Sage Journals: Discover world-class research

Abstract

Colorectal cancer is a very common cancer worldwide. Serum tumor-associated autoantibodies (TAAbs), especially the anti-p53 autoantibody, may be promising biomarkers to detect early-stage colorectal cancer. This study aimed to identify all known autoantibodies and their value in colorectal cancer diagnosis, as well as exploring the underlying connections and mechanisms through a bioinformatics analysis. Databases were used to select available articles of TAAbs in colorectal cancer. In a meta-analysis of the anti-p53 autoantibody, the diagnostic odds ratio and area under the curve (AUC) of the summary receiver-operating characteristic (SROC) curve were calculated using Stata 12.0 and Meta-Disc 1.4. We identified 73 articles including 199 single autoantibodies and 42 multiple autoantibodies. The maximum value of Youden’s index was 0.76, combining c-MYC, p53, cyclin B1, p62, Koc, IMP1, and survivin. The diagnostic odds ratio for anti-p53 autoantibody at all stages was 10.86 (95% CI 8.40, 14.06) with low heterogeneity (I² = 40.3%) and the AUC of the SROC curve was 0.82. For the anti-p53 autoantibody in early-stage colorectal cancer, the diagnostic odds ratio was 4.82 (95% CI 2.95, 7.87) with heterogeneity (I² = 7.9%) and the AUC of the SROC curve was 0.72. Eighty-seven autoantibodies were selected for bioinformatics analyses. We found that the most enriched functional terms and protein–protein interactions may relate to the mechanism of autoantibody generation. In summary, our study summarized the diagnostic value of TAAbs in colorectal cancer, either as single molecules or in combination. Bioinformatics analyses may be a new approach to explore the mechanism of autoantibody generation.

Keywords

Colorectal cancer autoantibody meta-analysis bioinformatics analysis

Introduction

Colorectal cancer (CRC) ranks third in terms of incidence but second in terms of mortality.¹ In men and women combined, the death rate has dropped by 51% at its peak of 28.6 per 100,000 in 1976 to 14.1 per 100,000 in 2014. Reductions in mortality are attributed to improvements in treatment (12%), changing patterns in CRC risk factors (35%), and screening uptake (53%).² In the European Group on Tumour Markers (EGTM) guidelines for the use of biomarkers in gastrointestinal cancer for screening of colorectal cancer, preoperative carcinoembryonic antigen (CEA) levels may be combined with clinical and histopathological criteria in determining prognosis in patients with newly diagnosed CRC. CEA and carbohydrate antigen 19-9 have an important role in monitoring the recurrence of disease, but they are not recommended for use as early biomarkers because of their low sensitivity and specificity. Genetic-based screening methods such as mRNA expression, gene mutations, or aberrantly methylated genes present in stool or body fluid samples have emerged and are changing CRC detection approaches. In October 2014, a non-invasive stool-based DNA test from Exact Sciences (Cologuard) for average-risk patients became the first US Food and Drug Administration (FDA)-approved test to be eligible for reimbursement. This test evaluates the presence of blood and DNA in a patient’s stool sample, which may be indicative of precancerous or cancerous polyps. However, it detects fewer than half of all large advanced adenomas (42%).³

The presence of autoantibodies precedes clinical findings by months or even years.⁴ The anti-p53 autoantibody was detected as early as 17–47 months prior to clinical tumor manifestation in uranium workers at high risk of lung cancer development. Furthermore, autoantibodies may be valuable biomarkers as they are stable serological proteins that are expressed at high levels in serum despite low levels of the corresponding antigen. CRC has the second highest anti-p53 autoantibody sero-positivity rates due, in part, to the high frequency of TP53 mutations.⁵ Panels of autoantibodies with the anti-p53 autoantibody are considered to have greater application value with high sensitivity and specificity.⁶ However, no detailed systematic review and meta-analysis has been performed to evaluate the diagnostic accuracy of autoantibodies in CRC. Although there are innumerable studies exploring their underlying value as biomarkers, the mechanisms of autoantibody generation remain unknown.

In this study, we first extensively searched for autoantibody studies conducted on CRC. Then, relevant information for the anti-p53 autoantibody regarding the diagnostic odds ratio (DOR), and area under curve (AUC) of the summary receiver-operating characteristic (SROC) curves were analyzed. Bioinformatics analyses—as a new method—were undertaken to analyze the corresponding genes and proteins of autoantibodies to reveal their underlying connections.

Methods

Search strategy

We performed a comprehensive search of all eligible literature using PubMed, EMBASE, the Cochrane library, and Web of Science databases until 11 January 2018. The following combination of terms was used: [colorectal (or) colon (or) rectum] (and) [cancer (or) neoplasm (or) carcinoma (or) adenoma (or) malignancy] (and) [autoantibodies (or) antibodies (or) immunoglobulin] (and) [detection OR diagnosis OR screen OR screening OR biomarker OR marker] (and) [serum carried (or) blood (or) plasma]. Related or additional articles were also identified by manual searches to ensure that all relevant articles were included. When the search was complete, duplicate articles were deleted.

Study selection

Studies that met the following criteria were included in our systematic review: (a) participants were evaluated for the presence of serum autoantibodies or antibodies; (b) studies provided both the sensitivity and specificity of the levels of mixed autoantibodies for the diagnosis of CRC; and (c) studies included a cancer-free group or information regarding a control group. Studies were excluded if they were: (a) conference abstracts and letters to journal editors; (b) reviews, meta-analyses, or proceedings; (c) studies concerning the function of autoantibodies in animal models; (d) studies with small sample sizes (n < 10) to avoid selection bias; and (e) studies without full texts available.

Data extraction and quality assessment

Two reviewers extracted the following information: first author, year of publication, location, autoantibodies, number of patients, test methods, cut-off value or AUC, and evaluation indexes (sensitivity, specificity).

The Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) was used to assess the quality of studies for the meta-analysis.⁷ This tool comprises four domains: patient selection, index test, reference standard, and flow and timing. Answers of “yes,” “no,” or “unclear” were given to each study. Only answers of “yes” were assigned a score.

Functional enrichment analysis

The genes of 92 corresponding non-repetitive autoantibodies were screened out for functional enrichment analysis using Metascape (http://metascape.org).⁸ Gene Ontology (GO) terms and Kyoto Encyclopaedia of Genes and Genomes (KEGG) pathways were enriched. Only terms with P-values of < 0.01 and the number of enriched genes ⩾ 3 were considered significant. All the resultant terms were grouped into clusters based on their similarities.

Integration of a protein–protein interaction network and module analysis

The STRING database is an online tool that can evaluate protein–protein interaction (PPI) information, and includes 18,838 human proteins with 25,914,693 core network interactions (STRING; https://string-db.org/).⁹ Identified autoantibodies were imported into STRING. The PPI network was built using Cytoscape (http://www.cytoscape.org/).¹⁰ In addition, higher-degree nodes were regarded as hub nodes. Sub-modules of the PPI network were analyzed by Molecular Complex Detection (MCODE; Version 1.5.1; Toronto, ON, Canada), with the criteria set as default.

Statistical analyses

Statistical analyses were undertaken utilizing Stata 12.0 (Stata Corp, College Station, TX, USA) and Meta-Disc 1.4 (Cochrane Colloquium, Barcelona, Spain). Diagnostic threshold effects were accessed by the Spearman rank correlation analysis. Heterogeneity was defined using the Cochrane-Q test (P < 0.05) and I² index (I² >50%). In the presence of significant heterogeneity, a random-effects model (DerSimonian-Laird) was used to calculate the pooled results. A fixed-effects model (Mantel-Haenszel) was used to calculate the pooled statistics if there was no statistical heterogeneity. The DOR and AUC with the corresponding 95% confidence interval (CI) were calculated using a bivariate binomial mixed model. Sensitivity analysis was conducted to assess the stability. Deeks funnel plot asymmetry analysis was performed to identify publication bias.

