to produce the greatest number of patients is a major objective in phar- of intron splicing and removal is splice variants is the Down syndrome macogenomics. A database was re- regulated by the spliceosome, a hetero- cell adhesion molecule (DSCAM) gene cently created to catalogue all the geneous complex comprised of RNAs in Drosophila.82 DSCAM can produce SNPs in miRNA target sites and link and proteins.71,72 The biology of the 38,016 putative splice forms using the these changes to complex traits and splicing mechanism has been recently combination of 4 clusters of mutually diseases.69 This database was termed elucidated, and some important cofac- exclusive exons and the usage of 20 polymorphism in miRNA target site tors in the selection of cryptic splice constitutive exons.82 DSCAM splice (PolymiRTS) and it integrates sequence sites have been described in normal variant proteins are responsible for of polymorphisms, phenotypes, gene and pathological states.73 One exam- axon guidance in brain development expression profiles in several microar- ple is the splicing factor SPF45 that was and are also expressed in specific ray data sets and characterizes the described with limited expression in olfactory receptor neuron cells in PolymiRTSs that are the potential normal tissue, but with a high expres- Drosophila.83,84 candidates responsible for quantitative sion in a many carcinomas associated Databases designed to store and trait locus (QTL) effects.69 The impact with drug resistance.74,75 provide access to reliable annotations of this database and of other upcom- Alternative splicing was postulated of the alternative splicing pattern of ing studies are of great importance for as one of the main cellular mechan- human genes and to the functional a better understanding on how SNPs isms for proteome diversity genera- annotation of predicted splicing iso- and other genetic variations will affect tion.76 Alternative splicing has also forms have been described by several the expression of protein-coding and emerged as a key mechanism respon- groups.85–87 Splice-site detection in noncoding genes in the genome by sible for the expansion of the tran- full-length transcripts have been car- changing the miRNA regulatory net- scriptome and proteome complexity ried out in genome-wide analyses work (Figure 1). Furthermore, the in humans and in other organisms.77 using specific bioinformatic algo- knowledge from these studies will be Recently, several studies have shown rithms, based on multiple alignments of relevance to evaluate drug metabo- how the alternative splicing process is of gene-related transcripts to the geno- lism and tolerance in patients suffer- controlled and how the expression of mic sequence. Alternative splicing da- ing from complex diseases. some splice variants might be asso- tabases will be able to provide ciated with diseases.78,79 Many of resources for functional interpretation these studies were performed using a of splicing variants for the human and Alternative splicing and proteome bioinformatics approach, as the mouse genomes and also for detection diversity amount of available transcriptome of splicing isoforms associated to and proteome data have been increas- diseases. Alternative splicing of pre-mRNAs was ing exponentially.80,81 The proteome Alternative splicing has become one proposed 30 years ago by the Nobel diversity generated by alternative spli- of the most elegant and important Prize winner Walter Gilbert as a way of cing of pre-mRNAs varies from protein mechanisms for proteome diversity generating different mRNAs from a to protein. To date, the gene that is generation. Growing evidence is also

The Pharmacogenomics Journal microRNAs and alternative splicing in phamacogenomics F Passetti et al 5

Figure 1 microRNAs and other noncoding RNAs as new players in pharmacogenomics. (a) Our current view of how genome variations such as single-nucleotide polymorphisms (SNPs) can affect drug metabolism. In this simplistic view, polymorphisms in one or a few protein-coding genes will affect drug resistance, efficacy and metabolism. (b) A proposed new view with microRNA variations, protein variations and noncoding RNA variations affecting the way drugs are metabolized is depicted. In this new integrated concept, all variations in the genome will have additive effects in pharmacokinetics and pharmacodynamics of drugs. indicating that defects in the alterna- affect the way patients metabolize ABL fusion protein.94 Patients with tive splicing pathways and generation drugs is still unexplored. Some exam- CML are commonly treated with Im- of wrong alternative variants is a ples of alternative splicing events with atinib mesylate that, in turn, inhibits common feature of complex dis- pharmacological relevance will be de- the activity of the BCR–ABL fusion eases.88 However, the implications of scribed in this section and listed in protein.95 Alternative splicing has splice variants in gene networks and Table 2. Our proposed model is de- been described as one of the possible how they can affect the metabolism of picted in Figure 2. reasons for drug resistance in some drugs in patients remains to be ex- One of the most important exam- patients treated with Imatinib mesy- plored. ples is the Philadelphia chromosome late;96 however, the majority of the which is generated by a translocation cases of drug resistance have been between the human chromosomes 9 described through the observation of Alternative splicing in drug and 22, leading to the fusion of two SNPs in the kinase domain of BCR– resistance, activation and genes: BCR and ABL. The BCR–ABL ABL. The first observation of alterna- metabolism fused gene is constitutively active and tive splicing variants in the BCR–ABL there are important consequences to fused gene and correlation with drug Several lines of evidence have already the cell: increase in cell proliferation resistance was in 2006, and a mutation suggested that protein diversity pro- and high rates of genomic instability. (L248V) in two CML patients with duced by alternative splicing might This chromosomal rearrangement is Imatinib mesylate resistance was de- affect important genes in pathways highly associated with chronic myelo- scribed.96 Sequence analysis was per- related to prodrug activation and drug genous leukemia (CML).93 In acute formed and it was found that this metabolism. Our understanding of lymphoblastic leukemia (ALL), around mutation produces two distinct how alternative splicing can act in 15–30% of patients have malignant mRNAs: one identical to the wild-type, drug resistance and on how it can transformation related to the BCR– differing only in one amino acid

The Pharmacogenomics Journal microRNAs and alternative splicing in phamacogenomics F Passetti et al 6

Table 2 Examples of alternative splicing events in protein-coding genes affecting drug response in several diseases

