quantitative cancer genomic analysis using droplet digital...
TRANSCRIPT
RainDance Technologies
Quantitative Cancer Genomic Analysis Using Droplet Digital PCR: Examples with Solid Tumors and Liquid Biopsies in Glioma, Breast, and Colon Cancer plus miRNA & RNA-Direct
Michael Samuels1, Leonora Balaj2, Xandra Breakefield2, Julia Beaver3, Ben Ho Park3, Manuel Krispin4, Saumya Das5, Valerie Taly6 , Pierre Laurent-Puig6 1RainDance Technologies, 2Massachusetts General Hospital, Boston MA, 3Johns Hopkins University, Baltimore MD, 4Zymo Research, Los Angeles CA, 5Beth Israel Deaconess, Boston MA, 6University Paris Descartes, FR
• High sensitivity
• Multiplex analysis
• Single pipetting step
www.raindancetech.com
Wide Dynamic Range With High Precision
Digital RNA Counting: 1-Step RT-dPCR Negative Control
Total Human RNA (1.04 ng)
GA
PD
H M
ole
cu
les C
ou
nte
d
Human Total RNA Input (ng)
Log Plot
1
10
100
1000
10000
100000
1000000
10000000
0.001 0.01 0.1 1 10 100 1000
• 4 different 1-Step RT-PCR Kits worked ‘right out of the box’
• mRNA and viral RNA demonstrated • Broad Dynamic Range (6 logs)
• Highly precise (%CV < 5% for >500 molecules per sample)
• Multiplexing capability demonstrated • True single RNA molecule counting enabled by droplet numbers
Sample #22
Negative for IDH1 mutation
WT
Dual
G395A
• One-Step qMethyl kit from Zymo Research used without optimization
• Kit Methylation Sensitive Restriction Enzymes eliminate the need for bisulfite treatment of DNA
• Linear counting of differences of 5-10% methylation across entire methylation range demonstrated
• Nanograms of input can be used directly without bisulfite treatment or additional purification steps
• Multiplex analysis works well for methylation determination using digital One Step qMethyl kit
Digital PCR Quantification of Breast Tumor Samples Using Normalized Duplex Assays and One Step qMethyl Kit
• Duplex assays using a methylation-independent REF assay enables normalized quantification
• Tumor CCDN2 methylation shows early stage increases; RARB methylation shows less change
• Digital quantification results confirmed by qPCR (data not shown) and consistent with literature
• Multiplexing of assays for methylation analysis demonstrated
4-plex Assay
CCDN2
RAB25
RARB
MGMT
STAGE II STAGE IV
Assay VIC/FAM REF/CCDN2 REF/CCDN2 REF/CCDN2 REF/CCDN2
Template Tumor Control Tumor Control
ng Input 20 20 20 20
ul Input 25 25 25 25
# Droplets 2821297 2695102 2777435 2459884
# NEG 2802492 2682575 2768332 2443456
# REF 6424 8066 7339 12673
# CCDN2 720 61 96 28
%CCDN2/REF 11.2 0.8 1.3 0.2
%RARB/REF 14.4 6.9 10.8 6.1
Cyclin D2
Target
RAR B
Target
Target
REF
Methylation Independent
Reference
Methylation
Target
Digital Methylation Analysis of Breast Cancer
Digital PCR of Glioma Spinal Fluid Exosomes
Digital PCR Analysis of Breast Cancer
Tumor #3 Tumor #4 Tumor #28 Positive Control
H1047 WT
E545K
H1047R
H1047 WT
E545K
H1047R
H1047 WT
E545K
H1047R
H1047 WT
E545K
H1047R
Precise Small-Fold-Change Measurements
Multiplexed 1-Step RT-dPCR Titration
Digital PCR of Colorectal Cancer using Plasma
Submitted for publication-Portions presented at ASCO (June 2013)
Detection of Cancer Specific Mutations in Plasma of Early Stage Breast Cancer Patients
Abstract
Sequencing of tumors identified seven PIK3CA exon 20 mutations (H1047R) and three exon 9 mutations (E545K). Analysis of
tumors by ddPCR confirmed these mutations and identified five additional mutations. Pre-surgery plasma samples (n=29)
were then analyzed for PIK3CA mutations using ddPCR. Of the fifteen PIK3CA mutations detected in tumor tissues by
ddPCR, fourteen of the corresponding mutations were detected in pre-surgical ptDNA specimens, while no mutations were
found in plasma from patients with PIK3CA wild type tumors (sensitivity 93.3%, specificity 100%). Ten patients with pre-
surgery mutation positive ptDNA had ddPCR analysis of post-surgery plasma, which identified five patients with detectable
ptDNA post-surgery.