Results

Study selection and characteristics

A total of 6106 potentially relevant publications were identified by the first search of the four databases. According to the criteria, 73 articles were included in our systematic analysis (Figure S1). These included 199 single autoantibodies, which also encompassed some autoantibodies studied with high frequency (Table 1). We also collected multiple autoantibodies in 42 different combinations for CRC (Table 2).^11-83 The research was mainly undertaken in China (n = 20) and the USA (n = 12), with the remainder in Japan, Germany, Spain, and elsewhere. The sample size of the populations with CRC ranged from 13 to 1068. Patients’ ages ranged from 21 to 95. The autoantibodies studied in more than three different trials were p53, MUC1, c-MYC, survivin, ANXA, CEA, and p62.

Table 1.

Studies on the single TAAbs in CRC.

Ref	First author	Year	TAAbs	Patient (AS/ES)	Control (BD/NH)	Method	Se (AS)	Se (ES)	Sp (NH)	QUADAS-2
83	Rohayem	2000	p53	49	300	ELISA	18.4%	/	100.0%	8
11	Hammel	1997	p53	54	BD = 24	ELISA	26.0%	/	100.0%	10
12	Fan	2017	p53	92/42	100	ELISA	41.3%	45.2%	80.0%	8
13	Zhang	2003	p53	45	82	ELISA	17.8%	/	97.6%	8
14	Liu	2008	p53	46	58	ELISA	23.9%	/	96.6%	8
15	Chan	2009	p53	94	54	ELISA	24.5%	/	98.1%	8
16	Broll	2001	p53	130	44	ELISA	15.0%	/	100.0%	10
17	Negm	2016	p53	131	131	Microarray	26.0%	/	95.0%	8
18	Kojima	2009	p53	45/31	22	ELISA	20.0%	13.0%	100.0%	9
19	Suppiah	2008	p53	92/46	28	ELISA	21.7%	21.7%	100.0%	10
20	Megliorino	2005	p53	45	103/82	ELISA	17.8%	/	97.6%	8
21	Wu	2011	p53	67/20	10/400	ELISA	46.3%	10.0%	95.0%	9
22	Scanlan	2002	p53	74	75	SADA	14.9%	/	100.0%	9
23	Pedersen	2013	p53-9	97	94	ELISA	22.0%	/	95.0%	9
23	Pedersen	2013	p53-10	97	94	ELISA	24.0%	/	95.0%	9
23	Pedersen	2013	p53-78	97	94	ELISA	21.0%	/	95.0%	9
23	Pedersen	2013	p53-25	97	94	ELISA	20.0%	/	95.0%	9
23	Pedersen	2013	p53-44	97	94	ELISA	14.0%	/	95.0%	9
23	Pedersen	2013	p53-5	97	94	ELISA	13.0%	/	95.0%	9
23	Pedersen	2013	p53-26	97	94	ELISA	13.0%	/	95.0%	9
23	Pedersen	2013	p53-43	97	94	ELISA	13.0%	/	95.0%	9
23	Pedersen	2013	p53-58	97	94	ELISA	13.0%	/	95.0%	9
23	Pedersen	2013	p53-34	97	94	ELISA	11.0%	/	95.0%	9
23	Pedersen	2013	p53-42	97	94	ELISA	11.0%	/	95.0%	9
23	Pedersen	2013	p53-27	97	94	ELISA	9.0%	/	95.0%	9
23	Pedersen	2013	p53-59	97	94	ELISA	8.0%	/	95.0%	9
23	Pedersen	2013	p53-4	97	94	ELISA	7.0%	/	95.0%	9
23	Pedersen	2013	p53-45	97	94	ELISA	7.0%	/	95.0%	9
23	Pedersen	2013	p53-14	97	94	ELISA	6.0%	/	95.0%	9
23	Pedersen	2013	p53-41	97	94	ELISA	5.0%	/	95.0%	9
23	Pedersen	2013	p53-39	97	94	ELISA	4.0%	/	95.0%	9
24	Looi	2006	p53	57	82	ELISA	8.8%	/	98.8%	8
25	Lechpammer	2004	p53	220/115	42	ELISA	18.2%	18.3%	100.0%	9
26	Tang	2001	p53	998	BD = 211	ELISA	13.0%	/	99.0%	10
27	Müller	2006	p53	197	436	ELISA	32.0%	/	100.0%	8
28	Wu	2010	p53	66	879	ELISA	16.7%	/	99.0%	8
29	Hallak	1998	p53	235	194	ELISA	18.0%	/	100.0%	8
30	Chen	2016	p53	352/198	124	Fluorescent bead-based GST	14.0%	13.0%	98.0%	9
31	Shiota	2000	p53	71	BD = 18	ELISA	25.0%	/	94.0%	9
32	Kobayashi	2016	p53	84/48	42	ELISA	25.0%	16.7%	90.5%	8
33	Ostendorff	2013	p53	47	47	VeraCode^TM Beads	11.0%	/	100.0%	8
33	Ostendorff	2013	p53	127	59	VeraCode^TM Beads	11.0%	/	100.0%	8
34	Sthoeger	1997	p53	65	50	ELISA	13.8%	/	100.0%	9
35	Chen	2007	p53	52	82	ELISA	9.6%	/	98.8%	10
36	Shimada	2002	p53	192	189/205	ELISA	20.4%	/	95.1%	9
42	Pedersen	2013	MUC1	157	40	Microarray	10.2%	/	95.0%	8
37	Nakamura	1998	MUC1	56	53/31	EIA	48.2%	/	83.9%	8
38	Pedersen	2011	MUC1	58	39/103	Glycopeptide array	6.9%	/	97.1%	8
39	Silk	2009	MUC1	20	47	ELISA	30.0%	/	83.0%	10
42	Pedersen	2013	MUC1-Tn	157	40	Microarray	16.6%	/	95.0%	8
42	Pedersen	2013	MUC1-STn	157	40	Microarray	42.0%	/	95.0%	8
42	Pedersen	2013	MUC1-Core3	157	40	Microarray	42.0%	/	95.0%	8
40	Wang	2012	MUC1 VNTR	135	95	ELISA	18.5%	/	96.8%	10
38	Pedersen	2011	MUC1 6Tn	58	39/103	Glycopeptide array	20.7%	/	98.1%	8
38	Pedersen	2011	MUC1 9Tn	58	39/103	Glycopeptide array	20.7%	/	98.1%	8
38	Pedersen	2011	MUC1 15Tn	58	39/103	Glycopeptide array	32.8%	/	98.1%	8
38	Pedersen	2011	MUC1 15Core3	58	39/103	Glycopeptide array	27.6%	/	98.1%	8