Gene Alternative splicing variant and drug effect Disease/mechanism References

CYP2C9 Inability to metabolize tolbutamide, diclofenac and warfarin NA 9,89 CYP2C18 Inability to metabolize tolbutamide and has low activity NA 90 under high doses of isosfamide CYP2C19 Lower metabolism of omeprazole and mephenytoin NA 91,92 BCR–ABL Resistance to Imatinib mesylate Chronic myelogenous leukemia (CML) 96,108 GR Resistance to glucocorticoids Asthma 113 GR Expression of the beta variant Lupus erythematosus 98 GR Resistance to Fluticasone Nasal polyposis 99 GR Insensitivity to glucocorticoid therapy Crohn’s disease 100 EGFR p110 sEGFR Human cancers 116,117 COX-1 Splice variant overexpressed in PBL and airway epithelial cells Asthma 103 COX-1 Acetaminophen can affect the enzyme activity of the splice variant NA 104 COX-2 Splice variant causes aspirin resistance Coronary artery bypass grafting (CABG) 105 PXR and CAR Isoforms can affect the pharmacokinetics and pharmacodynamics NA 106 of docetaxel and doxorubicin in Asian patients HMGCR Splice variants can cause statin metabolism changes LDL cholesterol level regulation 107

Abbreviations: NA, not analyzed; CYP, cytochrome P450; GR, glucocorticoid receptor; EGFR, epidermal growth factor receptor; COX-1, cyclooxygenase-1; PXR, pregnane Â receptor; CAR, constitutive androstane receptor. change (L248V), and another with the cing isoforms and drug resistance protein of 110 kDa that was termed usage of a cryptic splice site within the remains to be determined.109 soluble EGFR (sEGFR) or p110 exon 4, shortening in 81 nucleotides Another example of a gene that can sEGFR.116 In metastatic breast cancer its 30 portion (variant D248–274).96 have splicing variants is the gluco- patients under letrozole adjuvant ther- The leucine to valine change is located corticoid receptor (GR, NR3C1) (Figure apy, sEGFR showed decreased concen- in the ATP binding site and may 3a). NR3C1 (also known as GR) is a trations in 76% of the individuals and abrogate the kinase activity of the transcription factor that binds to cor- was also used as a biological marker.116 D248–274 splice variant protein.96 tisol and other glucocorticoids and has However, the authors suggest that Although the frequency of the alter- two main alternative splice products: additional studies are needed to under- native splicing variant D248–274 GRa and GRb.110,97 GRa is located in stand the role of p110 sEGFR in mRNA was low, the authors suggested the cytoplasm and is translocated to response to letrozole and the signifi- a dominant-negative role for this the nucleus after binding to glucocor- cance of EFGR as a cancer biomarker splice variant.96 Recently, another ticoids, whereas GRb cannot bind to for metastatic breast cancer. The p110 splice variant in the BCR–ABL fused its ligands.97 The GRb isoform has sEGFR splice isoform is depicted in gene was described in a cohort of 175 been associated with glucocorticoid Figure 3b. In another study, a cohort of patients in which 3 presented imati- resistance.97 The expression rate of 57 women with metastatic breast can- nib-resistance, and a new splice var- both GR isoforms has been studied cer was analyzed for the EGFR splice iant was also detected.108 In this case, and GRa is expressed up to 3000 times variant and reduced sEGFR concentra- the BCR–ABL isoform is produced by more than GRb.111,112 The role of GR tion was observed when compared to the usage of an alternative 35-nucleo- in many diseases is still unknown, but healthy individuals.117 On the other tide-long extra exon between exons 8 some associations of the splice variant hand, two studies trying to understand and 9.108 This new exon interrupts the GRb have been suggested in asth- the role of sEGFR in metastatic breast open reading frame of the BCR–ABL ma,113 systemic lupus erythemato- cancer patients under trastuzumab transcript and produces a truncated sus,98 Crohn’s disease100 and nasal therapy did not find any statistical protein. In a similar way that was polyposis99 (for more details see significant correlation between the described before,96 this splice variant Table 2). splice variant and drug metabo- has very low expression in 2 out of 3 A gene that also has splice variants lism.118,119 More studies are needed patients. In both cases, protein pro- and has been associated with many to understand the function and im- duction and/or activity will have to be types of cancer is the epidermal portance of EGFR receptor variants in better studied in order to understand growth factor receptor (EGFR).114 drug metabolism and activation. the role of these splice variants and Some studies have already shown that The cyclooxygenase gene family is how they can affect the mechanism of inhibiting the binding of EGFR to its represented in the human genome by imatinib resistance. Recently, the mi- ligands was associated with decreased two genes: PTGS1 (also known as COX- croRNA miR-203 was described as a cellular proliferation.101,102 The EGFR 1) and PTGS2 (also known as COX-2). regulator of the BCR–ABL gene in gene has several splice variants,115 and Cyclooxygenases (COXs) are key en- cancer but the connection with spli- one of the variants can produce a zymes in prostaglandin biosynthesis.

The Pharmacogenomics Journal microRNAs and alternative splicing in phamacogenomics F Passetti et al 7

Both protein products are activated by the pharmacokinetics and pharmaco- of an HMGCR splice isoform were acetylsalicylic acid (aspirin), acetami- dynamics of docetaxel and doxorubi- connected to reduced statin sensitivity nophen (paracetamol) and celecoxib, cin in Asian patients.106 and associated with interindividual among others drugs. Hence, the A recent study also showed that the differences in the metabolism of this knowledge of splicing variants of the NOVA2 gene interacts with a cis-acting drug.107 COX-1 and COX-2 transcripts is im- polymorphism to influence the There are several examples in litera- portant in order to better understand proportions of drug-responsive splice ture showing that splice variants of how drugs are metabolized. The COX- variants of SCN1A.106 The authors genes can change the way cells be- 1 gene has three splice variants already emphasize that genetic polymorph- come resistant to drugs, as well as the described.120,121,104 The first splice isms are important factors for modula- way drugs are metabolized but this is variant identified was one that does tion of the drug effect, illustrating probably just the tip of the iceberg. We not contain exons 1 and 2, and also alternative splicing as a potential are just starting to understand the has an alternative 50 end with which therapeutic target and the importance impact of proteome diversity on the uses part of intron 2 and it was termed of considering the activity of com- pharmacogenomics field. We propose COX1-SV. As there is a protein frame pounds at alternative splice isoforms a model in which alternative splicing shift after the translation of the main screening programs.128 In that is a player with big impact on phar- ture mRNA, this spliced variant was regard, variations in the production macogenomics. thought to produce a truncated protein. The action of acetaminophen (paracetamol) in inhibiting the action of this isoform of the protein in humans has already been shown in some studies (for review see Hersh et al.122). The splice variants have also been associated with the coronary artery bypass grafting (CABG), which is the most common medical intervention to treat heart failure. Aspirin (acetylsalicylic acid) is extensively used to inhibit platelet formation in this type of intervention, but up to 60% of the patients show some degree of resistance to aspirin.123,124 In addition, it has been described that after CABG surgery there is an overexpression of the COX-2 gene,125 and this could be the explanation for platelet formation in CABG patients.126 In a study to elucidate the function of COX-2,analter- native spliced product was discovered, named COX-2a, and it was associated with CABG and platelet formation.105 Splice variants for COX-1 and COX-2 are showninFigures3candd. Genetic variability in nuclear receptors can contribute to human variation at the magnitude of clinically significant drug–drug interactions. This is the case for the orphan nuclear receptors: pregnane Â receptor (PXR) and constitutive androstane receptor (CAR).127 These receptors are sensors that mediate drug-induced changes by Figure 2 Alternative splicing of protein-coding genes and the impact on pharmacogenomics. In our current and simplistic view, genetic variations (such as single nucleotide polymorphisms, increasing transcription of genes that SNPs) can change the sequence of proteins and affect the expression of protein-coding genes are involved in drug metabolism and that will be involved in drug metabolism. The proposed view suggests that alternative splicing is 127 clearance. It was already described an important player in drug resistance, activation and metabolism. In this model, different that genetic variants such as splicing isoforms produced from a specific protein can be responsible for different rates of drug isoforms of PXR and CAR can affect metabolism and clearance.

