Development and application of a pathology supported pharmacogenetic test for improved clinical management of South African patients with breast cancer and associated co-morbidities

Van der Merwe, Nicole (2016-03)

Thesis (PhD)--Stellenbosch University, 2016.

Thesis

ENGLISH ABSTRACT: Three major challenges in the field of breast cancer have been identified as research priorities for this study. The first is the need to combine genetic testing of high-risk patients with familial breast cancer with pharmacogenetics to reduce recurrence risk in cancer survivors due to drug failure as a consequence of anti-cancer treatment that does not match the patient’s genotype. The second is the delineation of key pathways through which genes implicated in breast cancer and associated co-morbidities can serve as nutritional and drug targets across diagnostic boundaries. The third is the discovery of genetic alterations underlying familial breast cancer not attributed to mutations in the two major tumour suppressor genes, BRCA1 and BRCA2. The study population consisted of 164 breast cancer patients (60 Coloured/Mixed Ancestry and 104 Caucasian), of whom 88 patients were selected from a total of 813 individuals who provided informed consent for inclusion of their data in a genomics database resource generated at the interface between the laboratory and routine clinical practice. In addition, DNA samples of 101 cancer-free individuals above the age of 65 years were available for clinical validation of potentially causative variants in an extended female control group. In the first phase of this study, real-time polymerase chain reaction (PCR) TaqMan© technology was used to confirm the potential value of adding pharmacogenetic testing (CYP2D6 allele 4) to standard immunohistochemistry (IHC)-based breast tumour subtyping complemented by BRCA mutation screening and/or microarray gene profiling in eligible patients. In phase two of the study, common genetic risk factors for cardiovascular disease (CVD) were shown to be significantly associated with earlier age (10 years on average) of breast cancer onset/diagnosis (APOE E4 allele p=0.003; 95% CI: 4-15) and body mass index (BMI) (MTHFR 1298 A>C; p=0.01; 95% CI: 3-14) in patients stratified according to estrogen receptor (ER) status, after adjustment for potential confounders. Age at diagnosis/onset of breast cancer was significantly lower in patients with ER-negative versus ER-positive tumours, after adjustment for ethnicity (p=0.022), while BMI was significantly higher in patiens with ER-positive compared to ERnegative tumours after adjustment for age, ethnicity, and family history of cancer (p=0.035). These findings contributed to the development of an exome pre-screening algorithm (EPA) used in part 3 of this study to select three genetically uncharacterized breast cancer patients for whole exome sequencing (WES) performed in comparison with three ethnically concordant cancer-free controls. WES followed by variant calling using both the standard human genome reference sequence (hg19) and an ethnically concordant major allele reference genome (MARS) revealed a more than 20% discrepancy in the number of gene variants identified in the same samples. After exclusion of a large number of false-positives caused by minor alleles in hg19, two rare missense mutations (<1%) were identified in a family with ER-positive breast cancer: RAD50 R385C and MUC1 Q67E. Three different bioinformatics tools were used to predict functionality and both mutations were confirmed by Sanger sequencing and/or real-time PCR in the Pathology Research Facility (PRF) laboratory. Neither the RAD50 nor the MUC1 missense mutation were identified in the exomes of an unrelated breast cancer patient with triple-negative breast cancer or three population-matched control individuals. This study led to the development of a pathology-supported genetic testing framework for WES beyond the limitations of single-gene BRCA mutation screening in South African breast cancer patients. Our findings support previous WES results indicating that the majority of genetically uncharacterised familial breast cancer may be caused by a combination of low-moderate penetrance mutations exerting their effect in a high-risk environment reflected by high BMI. WES enables identification of genetic risk factors of relevance to both cancer development and tailored therapeutic intervention in a single genetic test.