38	Pedersen	2011	MUC1 9Core4	58	39/103	Glycopeptide array	44.8%	/	98.1%	8
38	Pedersen	2011	MUC1 9STn	58	39/103	Glycopeptide array	39.7%	/	97.1%	8
38	Pedersen	2011	MUC1 15STn	58	39/103	Glycopeptide array	56.9%	/	97.1%	8
38	Pedersen	2011	MUC1 9C3	58	39/103	Glycopeptide array	44.8%	/	98.1%	8
38	Pedersen	2011	MUC1 15C3	58	39/103	Glycopeptide array	44.8%	/	98.1%	8
38	Pedersen	2011	MUC1 STn	58	39/103	Glycopeptide array	63.8%	/	96.1%	8
38	Pedersen	2011	MUC1 C3	58	39/103	Glycopeptide array	63.8%	/	96.1%	8
38	Pedersen	2011	recMUC4	58	39/103	Glycopeptide array	13.8%	/	95.1%	8
38	Pedersen	2011	Tn-recMUC4	58	39/103	Glycopeptide array	37.9%	/	88.3%	8
38	Pedersen	2011	Tn-MUC4-1-Tn-MUC4-5	58	39/103	Glycopeptide array	37.9%	/	98.1%	8
38	Pedersen	2011	Tn-MUC4-6-Tn-MUC4-20	58	39/103	Glycopeptide array	29.3%	/	96.1%	8
38	Pedersen	2011	Tn-MUC4-5	58	39/103	Glycopeptide array	15.5%	/	97.1%	8
38	Pedersen	2011	All Tn-MUC4	58	39/103	Glycopeptide array	53.4%	/	94.2%	8
13	Zhang	2003	c-MYC	45	82	ELISA	4.4%	/	100.0%	8
14	Liu	2008	c-MYC	46	58	ELISA	21.7%	/	94.8%	8
20	Megliorino	2005	c-MYC	45	103/82	ELISA	4.4%	/	100.0%	8
41	Ben-Mahrez	1990	c-MYC	214	62	WB	12.6%	/	98.4%	9
24	Looi	2006	c-MYC	57	82	ELISA	14.0%	/	100.0%	8
35	Chen	2007	c-MYC	52	82	ELISA	15.4%	/	100.0%	10
83	Rohayem	2000	Survivin	49	300	ELISA	8.2%	/	100.0%	8
43	Chen	2010	Survivin	232/99	365	ELISA	56.9%	58.6%	64.1%	10
40	Wang	2012	Survivin	135	95	ELISA	31.1%	/	98.5%	10
13	Zhang	2003	Survivin	45	82	ELISA	4.4%	/	97.6%	8
20	Megliorino	2005	Survivin	45	103/82	ELISA	4.4%	/	97.6%	8
17	Negm	2016	Annexin	131	131	Microarray	29.0%	/	94.0%	8
45	Chen	2011	ANXA	220/91	216	ELISA	85.0%	81.3%	61.6%	11
12	Fan	2017	ANXA4	92	100	ELISA	15.2%	/	90.0%	8
30	Chen	2016	ANXA4	352/198	124	Fluorescent bead-based GST	6.0%	6.0%	98.0%	9
47	Song	2009	ANXA11	19	77	ELISA	5.3%	/	97.4%	8
48	Babel	2009	CEA	52	42	ELISA	63.5%	/	59.5%	8
49	Wu	2011	CEA	69	28/37	ELISA	63.8%	/	89.2%	11
50	Albanopoulos	2000	CEA	58	28	ELISA	32.7%	/	96.4%	9
51	Konstadoulakis	1994	CEA	25	35/28	ELISA	56.0%	/	89.4%	8
33	Ostendorff	2013	CEA	127	59	VeraCode^TM Beads	21.0%	/	98.0%	8
13	Zhang	2003	p62	45	82	ELISA	11.1%	/	98.8%	8
14	Liu	2008	p62	46	58	ELISA	21.7%	/	98.0%	8
52	Liu	2013	p62	64	42/34	ELISA	23.4%	/	97.1%	8
53	Zhang	2001	p62	65	139/112	ELISA	9.2%	/	98.8%	8
54	Chen	2011	RPH3AL	84/34	63	WB	72.6%	/	84.1%	9
12	Fan	2017	RPH3AL	92	100	ELISA	22.8%	/	90.0%	8
30	Chen	2016	RPH3AL	352/198	124	Fluorescent bead-based GST	7.0%	7.0%	98.0%	9
12	Fan	2017	NY-CO-16	92	100	ELISA	20.7%	/	90.0%	8
17	Negm	2016	NY-CO-16	131	131	Microarray	41.0%	/	95.0%	8
15	Chan	2009	NY-CO-16	94	54	ELISA	18.1%	/	100.0%	8
13	Zhang	2003	Koc	45	82	ELISA	8.9%	/	98.8%	8
53	Zhang	2001	Koc	65	139/113	ELISA	13.8%	/	98.8%	8
14	Liu	2008	Koc	46	58	ELISA	15.2%	/	100.0%	8
15	Chan	2009	HDAC5	94	54	ELISA	20.2%	/	96.3%	8
12	Fan	2017	HDAC5	92	100	ELISA	21.7%	/	90.0%	8
22	Scanlan	2002	HDAC5	74	75	SADA	2.7%	/	100.0%	9
13	Zhang	2003	IMP1	45	82	ELISA	13.3%	/	97.6%	8
14	Liu	2008	IMP1	46	58	ELISA	21.7%	/	100.0%	8
30	Chen	2016	IGF2BP1	352/198	124	Fluorescent bead-based GST	5.0%	6.0%	98.0%	9
56	Li	2017	NY-ESO-1	236/137	89	ELISA	16.9%	8.8%	97.8%	9
22	Scanlan	2002	NY-ESO-1	74	75	SADA	6.7%	/	100.0%	9
30	Chen	2016	CTAG1	352/198	124	Fluorescent bead-based GST	6.0%	7.0%	98.0%	9
30	Chen	2016	CTAG2	352/198	124	Fluorescent bead-based GST	5.0%	4.0%	98.0%	9
22	Scanlan	2002	MAGEA3	74	75	SADA	8.1%	/	100.0%	9
30	Chen	2016	MAGEA3	352/198	124	Fluorescent bead-based GST	3.0%	4.0%	98.0%	9
30	Chen	2016	MAGEA4	352/198	124	Fluorescent bead-based GST	9.0%	9.0%	98.0%	9
35	Chen	2007	Cyclin B1	52	82	ELISA	32.7%	/	97.6%	10
13	Zhang	2003	Cyclin B1	45	82	ELISA	15.6%	/	97.6%	8
35	Chen	2007	Cyclin D1	52	82	ELISA	21.2%	/	98.8%	10
12	Fan	2017	SEC61β	92/42	100	ELISA	30.4%	35.7%	80.0%	8
57	Fan	2011	SEC61β	86	72	WB	79.0%	/	75.0%	9
12	Fan	2017	CCCAP	92	100	ELISA	21.7%	/	90.0%	8
15	Chan	2009	CCCAP	94	54	ELISA	35.1%	/	96.3%	8
58	He	2009	IMPDH2	25	15	WB	32.0%	/	100.0%	8
30	Chen	2016	IMPDH2	352/198	124	Fluorescent bead-based GST	8.0%	10.0%	98.0%	9
30	Chen	2016	ERBB2-C-term	352/198	124	Fluorescent bead-based GST	6.0%	6.0%	98.0%	9
30	Chen	2016	ERBB2-N-term	352/198	124	Fluorescent bead-based GST	3.0%	4.0%	98.0%	9
59	O’Reilly	2015	ZNF346	96/52	35	ELISA	15.6%	11.5%	94.3%	9
59	O’Reilly	2015	ZNF638	96/52	35	ELISA	10.4%	5.8%	97.1%	9
59	O’Reilly	2015	ZNF700	96/52	35	ELISA	19.8%	21.2%	94.3%	9
59	O’Reilly	2015	ZNF768	96/52	35	ELISA	15.6%	11.5%	100.0%	9
60	Tsunemi	2010	GRP78	15	20	WB	20.0%	/	100.0%	8
61	Raiter	2014	GRP78	27	28/30	ELISA	63.0%	/	67.0%	9
17	Negm	2016	K-ras	131	131	Microarray	27.0%	/	96.0%	8
62	Takahashi	1995	p21	160	40	ELISA	32.0%	/	97.5%	7
15	Chan	2009	NMDA	94	54	ELISA	20.2%	/	96.3%	8
12	Fan	2017	NMDAR	92	100	ELISA	15.2%	/	90.0%	8
12	Fan	2017	RPL36	92	100	ELISA	38.0%	/	90.0%	8
30	Chen	2016	RPL13	352/198	124	Fluorescent bead-based GST	8.0%	9.0%	98.0%	9
55	Xia	2005	DDX48	30	60	ELISA	10.0%	/	100.0%	8
30	Chen	2016	DDX53	352/198	124	Fluorescent bead-based GST	5.0%	5.0%	98.0%	9
12	Fan	2017	SLP2	92/42	100	ELISA	51.1%	47.6%	80.0%	8
12	Fan	2017	SUR	92	100	ELISA	22.8%	/	90.0%	8
12	Fan	2017	PLSCR1	92/42	100	ELISA	35.9%	38.1%	80.0%	8
63	Jagadish	2016	AKAP4	172/45	138	ELISA	82.0%	86.0%	100.0%	8
35	Chen	2007	Calnuc	52	82	ELISA	11.5%	/	98.8%	10
64	Hector	2012	TRIM28/KAP1	58/22	55	Reverse-ELISA	12.1%	18.2%	100.0%	8
65	Li	2012	IGFBP-2	70/38	141	ELISA	66.0%	40.0%	81.0%	10
66	Wang	2016	F. nucleatum-IgA	258/55	200	ELISA	30.6%	27.3%	96.2%	10
17	Negm	2016	AFP	131	131	Microarray	27.0%	/	95.0%	8
17	Negm	2016	RAF1	131	131	Microarray	18.0%	/	95.0%	8
67	Kocer	2006	MUC5AC	30/8	20/22	ELISA	60.0%	50.0%	72.7%	8
48	Babel	2009	ACVR2B	52	42	ELISA	60.0%	/	76.2%	8
48	Babel	2009	PIM1	52	42	ELISA	48.1%	/	83.3%	8
48	Babel	2009	MAPKAPK3	52	42	ELISA	72.7%	/	74.0%	8
68	Mosolits	1999	GA733-2/EGP	1068/519	20/45	ELISA	14.5%	15.0%	100.0%	9
69	Syrigos	1998	CRLS1	43/19	26/40	ELISA	33.0%	36.8%	95.0%	8
70	Luo	2013	OY-TES-1	73/43	76	ELISA	9.6%	9.3%	100.0%	8
80	Rimm	1995	Villin/Vil1	25	55/52	WB	44.0%	/	90.4%	8
22	Scanlan	2002	MBD2	74	75	SADA	2.7%	/	100.0%	9
22	Scanlan	2002	TRIP4	74	75	SADA	8.1%	/	100.0%	9
22	Scanlan	2002	NY-CO-45	74	75	SADA	4.0%	/	100.0%	9
22	Scanlan	2002	KNSL6	74	75	SADA	4.0%	/	100.0%	9
22	Scanlan	2002	HIPIR	74	75	SADA	6.8%	/	100.0%	9
22	Scanlan	2002	SEB4D	74	75	SADA	5.4%	/	100.0%	9
22	Scanlan	2002	KIAA1416	74	75	SADA	5.4%	/	100.0%	9
22	Scanlan	2002	LMNA	74	75	SADA	4.1%	/	100.0%	9
22	Scanlan	2002	SSX-2	74	75	SADA	2.7%	/	100.0%	9
24	Looi	2006	p16	57	82	ELISA	10.5%	/	98.8%	8
71	Jiang	2010	ARD1A	198	200	ELISA	14.7%	/	98.0%	9
72	Kanojia	2010	SPAG9	54/12	50	ELISA + WB	70.0%	100.0%	100.0%	8
73	Hanafusa	2012	Tektin5	44	16	ELISA	9.1%	/	100.0%	8
74	Song	2012	CCDC83	37	21	WB	70.0%	/	81.0%	9
75	Kiyamova	2012	FKBP4	13	69	SEREX	7.7%	/	100.0%	8
75	Kiyamova	2012	XRCC4	13	69	SEREX	7.7%	/	100.0%	8
75	Kiyamova	2012	ACTR1A	13	69	SEREX	7.7%	/	100.0%	8
75	Kiyamova	2012	TSGA2	13	69	SEREX	7.7%	/	100.0%	8
76	Babel	2011	SULF1	50	103	ELISA	50.0%	/	73.9%	9
76	Babel	2011	NHSL1	50	103	ELISA	52.0%	/	52.2%	9
76	Babel	2011	MST1	50	103	ELISA	46.0%	/	71.7%	9
76	Babel	2011	GTF2i	50	103	ELISA	58.0%	/	52.2%	9
76	Babel	2011	SREBF2	50	103	ELISA	48.0%	/	60.9%	9
76	Babel	2011	GRN	50	103	ELISA	58.0%	/	58.7%	9
77	Song	2011	FBXO39	50	50	ELISA	6.0%	/	94.0%	9
78	Nam	2003	UCHL3	43	35/15	WB	44.2%	/	100.0%	8
30	Chen	2016	MDM2	352/198	124	Fluorescent bead-based GST	5.0%	8.0%	98.0%	9
30	Chen	2016	MTDH	352/198	124	Fluorescent bead-based GST	4.0%	6.0%	98.0%	9
30	Chen	2016	TPM3	352/198	124	Fluorescent bead-based GST	7.0%	6.0%	98.0%	9
30	Chen	2016	MIA	352/198	124	Fluorescent bead-based GST	4.0%	5.0%	98.0%	9
30	Chen	2016	HMGN3	352/198	124	Fluorescent bead-based GST	5.0%	5.0%	98.0%	9
30	Chen	2016	CALU	352/198	124	Fluorescent bead-based GST	4.0%	4.0%	98.0%	9
30	Chen	2016	KLK3	352/198	124	Fluorescent bead-based GST	4.0%	4.0%	98.0%	9
30	Chen	2016	REG3A	352/198	124	Fluorescent bead-based GST	3.0%	4.0%	98.0%	9
44	Yie	2016	TOPO48	87/47	764	ELISA	51.7%	27.7%	100.0%	8
32	Kobayashi	2016	FIRs	87/49	42	AlphaLISA	32.2%	36.7%	100.0%	8
46	Benvenuto	2015	Rib-P	67	5/73	ELISA	11.7%	/	100.0%	9
33	Ostendorff	2013	GDF15	127	59	VeraCode^TM Beads	38.0%	/	100.0%	8
79	Ran	2008	Six phage-peptides	24	24	Phage plaque assay	83.0%	/	88.0%	8