The Pharmacogenomics Journal microRNAs and alternative splicing in phamacogenomics F Passetti et al 8

Figure 3 Schematic representation of genes with alternative splicing described here according to the Human Genome Browser (UCSC, Human genome release of March 2006, http://genome.ucsc.edu). (a) Gene structure of the glucocorticoid receptor (GR). Red arrow indicates the alternative splicing event of exon 9 that produces the isoform GRb as described by Oakley et al.112 (b) Gene structure of the EGFR gene. The alternative splicing of EGFR encoding sEGFR p110 is indicated with a red arrow as described by Reiter et al.115 (c) Gene structure of the COX-1 gene. Red arrow indicates the splicing event of partial usage of exon 9 as previously described by Diaz et al.120 (d) Gene structure of the COX-2 gene. Red arrow indicates the partial usage of exon 5 generating the COX-2a isoform as described by Censarek et al.105 Boxes represent the alignment of exons with the human genome sequence. RefSeq mRNAs are represented in blue and known mRNAs are represented in black.

The Pharmacogenomics Journal microRNAs and alternative splicing in phamacogenomics F Passetti et al 9

microRNAs and alternative splicing: works.131 It was also shown that drugs are metabolized by patients, a new emerging field? microRNAs such as miR-124 are able and this will have major implications to change neuron fate by affecting for both drug design and personalized The relationship between alternative brain-specific pre-mRNA alternative medicine in the future. splicing, the 30UTR of protein-coding splicing.132 In addition, a study has genes and how alternative splicing can recently shown that miR-148 is able to affect the binding of microRNAs and regulate specific DNA methyltransfer- Conclusions and future prospects change gene regulation in the cells is ase (DNMT) protein isoforms, provid- an unexplored field. It is already ing evidence that this type of miRNAs are noncoding RNAs that can accepted that alternative splicing can mechanism might be involved in regulate gene expression by Watson– generate transcripts with different determining the relative abundance Crick base pairing to target several 30UTR and 50UTR,129 but the impact of different splice variants.133 Thus, mRNAs in a gene regulatory network. on gene regulation by microRNAs and as microRNAs are able to regulate They are involved in several important how this will affect drug metabolism is hundreds of effector genes in a multi- biological and pathological processes. unknown. One group was already able level regulatory mechanism that allow The binding of miRNAs to their target to show that retained introns in 30UTR individual miRNAs to profoundly af- mRNAs is critical for regulating mRNA of genes can increase putative miRNA fect the gene expression program in levels and therefore protein expres- targets in human mRNAs.130 More the cells,134 we propose that alterna- sion. It has also been shown that they recently, a study has shown that short- tive splicing might affect microRNA can regulate the expression of longer er 30-UTR isoforms can affect the regulation (Figure 4). We also propose non-coding RNAs. It is becoming clear regulation by microRNAs suggesting that changes in proteome diversity by that the binding of miRNAs to their that alternative splicing can have a both microRNA regulation and alter- targets can be affected by polymorph- major effect in gene regulatory net- native splicing can affect the way isms such as SNPs and genomic varia-

Figure 4 The impact of alternative splicing events in microRNA regulation. As most microRNAs target the 30UTR of mRNAs, if a gene has different isoforms or splice variants it can affect regulation by microRNAs. In this proposed model, several differences in drug metabolism could be explained, at least in part, by loss of microRNA control in shorter isoforms of mRNA targets. Other variations are also exemplified as potential players in pharmacogenomic differences.

The Pharmacogenomics Journal microRNAs and alternative splicing in phamacogenomics F Passetti et al 10

