Subscriber access provided by BUFFALO STATE
Article
Discovery and Qualification of Serum Protein Biomarker Candidates for Cholangiocarcinoma Diagnosis Kassaporn Duangkumpha, Thomas Stoll, Jutarop Phetcharaburanin, Puangrat Yongvanit, Raynoo Thanan, Anchalee Techasen, Nisana Namwat, Narong Khuntikeo, Nittaya Chamadol, Sittiruk Roytrakul, Jason Mulvenna, Ahmed Mohamed, Alok K. Shah, Michelle M. Hill, and Watcharin Loilome J. Proteome Res., Just Accepted Manuscript • DOI: 10.1021/acs.jproteome.9b00242 • Publication Date (Web): 16 Jul 2019 Downloaded from pubs.acs.org on July 17, 2019
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Journal of Proteome Research
1
Discovery and Qualification of Serum Protein
2
Biomarker Candidates for Cholangiocarcinoma
3
Diagnosis
4 5
Kassaporn Duangkumpha, †,‡ Thomas Stoll, ¥ Jutarop Phetcharaburanin, †,‡Puangrat
6
Yongvanit, ‡ Raynoo Thanan, † Anchalee Techasen, ‡,ǁ Nisana Namwat, †,‡ Narong Khuntikeo,
7
‡,€
Nittaya Chamadol, ‡,Ꜫ Sittiruk Roytrakul, Ω Jason Mulvenna, ¥ Ahmed Mohamed, ¥ Alok K. Shah, ¥ Michelle M. Hill, ¥,* and Watcharin Loilome †,‡,*
8 9 †
10
Department of Biochemistry, Faculty of Medicine, Khon Kaen University, Khon Kaen,
11
Thailand ‡
12
Cholangiocarcinoma Research Institute, Khon Kaen University, Khon Kaen, Thailand; ¥
13 ǁ
14 15 16
€
QIMR Berghofer Medical Research Institute, Queensland, Australia
Faculty of Associated Medical Sciences, Khon Kaen University, Khon Kaen, Thailand
Department of Surgery, Faculty of Medicine, Khon Kaen University, Khon Kaen, Thailand Ꜫ
Department of Radiology, Faculty of Medicine, Khon Kaen University, Khon Kaen,
17 18
Thailand Ω
Proteomics Research Laboratory, Genome Institute, National Center for Genetic
19
Engineering and Biotechnology, National Science and Technology Development Agency,
20
Pathum Thani, Thailand
21 22
Corresponding authors:
23
Associate Professor Watcharin Loilome, Email:
[email protected] 24
Associate Professor Michelle Hill, Email:
[email protected] 1 ACS Paragon Plus Environment
Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
25
Page 2 of 37
ABSTRACT
26
Cholangiocarcinoma ( CCA) is a major health problem in northeastern Thailand. The
27
majority of CCA cases are clinically silent and difficult to detect at an early stage. Although
28
abdominal ultrasonography ( US) can detect pre- malignant periductal fibrosis ( PDF) , this
29
method is not suitable for screening populations in remote areas. With the goal of developing
30
a blood test for detecting CCA in the at-risk population, we carried out serum protein biomarker
31
discovery and qualification. Label-free shotgun proteomics was performed on depleted serum
32
samples from 30 participants (n=10 for US-normal, US-PDF and CCA groups). Of 40 protein
33
candidates selected using multiple reactions monitoring on 90 additional serum samples (n=30
34
per group) , 11 discriminatory proteins were obtained using supervised multivariate statistical
35
analysis. We further evaluated 3 candidates using ELISA and immunohistochemistry (IHC) .
36
S100A9, thioredoxin ( TRX) and cadherin- related family member 2 ( CDHR2) were
37
significantly different between CCA and normal, and CCA and PDF groups when measured in
38
an additional 247 serum samples (p0.9 Pearson’s correlation. Then, normalized peptide intensities were averaged to protein
162
intensity, and a log2 transformation was performed to obtain a near-normal distribution needed
163
for statistical tests using in the computing environment R (Table S6). For statistical analysis,
164
missing values were replaced by the minimum detected intensity for each peptide 27.
165 166
Bioinformatics and statistical analysis
167
Protein interaction network analysis was generated using STRING software based on
168
the STRING database and Gene Oncology ( GO) term. All of analyses were constructed and
169
visualized in SPSS 19. 0 ( IBM, USA) , GraphPad Prism 5 and R statistical software. The
170
ANOVA test was conducted with Shiny MixOmics online software ( http: / / mixomics-
7 ACS Paragon Plus Environment
Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 8 of 37
171
projects. di. uq. edu. au/ Shiny) . Principal component analysis ( PCA) and orthogonal signal
172
correction projection to latent structures discriminant analysis (O-PLS-DA) was conducted in
173
SIMCA 15.0 (Umetrics, Sweden). Described and detailed of statistical analysis are available
174
in the Supplementary Methods.
175 176
Antibody-based methods
177
The following primary antibodies were used for indirect enzyme- linked
178
immunosorbent assay ( ELISA) and immunohistochemical staining ( IHC) : S100A9 ( Cat.
179
#ab24111) , TRX ( Cat. #ab185329) purchased from Abcam ( Cambridge, MA) and
180
CDHR2/PCLKC (Cat. #orb158119) purchased from Biorbyt (San Fransisco, CA). Detailed
181
methods for ELISA and IHC are available in the Supplementary Methods.
182 183
RESULTS
184
Overview of biomarker workflow and baseline characteristics of samples
185
We performed a multi- phased biomarker discovery and development study as
186
illustrated in Figure 1. Participant recruitment, ultrasound, blood sample collection and
187
biobanking were completed before biomarker discovery.
188
‘triangular’ biomarker study design for development of early cancer biomarkers
189
small cohort for discovery using shotgun proteomics to measure a broad range of proteins, then
190
increased the sample size while reducing the number of biomarker candidates. For the
191
discovery and qualification phases, we selected age-sex matched participants with ultrasound-
192
confirmed normal liver and PDF pathology. CCA cases were confirmed by pathology
193
diagnosis. In the qualification cohort, smoking status and alcohol consumption were
194
significantly different (Table 1 and Table S9). Orthogonal investigation using antibody-based
According to the proposed 28
, we used a
8 ACS Paragon Plus Environment
Page 9 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Journal of Proteome Research
195
methods were conducted in independent samples using ELISA and IHC techniques on serum
196
samples and tumor micro tissue arrays, respectively.
197 198
Figure 1 Generalized workflow diagram for serum protein biomarker discovery. A; Serum
199
samples from respective patient groups were stored at - 80ºc until analysis. B; Discovery
200
samples (n=30) were depleted for the top 12 serum proteins and spiked with internal standard
201
protein. Tryptic peptides were analysed by label- free proteomics and MaxQuant software.
202
Biomarker candidates were selected after analysis with Shiny MixOmics. C; A custom multiple
203
reaction monitoring-mass spectrometry (MRM-MS) was developed for biomarker verification
204
in an independent cohort of 90 participants. Data processing and analysis used Skyline, R,
205
SIMCA and SPSS. D; Antibody-based assays were used to validate peptide level MS data for
206
selected candidates at the protein level in additional independent cohorts, n= 247 for ELISA,
207
n=208 for IHC.
208 209 210
9 ACS Paragon Plus Environment
Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
211
Page 10 of 37
Table 1 Baseline characteristics of serum samples in the discovery and qualification phases Discovery phase (N=30) Normal US
PDF US
CCA
Sample size
10
10
10
Gender Male/Female
5/5
5/5
5/5
Age in year (Median ± SD)
60 ± 11
60 ± 12
61 ± 11
Diagnosed with Diabetes Yes No Unknown*
1 (10%) 5 (50%) 4 (40%)
2 (20%) 8 (80%) -
Smoking status Yes No Unknown*
1 (10%) 4 (40%) 5 (50%)
Alcohol consumption Yes No Unknown*
2 (20%) 3 (30%) 5 (50%)
Qualification phase (N=90) p-value
Normal US
PDF US
CCA
30
30
30
1.000
15/15
15/15
15/15
1.000
1.000
63 ± 5
65 ± 6
64 ± 5
0.497
9 (90%) 1 (10%)
0.376 6 (20%) 4 (13%) 6 (20%) 24(80%) 26 (87%) 18 (60%) 6 (20%)
0.738
3 (30%) 7 (70%) -
4 (40%) 5 (50%) 1 (10%)
0.649 13 (43%) 6 (20%) 14 (47%) 13 (43%) 22 (73%) 10 (33%) 4 (14%) 2 (7%) 6 (20%)
0.042
4 (40%) 6 (60%) -
6 (60%) 3 (30%) 1 (10%)
0.449 18(60%) 8 (27%) 21 (70%) 8 (27%) 20 (67%) 3 (10%) 4 (13%) 2 (6%) 6 (20%)
P0. 9 ( Pearson’ s correlation test) were selected and 12 ACS Paragon Plus Environment
Page 13 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Journal of Proteome Research
251
converted to protein intensities and log2 transformed. The qualification data set was then
252
subjected to multivariate statistical analysis (Figure 3). PCA with Pareto scaling showed a
253
clear separation of the CCA group from the normal and PDF groups (Figure 3A). The first two
254
principal components expressed 35.9% for PC1 and 21.9% for PC2 in the PCA model (Figure
255
3B). Orthogonal signal correction projection to latent structures discriminant analysis (O-PLS-
256
DA) was applied to conduct pair-wise comparisons of normal versus PDF, normal versus CCA,
257
and PDF versus CCA ( Figure S2) . Although normal and PDF groups were unable to be
258
separated, the CCA group showed clear separation from the normal as well as PDF groups. The
259
O- PLS- DA regression model confirmed the above visual observations, with significant CV-
260
ANOVA for the normal versus CCA comparison, PDF versus CCA comparison, but not for
261
the normal versus PDF comparison (Table 2).
262
The loadings of the pairwise O- PLS- DA model identified significantly higher
263
normalized protein intensities for 11 proteins: Haptoglobin ( HP) , Alpha- 1- antichymotrypsin
264
( A1AC / SERPIN3) , Complement component C9 ( C9) , Intercellular adhesion molecule 1
265
(ICAM1), Protein S100A9 (S100A9), Thioredoxin (TRX/TXN), Aminopeptidase N (ANPEP),
266
Fumarylacetoacetase (FAH), Lipopolysaccharide-binding protein (LBP), Inter-alpha-trypsin
267
inhibitor heavy chain H3 (ITIH3), Cadherin-related family member 2 (CHDR2/PCLKC) in the
268
CCA group when compared with the normal and PDF groups (Table 2). Data for all measured
269
proteins are provided in Table S7.
270
13 ACS Paragon Plus Environment
Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 14 of 37
271 272 273
Figure 3 Biomarker qualification and multivariate analysis. A and B; PCA score plot and
274
loading plot of MRM results of candidates that shows sample differentiation. The three
275
significant candidate proteins are shown in red circles.
276 277 278 279
14 ACS Paragon Plus Environment
Page 15 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46
280
Journal of Proteome Research
Table 2 Summary of O-PLS-DA qualified serum biomarker candidate proteins O-PLS-DA Model CCA (-) vs Normal (+) R2X = 56.7%; Q2Y = 0.645; p