Proteomic Profiling of Prostate Cancer Plasma Specimens Using Proteograph and TIMS Technology
Matthew E.K. Chang1, Jessie May Cartier1, Ryan W. Benz2, Max R. Mahoney2, Michael Krawitzky3, Ryan Kopp1, Mark R. Flory1
1Oregon Health and Science University (OHSU), Knight Cancer Institute (KCI), Cancer Early Detection Advanced Research Center (CEDAR); 2720 SW Moody Avenue, Portland, OR 97201 2Seer Inc., 3800 Bridge Parkway, Suite 102, Redwood City, CA 94065
3Bruker Daltonics, 61 Daggett Drive, San Jose, CA 95134
OVERVIEW
The Seer Proteograph platform affords a unique combination of deep proteomic sampling and study scalability, a breakthrough development for proteomic biomarker discovery studies.
A proof-of-concept pilot study was initiated on 32 prostate cancer plasma specimens retrospectively collected from patients with high and low tumor grades.
While admittedly underpowered for biomarker discovery, this study provided results motivating multiple scaled biomarker discovery studies currently in progress.
SPECIMENS AND WORKFLOW
Existing double-spun prostate cancer patient plasma selected from available set in our internal biorepository. A small set (n = 32, 16 per group) were selected as “high” or “low” grade based on several factors and under the consult of Dr. Ryan Kopp.
OHSU CEDAR specimen biorepository sourcing from OHSU and VA hospitals
“Low-Grade” Study Arm, n =16 “High-Grade” Study Arm, n =16
ProteographTM Product Suite
1 hr setup
7 hr automated workflow
16 experimental samples per Proteograph plate + multiple internal controls
1
1
• Benign prostate hyperplasia
• Prostatic intraepithelial neoplasia • Age range 52-74 (avg. 66)
• Adenocarcinoma
• Gleason grade ≥ 4+3
• Age range 56-72 (avg. 67)
Plasma 250 uL per sample
On-bead digestion
1 2 3 4 5
Peptide Preparation
• Manual fluorescence peptide assay (Pierce/ThermoScientific)
• Standard mixture of internal peptide standards (PepCal, SCIEX)
timsTOF Pro LC-MS Analysis
• nanoELUTE LC (Bruker)
• CaptiveSpray source (Bruker) • Aurora C18 column, 25-cm
(IonOpticks)
• nanoEase M/Z Symmetry
C18 trap (Waters)
• DDA-PASEF acquisition mode
Protein Corona
Data Analysis
• MaxQuant v1.6.17.0 • Reference proteome:
Homo sapiens, Aug 2019 • Peptide and protein FDR
for identification at 1%. • Output analyzed in R.
Proteins
5-nanoparticle (NP) panel for protein enrichment
DYNAMIC RANGE
1689
0
-10
-20
OT Score
0.5 0.4 0.3 0.2 0.1
0 2000 Rank 4000
Measurement Depth in Plasma
Identified proteins spanned a wide dynamic range across the plasma proteome as displayed with reference to normalized intensities from Keshishian et al.2 The mean rank of mapped proteins is displayed as a vertical line.
Proteins previously demonstrated as associated with prostate cancer based on the Open Targets (OT) Platform are colored according to their OT score. Points colored white mapped back to the plasma proteome reference but did not match the OT query.
PROTEOMIC COVERAGE
1500
1000
500
0
15000
10000
5000
0
750 500
Status
250
High-Grade Low-Grade
0
All
39.1
35.6
47.4
45.2
4000
3000
2000
Status
All
1000
High-Grade Low-Grade
0
High-Grade Low-Grade Combined
Protein and Peptide Identification Rates and Variability
Protein group identifications per sample are displayed in the top left panel with an average of 947 in high-grade and 960 in low-grade samples. Variation in protein intensities for both groups is shown in the adjacent top right panel with median CV displayed.
Peptide identifications per sample are displayed in the lower left panel with an average of 6960 in high-grade and 7047 in low-grade samples. Variation in peptide intensities for both groups is shown in the adjacent lower right panel along with median CV.
Principal Component Analysis
Samples clustered based on protein intensities with the exception of three potential outliers that were not removed from the overall analysis here.
0.50 0.25 0.00
-0.25
Status
High-Grade Low-Grade
0.165 0.170 PC1 0.175
0.180
STUDY POWER ESTIMATE
Assumes 80% power and Bonferroni correction for significance (0.05/n = 1,199 )
200
100
0
20%
50% 100% 200%
284
49
16
7
Percent Change in log2(Intensity)
Total features includes features present in at least 25% of at least one study class
Statistical Power Curve and Sample Size Estimation
Relationship between increasing sample number and ability to detect more subtle fold changes across study arms within the context of this assay.
CURRENT WORK
PILOT
SCALE UP
BIOPSY
MRI
?
STAGE
TREATMENT
CLINICAL CONTEXT
SCREENING
RECURRENCE
Future work includes highly scaled studies in solid and liquid tumor indications
• Careful study contextualization based on clinical need.
• Large sample numbers for reducing patient-patient variance.
• PRoBE biomarker discovery framework3 for improved study design.
Evolving improvements in our workflow
• Exploration of alternative specimen types (e.g. serum).
• diaPASEF (data-independent acquisition PASEF) for increased sampling depth, reproducibility,
and lower acquisition times.
• Broad survey of analytical column types for optimized blend of sensitivity and ruggedness. • Post-acquisition processing using a variety of spectral library-based and library-free
approaches with DIA-NN4 and Seer Proteograph Analysis Suite (PAS).
References
1. Nickel, J. Curtis. “Prostate Inflammation and Prostate Cancer: What do I need to know?” January 27, 2018. Accessed Oct 2021. https://grandroundsinurology.com/Prostate-Inflammation-and-Prostate-Cancer/
2. Keshishian, H. et al. Quantitative, multiplexed workflow for deep analysis of human blood plasma and biomarker discovery by mass spectrometry. Nat. Protoc. 12, 1683–1701 (2017).
3. Pepe, M. S., Li, C. I., & Feng, Z. Improving the Quality of Biomarker Discovery Research: The Right Samples and Enough of Them. Cancer Epidemiol Biomarkers Prev. 24(6), 944-950 (2015).
4. Demichev et al. DIA-NN: neural networks and interference correction enable deep proteome coverage in high throughput. Nat. Methods. 17(1), 41-44 (2020).
High-Grade Low-Grade
High-Grade Low-Grade
PC2
14594
995
1545
Peptides
Proteins
6642 5862
866967 915
6776 6688
983 968
6570 6405
860965 892
7437 7490
1032 985 1026
7908 7809
1007 1014
6981 7431
1056 814 9471094
6201 6914
6548 7041
960 952 956
7579 7826
7569 7286
1059 989
7516
7358
7327
1002
7205 7956
951073 938
7358 6573
942 893
7201 6264
916 944
6484
6645
6624
950
CV
CV
Samples Required Per Study Arm
log2(Normalized Mean Intensity)