tions. Hence, polymorphisms in miR- para Pesquisa e Controle do Caˆncer. FFC is towards individualized therapy. Pharmaco- NAs represent a newly identified type supported by the Maeve McNicholas Mem- genomics 2007; 12: 1625–1627. of genetic variability that can influ- orial Foundation and Children’s Memorial 14 Lander ES, Linton LM, Birren B, Nusbaum Research Center. C, Zody MC, Baldwin J et al. Initial ence the risk of certain human diseases sequencing and analysis of the human and also influence how drugs can be genome. Nature 2001; 409: 860–921. Correspondence: Dr FF Costa, Cancer activated and metabolized by patients. 15 Venter JC, Adams MD, Myers EW, Li PW, Biology and Epigenomics Program, Depending on phenotypic differences Mural RJ, Sutton GG et al. The sequence of Children’s Memorial Research Center, the human genome. Science 2001; 291: in sequences coding for miRNAs, levels Northwestern University’s Feinberg School 1304–1351. of cellular effectors such as enzymes 16 Kidd JM, Cooper GM, Donahue WF, of Medicine, 2300 Children’s Plaza, Box and other proteins will be different Hayden HS, Sampas N, Graves T et al. 220, Chicago, IL 60614-3394, USA. and patients can have different phar- Mapping and sequencing of structural Tel.: þ 1 773 8804000x57312. variation from eight human genomes. macokinetics and pharmacodynamics. Fax: þ 1 773 7556551. Nature 2008; 453: 56–64. Another layer of gene regulation that 17 ENCODE Project Consortium et al. Identi- E-mail: [email protected] is still unexplored is alternative spli- fication and analysis of functional elements cing in protein-coding genes. Depend- in 1% of the human genome by the ENCODE pilot project. Nature 2007; 447: ing on the splice variant of a specific References 799–816. enzyme related to drug metabolism, 1 Vogel F. Moderne probleme der human- 18 Costa FF. Non-coding RNAs: new players in patients will respond better or worse to genetik. Ergebnisse Inneren Medizin und eukaryotic biology. Gene 2005; 2: 83–94. therapies. Several examples described Keinderheilkunde 1959; 12: 52–125. 19 Costa FF. Non-coding RNAs: lost in translation? Gene 2007; 1-2: 1–10. here support the hypothesis that alter- 2 Meyer UA. Pharmacogenetics—five dec- ades of therapeutic lessons from genetic 20 Costa FF. Non-coding RNAs, epigenetics native splicing can modify drug resis- diversity. Nat Rev Genet 2004; 5: 669–676. and complexity. Gene 2008; 1: 9–17. tance and metabolism suggesting that 3 Roses AD. Genome-based pharmacoge- 21 Bartel D. MicroRNAs: genomics, biogen- these events could be one of the netics and the pharmaceutical industry. esis, mechanism, and function. Cell 2004; 2: 281–297. factors impacting in interpatient varia- Nat Rev Drug Discov 2002; 1: 541–549. 4 Giacomini KM, Brett CM, Altman RB, 22 Calin GA, Liu CG, Ferracin M, Hyslop T, bility in drug response. The final Benowitz NL, Dolan ME, Flockhart DA et Spizzo R, Sevignani C et al. Ultraconserved model that we propose here is that al. Pharmacogenetics Research Network. regions encoding ncRNAs are altered in genetic variations in the sequence of The pharmacogenetics research network: human leukemias and carcinomas. Cancer Cell 2007; 3: 215–229. miRNAs, target sites of miRNAs and from SNP discovery to clinical drug response. Clin Pharmacol Ther 2007; 81: 328–345. 23 Hobert O. Gene regulation by transcription alternative splicing will result in phar- 5 Ingelman-Sundberg M, Sim SC, Gomez A, factors and microRNAs. Science 2008; 319: macogenomic differences. As miRNAs Rodriguez-Antona C. Influence of cyto- 1785–1786. and isoforms of the same protein will chrome P450 polymorphisms on drug thera- 24 Gartel AL, Kandel ES. miRNAs: little known mediators of oncogenesis. Semin Cancer affect gene networks in the cells, pies: pharmacogenetic, pharmacoepigenetic and clinical aspects. Pharmacol Ther 2007; Biol 2008; 18: 103–110. changes in these mechanisms will be 116: 496–526. 25 Clark TA, Schweitzer AC, Chen TX, Staples important in clinics. Personalized 6 Gardiner SJ, Begg EJ. Pharmacogenetics, MK, Lu G, Wang H et al. Discovery of medicine must take all of these drug-metabolizing enzymes, and clinical tissue-specific exons using comprehensive human exon microarrays. Genome Biol variations into account so we can practice. Pharmacol Rev 2006; 58: 521–590. 2007; 8: R64. administer the correct dosage of 7 Chauhan N, Padh SR. Pharmacogenetics: 26 Tress ML, Martelli PL, Frankish A, Reeves drugs and evaluate drug efficacy. We genetic basis for rational drug therapy. GA, Wesselink JJ, Yeats C. The implications believe that a map of all variations in Indian J Pharm Sci 2007; 69: 180–189. of alternative splicing in the ENCODE protein complement. Proc Natl Acad Sci miRNAs and alternative variants 8 Rodriguez-Antona C, Ingelman-Sundberg M. Cytochrome P450 pharmacogenetics USA 2007; 104: 5495–5500. of protein-coding genes should be and cancer. Oncogene 2006; 25: 1679–1691. 27 Romero PR, Zaidi S, Fang YY, Uversky VN, generated. A lot of work is still 9 Oner Ozgon G, Langaee TY, Feng H, Buyru Radivojac P, Oldfield CJ et al. Alternative needed to completely understand the N, Ulutin T, Hatemi AC. VKORC and splicing in concert with protein intrinsic disorder enables increased functional di- consequences of variations in gene CYP2C9 polymorphisms are associated with warfarin dose requirements in Turkish versity in multicellular organisms. Proc Natl regulatory networks and pathways patients. Eur J Clin Pharmacol 2008; 64: Acad Sci USA 2006; 103: 8390–8395. responsible for drug metabolism and 889–894. 28 Wang G, van der Walt JM, Mayhew G, Li YJ, resistance, but we believe that these 10 Wadelius M, Chen LY, Downes K, Ghori J, Zu¨chner S, Scott WK et al. Variation in the miRNA-433 binding site of FGF20 confers new paradigms are clearly redefining Hunt S, Eriksson N et al. Common VKORC1 and GGCX polymorphisms associated with risk for Parkinson disease by overexpression the pharmacogenomics field. warfarin dose. Pharmacogenomics J 2005; of alpha-synuclein. Am J Hum Genet 2008; 5: 262–270. 82: 283–289. 11 International HapMap consortium. A sec- 29 Jazdzewski K, Murray EL, Franssila K, Jarzab Acknowledgments ond generation human haplotype map of B, Schoenberg DR, de la Chapelle A. over 3.1 million SNPs. Nature 2007; 449: Common SNP in pre-miR-146a decreases We thank the insights and suggestions from 851–861. mature miR expression and predisposes to 12 Chen Y, Zhu J, Lum PY, Yang X, Pinto S, papillary thyroid carcinoma. Proc Natl Acad Dr Tom Wurdinger, Dr Patricia Savio de MacNeil DJ et al. Variations in DNA Sci USA 2008; 105: 7269–7274. Araujo Souza, Dr Alexandre da Costa Per- elucidate molecular networks that cause 30 Georges M, Clop A, Marcq F, Takeda H, eira, Dr Elio Vanin and Kelly Arndt. FP and disease. Nature 2008; 452: 429–435. Pirottin D, Hiard S et al. SNPs can affect the CGF are supported by the Swiss Bridge 13 Bertino JR, Banerjee D, Mishra PJ. Pharma- way miRNAs regulate their targets pointing Foundation and Fundac¸a˜o Ary Frauzino cogenomics of microRNA: a miRSNP to a function in drug metabolism and in

The Pharmacogenomics Journal microRNAs and alternative splicing in phamacogenomics F Passetti et al 11

phenotypic variation. Cold Spring Harb activation of microRNA-127 with down- 62 Tan Z, Randall G, Fan J, Camoretti-Merca- Symp Quant Biol 2006; 71: 343–350. regulation of the proto-oncogene BCL6 by do B, Brockman-Schneider R, Pan L et al. 31 Akgul C, Moulding DA, Edwards SW. chromatin-modifying drugs in human can- Allele-specific targeting of microRNAs to Alternative splicing of Bcl-2-related genes: cer cells. Cancer Cell 2006; 9: 435–443. HLA-G and risk of asthma. Am J Hum Genet functional consequences and potential 49 Brueckner B, Stresemann C, Kuner R, 2007; 81: 829–834. therapeutic applications. Cell Mol Life Sci Mund C, Musch T, Meister M et al. The 63 Martin MM, Buckenberger JA, Jiang J, 2004; 61: 2189–2199. human let-7a-3 locus contains an epigen- Malana GE, Nuovo GJ, Chotani M et al. 32 Eulalio A, Huntzinger E, Izaurralde E. Get- etically regulated microRNA gene with The human angiotensin II type 1 receptor ting to the root of miRNA-mediated gene oncogenic function. Cancer Res 2007; 67: +1166 A/C polymorphism attenuates mi- silencing. Cell 2008; 132:9–14. 1419–1423. crorna-155 binding. J Biol Chem 2007; 282: 33 Lee RC, Feinbaum RL, Ambros V. The C. 50 Hung RJ, McKay JD, Gaborieau V, Boffetta 24262–24269. elegans heterochronic gene lin-4 encodes P, Hashibe M, Zaridze D et al. A suscept- 64 Mishra PJ, Humeniuk R, Mishra PJ, Longo- small RNAs with antisense complementar- ibility locus for lung cancer maps to Sorbello GS, Banerjee D, Bertino JR. A miR- ity to lin-14. Cell 1993; 75: 843–854. nicotinic acetylcholine receptor subunit genes 24 microRNA binding-site polymorphism 34 Rajewsky N. microRNA target predictions in on 15q25. Nature 2008; 452: 633–637. in dihydrofolate reductase gene leads to animals. Nat Genet 2006; 38(Suppl): S8–13. 51 Carlson CS, Eberle MA, Kruglyak L, Nick- methotrexate resistance. Proc Natl Acad Sci 35 Filipowicz W, Bhattacharyya SN, Sonen- erson DA. Mapping complex disease loci in USA 2007; 104: 13513–13518. berg N. Mechanisms of post-transcriptional whole-genome association studies. Nature 65 Lal A, Kim HH, Abdelmohsen K, Kuwano Y, regulation by microRNAs: are the answers 2004; 429: 446–452. Pullmann Jr R, Srikantan S et al. p16(INK4a) in sight? Nat Rev Genet 2008; 9: 102–114. 52 Dunckley T, Huentelman MJ, Craig DW, translation suppressed by miR-24.PLoS 36 Bartel D. MicroRNAs: genomics, biogen- Pearson JV, Szelinger S, Joshipura K et al. ONE. 2008; 3: e1864. esis, mechanism, and function. Cell 2004; Whole-genome analysis of sporadic amyo- 66 Wang Q, Huang Z, Xue H, Jin C, Ju XL, Han 116: 281–297. trophic lateral sclerosis. N Engl J Med 2007; JD et al. MicroRNA miR-24 inhibits erythro- 37 Wurdinger T, Costa FF. Molecular therapy 357: 775–788. poiesis by targeting activin type I receptor in the microRNA era. Pharmacogenomics 53 Fellay J, Shianna KV, Ge D, Colombo S, ALK4. Blood 2008; 111: 588–595. J 2007; 7: 297–304. Ledergerber B, Weale M et al. A whole- 67 Sun Q, Zhang Y, Yang G, Chen X, Zhang Y, 38 Esau C, Davis S, Murray SF, Yu XX, Pandey genome association study of major deter- Cao G et al. Transforming growth factor- SK, Pear M et al. miR-122 regulation of minants for host control of HIV-1. Science beta-regulated miR-24 promotes skeletal lipid metabolism revealed by in vivo 2007; 317: 944–947. muscle differentiation. Nucleic Acids Res antisense targeting. Cell Metab 2006; 3: 54 Yu Z, Li Z, Jolicoeur N, Zhang L, Fortin Y, 2008; 36: 2690–2699. 87–98. Wang E et al. Aberrant allele frequencies of 68 Duan R, Pak C, Jin P. Single nucleotide 39 Poy MN, Eliasson L, Krutzfeldt J, Kuwajima the SNPs located in microRNA target sites polymorphism associated with mature S, Ma X, Macdonald PE et al. A pancreatic are potentially associated with human miR-125a alters the processing of islet-specific microRNA regulates insulin cancers. Nucleic Acids Res 2007; 35: pri-miRNA. Hum Mol Genet 2007; 16: secretion. Nature 2004; 432: 226–230. 4535–4541. 1124–1131. 