Identification of potential cervical cancer serum biomarkers in Thai patients

Abstract

Cervical cancer is one of the most common causes of cancer-associated mortality in females worldwide. Serum biomarkers are important tools for diagnosis, disease staging, monitoring treatment and detecting recurrence in different types of cancer. However, only a small number of established biomarkers have been used for clinical diagnosis of cervical cancer. Therefore, the identification of minimally invasive, sensitive and highly specific biomarkers for detection of cervical cancer may improve outcomes. In the present pilot study, changes in disease-relevant proteins in 31 patients with cervical cancer were compared with 16 healthy controls. The Human 14 Multiple Affinity Removal system was used to deplete the 14 most abundant serum proteins to decrease sample complexity and to enrich proteins that exhibited decreased levels of abundance in the serum samples. Immunoaffinity-depleted serum samples were analyzed by in-gel digestion, followed by liquid chromatography mass spectrometry analysis and data processing. Automated quantitative western blot assays and receiver operating characteristic (ROC) curves were used to evaluate the differential protein expression levels between the two groups. Capillary electrophoresis-based western blot analysis was performed to quantitatively determine serum levels of the candidate biomarkers. Significantly increased levels of α-1-antitrypsin (A1AT) and pyrroline-5-carboxylate reductase 2 (PYCR2) were detected, whereas the levels of transthyretin (TTR), apolipoprotein A-I (ApoA-I), vitamin D binding protein (VDBP) and multimerin-1 (MMRN1) were significantly decreased in patients with cervical cancer compared with the healthy controls. ROC curve analysis indicated that the sensitivity and specificity was improved through the combination of the 6 candidate biomarkers. In summary, the results demonstrated that 6 candidate biomarkers (A1AT, PYCR2, TTR, ApoA-I, VDBP and MMRN1) exhibited significantly different expression between serum samples from healthy controls and patients with cervical cancer. These proteins may represent potential biomarkers for distinguishing patients with cervical cancer from healthy controls and for differentiation of patient subgroups.

Keywords: Human 14 Multiple Affinity Removal System; apolipoprotein A-I; cervical cancer; multimerin-1; pyrroline-5-carboxylate reductase 2; serum biomarkers; transthyretin; vitamin D binding protein; α-1-antitrypsin.

Figure 1.

Gene Ontology analysis of the proteins identified by liquid chromatography mass spectrometry in the bound and flow-through fractions. (A and B) Molecular functions of the (A) bound and (B) flow-through fractions. (C and D) Biological processes of the (C) bound and (D) flow-through fractions. (E and F) Cellular compartments of proteins in the (E) bound and (F) flow-through fractions were determined using the Protein Analysis Through Evolutionary Relationships classification system.

Figure 2.

Venn diagrams of the proteins identified by liquid chromatography mass spectrometry in the bound and flow-through fractions. Venn diagram showing unique and overlapping proteins in (A) bound and (B) flow-through fractions between the three sample groups: N negative; N positive; and patients with cervical cancer. N negative, Normal controls with HPV negative; N positive, normal controls with HPV positive.

Figure 3.

Western blot analysis. (A) Levels of A1AT, VDBP, TTR, ApoA-I, MMRN1 and PYCR2 in pooled serum samples from N negative, N positive and stage I–IV cervical cancer patients. (B) A Coomassie-stained western blot membrane indicating the total protein profile as a loading control. N negative, Normal controls with HPV negative; N positive, normal controls with HPV positive; A1AT, α-1-antitrypsin; PYCR2, pyrroline-5-carboxylate reductase 2; TTR, transthyretin; ApoA-I, apolipoprotein A-I; VDBP, vitamin D binding protein; MMRN1, multimerin-1.

Figure 4.

Quantitative measurement of protein expression levels using capillary electrophoresis. (A) Electropherogram of the A1AT, PYCR2, TTR, ApoA-I, VDBP and isoform-2 of MMRN1 levels in serum samples from the HPV negative normal controls (N1-N5), HPV positive normal controls (N6-N15), and patients with stage I (C1-C5), stage II (C6, C7, C10-C14 and C16), stage III (C23-C27) and stage IV (C28-C31) cervical cancer. (B) Box plots demonstrating the distribution of protein expression in serum samples from the normal and stage I–IV cervical patients. *P<0.05, **P<0.01 and ***P<0.001. A1AT, α-1-antitrypsin; PYCR2, pyrroline-5-carboxylate reductase 2; TTR, transthyretin; ApoA-I, apolipoprotein A-I; VDBP, vitamin D binding protein; MMRN1, multimerin-1.

Figure 5.

Performance of the best marker models for each comparison. AUC, sensitivity with 95% CI and specificity with 95% CI are stated. (A) Comparison of normal controls with HPV-negative vs. early-stage (N- vs. E). (B) Comparison of normal controls with HPV-negative vs. late stage (N- vs. L). (C) Comparison of normal controls with HPV-positive vs. early-stage (N+ vs. E). (D) Comparison of normal controls with HPV-positive vs. late stage (N+ vs. L). (E) Comparison of normal controls with HPV-negative vs. normal controls with HPV-positive (N- vs. N+). (F) Comparison of early-stage vs. late stage cervical cancer (E vs. L). ROC, receiver operating characteristic; AUC, area under the ROC curve; CI, confidence interval; A1AT, α-1-antitrypsin; PYCR2, pyrroline-5-carboxylate reductase 2; TTR, transthyretin; ApoA-I, apolipoprotein A-I; VDBP, vitamin D binding protein; MMRN1, multimerin-1.