Development of new tools for genomic biomarker investigations

During drug research and development, biomarkers are broadly used to improve the understanding of drug mechanism of action, to investigate drug efficacy and safety, to support the selection of target patient population and to optimize treatment schedule. Among different classes of biomarkers, genomic biomarkers are defined as a measurable DNA or RNA characteristics that are indicator of normal biologic processes, pathogenic processes, and/or response to therapeutic or other intervention. A genomic biomarker can consist, for example, in one or more DNA characteristics such as single nucleotide polymorphisms (SNPs), insertions, deletions or RNA characteristics such as RNA expression levels, RNA processing (splicing and editing) and microRNA levels. The present research aimed at developing new genomics-based tools using non-conventional biological samples that might support biomarker investigations in clinical settings. Concerning DNA biomarkers, Single Nucleotide Polymorphisms (SNP) analysis was validated on DNA extracted from Formalin-Fixed Paraffin-Embedded (FFPE) tissue, from Hematoxylin and Eosin (H&E) stained FFPE slides, and from serum samples. Unconventional samples represent a challenge for genetic analysis due to limitation of biological material and/or the poor quality of the DNA extracted. For example, due to fixation effect and formaldehyde interaction DNA extracted from FFPE samples are characterized by degradation, cross-link, limitation of material, methylol derivatives and PCR inhibitors presence. Therefore before analyzing these samples a method validation is necessary to prove data reliability in accordance with Regulatory Agencies guidelines that encourage the scientific community to perform a fit-for-purpose method validation to support any pharmacogenetic data submission. SNPs were investigated by Real-Time PCR using TaqMan SNP genotyping assays. Polymorphisms in a panel of genes involved in EGFR pathway, which is directly associated with many type of cancer, were evaluated. In particular, each single assay was first validated for accuracy, intra-assay precision (repeatability under the same operating conditions) and ruggedness (reproducibility with different operators, different batches). These parameters were tested first on good quality DNA such as DNA extracted from cell lines and then on real sample to evaluate the non-conventional matrix. On this purpose, the impact of fragmented DNA (FFPE samples) and H&E staining on FFPE samples was evaluated. DNA was extracted from different tissues of 10 commercial donors. DNA genotyping results of unstained FFPE and H&E staining were compared with the genotype obtained from high quality genomic DNA extracted from Fresh Frozen (FF) tissues obtained from the same donors (used as reference samples). Overall, these results demonstrate that SNP genotyping can be performed on archived FFPE tissues providing reliable results. As additional test serum was used as source of DNA to perform SNPs analyses. Serum is usually used to investigate protein biomarker and is generally collected in most of the clinical trials. It has been demonstrated indeed that free circulating DNA is present in serum: in particular, DNA is present in healthy individual at low concentration while levels are higher in cancer patients, in arthritis, hepatitis (Board et al., 2008; Gahana et al., 2008; Gormally et al., 2007). To validate SNPs analysis on serum, two aliquots of whole blood were obtained from 35 healthy volunteers. For each subject one aliquot was used to extract good quality DNA, the other was used to prepare serum prior to DNA extraction. As expected DNA quantity was very low for serum samples. As result, even though DNA was not degraded, genotype analysis was successful only on 70% of the samples. Overall, the validation conducted showed that serum could be used as source of biological material to conduct genetic analyses. However limitation of DNA does not consent to perform a large panel of analysis. These could be further explored in patients since circulating DNA is present at higher levels in several diseases. Part of the present thesis focused also on the validation of methods for KRAS mutation analysis. This gene encodes for a G-protein which plays a key role in the Ras/mitogen-activated protein kinase (MAPK) signaling pathway and located downstream Epidermal Growth Factor Receptor (EGFR) which is involved in colorectal cancer (CRC). KRAS status can predict which patients benefit (KRAS wild-type) or do not benefit (KRAS mutated) from anti-EGFR therapy. Since KRAS analysis is also used for diagnostic analysis, an accurate validation of the method was required. The aim was to compare and validate two different methods for KRAS mutation detection on FFPE tumor specimens, and on H&E stained FFPE which represent an unconventional source of samples for this type of analysis. In particular, DxS ThreraScreen KRAS mutation kit, a Real-Time PCR assay, was compared to the PyroMark KRAS Kit, based on pyrosequencing technology. The DxS ThreraScreen KRAS mutation test kit is able to detect 7 different mutations present in codons 12 and 13 of the KRAS gene while PyroMark KRAS Kit is able to detect 9 KRAS mutations in codon 12-13 and 5 mutations in codon 61. Results from validation showed that both Pyrosequencing assay both DxS ThreraScreen assay are accurate and reproducible. Moreover no impact of degraded DNA obtained from FFPE or influence due to H&E staining was observed in both methods. In conclusion PyroMark KRAS Kit showed advantages such as lower amount of DNA needed for analysis, detection of additional mutations in cod.12/13 and codon 61 than DxS TheraScreen KRAS kit; on the other hand DxS TheraScreen KRAS resulted more sensitive than pyrosequencing assays and less time consuming. This thesis focused also on establishing a simple method to perform gene expression investigation on hair follicles (HF) and to evaluate its applicability in clinical trials. Despite 80% of solid cancers arising from epithelial tissues, blood is still one of the most common peripheral tissues used for biomarkers and pharmacogenomic investigations in oncology. Hair follicles may offer a viable alternative since they can reflect biological response in epithelial tissue, they are easy to collect (non-invasive) and available from most individuals. After the establishment of sample collection and RNA extraction, HFs were collected from 23 health donors to evaluate inter-individual variability of RNA yield and quality. Gene expression analysis was then conducted on the extracted RNA. First it was evaluated a panel of 16 housekeeping genes to assess the feasibility of the analysis. Then it was shown that in HF a panel of epithelial specific genes were expressed. Indeed, Realtime PCR analyses showed that EGFR, Keratin 19 (KRT19), Collagen, Melan-A were expressed in HFs but not in RNA derived from blood. On the opposite, FPR1 and PRF1 genes were expressed only in blood. These results suggest that HF represents a valuable biological source to study pathways active in epithelial tissue. Finally, gene expression analysis was conducted on an in vivo experiment to evaluate if a response to treatment could be observed in HFs. In particular, PD markers of Interferon treatment were investigated after in vivo subministration of Interferon-beta (IFN-β) in Macaca fascicularis. The expression of the known IFN-β responding genes MxA was investigated both in blood and in anagen HFs. Results showed that MxA induction was observed both in blood and HF: gene induction in blood was observed at 6 hours after subministration while in HF at 24 hours probably due to a different IFN-β distribution. These data suggest that gene expression analysis can be carried out in HF samples. However, it is important to highlight that in HF the response had a lower degree of induction and higher variability than in blood. However this preliminary observations need to be further explored in pilot clinical studies to evaluate its applicability. Overall the validation of different genomic analysis on unconventional sampled opens the possibility to conduct biomarker investigations on several clinical trials conducted in the past or to plan new investigations with non invasive methods. In addition, from the deep evaluation of the current guidelines from the Regulatory Agencies (and from the open debate in the scientific community) a proper strategy to validate genomic analytical assays was proposed according to fit-for-purpose criteria.