40 Hansen T, Olsen L, Lindow M, Jakobsen 55 Saunders MA, Liang H, Li WH. Human 69 Bao L, Zhou M, Wu L, Lu L, Goldowitz D, KD, Ullum H, Jonsson E et al. Brain polymorphism at microRNAs and micro- Williams RW et al. PolymiRTS Database: expressed microRNAs implicated in schizo- RNA target sites. Proc Natl Acad Sci USA linking polymorphisms in microRNA target phrenia etiology. PLoS ONE 2007; 2: e873. 2007; 104: 3300–3305. sites with complex traits. Nucleic Acids Res 41 Garzon R, Fabbri M, Cimmino A, Calin GA, 56 Wu M, Jolicoeur N, Li Z, Zhang L, Fortin Y, 2007; 35(Database issue): D51–D54. Croce CM. MicroRNA expression and Denis L et al. Genetic variations of micro- 70 Gilbert W. Why genes in pieces? Nature function in cancer. Trends Mol Med 2006; RNAs in human cancer and their effects on 1978; 271: 501. 12: 580–587. the expression of miRNAs. Carcinogenesis 71 Smith CW, Valca´rcel J. Alternative pre- 42 Mayr C, Hemann MT, Bartel DP. Disrupting 2008; 29: 1710–1716. mRNA splicing: the logic of combinatorial the pairing between let-7 and Hmga2 57 Yang H, Dinney CP, Ye Y, Zhu Y, Grossman control. Trends Biochem Sci 2000; 25: enhances oncogenic transformation. HB, Wu X. Evaluation of genetic variants in 381–388. Science 2007; 315: 1576–1579. microRNA-related genes and risk of blad- 72 Valadkhan S. The spliceosome: caught in a 43 He L, He X, Lim LP, de Stanchina E, Xuan Z, der cancer. Cancer Res 2008; 68: 2530– web of shifting interactions. Curr Opin Liang Y et al. A microRNA component of 2537. Struct Biol 2007; 17: 310–315. the p53 tumour suppressor network. Nat- 58 Brendle A, Lei H, Brandt A, Johansson R, 73 House AE, Lynch KW. Regulation of alter- ure 2007; 447: 1130–1134. Enquist K, Henriksson R. Polymorphisms in native splicing: more than just the ABCs. 44 Ma L, Teruya-Feldstein J, Weinberg RA. predicted microRNA binding sites in integ- J Biol Chem 2008; 283: 1217–1221. Tumour invasion and metastasis initiated rin genes and breast cancer: ITGB4 as 74 Sampath J, Long PR, Shepard RL, Xia X, by microRNA-10b in breast cancer. Nature prognostic marker. Carcinogenesis 2008; Devanarayan V, Sandusky GE et al. 2007; 449(7163): 682–688. 29: 1394–1399. Human SPF45, a splicing factor, has limited 45 Tavazoie SF, Alarco´n C, Oskarsson T, Padua 59 Landi D, Gemignani F, Naccarati A, Pardini expression in normal tissues, is D, Wang Q, Bos PD et al. Endogenous B, Vodicka P, Vodickova L. Polymorphisms overexpressed in many tumors, and can human microRNAs that suppress breast within micro-RNA-binding sites and risk of confer a multidrug-resistant phenotype cancer metastasis. Nature 2008; 451: sporadic colorectal cancer. Carcinogenesis to cells. Am J Pathol 2003; 163: 147–152. 2008; 29: 579–584. 1781–1790. 46 Schetter AJ, Leung SY, Sohn JJ, Zanetti KA, 60 Hu Z, Chen J, Tian T, Zhou X, Gu H, Xu L. 75 Perry III WL, Shepard RL, Sampath J, Yaden Bowman ED, Yanaihara N et al. MicroRNA Genetic variants of miRNA sequences and B, Chin WW, Iversen PW et al. Human expression profiles associated with non-small cell lung cancer survival. J Clin splicing factor SPF45 (RBM17) confers prognosis and therapeutic outcome in Invest 2008; 118: 2600–2608. broad multidrug resistance to anticancer colon adenocarcinoma. JAMA 2008; 299: 61 Sethupathy P, Borel C, Gagnebin M, Grant drugs when overexpressed—a phenotype 425–436. GR, Deutsch S, Elton TS et al. Human partially reversed by selective estrogen 47 Yu SL, Chen HY, Chang GC, Chen CY, microRNA-155 on chromosome 21 differ- receptor modulators. Cancer Res 2005; Chen HW, Singh S et al. MicroRNA entially interacts with its polymorphic 65: 6593–6600. signature predicts survival and relapse in target in the AGTR1 30 untranslated region: 76 Black DL. Protein diversity from alternative lung cancer. Cancer Cell 2008; 13: 48–57. a mechanism for functional single-nucleo- splicing: a challenge for bioinformatics 48 Saito Y, Liang G, Egger G, Friedman JM, tide polymorphisms related to phenotypes. and post-genome biology. Cell 2000; Chuang JC, Coetzee GA et al. Specific Am J Hum Genet 2007; 81: 405–413. 103: 367–370.

The Pharmacogenomics Journal microRNAs and alternative splicing in phamacogenomics F Passetti et al 12

77 Kim E, Magen A, Ast G. Different levels of 92 Roh HK, Dahl ML, Tybring G, Yamada H, 105 Censarek P, Freidel K, Udelhoven M, Ku SJ, alternative splicing among eukaryotes. Cha YN, Bertilsson L. CYP2C19 genotype Hohlfeld T, Meyer-Kirchrath J et al. Cy- Nucleic Acids Res 2007; 35: 125–131. and phenotype determined by omeprazole clooxygenase COX-2a, a novel COX-2 78 Wang GS, Cooper TA. Splicing in disease: in a Korean population. Pharmacogenetics mRNA variant, in platelets from patients disruption of the splicing code and the 1996; 6: 547–551. after coronary artery bypass grafting. decoding machinery. Nat Rev Genet 2007; 93 Daley GQ, Van Etten RA, Baltimore D. Thromb Haemost 2004; 92: 925–928. 8: 749–761. Induction of chronic myelogenous leukae- 106 Hor SY, Lee SC, Wong CI, Lim YW, Lim RC, 79 Skotheim RI, Nees M. Alternative splicing mia in mice by the P210 bcr/abl gene of Wang LZ. PXR, CAR and HNF4alpha in cancer: noise, functional, or systematic? the Philadelphia chromosome. Science genotypes and their association with phar- Int J Biochem Cell Biol 2007; 39: 1432– 1990; 247: 824–830. macokinetics and pharmacodynamics of 1449. 94 Faderl S, Kantarjian HM, Thomas DA, docetaxel and doxorubicin in Asian 80 Thorsen K, Sørensen KD, Brems-Eskildsen Cortes J, Giles F, Pierce S et al. Outcome patients. Pharmacogenomics J 2008; 2: AS, Modin C, Gaustadnes M, Hein AM et of Philadelphia chromosome- 139–146. al. Alternative splicing in colon, bladder, positive adult acute lymphoblastic 107 Medina MW, Gao F, Ruan W, Rotter JI, and prostate cancer identified by exon- leukemia. Leuk Lymphoma 2000; 36: Krauss RM. Alternative Splicing of 3-Hydro- array analysis. Mol Cell Proteomics 2008; 7: 263–273. xy-3-Methylglutaryl Coenzyme A Reduc- 1214–1224. 95 Druker BJ, Talpaz M, Resta DJ, Peng B, tase Is Associated With Plasma Low-Density 81 Ben-Dov C, Hartmann B, Lundgren J, Buchdunger E, Ford JM et al. Efficacy and Lipoprotein Cholesterol Response to Valca´rcel J. Genome-wide analysis of alter- safety of a specific inhibitor of the BCR-ABL Simvastatin. Circulation 2008; 118: native pre-mRNA splicing. J Biol Chem tyrosine kinase in chronic myeloid leuke- 355–362. 2008; 283: 1229–1233. mia. N Engl JMed 2001; 344: 1031–1037. 108 Laudadio J, Deininger MW, Mauro MJ, 82 Schmucker D, Clemens JC, Shu H, Worby 96 Gruber FX, Hjorth-Hansen H, Mikkola I, Druker BJ, Press RD. An intron-derived CA, Xiao J, Muda M et al. Drosophila Stenke L, Johansen T. A novel Bcr-Abl splice insertion/truncation mutation in the BCR- Dscam is an axon guidance receptor isoform is associated with the L248V ABL kinase domain in chronic myeloid exhibiting extraordinary molecular diver- mutation in CML patients with acquired leukemia patients undergoing kinase in- sity. Cell 2000; 101: 671–684. resistance to imatinib. Leukemia 2006; 20: hibitor therapy. J Mol Diagn 2008; 10: 83 Schmucker D. Molecular diversity of 2057–2060. 177–180. Dscam: recognition of molecular identity in 97 Lewis-Tuffin LJ, Cidlowski JA. The physiol- 109 Bueno MJ, Pe´rez de Castro I, Go´mez de neuronal wiring. Nat Rev Neurosci 2007; 8: ogy of human glucocorticoid receptor beta Cedro´n M, Santos J, Calin GA, Cigudosa JC 915–920. (hGRbeta) and glucocorticoid resistance. et al. Genetic and epigenetic silencing of 84 Hummel T, Vasconcelos ML, Clemens JC, Ann NY Acad Sci 2006; 1069: 1–9. microRNA-203 enhances ABL1 and BCR- Fishilevich Y, Vosshall LB, Zipursky SL. 98 Piotrowski P, Burzynski M, Lianeri M, ABL1 oncogene expression. Cancer Cell Axonal targeting of olfactory receptor Mostowska M, Wudarski M, Chwalinska- 2008; 6: 496–506. neurons in Drosophila is controlled by Sadowska H. Glucocorticoid receptor beta 110 Lu NZ, Cidlowski JA. Glucocorticoid recep- Dscam. Neuron 2003; 37: 221–231. splice variant expression in patients with tor isoforms generate transcription specifi- 85 Kim N, Alekseyenko AV, Roy M, Lee C. The high and low activity of systemic lupus city. Trends Cell Biol 2006; 16: 301–307. ASAP II database: analysis and comparative erythematosus. Folia Histochem Cytobiol 111 Pujols L, Mullol J, Picado C. Alpha and beta genomics of alternative splicing in 15 2007; 45: 339–342. glucocorticoid receptors: relevance in air- animal species. Nucleic Acids Res 2007; 99 Hamilos DL, Leung DY, Muro S, Kahn AM, way diseases. Curr Allergy Asthma Rep 35(Database issue): D93–D98. Hamilos SS, Thawley SE et al. GRbeta 2007; 7: 93–99. 86 Lee Y, Lee Y, Kim B, Shin Y, Nam S, Kim P et expression in nasal polyp inflammatory 112 Oakley RH, Sar M, Cidlowski JA. The al. ECgene: an alternative splicing database cells and its relationship to the anti- human glucocorticoid receptor beta iso- update. Nucleic Acids Res 2007; 35(Data- inflammatory effects of intranasal flutica- form. Expression, biochemical properties, base issue): D99–D103. sone. J Allergy Clin Immunol 2001; 108: and putative function. J Biol Chem 1996; 87 Holste D, Huo G, Tung V, Burge CB. 59–68. 271: 9550–9559. HOLLYWOOD: a comparative relational 100 Towers R, Naftali T, Gabay G, Carlebach 113 Goleva E, Li LB, Eves PT, Strand MJ, Martin database of alternative splicing. Nucleic Acids M, Klein A, Novis B. High levels of RJ, Leung DY. Increased glucocorticoid Res 2006; 34(Database issue): D56–D62. glucocorticoid receptors in patients with receptor beta alters steroid response in 88 Dahmcke CM, Bu¨chmann-Møller S, Jensen active Crohn’s disease may predict steroid glucocorticoid-insensitive asthma. Am J NA, Mitchelmore C. Altered splicing in resistance. Clin Exp Immunol 2005; 141: Respir Crit Care Med 2006; 173: 607–616. exon 8 of the DNA replication factor CIZ1 357–362. 114 Yarden Y, Sliwkowski MX. Untangling the affects subnuclear distribution and is asso- 101 Masui H, Kawamoto T, Sato JD, Wolf B, ErbB signalling network. Nat Rev Mol Cell ciated with Alzheimer’s disease. Mol Cell Sato G, Mendelsohn J. Growth inhibition Biol 2001; 2: 127–137. Neurosci 2008; 38: 589–594. of human tumor cells in athymic mice by 115 Reiter JL, Threadgill DW, Eley GD, Strunk 89 Ariyoshi N, Shimizu Y, Kobayashi Y, Naka- anti-epidermal growth factor receptor KE, Danielsen AJ, Sinclair CS et al. Com- mura H, Nakasa H, Nakazawa K. Identifica- monoclonal antibodies. Cancer Res 1984; parative genomic sequence analysis and tion and partial characterization of a novel 44: 1002–1007. isolation of human and mouse alternative CYP2C9 splicing variant encoding a pro- 102 Mendelsohn J. Anti-EGF receptor mono- EGFR transcripts encoding truncated re- tein lacking eight amino acid residues. clonal antibodies: biological studies and ceptor isoforms. Genomics 2001; 71: 1–20. Drug Metab Pharmacokinet 2007; 22: potential clinical applications. Trans Am 116 Lafky JM, Baron AT, Cora EM, Hillman DW, 187–194. Clin Climatol Assoc 1989; 100: 31–38. Suman VJ, Perez EA et al. Serum soluble 90 Zhu-Ge J, Yu YN. Enzyme activity analysis 103 Kowalski ML, Borowiec M, Kurowski M, epidermal growth factor receptor concen- of CYP2C18 with exon 5 skipped. Acta Pawliczak R. Alternative splicing of cycloox- trations decrease in postmenopausal Pharmacol Sin 2004; 25: 1065–1069. ygenase-1 gene: altered expression in metastatic breast cancer patients treated 91 Ibeanu GC, Blaisdell J, Ferguson RJ, Gha- leucocytes from patients with bronchial with letrozole. Cancer Res 2005; 65: nayem BI, Brosen K, Benhamou S et al. A asthma and association with aspirin-in- 3059–3062. novel transversion in the intron 5 donor duced 15-HETE release. Allergy 2007; 62: 117 Asgeirsson KS, Agrawal A, Allen C, Hitch A, splice junction of CYP2C19 and a sequence 628–634. Ellis IO, Chapman C et al. Serum epidermal polymorphism in exon 3 contribute to the 104 Censarek P, Freidel K, Hohlfeld T, Schro¨r K, growth factor receptor and HER2 expres- poor metabolizer phenotype for the anti- Weber AA. Human cyclooxygenase-1b is sion in primary and metastatic breast convulsant drug S-mephenytoin. J Pharma- not the elusive target of acetaminophen. cancer patients. Breast Cancer Res 2007; col Exp Ther 1999; 290: 635–640. Eur J Pharmacol 2006; 551: 50–53. 9: R75.