AS: all stage; BD: benign disease; CRC: colorectal cancer; EIA: enzyme immunoassay; ELISA: enzyme-linked immunosorbent assay; ES: early stage; NH: normal healthy control; QUADAS: diagnostic accuracy studies; Ref: reference; SADA: serum antibody detection array; Se: sensitivity; SEREX: serological analysis of recombinant cDNA expression library; Sp: specificity; TAAbs: tumor-associated autoantibodies; WB: western blotting.

Table 2.

Studies on the multiple TAAbs in CRC.

Ref	First author	Year	TAAbs	Patient (AS/ES)	Control (BD/NH)	Se (AS)	Se (ES)	Sp (NH)	QUADAS-2
35	Chen	2007	p53, c-MYC, cyclin B1, cyclin D1, Calnuc	52	82	65.4%	/	93.7%	10
10	Rohayem	2000	Survivin, p53	49	300	26.6%	/	100.0%	8
42	Pedersen	2013	MUC1-Core3, MUC1-STn	157	40	44.6%	/	95.0%	8
42	Pedersen	2013	MUC1-STn, p53-10, p53-25	157	40	20.4%	/	95.0%	8
42	Pedersen	2013	MUC1-Core3, p53-10, p53-26	157	40	35.0%	/	95.0%	8
42	Pedersen	2013	MUC1-Core3, MUC1-STn, p53-10, p53-26	157	40	47.8%	/	95.0%	8
42	Pedersen	2013	MUC1-Core3, MUC1-STn, p53-58	157	40	47.8%	/	95.0%	8
42	Pedersen	2013	MUC1-Core3, MUC1-STn, p53-43	157	40	54.8%	/	95.0%	8
40	Wang	2012	Survivin, MUC1 VNTR	135	95	36.3%	/	95.8%	10
12	Fan	2017	SLP2, p53, SEC61β, PLSCR1	92/42	100	64.1%	66.7%	80.0%	8
81	Vazquez	2016	GTF2B, EDIL3, HCK, PIM1, STK4, p53	135/60	93	65.7%	61.7%	90.0%	8
59	O’Reilly	2015	ZNF346, ZNF638	96	35	25.0%	/	91.4%	9
59	O’Reilly	2015	ZNF346, ZNF638, ZNF700	96	35	33.3%	/	91.4%	9
59	O’Reilly	2015	ZNF346, ZNF638, ZNF700, ZNF768	96/52	35	41.7%	36.5%	91.4%	9
13	Zhang	2003	c-MYC, p53, Cyclin B1, p62, Koc, IMP1, Survivin	45	82	51.1%	/	89.0%	8
14	Liu	2008	Koc, p62	46	58	32.6%	/	98.3%	8
14	Liu	2008	Koc, p63, IMP1	46	58	43.5%	/	98.3%	8
14	Liu	2008	Koc, p63, IMP2, c-MYC	46	58	52.2%	/	93.1%	8
14	Liu	2008	Koc, p63, IMP3, c-MYC, p53	46	58	60.9%	/	89.7%	8
17	Negm	2016	AFP, p53, K-ras, NY-CO-16, RAF1, Annexin	131	131	75.0%	/	78.0%	8
48	Babel	2009	MAPKAPK3, ACVR2B, PIM-1	52	42	84.4%	/	71.4%	8
48	Babel	2009	MAPKAPK3, ACVR2B	52	42	83.3%	/	73.9%	8
15	Chan	2009	CCCAP, HDAC5, p53, NY-CO-16, NMDA	94/41	54	58.5%	/	92.6%	8
20	Megliorino	2005	p53, c-MYC, Survivin	45	103/82	24.4%	/	95.1%	8
53	Zhang	2001	p62, Koc	65	139/112	20.0%	/	97.6%	8
24	Looi	2006	p16, c-MYC, p53	57	82	28.1%	/	87.6%	8
38	Pedersen	2011	Tn, STn-, Core3-MUC1, Tn-MUC4	58	39/103	86.2%	/	89.3%	8
38	Pedersen	2011	STn-MUC1, Tn-MUC4-1-Tn-MUC4-5	58	39/103	79.3%	/	95.1%	8
82	Koziol	2003	c-MYC, p53, cyclinB1, p62, Koc, IMP1, Survivin	45	346	88.0%	/	88.0%	8
76	Babel	2011	SULF1, NHSL1, MST1, GTF2i, SREBF2, GRN	50	103	72.0%	/	73.9%	9
30	Chen	2016	p53, IMPDH2	352/198	124	22.0%	23.0%	95.0%	9
30	Chen	2016	p53, IMPDH2, MDM2	352/198	124	26.0%	30.0%	93.0%	9
30	Chen	2016	p53, IMPDH2, MDM2, MAGEA4	352/198	124	33.0%	37.0%	90.0%	9
30	Chen	2016	p53, IMPDH2, MDM2, MAGEA4, CTAG4	352/198	124	38.0%	43.0%	89.0%	9
30	Chen	2016	p53, IMPDH2, MDM2, MAGEA4, CTAG5, MTDH	352/198	124	40.0%	46.0%	88.0%	9
30	Chen	2016	p53, IMPDH2, MDM2, MAGEA4, CTAG1, TPM3_iso1	352/198	124	42.0%	48.0%	86.0%	9
30	Chen	2016	p53, IMPDH2, MDM2, MAGEA4, RPL13, SAG	352/198	124	39.0%	44.0%	89.0%	9
30	Chen	2016	p53, IMPDH2, MDM2, MAGEA4, TPM3-iso1, CTAG1	352/198	124	39.0%	44.0%	89.0%	9
30	Chen	2016	p53, IMPDH2, MDM2, MAGEA4, CTAG1, SAG	352/198	124	39.0%	45.0%	88.0%	9
30	Chen	2016	p53, IMPDH2, MDM2, MAGEA4, CTAG1, IGF2BP1	352/198	124	41.0%	46.0%	87.0%	9
30	Chen	2016	p53, IMPDH2, MDM2, MAGEA4, CTAG1, RPL13	352/198	124	41.0%	46.0%	87.0%	9
30	Chen	2016	p53, IMPDH2, MDM2, MAGEA4, TPM3-iso1, RPL13	352/198	124	42.0%	47.0%	86.0%	9

AS: all stage; BD: benign disease; CRC: colorectal cancer; ES: early stage; NH: normal healthy control; QUADAS: diagnostic accuracy studies; Ref: reference; Se: sensitivity; Sp: specificity; TAAbs: tumor-associated autoantibodies.

Diagnostic value of single tumor-associated autoantibodies for any stage of CRC

A total of 199 individual autoantibodies were selected from 73 articles. The sensitivity ranged from 3.0% to 85.0%, with a mean of 24.0% and a median of 18.0%. The specificity ranged from 52.2% to 100%, with a mean of 94.2% and a median of 97.6%. When the sensitivity was < 50%, 173 single autoantibodies were selected with a mean of 17.9% and a median of 15.2%, but the specificity still ranged from 61.0% to 100% (mean, 96.4%; median, 98%). When sensitivity was not lower than 50.0%, with a mean of 63.9% and a median of 63.3%, the specificity ranged from 52.0% to 100% (mean, 79.4%; median, 80.5%).

Diagnostic value of multiple autoantibodies for patients at all stages of CRC

There were 42 test results for multiple autoantibodies from 19 articles. The sensitivity ranged from 15.5% to 88.0% (mean, 47.0%; median: 41.7%); at the same time, the specificity ranged from 71.4% to 100% (mean, 90.5%; median, 91.4%). The maximum value of the Youden index was 0.76 from an article combining c-MYC, p53, cyclin B1, p62, Koc, IMP1, and survivin, with a sensitivity of 88% and a specificity of 88%. A combination of multiple autoantibodies is a better choice to detect CRC than a single autoantibody.

Diagnostic value for early stages of CRC

The diagnostic value of a single autoantibody for early-stage CRC was assessed for 48 autoantibodies. The sensitivity was 4.0% to 100% (mean, 21.3%; median, 10.8%) and the specificity was 61.6% to 100% (mean, 94.1%; median, 98.0%). Sixteen combinations of autoantibodies were used to assess early-stage CRC. The sensitivity was 22.0% to 65.7% (mean, 42.0%; median, 40.5%) in early-stage CRC and the specificity was 80.0% to 95.0% (mean, 88.8%; median, 89.0%).

Meta-analysis for the anti-p53 autoantibody

The anti-p53 autoantibody was described 27 times. A study by Pedersen et al.²³ on the diagnosis value of p53 peptides was excluded because of huge heterogeneity. For the anti-p53 autoantibody in all stages, the Spearman rank correlation coefficient was 0.633 (P < 0.05), suggesting that there were diagnostic threshold effects. The best summary of study results is an ROC curve rather than a single pair of pooled indexes. Parameter b = 0.52 (P > 0.05) including the ROC curve was symmetrical. The DOR and AUC of the SROC were used in the meta-analysis using a fixed-effects model (Mantel-Haenszel). The DOR was 10.86 (95% CI 8.40, 14.06) with low heterogeneity (I² = 40.3%) (Figure 1(a)). AUC of the SROC curve was 0.82 (Figure 1(b)). Sensitivity analysis was performed by eliminating studies one by one (Figure S2). The results showed that the studies by Fan et al.¹² and Müller et al.²⁷ accounted for 25.5% and 12.7% of the heterogeneity, respectively. When these two studies were removed, there was no heterogeneity (I² = 0). The DOR was 11.01 (95% CI 8.15, 14.86) and the AUC of the SROC curve was 0.83 (Figure S3).

Figure 1.

Forest plot of diagnostic odds ratio (DOR) (a) and summary receiver-operating characteristic (SROC) (b) for anti-p53 autoantibodies in all stage of colorectal cancer. Forest plot of DOR (c) and summary receiver-operating characteristic (SROC) (d) for anti-p53 autoantibodies in early stage of colorectal cancer.

For the anti-p53 autoantibody in early-stage disease, the Spearman rank correlation coefficient was 0.964 (P < 0.05) with parameter b = 0.15 (P > 0.05). The DOR and AUC were researched in the meta-analysis by a fixed-effects model. The DOR was 4.82 (95% CI 2.95, 7.87) with low heterogeneity (I² = 7.9%) (Figure 1(c)). The AUC of the SROC curve was 0.72 (Figure 1(d)).

Publication bias

As shown in Deeks funnel plots (Figure S4), publication biases were likely absent with symmetrical funnel shapes. The P-values for the funnel plot asymmetry tests (P = 0.43, P = 0.79) indicated no evidence of publication bias.

Functional characterization of autoantibodies

In total, 87 genes were identified, among which 5 could not be identified, including F. nucleatum, HIPIR, KIAA1416, SSX-2, and TOPO48. The analysis was performed using Metascape, which was carried out with the four GO sources: KEGG Pathway, GO Biological Processes, Reactome Gene Sets, and CORUM. This analysis revealed that “bladder cancer” was the most significantly enriched functional cluster. This finding may illustrate the low specificity associated with autoantibodies. It was remarkable to find that the GO analysis revealed no association with inflammation. This may be the reason why autoantibodies could be biomarkers for cancers. “Cell cycle” may provide a new clue in the study of autoantibody generation. “Regulation of protein serine/threonine kinase activity” and “Regulation of cellular protein location” were consistent with the theory of abnormal protein localization. Molecules in canonical pathways were also related to the production of autoantibodies. The top 20 clusters of significantly enriched terms are shown in Figure 2.

Figure 2.

Functional enrichment results: a network of top 20 clusters of enriched terms. Each node represents one enriched term colored by cluster ID, nodes share the same cluster are typically close to each other. Terms with Kappa similarity above 0.3 were connected. The thicker edge displayed, the higher the similarity is. The term with the best P-value was selected to represent each cluster; B heatmap of top 20 clusters, colored by P-values.

PPI network construction and module screening

The PPI network of corresponding to autoantibodies is presented in Figure 3(a). The network is composed of 83 nodes and 125 edges. Degrees > 2 were set as the cut-off criterion, from which nine genes were selected as hub genes. Hub nodes included TP53, MDM2, RAF1, MYC, CDKN1A, ERBB2, BIRC5, HRAS, and CDKN2A. The significant module with the highest score was selected using the plug-in MCODE program. This module included nine nodes and 34 edges (Figure 3(b)).

Figure 3.

PPI network and hubs.

Discussion

Autoantibodies can be detected in sera from many kinds of cancers at the asymptomatic stage, so they usually serve as biomarkers for early cancer diagnosis. Many researchers devote themselves to the discovery of new autoantibodies, but some autoantibodies are limited in clinical practice because of their poor specificity and sensitivity. In our research, 199 single autoantibodies and 42 multiple autoantibodies were selected and analyzed. For a single autoantibody, the sensitivity and specificity are usually lower than that for an ideal cancer biomarker. The sensitivity ranged from 3.0% to 85.0%, while the specificity ranged from 52.2% to 100%. Combining multiple autoantibodies can elevate the sensitivity and specificity. The sensitivity ranged from 15.5% to 88.0%; at the same time, the specificity ranged from 71.4% to 100%.

In a review of the anti-p53 autoantibody, the researchers found that the anti-p53 autoantibody had low (13%–32%) sensitivity in CRC but was nearly 100% specific for malignancy.⁶ In our research, the DOR was 10.86. The AUC of the SROC curve was 0.82. Two studies accounted for the heterogeneity, and when these were removed the DOR was 11.01 and the AUC of the SROC curve was 0.83. However, the diagnostic threshold effects could not be eliminated. The diagnostic methods varied between different studies, which included mean absorbance or levels of the autoantibodies in the control group plus two or three SD or maximum levels, ROC curve, and positive proportion. This is the main barrier to replicating the same autoantibody or comparing the different autoantibodies. For the anti-p53 autoantibody in early-stage CRC, the DOR was 4.82 and the AUC was 0.72. There may be a promising future role of the anti-p53 autoantibody in screening.

The production mechanism of autoantibodies is so complex that a large number of researchers have only studied their clinical application. In turn, the lack of a known mechanism limits the clinical application of autoantibodies. In our research, we attempted to identify a new approach to the study of these mechanisms. Bioinformatics can be used to rapidly analyze big data, and it can provide guidelines in medical research that can save money and time. We used databases and software to analyze the corresponding genes and proteins of all identified autoantibodies. The results showed that the gene functions and pathways may relate to the connections of generation. PPIs indicate a relationship between proteins that can provide clues when studying autoantibodies at the protein level.

Conclusion

Our study summarized and presented the diagnostic value of autoantibodies in CRC, either as single molecules or in combination. Bioinformatics analysis may be a new approach to explore the connections of autoantibody generation.

Supplemental Material

Figure_S1 – Supplemental material for Autoantibodies as biomarkers for colorectal cancer: A systematic review, meta-analysis, and bioinformatics analysis

Supplemental material, Figure_S1 for Autoantibodies as biomarkers for colorectal cancer: A systematic review, meta-analysis, and bioinformatics analysis by Hejing Wang, Xiaojin Li, Donghu Zhou and Jian Huang in The International Journal of Biological Markers

Supplemental Material

Figure_S2 – Supplemental material for Autoantibodies as biomarkers for colorectal cancer: A systematic review, meta-analysis, and bioinformatics analysis

Supplemental material, Figure_S2 for Autoantibodies as biomarkers for colorectal cancer: A systematic review, meta-analysis, and bioinformatics analysis by Hejing Wang, Xiaojin Li, Donghu Zhou and Jian Huang in The International Journal of Biological Markers

Supplemental Material

Figure_S3 – Supplemental material for Autoantibodies as biomarkers for colorectal cancer: A systematic review, meta-analysis, and bioinformatics analysis

Supplemental material, Figure_S3 for Autoantibodies as biomarkers for colorectal cancer: A systematic review, meta-analysis, and bioinformatics analysis by Hejing Wang, Xiaojin Li, Donghu Zhou and Jian Huang in The International Journal of Biological Markers

Supplemental Material

Figure_S4 – Supplemental material for Autoantibodies as biomarkers for colorectal cancer: A systematic review, meta-analysis, and bioinformatics analysis

Supplemental material, Figure_S4 for Autoantibodies as biomarkers for colorectal cancer: A systematic review, meta-analysis, and bioinformatics analysis by Hejing Wang, Xiaojin Li, Donghu Zhou and Jian Huang in The International Journal of Biological Markers

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The current study was supported by grants from the Nature Science Foundation of China (grant no. 81650014).

ORCID iD

Hejing Wang

Supplemental material

Supplemental material for this article is available online.

References

Bray

Ferlay

Soerjomataram

, et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2018; 68: 394–324.

Edwards

Ward

Kohler

, et al. Annual report to the nation on the status of cancer, 1975–2006, featuring colorectal cancer trends and impact of interventions (risk factors, screening, and treatment) to reduce future rates. Cancer 2010; 116: 544–573.

Karim

MA.

Cologuard test: A possible alternate to colonoscopy. J Ayub Med Coll Abbottabad 2017; 29: 176.

Caron

Choquet-Kastylevsky

Joubert-Caron

Cancer immunomics using autoantibody signatures for biomarker discovery. Mol Cell Proteomics 2007; 6: 1115–1112.

Kunizaki

Sawai

Takeshita

, et al. Clinical value of serum p53 antibody in the diagnosis and prognosis of colorectal cancer. Anticancer Res 2016; 36: 4171–4175.

Song

, et al. The roles and applications of autoantibodies in progression, diagnosis, treatment and prognosis of human malignant tumours. Autoimmun Rev 2017; 16: 1270–1281.

Whiting

Rutjes

Westwood

, et al. QUADAS–2: A revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med 2011; 155: 529–536.

Zhou

Pache

, et al. Metascape provides a biologist-oriented resource for the analysis of systems-level datasets. Nat Commun 2019; 10: 1523.

Szklarczyk

Morris

Cook

, et al. The STRING database in 2017: quality-controlled protein–protein association networks, made broadly accessible. Nucleic Acids Res 2017; 45: 362–368.

10.

Doerks

Copley

Schultz

, et al. Systematic identification of novel protein domain families associated with nuclear functions. Genome Res 2002; 12: 47–56.

11.

Hammel

Boissier

Chaumette

, et al. Detection and monitoring of serum p53 antibodies in patients with colorectal cancer. Gut 1997; 40: 356–361.

12.

Fan

Kuo

Lin

, et al. Development of a multiplexed tumor-associated autoantibody-based blood test for the detection of colorectal cancer. Clin Chim Acta 2017; 475: 157–163.

13.

Zhang

Casiano

Peng

, et al. Enhancement of antibody detection in cancer using panel of recombinant tumor-associated antigens. Cancer Epidemiol Biomarkers Prev 2003; 12: 136–143.

14.