AFRIKAANSE OPSOMMING: Drie belangrike uitdagings in die veld van borskanker-navorsing is geïdentifiseer as die fokus vir navorsing in hierdie studie. Die eerste is die nodigheid om genetiese toetsing van hoë-risiko pasiënte met familiële borskanker te kombineer met farmakogenetika, om die risiko van siekteherhaling in borskanker oorlewendes as gevolg van mislukking van kankerbehandeling wat nie in ooreenstemming is met die pasiënt se genotipe nie, te voorkom. Die tweede is die afbakening van sleutel paaie waardeur gene wat betrek word by borskanker en verwante komorbiditeite, voedings- en medikasie behandelingsteikens oor diagnostiese grense heen kan voorsien. Die derde is die ontdekking van genetiese veranderinge onderliggend aan familiële borskanker wat nie verklaar kan word deur mutasies in die twee belangrikste kankeronderdrukker gene, BRCA1 en BRCA2 nie. Die studiepopulasie het bestaan uit 164 borskankerpasiënte (60 Kleurling/Gemengde Afkoms en 104 Koukasiers), van wie 88 geselekteer is uit 'n totaal van 813 individue wat ingeligte toestemming verleen het vir insluiting van hulle data in ‘n genomiese databasishulpbron, wat ontwikkel is op die interfase van die laboratorium en roetine kliniese praktyk. DNA van ‘n verdere 101 kankervrye individue bokant die ouderdom van 65 jaar was beskikbaar vir kliniese validasie van potensieel kanker-verwante variante in ‘n uitgebreide vroulike kontrole groep. In die eerste fase van die studie is reël-tyd polimerase ketting reaksie (PKR) TaqMan© tegnologie gebruik om die potensiële waarde te bepaal om farmakogenetiese toetsing (CYP2D6 alleel 4) by te voeg tot standaard immunohistochemie (IHC)-gebasseerde borstumour subtipering wat gekomplementeer is deur BRCA mutasie sifting en/of mikroarray-geen profilering in toepaslike pasiënte. In fase twee van die studie is algemene genetiese risikofaktore vir kardiovaskulêre siekte (KVS) aangetoon om betekenisvol geassosieerd te wees met vroeër ouderdom (10 jaar gemiddeld) van borskanker aanvang/diagnose (APOE E4 alleel, p=0.003; 95% CI: 4-15) en liggaamsmassa indeks (BMI) (MTHFR 1298 A>C; p=0.01; 95% CI: 3-14) in pasiënte wat gestratifiseer is volgens estrogeen reseptor (ER) status, na aanpassing vir potensiële strengelveranderlikes. Ouderdom by diagnose/aanvang van borskanker was betekenisvol laer in ER-negatiewe as ER-positiewe pasiënte na statistiese aanpassing vir etnisiteit (p=0.022), terwyl BMI beduidend hoër was in ER-positiewe as ER-negatiewe borskanker pasiënte na aanpassing vir ouderdom, etnisiteit en familiegeskiedenis van kanker (p=0.035). Hierdie bevindinge het bygedra tot die ontwikkeling van ‘n eksoom pre-siftings algoritme (EPA) wat gebruik is in deel 3 van hierdie studie om drie geneties ongekarakteriseerde borskanker pasiente vir WES te selekteer. Die resultate is vergelyk met die bevindings in drie kankervrye kontroles uit dieselfde populasiegroep. WES, gevolg deur variant roeping met beide die standaard mensgenoom verwysingsvolgorde (hg19) en ‘n etnies verenigbare major alleel verwysings genoom (MARS) het meer as 20% teenstrydigheid getoon in die aantal geen variante wat geïdentifiseer is in dieselfde DNS monsters. Nadat ‘n groot aantal vals-positiewe resultate uitgesluit is wat deur minor allele in hg19 veroorsaak is, is twee raar missens mutasies (<1%) geïdentiseer in ‘n familie met ER-positiewe borskanker: RAD50 R385C en MUC1 Q67E. Drie verskillende bioinformatika programme is gebruik om funksionaliteit te voorspel en albei mutasies is bevestig deur Sanger volgorderbepaling en/of RT-PKR. Nie een van die twee missens mutasies in die RAD50 of MUC1 gene is gevind in die eksome van ‘n onverwante borskanker pasiënt met trippel-negatiewe borskanker of in die eksome van die drie kontrole indiwidue van dieselfde populasiegroep nie. Hierdie studie het gelei tot die ontwikkeling van 'n raamwerk vir WES wat n verbetering is op die beperkings van enkelgeen BRCA-toetsing van Suid-Afrikaanse borskankerpasiënte. Ons bevindinge ondersteun vorige WES-resultate wat aangedui het dat die meeste genetiesongekarakteriseerde familiere borskankers waarskynlik veroorsaak word deur 'n kombinasie van lae tot matige penetrasie-mutasies wat tot uitdrukking kom in 'n hoë-risiko omgewing soos gereflekteer word deur hoë BMI. WES maak dit moontlik om genetiese risikofaktore te identifiseer wat relevant is tot beide kankerontwikkeling en geteikende terapeutiese intervensie met dieselfde genetiese toets.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/98481
This item appears in the following collections: