MSACL 2017 EU Plenary Abstract

Proteogenomic Analysis of Clinical Samples Using a Unified Extraction Method

Jared Isaac
Thermo Fisher Scientific

Bio: My passion is innovative product development and commercialization of disruptive technologies in the biotechnology market to achieve strategic growth and transform patient care. I have progressed from an individual contributor, to project leader to project manager. I have completed technical milestones and managed projects and duties of greater responsibility. I completed my MBA to refine my competencies in people, financial and commercial management. My experience spans new product research and development, sustaining engineering, project management, marketing, product management, and business development. I have an optimistic attitude and I am eager to take on new responsibilities to ensure projects stay on track for product launch.

Authorship: Jared Isaac(1), Craig Dufresne(2), Mousumi Rath(3), Kirk Elliott(1), Robert Brown(1), Jennifer Freeland(1), Kristin Aguinaldo(4), Iris Casuga(5), Vinnay Mital(6), Haiping Liu(7), Ferda Filiz(7) Shayna Donoghue(1), Jason Lai(3), Stanley Hamilton(8), and Savitri Krishnamurthy(8)
Thermo Fisher Scientific, Molecular Pathology(1), Chromatography Mass Spectrometry(2) ,Genetic Sciences Division, (GSD)(3), Clinical Sequencing Division (CSD)(4-6), and Anatomical Pathology Division(7), and The University of Texas MD Anderson Cancer Center, Houston, TX(8)

Short Abstract

The Thermo Scientific™ KingFisher™ Flex Purification System and MagMAX FFPE DNA/RNA kit were optimized for DNA, RNA and protein extraction for proteogenomics in a Unified Method (UM). Fresh frozen and formalin fixed paraffin embedded tissues from breast and lung cancer tissues were fixed and processed and/or sectioned followed by nucleic acid and peptide extraction, QC analyses, and “omics” techniques. NGS and LC-MS were performed on UM biomolecules using panels of cancer biomarkers producing quantitative data. This study shows that Proteogenomics in Pathology can provide Quantitative data to supplement Qualitative interpretation and speculation.

Long Abstract

Introduction

Correlation of RNA expression and DNA mutations by Next Generation Sequencing (NGS) and protein expression by liquid chromatography mass spectrometry (LC-MS) from the same tissue section empowers clinical researchers with confirmatory, quantitative data. The MagMAX FFPE DNA/RNA kit (Thermo Fisher Scientific, Kalamazoo, MI) was modified for isolating DNA, RNA, and Proteins from a single FFPE section to analyze RNA expression, DNA mutations and protein expression from a single clinical sample. This study shows the capability of the KingFisher Flex to extract DNA, RNA and protein from clinical lung and breast cancer specimens using the Unified Method (UM) that were analyzed with Ion AmpliSeq NGS and Thermo Scientific LC-MS instruments. Analysis of breast cancer specimens showed potential false positive IHC scores that were clarified by NGS and LC-MS and is applicable to clinical breast cancer samples 20 months old. The KingFisher Flex and MagMAX FFPE DNA/RNA kit combined with targeted NGS and LC-MS provides cancer researchers a method to develop quantitative tests for validation in larger and older sample studies.

Methods

Sample Preparation

Fresh frozen (FF) human tissue samples were purchased from a third-party tissue biobank (Asterand Biosciences, Detroit, MI) for preparation of formaldehyde fixed paraffin embedded (FFPE) blocks. FF tissues were quickly cut into 3 cm3 sections on dry ice and fixed with 10% neutral buffered formalin (NBF) for 24 hours at room temperature while retaining one 3 cm3 section at -80°C as a FF control. After fixation, tissues were processed upon a Excelsior AS tissue processor (Thermo Fisher Scientific). FFPE BCa blocks of IDC subtype and matching FF tissues were collected from the MD Anderson Cancer Center (MDACC, Houston, TX) biobank of the following sample ages: 4, 6, 9 and 20 months.

UM combined NA and Protein Extraction

7 ìm curls of ~100 µm2 of FFPE or FF tissue were into Modified MagMAX lysis buffer, heated to 75°C with DTTl, alkylated with iodoacetamide at RT for 1 hour, diluted with 50mM Tris-HCl pH 8.0, 5mM CaCl2 and digested overnight with 20 µg MS grade trypsin (Thermo Fisher Scientific). DNA, RNA and peptides were isolated with a Kingfisher Flex magnetic bead processor (Thermo Fisher Scientific). Peptides were concentrated using a Savant 2120 Speedvac (Thermo Fisher Scientific) and reconstituted in 1 mL of 0.1% Formic Acid (FA) in MS grade water (Thermo Fisher Scientific). Peptides were purified by solid phase extraction using SOLA Cx 2mL 96 well SPE plates (Thermo Fisher Scientific) under vacuum pressure and resuspended in 50 µL of 0.1% FA and heated at 98°C for 2 minutes at 900 rpm in a ThermoMixer C (Eppendorf).

