Affymetrix Gene Expression Microarrays Application to Pulmonary Arterial

Affymetrix Gene Expression Microarrays Application to Pulmonary Arterial Hypertension Bob Stearman 02/24/2014

Pulmonary Arterial Hypertension (PAH) • Rare Disease: USA ~1000 new cases/yr • More Common in Females (2: 1) with Earlier Presentation (30 s-40 s vs 40 s-50 s) • BMPR 2 Mutations Found in HPAH with Incomplete Penetrance • Combination Drug Therapy Common • • PDE 5 and Endothelin Inhibitors plus Prostacyclin Slows Progression But No Cure (5 -yr 66%) • Elevations of Pulmonary Hemodynamics

Pulmonary Arterial Hypertension (PAH) • Hyperproliferation of PAECs and SMCs • Structural Changes in Lung Vasculature Vessel Thickening Muscularization (SMCs) Plexiform Lesion (ECs) Stacher et al. , AJRCCM 186, 261 (2012)

Pulmonary Arterial Hypertension (PAH) • Elevation in Pulmonary Hemodynamics • Echocardiogram and Right Heart Catherization Studies • Increased Inflammation Vessel Thickening vs m. PAP (p-value 0. 03; r=0. 27) Perivascular Inflammation Stacher et al. , AJRCCM 186, 261 (2012)

Pulmonary Hypertension Breakthrough Initiative • Multi-Center Lung Tissue Procurement • 9 Geographically Spread Procurement Centers • • “Excessively” Complex Tissue Dissection 3+ Hours To Carryout • 1 Central Tissue Repository • Genomics Core & Secondary Tissue Repository (Frozen and RNALater) • Rare Disease = Slow Recruitment

PHBI Clinical Information Stacher et al. , AJRCCM 186, 261 (2012)

PHBI Pathological Information • Vessel Thickness (measured and fractional) • Total, Intima, Media, Internal Diameter • Plexiform Lesions • Number and Density • Inflammation • Thrombi • Pulmonary Vein Remodeling Stacher et al. , AJRCCM 186, 261 (2012)

PHBI Lung Tissue Harvest RUL = right upper lobe RML = right middle lobe RLL = right lower lobe H= Hilum. Stacher et al. , AJRCCM 186, 261 (2012)

Pulmonary Hypertension Breakthrough Initiative: Sample Summary • PAH Patients: End-stage disease getting transplant • 4 General Types of PAH: HPAH, IPAH, APAH, Other • Failed Donors (FDs; controls): No match for lungs but available for research • Slow Recruitment: MAs run in 3 “annual” batches • PAH: Female Bias (2: 1 F: M) and older • FDs: Male Bias (1: 2 F: M) and younger compared to PAH

Workflow of Gene Expression Biological question Experimental design QC Tissue / sample preparation Extraction of Total RNA QC Probe amplification & labeling QC Microarray hybridization & processing Image analysis Data analysis Biological Verification QC QC Expression measures - Normalization Reddy Gali, Ph. D. Statistical Filtering - Clustering Microarray Analysis Using Pathway analysis R/Bioconductor

PHBI: Experimental Design • PAH Patients vs. FD • • • Disease-specific gene expression patterns? 4 General Types of PAH: HPAH, IPAH, APAH, Other Will Gene Expression Patterns Detect Clinical Subclasses? • PAH Patient Gender Bias • PAH vs FD Gender and Age (? ) Bias • Microarrays run in 3 “annual” batches • • Different lots of arrays and reagents Changes in labeling methodology

PHBI Lung RNA Samples (in RNALater) Are Variable

PHBI: Normalization Uncorrected PHBI Data

PHBI: Normalization Corrected PHBI Data By Batch and Gender

PHBI: Normalization Corrected PHBI Data By Batch and Gender Low Variance Genes Removed (33 K->6 K)

PHBI: Clustering

PHBI: Expression Correlation With Clinical Parameters

PHBI: Expression Correlation With Clinical Parameters

Sources of Reproducibility (Experimental Design- Randomization) C 4 Hybridized Sample B & A 1 (Day 1); A 6 & A 15 (Day 2) [while same day samples had very similar #DE genes as C 1, C 2, and C 3] C 5 Hybridized Samples B & A 1 (Day 1); A 6 (Day 2); A 15 (Day 3) C 2 and C 3 Stored Male Arrays ON, 4 o. C while Female Arrays were Wash/Stain (Yang et al 2008)

Sources of Reproducibility (Experimental Design- Randomization) Each dataset is self-consistent Training on Center 3 generally groups M and F by sample type Training on Center 4 produces poor correlation between M and F by sample type (Yang et al 2008)

Stability of Results Critical Review of Published Microarray Studies for Cancer Outcome and Guidelines on Statistical Analysis and Reporting Alain Dupuy , Richard M. Simon (JNCI 2007) “Simon Says Top 40”

Nature Biotechnology 28, 827 (2010)

Affymetrix Gene Expression Microarrays Report Expectations Experimental Design and Questions Summary Signal Intensities Before and After Normalization PCA plots Significant Genes Listing including FC, p-value, FDR analysis • Cluster Diagram (+/- Signal Intensities) • Over-represented Gene Ontology Categories • • • RDAVID? ? Experimental Conclusions and Discussion

Affymetrix Gene Expression Microarrays Report Expectations