Julia A. Beaver MD*1, Danijela Jelovac MD*1, Sasidharan Balukrishna MD2, Rory Cochran BS1, Sarah Croessmann BS1, Daniel J. Zabransky BS1, Hong Yuen Wong BS1,
Patricia Valda Toro BS1, Justin Cidado BS1, Brian G. Blair PhD1, David Chu BS1, Timothy Burns MD PhD3, Michaela J. Higgins MB BCh MD4, Vered Stearns MD1, Lisa
Jacobs MD1, Mehran Habibi MD1, Julie Lange MD1, Paula J. Hurley PhD1, Josh Lauring MD PhD1, Dustin VanDenBerg BS1, Jill Kessler BS1, Stacie Jeter BS1, Michael L.
Samuels PhD5, Dianna Maar PhD6, Leslie Cope PhD1, Ashley Cimino-Mathews MD1, Pedram Argani MD1, Antonio C. Wolff MD1¥ and Ben H. Park MD PhD1¥
1The Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins 1650 Orleans Street Baltimore, MD 21287 2Christian Medical College Vellore Tamil Nadu, India 632004 3University of Pittsburgh
Hillman Cancer Center Research Pavilion 5117 Centre Avenue Pittsburgh, PA 15213-1863 4Massachusetts General Hospital 55 Fruit Street Boston, MA 02114-2696 5RainDance Technologies, 749
Middlesex Turnpike Billerica, MA 01821 6Bio-Rad Laboratories, Digital Biology Center 7068 Koll Center Pkwy, Suite 401 Pleasanton, CA 94566
RainDrop dPCR Platform
• Contamination-free design
• Simple and flexible workflow
• Robust open reagent platform
Example RainDrop dPCR Data
IDH1 Duplex Analysis
WT
G395A
Sample #11
Positive for IDH1 mutation
Example RainDrop dPCR Data: PIK3CA Triplex Analysis
Example Multiplex Panels:
KRAS Codon 12 and 13
KRAS Panel#1: WT+3 MUTs
G12R
G13D
W T
PCR (-)
G12D
KRAS Panel#2: WT+4 MUTs
G12S
G12C
G12A
WT
G12V PCR (-)
Source Sense Disposable
Chips
*Xeno RNA is a synthetic template spiked-in at a known concentration
R² = 0.9981
1
2
3
4
5
6
7
1 3 5 7
Lo
g o
bs
erv
ed
co
un
ts
Log expected input
Abstract Picoliter droplet digital PCR was used in separate studies for quantification of mutant IDH1 mRNA in glioma patient
cerebrospinal fluid extracellular vesicles, PIK3CA mutations and methylation of CyclinB2 and Retinoic Acid Receptor
promoters in breast tumors, KRAS and BRAF mutations in colorectal cancer plasma, and multiplexed miRNA
biomarkers from plasma. In addition, we show quantification of RNA molecules directly loaded on the RainDrop
provides highly precise multiplex One-Step RT-PCR measurements across a wide dynamic range.
RainStormTM Droplet Microfluidics
Divide and Count: Single volume divided into countable volume elements
RainDropTM Droplet Digital PCR
Rapid and reproducible processing of millions of reactions is enabled by replacing traditional assay plates and automation
systems with microscopic droplets and disposable fluidic chips. Aqueous samples (beads, cells, enzymes, antibodies, DNA)
can be encapsulated within each droplet, surrounded by an immiscible carrier oil. The droplets are stabilized with bio-
compatible surfactants, allowing for robust manipulations both on and off chip. Droplet fluorescence can be measured by
flowing the droplets through a laser spot positioned in the microfluidic channel.