Genomics and Transcriptomics

Sequencing libraries were prepared using the Ion AmpliSeq Cancer Hotspot Panel v2 for DNA and Ion AmpliSeq RNA Cancer Panel for RNA using the Ion Chef system including automated library preparation, templating, and chip loading (Thermo Fisher Scientific). Sequencing was performed using Ion 540 chips with a read length of 200 bps upon an Ion S5 DNA sequencer (Thermo Fisher Scientific). Raw NGS files were imported as .vcf and .bam files into Ion Reporter software (Thermo Fisher Scientific). DNA mutations were reported as Total Variant Calls (TVC) and RNA expression were reported as reads per million (RPM).

Proteomics

To confirm transcriptomic data, a targeted LC-MS method and a selective FASTA protein primary sequence database was developed to quantitate endogenous peptides in FFPE cancer tissue. Known variants of the following proteins were selected for quantification: PDL-1, CTLA-4, EGFR, HER2, PI3K, KRAS, BRAF, AKT, MTOR, NFêB/p65/p100/RelB, MYC, ERá, PR, HER3, MEK1, 4EBP1, MET, FGFR1/2/4, AR, RhoA/C, ERK, MAPK, Ki67, PTEN, Caspase 3/6/8/9, BCL2, and S6K.

Targeted LC-MS PRM Method Development

Purified endogenous and recombinant proteins were purchased (Origene), tryptic digested and injected singly into a Q Exactive Plus Orbitrap Mass Spectrometer (OMS) in 2%ACN/0.1%FA/98%H2O into an EASY-nLC 1200 UHPLC system with a PEPMAP100 C18 2 cm long 75ìm ID 3ìm particle size nanoViper trap followed by an EASY-Spray C18 75ìm ID 2ìm particle size 25 cm column (Thermo Fisher Scientific). Discovery data were acquired in 90 minute runs with data dependent analysis (DDA) mode on an OMS with the following settings: ITmax= 100 ms, AGC=1e6, 70,000 resolution, and a mass range of 300-2000 m/z.  Proteins were identified from. raw files using Proteome Discoverer 3.1 (Thermo Fisher Scientific) to search against the limited FASTA database with a (FDR) of <1% and rated with High, Medium, or Low confidence criteria for protein identification. Heavy labeled peptides were prepared according to peptide sequences unique to each of their corresponding target proteins. With the identification of 3-5 high confidence peptides per target protein, a targeted Parallel Reaction Monitoring (PRM) quantitative LC-MS method was developed.

Relative and Absolute Peptide Quantification

Two methods were developed to quantify endogenous proteins in FFPE tissues. The first method involved quantification of target proteins against housekeeping proteins by comparing the average of target peptide peak areas to the average of housekeeping protein peak areas which were selected by ranking highest delta score (ÄCn) then Xcorr values. The second method involved isotopically labeled peptide (ILP) standards for each target quantification peptide. Precedence was given to peptides containing pathological mutations or sites of post-translational modification, and selected peptides for each protein were confirmed in FFPE BCa of IDC subtype prior to manufacture of synthetic isotopically labeled AQUA Ultimate (Thermo Fisher Scientific) peptides (ILP).. ILP standards of known concentration were combined with FFPE sample peptides injected into an OMS and target proteins were quantified.

Results

Figure 1. UM NA Yield from BCa and LCa FFPE Tissues

BCa and LCa 7 µm FFPE sections were run in triplicate through the UM. DNA (A) and RNA (B) concentration were quantitated using the Qubit 3.0 flourometer. N=9.

Table 1. Indominus LC-MS Panel (IP)

List of Proteins Targeted in the Indominus LC-MS Panel (IP). Proteogenomic overlap between the Cancer 50 DNA and RNA panels with the IP are as follows: AKT, BRAF, EGFR, HER2, FGFR, KRAS, MET, PI3K, PTEN.

Figure 2. Protein and Peptides Identified Using the UM with the IP

A representative chromatogram (A) and MS1 (B) and MS2 (C) spectra from peptides generated by the UM LC-MS method. NL values were as follows: Chromatogram total ion count (TIC)= 2e9; MS1= 4E7; MS2=1E5. Using the IP library of 30 target proteins, fresh frozen (FF) and (FFPE) samples were run through the UM LC-MS method and Venn diagrams show the overlap of 86% proteins (D) and 74% peptides (E) found in common between both sample types, respectively.

Figure 3. UM MS2 Spectra of IP proteins

OMS MS2 spectra generated by UM BCa samples showing KRAS, EGFR, HER2 and HGFR (MET) peptide fragmentation. OMS data were searched using Proteome Discoverer with a false discovery rate (FDR) of <0.1% and rated with High, Medium, or Low confidence criteria for identification of EGFR, HER2, KRAS and HGFR peptide mass spectra.