Liu

Wang

, et al. Evaluation of tumour–associated antigen (TAA) miniarray in immunodiagnosis of colon cancer. Scand J Immunol 2009; 69: 57–63.

15.

Chan

Fan

Kuo

, et al. Multiple serological biomarkers for colorectal cancer detection. Int J Cancer 2010; 126: 1683–1690.

16.

Broll

Duchrow

Oevermann

, et al. p53 autoantibodies in sera of patients with a colorectal cancer and their association to p53 protein concentration and p53 immunohistochemistry in tumor tissue. Int J Colorectal Dis 2001; 16: 22–27.

17.

Negm

Hamed

Schoen

, et al. Human blood autoantibodies in the detection of colorectal cancer. PloS one 2016; 11: e0156971.

18.

Kojima

Yoshikawa

Matsui

, et al. Titration of serum CEA, p53 antibodies and CEA-IgM complexes in patients with colorectal cancer. Mol Med Rep 2009; 2: 477–480.

19.

Suppiah

Alabi

Madden

, et al. Anti-p53 autoantibody in colorectal cancer: prognostic significance in long-term follow-up. Int J Colorectal Dis 2008; 23: 595–600.

20.

Megliorino

Shi

Peng

, et al. Autoimmune response to anti-apoptotic protein survivin and its association with antibodies to p53 and c-myc in cancer detection. Cancer Detect Prev 2005; 29: 241–248.

21.

Qiu

Pan

, et al. Detection of serum anti-P53 antibodies from patients with colorectal cancer in china using a combination of P53– and phage–ELISA: Correlation to clinical parameters. Asian Pac J Cancer Prev 2011; 12: 2921–2924.

22.

Scanlan

Welt

Gordon

, et al. Cancer-related serological recognition of human colon cancer-identification of potential diagnostic and immunotherapeutic targets. Cancer Res 2002; 62: 4041–4047.

23.

Pedersen

Gentry–Maharaj

Fourkala

, et al. Early detection of cancer in the general population: a blinded case-control study of p53 autoantibodies in colorectal cancer. Br J Cancer 2013; 108: 107–114.

24.

Looi

Megliorino

Shi

, et al. Humoral immune response to p16, a cyclin-dependent kinase inhibitor in human malignancies. Oncol Rep 2006; 16: 1105–1110.

25.

Lechpammer

Lukac

Lechpammer

, et al. Humoral immune response to p53 correlates with clinical course in colorectal cancer patients during adjuvant chemotherapy. Int J Colorectal Dis 2004; 19: 114–120.

26.

Tang

Wang

, et al. Humoral response to p53 in human colorectal tumors: a prospective study of 1209 patients. Int J Cancer 2001; 94: 859–863.

27.

Müller

Meyer

Schilling

, et al. Testing for anti-p53 antibodies increases the diagnostic sensitivity of conventional tumor markers. Int J Oncol 2006; 29: 973–980.

28.

Mao

Chen

, et al. Serum p53 protein and anti-p53 antibodies are associated with increased cancer risk: a case-control study of 569 patients and 879 healthy controls. Mol Biol Rep 2010; 37: 339–343.

29.

Hallak

Mueller

Lotter

, et al. p53 genetic alterations, protein expression and autoantibodies in human colorectal carcinoma: A comparative study. Int J Oncol 1998; 12: 7857–7891.

30.

Chen

Werner

Butt

, et al. Prospective evaluation of 64 serum autoantibodies as biomarkers for early detection of colorectal cancer in a true screening setting. Oncotarget 2016; 7: 16420–16432.

31.

Shiota

Ishida

Noguchi

, et al. Circulating p53 antibody in patients with colorectal cancer – Relation to clinicopathologic features and survival. Dig Dis Sci 2000; 45: 122–128.

32.

Kobayashi

Hoshino

Hiwasa

, et al. Anti-FIRs (PUF60) auto-antibodies are detected in the sera of early-stage colon cancer patients. Oncotarget 2016; 7: 82493–82403.

33.

Ostendorff

Awad

Braunschweiger

, et al. Multiplexed VeraCode bead-based serological immunoassay for colorectal cancer. J Immunol Methods 2013; 400–401: 58–69.

34.

Sthoeger

Evron

Goland

, et al. Anti-p53 autoantibodies in colon cancer patients. Ann N Y Acad Sci 1997; 815: 496–498.

35.

Chen

Lin

Qiu

, et al: Autoantibodies to Ca²⁺ binding protein Calnucis a potential marker in colon cancer detection. Int J Oncol 2007; 30: 1137–1144.

36.

Shimada

Ochiai

Nomura

, Japan p53 Antibody Research Group. Titration of serum p53 antibodies in 1085 patients with various types of malignant tumors: a multiinstitutional analysis by the Japan p53 Antibody Research Group. Cancer 2003; 97: 682–689.

37.

Nakamura

Hinoda

Nakagawa

, et al. Detection of circulating anti–MUC1 mucin core protein antibodies in patients with colorectal cancer. J Gastroenterol 1998; 33: 354–361.

38.

Pedersen

Blixt

Bennett

, et al. Seromic profiling of colorectal cancer patients with novel glycopeptide microarray. Int J Cancer 2011; 128: 1860–1871.

39.

Silk

Schoen

Potter

, et al. Humoral immune response to abnormal MUC1 in subjects with colorectal adenoma and cancer. Mol Immunol 2009; 47: 52–56.

40.

Wang

Zhang

Liu

, et al. Correlation between auto-antibodies to survivin and MUC1 variable number tandem repeats in colorectal cancer. Asian Pac J Cancer Prev 2012; 13: 5557–5562.

41.

Ben–Mahrez

Sorokine

Thierry

, et al. Circulating antibodies against c-myc oncogene product in sera of colorectal cancer patients. Int J Cancer 1990; 46: 35–38.

42.

Pedersen

Gentry-Maharaj

Nøstdal

, et al. Cancer-associated autoantibodies to MUC1 and MUC4—a blinded case-control study of colorectal cancer in UK collaborative trial of ovarian cancer screening. Int J Cancer 2014; 134: 2180–2188.

43.

Chen

Fan

, et al. Combined analysis of survivin autoantibody and carcinoembryonic antigen biomarkers for improved detection of colorectal cancer. Clin Chem Lab Med 2010; 48: 719–725.

44.

Yie

, et al. A protein fragment derived from DNA-topoisomerase I as a novel tumour-associated antigen for the detection of early stage carcinoma. Br J Cancer 2016; 115: 1555–1564.

45.

Chen

Chou

Chen

, et al. Detection of annexin A autoantibodies in sera from colorectal cancer patients. J Clin Gastroenterol 2011; 45: 125–132.

46.

Benvenuto

Sileri

Rossi

, et al. Natural humoral immune response to ribosomal P0 protein in colorectal cancer patients. J Transl Med 2015; 13: 101.

47.

Negm

Hamed

Schoen

, et al. Detection of autoantibodies to annexin A11 in different types of human cancer. Clin Proteomics 2016; 11: 125–131.

48.

Babel

Barderas

Diaz-Uriarte

, et al. Identification of MST1/STK4 and SULF1 proteins as autoantibody targets for the diagnosis of colorectal cancer by using phage microarrays. Mol Cell Proteomics 2011; 10: M110.001784.

49.

Fang

Lin

, et al. Circulating antibodies to carcinoembryonic antigen related to improved recurrence-free survival of patients with colorectal carcinoma. J Int Med Res 2011; 39: 838–845.

50.

Albanopoulos

Armakolas

Konstadoulakis

, et al. Prognostic significance of circulating antibodies against carcinoembryonic antigen (Anti-CEA) in patients with colon cancer. Am J Gastroenterol 2000; 95: 1056–1061.

51.

Konstadoulakis

Syrigos

Albanopoulos

, et al. The presence of anti-carcinoembryonic antigen (CEA) antibodies in the sera of patients with gastrointestinal malignancies. J Clin Immunol 1994; 14: 310–313.

52.

Liu

, et al. Humoral autoimmune response to IGF2 mRNA-binding protein (IMP2/p62) and its tissue-specific expression in colon cancer. Scand J Immunol 2013; 77: 255–260.