oil
aqueous
Droplet Fluorescence Readout Droplet Generation
Laser spot
Droplets Flowing In Oil
B: Multiplex with intensity
Different intensity for different targets
Target 1 Target 2
A: Multiplex with color
Different color for different targets
Target 1 Target 2
Probe Concentration sets
Endpoint Fluorescence
oil
Droplet Schematic
surfactant
molecules
fluorocarbon oil
exterior
DNA/RNA Protein/
Antibodies
Single cells
aqueous
interior
Droplets Stable for Off-Chip Collection,
Incubation and Re-injection
Digital PCR with droplet microfluidics. A) Sample containing Target nucleic acids is mixed with assay reagents in 50ul; B) A
microfluidic device is used to divide the sample with assays into 10 million discrete 5 pl droplets such that only a single
target molecule is present in any droplet; C) Hydrolysis of the assay probe during PCR amplification makes droplets
containing specific sequences fluorescent; D) The fluorescence signal intensity is measured as droplets pass one at a time
through a laser spot positioned in a microfluidic channel on the readout chip.
No Target PCR-
droplet
Divide & Collect
PCR Amplification
Count
Background
Target
Sample+ Assay
PCR+ “bright”
droplets
PCR- “dark”
droplets
Target PCR+
droplet A B C D
Single molecule endpoint PCR enables easy multiplexing
Digital Multiplex Analysis With Endpoint PCR
Multiplexing enabled by creating ‘digital’ partitions containing either single target molecules or no targets. A) Each Target
molecule is assayed with a different ‘color’; B) Each Target molecule is assayed with a different ‘endpoint intensity’, with the
fluorescence at PCR endpoint determined by the probe concentration added for each target type (see plot); C) Multiplex
analysis of multiple target types in every sample is performed by combining assays based on color and/or intensity of the
added probes (e.g. Target 1 and 2 use different FAM probe concentrations, Target 3 uses a VIC probe only, Targets 4 and 5
use mixtures of FAM and VIC probes for each target, with the Target 5 probe mixture weighted more with FAM than Target 4).
Data is presented in a 2-D “Cluster Plot” of fluorescence intensity (VIC y-axis; FAM x-axis) with gates used to count droplets.
FAM Intensity
VIC
In
ten
sit
y
C: Multiplex with color and intensity
“Cluster Plot”
EvaGreen® Assays on the RainDrop System
RNA Dilution Series Shows Linearity with 3 Endogenous Targets and Xeno Control
Human Total RNA Input (ng)
Targ
et
Mo
lecu
les
R² = 0.9993
0
5000
10000
15000
20000
25000
30000
35000
0.00 5.00 10.00 15.00
R² = 0.9991
0
5000
10000
15000
20000
25000
30000
35000
0.00 5.00 10.00 15.00
R² = 0.9973
0
500
1000
1500
2000
2500
3000
0.00 5.00 10.00 15.00
POLR2A GAPDH PPI
Human Total RNA Input (ng)
Targ
et
Mo
lecu
les
Human Total RNA Input (ng)
Targ
et
Mo
lecu
les
Intercalating dye-based assay worked on the RainDrop “off the shelf”
Chen and Balaj, et.al. Molecular Therapy—Nucleic Acids (2013) 2, e109; doi:10.1038/mtna.2013.28
BEAMing and Droplet Digital PCR Analysis of Mutant IDH1 mRNA in
Glioma Patient Serum and Cerebrospinal Fluid Extracellular Vesicles
Abstract
Development of biofluid-based molecular diagnostic tests for cancer is an important step
towards tumor characterization and real-time monitoring in a minimally invasive fashion.
Extracellular vesicles (EVs) are released from tumor cells into body fluids and can provide a
powerful platform for tumor biomarkers because they carry tumor proteins and nucleic acids.
Detecting rare point mutations in the background of wild-type sequences in biofluids such as
blood and cerebrospinal fluid (CSF) remains a major challenge. Techniques such as BEAMing
(beads, emulsion, amplification, magnetics) PCR and droplet digital PCR (ddPCR) are
substantially more sensitive than many other assays for mutant sequence detection. Here, we
describe a novel approach that combines biofluid EV RNA and BEAMing RT-PCR (EV-
BEAMing), as well droplet digital PCR to interrogate mutations from glioma tumors. EVs from
CSF of patients with glioma were shown to contain mutant IDH1 transcripts, and we were
able to reliably detect and quantify mutant and wild-type IDH1 RNA transcripts in CSF of
patients with gliomas. EV-BEAMing and EV-ddPCR represent a valuable new strategy for
cancer diagnostics, which can be applied to a variety of biofluids and neoplasms.