The Pharmacogenomics Journal microRNAs and alternative splicing in phamacogenomics F Passetti et al 13

118 Hudelist G, Ko¨stler WJ, Gschwantler-Kau- 122 Hersh EV, Lally ET, Moore PA. Update on cis-acting polymorphism to influence the lich D, Czerwenka K, Kubista E, Mu¨ller R et cyclooxygenase inhibitors: has a third COX proportions of drug-responsive splice var- al. Serum EGFR levels and efficacy of isoform entered the fray? Curr Med Res iants of SCN1A. Am J Hum Genet 2007; 5: trastuzumab-based therapy in patients Opin 2005; 8: 1217–1226. 876–883. with metastatic breast cancer. Eur J Cancer 123 Zimmermann N, Kienzle P, Weber AA, 129 Kwan T, Benovoy D, Dias C, Gurd S, 2006; 42: 186–192. Winter J, Gams E, Schro¨r K et al. Aspirin Provencher C, Beaulieu P et al. Genome- 119 Cameron D, Casey M, Press M, Lindquist resistance after coronary artery bypass wide analysis of transcript isoform variation D, Pienkowski T, Romieu CG et al. A phase grafting. J Thorac Cardiovasc Surg 2001; in humans. Nat Genet 2008; 40: 225–231. III randomized comparison of lapatinib 121: 982–984. 130 Tan S, Guo J, Huang Q, Chen X, Li-Ling J, Li plus capecitabine versus capecitabine 124 Gasparyan AY, Watson T, Lip GY. The role Q et al. Retained introns increase putative alone in women with advanced breast of aspirin in cardiovascular prevention: microRNA targets within 30 UTRs of human cancer that has progressed on trastuzu- implications of aspirin resistance. J Am Coll mRNA. FEBS Lett 2007; 6: 1081–1086. mab: updated efficacy and biomarker Cardiol 2008; 51: 1829–1843. 131 Sandberg R, Neilson JR, Sarma A, Sharp PA, analyses. Breast Cancer Res Treat 2008, 125 Weber AA, Zimmermann KC, Meyer-Kir- Burge CB. Proliferating cells express e-pub ahead of print. chrath J, Schro¨r K. Cyclooxygenase-2 in mRNAs with shortened 30 untranslated 120 Diaz A, Reginato AM, Jimenez SA. Alter- human platelets as a possible factor in regions and fewer microRNA target sites. native splicing of human prostaglandin G/ aspirin resistance. Lancet 1999; 353: 900. Science 2008; 5883: 1643–1647. H synthase mRNA and evidence of differ- 126 Zimmermann N, Wenk A, Kim U, Kienzle P, 132 Makeyev EV, Zhang J, Carrasco MA, ential regulation of the resulting transcripts Weber AA, Gams E et al. Functional and Maniatis T. The MicroRNA miR-124 pro- by transforming growth factor beta 1, biochemical evaluation of platelet aspirin motes neuronal differentiation by trigger- interleukin 1 beta, and tumor necrosis resistance after coronary artery bypass ing brain-specific alternative pre-mRNA factor alpha. J Biol Chem 1992; 267: surgery. Circulation 2003; 108: 542–547. splicing. Mol Cell 2007; 3: 435–448. 10816–10822. 127 Lamba J, Lamba V, Schuetz E. Genetics 133 Duursma AM, Kedde M, Schrier M, le Sage 121 Kitzler J, Hill E, Hardman R, Reddy N, variants of PXR (NR1I2) and CAR (NR1I3) C, Agami R. miR-148 targets human Philpot R, Eling TE. Analysis and quantita- and their implications in drug metabolism DNMT3b protein coding region. RNA tion of splicing variants of the TPA-induci- and pharmacogenetics. Curr Drug Metab 2008; 5: 872–877. ble PGHS-1 mRNA in rat tracheal epithelial 2005; 4: 369–383. 134 Makeyev EV, Maniatis T. Multilevel regula- cells. Arch Biochem Biophys 1995; 316: 128 Heinzen EL, Yoon W, Tate SK, Sen A, Wood tion of gene expression by microRNAs. 856–863. NW, Sisodiya SM. Nova2 interacts with a Science 2008; 5871: 1789–1790.

The Pharmacogenomics Journal