53.

Zhang

Chan

Peng

, et al. Autoimmune responses to mRNA binding proteins p62 and Koc in diverse malignancies. Clin Immunol 2001; 100: 149–156.

54.

Chen

Kuo

Chou

, et al. Detection of autoantibodies against Rabphilin-3A-like protein as a potential biomarker in patient’s sera of colorectal cancer. Clin Chim Acta 2011; 412: 1417–1422.

55.

Xia

Kong

Zhang

, et al. Proteomics-based identification of DEAD-box protein 48 as a novel autoantigen, a prospective serum marker for pancreatic cancer. Biochem Biophys Res Commun 2005; 330: 526–532.

56.

Song

, et al. Expression and immunogenicity of NY-ESO-1 in colorectal cancer. Exp Ther Med 2017; 13: 3581–3585.

57.

Fan

Chan

Chen

, et al. Identification of SEC61beta and its autoantibody as biomarkers for colorectal cancer. Clin Chim Acta 2011; 412: 887–893.

58.

Mou

, et al. Identification of IMPDH2 as a tumor-associated antigen in colorectal cancer using immunoproteomics analysis. Int J Colorectal Dis 2009; 24: 1271–1279.

59.

O’Reilly

Fitzgerald

, et al. Diagnostic potential of zinc finger protein-specific autoantibodies and associated linear B-cell epitopes in colorectal cancer. PloS one 2015; 10: e0123469.

60.

Tsunemi

Nakanishi

Fujita

, et al. Proteomics-based identification of a tumor-associated antigen and its corresponding autoantibody in gastric cancer. Oncol Rep 2010; 23: 946–956.

61.

Raiter

Vilkin

Gingold

, et al. The presence of anti-GRP78 antibodies in the serum of patients with colorectal carcinoma: a potential biomarker for early cancer detection. Int J Biol Markers 2014; 29: e431–435.

62.

Takahashi

Chen

Byrd

, et al. Antibody to ras proteins in patients with colon cancer. Clin Cancer Res 1995; 1: 1071–1077.

63.

Jagadish

Parashar

Gupta

, et al. A novel cancer testis antigen target A-kinase anchor protein (AKAP4) for the early diagnosis and immunotherapy of colon cancer. Oncoimmunology 2016; 5: e1078965.

64.

Hector

Chen

Kijanka

, et al. A reverse-ELISA for the detection of TRIM28/KAP1 serum autoantibodies in colorectal cancer patients. Acta Oncologica 2012; 51: 394–396.

65.

Jiang

Zhang

, et al. Elevated serum antibodies against insulin-like growth factor-binding protein-2 allow detecting early-stage cancers: Evidences from glioma and colorectal carcinoma studies. Ann Oncol 2012; 23: 2415–2422.

66.

Wang

Guo

, et al. Evaluation of antibody level against Fusobacterium nucleatum in the serological diagnosis of colorectal cancer. Sci Rep 2016; 6: 33440.

67.

Kocer

McKolanis

Soran

Humoral immune response to MUC5AC in patients with colorectal polyps and colorectal carcinoma. BMC Gastroenterol 2006; 6: 4.

68.

Mosolits

Harmenberg

Rudén

, et al. Autoantibodies against the tumour-associated antigen GA733-2 in patients with colorectal carcinoma. Cancer Immunol Immunother 1999; 47: 315–320.

69.

Syrigos

Charalampopoulos

Konstantoulakis

, et al. Autoantibodies against cardiolipin in the serum of patients with colorectal adenocarcinoma: Their prognostic significance. Eur Surg Res 1998; 30: 55–60.

70.

Luo

Yun

Fan

, et al. Cancer testis antigen OY-TES-1 expression and serum immunogenicity in colorectal cancer: Its relationship to clinicopathological parameters. Int J Clin Exp Pathol 2013; 6: 2835–2845.

71.

Jiang

Ren

Dong

, et al. Peptide mimic isolated by autoantibody reveals human arrest defective 1 overexpression is associated with poor prognosis for colon cancer patients. Am J Pathol 2010; 177: 1095–10003.

72.

Kanojia

Garg

Gupta

, et al. Sperm-associated antigen 9 is a novel biomarker for colorectal cancer and is involved in tumor growth and tumorigenicity. Am J Pathol 2011; 178: 1009–1020.

73.

Hanafusa

Mohamed

Domae

, et al. Serological identification of Tektin5 as a cancer/testis antigen and its immunogenicity. BMC Cancer 2012; 12: 520.

74.

Song

Lee

, et al. Identification of BCP-20 (FBXO39) as a cancer/testis antigen from colon cancer patients by SEREX. Biochem Biophys Res Commun 2011; 408: 195–201.

75.

Kiyamova

Garifulin

Gryshkova

, et al. Preliminary study of thyroid and colon cancers-associated antigens and their cognate autoantibodies as potential cancer biomarkers. Biomarkers 2012; 17: 362–371.

76.

Babel

Barderas

Díaz-Uriarte

, et al. Identification of tumor-associated autoantigens for the diagnosis of colorectal cancer in serum using high density protein microarrays. Mol Cell Proteomics 2009; 8: 2382–2395.

77.

Song

Shin

, et al. KP-CoT-23 (CCDC83) is a novel immunogenic cancer/testis antigen in colon cancer. Int J Oncol 2012; 41: 1820–1826.

78.

Nam

Madoz–Gurpide

Wang

, et al. Molecular profiling of the immune response in colon cancer using protein microarrays: Occurrence of autoantibodies to ubiquitin C-terminal hydrolase L3. Proteomics 2003; 3: 2108–2115.

79.

Ran

Zhou

, et al. Profiling tumor-associated autoantibodies for the detection of colon cancer. Clin Cancer Res 2008; 14: 2696–2700.

80.

Rimm

Holland

Morrow

, et al. Autoantibodies specific for villin found in patients with colon cancer and other colitides. Dig Dis Sci 1995; 40: 389–395.

81.

Villar-Vázquez

Padilla

Fernández-Aceñero

, et al. Development of a novel multiplex beads-based assay for autoantibody detection for colorectal cancer diagnosis. Proteomics 2016; 16: 1280–1290.

82.

Koziol

Zhang

Casiano

, et al. Recursive partitioning as an approach to selection of immune markers for tumor diagnosis. Clin Cancer Res 2003; 9: 5120–5126.

83.

Rohayem

Diestelkoetter

Weigle

, et al. Antibody response to the tumor-associated inhibitor of apoptosis protein survivin in cancer patients. Cancer Res 2000; 60: 1815–1817.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.55 MB

0.58 MB

0.46 MB

0.51 MB

Autoantibodies as biomarkers for colorectal cancer: A systematic review,meta-analysis,and bioinformatics analysis

Abstract

Keywords

Introduction

Methods

Search strategy

Study selection

Data extraction and quality assessment

Functional enrichment analysis

Integration of a protein–protein interaction network and module analysis

Statistical analyses

Results

Study selection and characteristics

Diagnostic value of single tumor-associated autoantibodies for any stage of CRC

Diagnostic value of multiple autoantibodies for patients at all stages of CRC

Diagnostic value for early stages of CRC

Meta-analysis for the anti-p53 autoantibody

Publication bias

Functional characterization of autoantibodies

PPI network construction and module screening

Discussion

Conclusion

Supplemental Material

Figure_S1 – Supplemental material for Autoantibodies as biomarkers for colorectal cancer: A systematic review, meta-analysis, and bioinformatics analysis

Supplemental Material

Figure_S2 – Supplemental material for Autoantibodies as biomarkers for colorectal cancer: A systematic review, meta-analysis, and bioinformatics analysis

Supplemental Material

Figure_S3 – Supplemental material for Autoantibodies as biomarkers for colorectal cancer: A systematic review, meta-analysis, and bioinformatics analysis

Supplemental Material

Figure_S4 – Supplemental material for Autoantibodies as biomarkers for colorectal cancer: A systematic review, meta-analysis, and bioinformatics analysis

Footnotes

Declaration of conflicting interests

Funding

ORCID iD

Supplemental material

References

Supplementary Material