Digital Liquid Biopsy of Cancer using Urine*
Taly V , et. al. Clin Chem. 2013 Aug 12. [Epub ahead of print]
Multiplex Picodroplet Digital PCR to Detect KRAS Mutations in
Circulating DNA from the Plasma of Colorectal Cancer Patients.
Abstract
BACKGROUND:Multiplex digital PCR (dPCR) enables noninvasive and sensitive detection of
circulating tumor DNA with performance unachievable by current molecular-detection
approaches. Furthermore, picodroplet dPCR facilitates simultaneous screening for multiple
mutations from the same sample.METHODS: We investigated the utility of multiplex dPCR to
screen for the 7 most common mutations in codons 12 and 13 of the KRAS (Kirsten rat
sarcoma viral oncogene homolog) oncogene from plasma samples of patients with metastatic
colorectal cancer. Fifty plasma samples were tested from patients for whom the primary tumor
biopsy tissue DNA had been characterized by quantitative PCR.RESULTS: Tumor
characterization revealed that 19 patient tumors had KRAS mutations. Multiplex dPCR analysis
of the plasma DNA prepared from these samples identified 14 samples that matched the
mutation identified in the tumor, 1 sample contained a different KRAS mutation, and 4 samples
had no detectable mutation. Among the tumors samples that were wild type for KRAS, 2 KRAS
mutations were identified in the corresponding plasma samples. Duplex dPCR (i.e., wild-type
and single-mutation assay) was also used to analyze plasma samples from patients with
KRAS-mutated tumors and 5 samples expected to contain the BRAF (v-raf murine sarcoma
viral oncogene homolog B) V600E mutation. The results for the duplex analysis matched those
for the multiplex analysis for KRAS-mutated samples and, owing to its higher sensitivity,
enabled detection of 2 additional samples with low levels of KRAS-mutated DNA. All 5 samples
with BRAF mutations were detected.CONCLUSIONS: This work demonstrates the clinical utility
of multiplex dPCR to screen for multiple mutations simultaneously with a sensitivity sufficient to
detect mutations in circulating DNA obtained by noninvasive blood collection.
Normal Patient Plasma Affected Patient Plasma
miRNA #1 2504 molecules
miRNA #2 1011 molecules
PCR - PCR -
Multiplexed FAM-Probes for Counting Cardiomyopathy-specific Plasma miRNA*
Digital PCR of Plasma miRNA Biomarkers
Rare Mutation Detection
Viral Load
Aneuploidy Detection
Copy Number Variation
Applications
Methylation Quantification
Many Others
miRNA Biomarker Counting
RNA Counting
Xeno
PPI
Neg
GAPDH
POLR2A
Xeno
GAPDH Neg
Xeno
GAPDH Neg
R² = 0.9999
0
10000
20000
30000
40000
50000
60000
70000
80000
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00
%CV = 1.3
1.1 fold change
%CV = 1.5
1.2 fold change
GA
PD
H M
ole
cu
les C
ou
nte
d
Human Total RNA Input (ng)
Total Human RNA (1.04 ng)
Predesigned TaqMan® MGB gene expression assays – off the shelf!
• Data shows SMN assay primers with EvaGreen detection
• Highly precise quantification (%CV from 0.9 - 8%)
• Over 3 logs of dynamic range
• Demonstrates miRNA biomarker counting from plasma
• Duplexed FAM assays (TaqMan)
• cDNA counts agree with miRNA inputs
*Information reproduced from Trovagene Corporate Slide Deck
Filip Janku MD PhD; MD Anderson Cancer Center
1 Janku et al, AACR-NCI-EROTC International Conference, 2013
*1
• EvaGreen dye purchased from Biotium
CV = 1.5%
CV = 3.3%
CV = 7.6%
Linear Plot
• miRNA cDNA analyzed
*Collaboration with Dr. Saumya Das, Beth Israel Deaconess Medical Center