Table 2. Proteogenomic Analysis of BCa B

Nine individual successive curls were processed through the UM, analyzed by NGS and LC-MS, and four sections were stained by IHC and scored by a staff pathologist. Data are averages presented as absence or presence, N or Y, of any DNA variants, relative RNA expression by transcripts per million (TPM) to HER2 RPM of BCa A, and relative Protein quantitation to TubA. (-) indicates IHC was not performed for these protein biomarkers.

Figure 4. LC-MS Relative Quantitation of Biomarkers in BCa

Protein biomarkers from three cases of infiltrating ductal carcinoma of the breast (BCa) were analyzed by LC-MS relative quantitation to TubA average peak areas set as 1.

Figure 5. UM vs. IHC: HER2

Three patient cases of infiltrating ductal carcinoma of the breast (IDC) were analyzed by immunohistochemistry and LC-MS and RNA NGS for quantitation of HER2. RNA expression is shown as reads/million base pairs (RPM), IHC stained tissues were scored by a staff pathologist, and protein level was calculated as relative expression to TubA.

Figure 6. Yield of Nucleic Acids Extracted from FFPE Lung and Breast Tissues

Using the UM, DNA (A) and RNA (B) species were extracted from lung and breast tissues using the UM and quantified fluorometrically, n=9/bar.

Figure 7. UM RNA QC Metrics and DNA Variants.

RNA collected from BCa FFPE tissue by the UM were sequenced and quality control metrics are reported for Average Mapped Reads (A), % Targets Detected (B), and % Valid Reads (C) from FFPE breast tissues up to 20 months of age. Unfiltered average number of detected DNA variants per sample (D) are reported from FFPE breast tissues up to 20 months of age, n=9/bar.

Table 3. UM Absolute Protein Quantification of BCa Time Course.

The IP quantified the level of eight biomarkers using the UM and LC-MS. Protein concentration was determined using absolute quantitation to HLP standards in fmol/µL.

Table 4. UM Proteogenomics of BCa time course for FGFR1.

Nine individual successive curls each were processed through the UM for 6, 9, and 20-month-old BCa FFPE samples. Data are averages presented as absence or presence, N or Y, of any DNA variants, relative RNA expression by transcripts per million (TPM), and absolute quantitation to HLP standards in fmol/µL, n=9/bar.

Conclusions & Discussion

Reflex testing using FISH and IHC is common practice to diagnose difficult cancer cases (1). However, when biopsy samples are limited and gene amplification status is equivocal a second surgery may be needed causing delayed time to diagnosis (2). Failure to diagnose quickly leads to increased workload on technicians and pathologists, cost for healthcare providers, and physical and psychological toll on patients. Gene sequencing provides DNA mutation information of whether patients may be respond to targeted therapy, however this correlation is not direct as the protein targets may or not be present (3). Most recently approved targeted cancer therapies are antibodies that interact with proteins thus making the correlation between gene status and effector proteins imperative (4).

The current state of clinical proteogenomics requires confirmatory testing between in-situ hybridization (ISH) and IHC which have been static from the 1990s. These methods are limited in that they require visual examination and individual slides are needed per technique and biomarker. Proteogenomics using NGS and LC-MS with the UM enables clinicians to multiplex biomarker quantification using less patient sample. This study serves as a baseline for future work to expand the use of the UM combined with proteogenomics in larger sample cohorts with longer storage times.

References

Owens V., et al. HER2 Amplification Ratios by Fluorescence In Situ Hybridization and Correlation with Immunohistochemistry in a Cohort of 6556 Breast Cancer Tissues. Clinical Breast Cancer. 2004

Tsuda H., et al. HER2 testing on core needle biopsy specimens from primary breast cancers: interobserver reproducibility and concordance with surgically resected specimens BMC Cancer 2010

Lampros D., et al. Onco-proteogenomics: Multi-omics level data integration for accurate phenotype prediction. Critical Reviews in Clinical Laboratory Sciences 2017

Beck A., et al. Strategies and challenges for the next generation of antibody–drug conjugates Nature Reviews Drug Discovery 2017

David Sarracino9, Scott Peterman9, Mark Shannon3, Kate Rhodes4, Michael Hogan6, Seth Sadis6, Sharmini Muralitharan7, Colin Toombs1, Brad Hart2, Shen Luan9, Jon Ferguson2, Mazi Mohiuddin9, Junko Stevens3, Angie Cheng10, and Amol Prakash9

Chromatography Mass Spectrometry Division, Cambridge, MA9

GSD, Austin, TX10


Financial Disclosure

DescriptionY/NSource
Grantsno
SalaryyesThermo Fisher Scientific
Board Memberno
Stockyes Thermo Fisher Scientific
Expensesno

IP Royalty: no

Planning to mention or discuss specific products or technology of the company(ies) listed above:

yes