Circulating nucleosomes as new blood-based biomarkers for detection of colorectal cancer

Abstract

Background: Colonoscopy is currently widely accepted as the gold standard for detection of colorectal cancer (CRC) providing detection of up to 95% of pre-cancerous lesions during the procedure. However, certain limitations exist in most countries including cost and access to the procedure. Moreover, colonoscopy is an invasive technique with risk inherent to the endoscopic procedure. For this reason, alternative screening tests, in particular, fecal occult blood-based tests, have been widely adopted for frontline screening. Limited compliance to colonoscopy and fecal screening approaches has prompted research on blood-based tests as an alternative approach to identifying individuals at risk who could then be referred for colonoscopy. Increased total levels of nucleosomes in the blood have been associated with tumor burden and malignancy progression. Here, we report for the first time, CRC-associated epigenetic profiles of circulating cell-free nucleosomes (cf-nucleosomes).

Methods: Levels of 12 epigenetic cf-nucleosome epitopes were measured in the sera of 58 individuals referred for endoscopic screening for CRC.

Results: Multivariate analysis defined an age-adjusted panel of four cf-nucleosomes that provided an AUC of 0.97 for the discrimination of CRC from healthy controls with high sensitivity at early stages (sensitivity of 75 and 86 at 90% specificity for stages I and II, respectively). A second combination of four cf-nucleosome biomarkers provided an AUC of 0.72 for the discrimination of polyps from the healthy group.

Conclusions: This study suggests that a combination of different cf-nucleosome structures analyzed in serum samples by a simple ELISA is a promising approach to identify patients at risk of CRC.

Keywords: Blood-based screening test; Colorectal cancer; Epigenetics; Nucleosomes.

Fig. 1

Combination of four cf-nucleosome biomarkers in CRC patients, patients with polyps, and healthy controls. Box plot demonstrating significantly higher score in patients with a CRC (n = 23) compared with healthy controls (n = 19) (p < 0.001) and between CRC patients (n = 23) and patients with polyps (n = 16) (p = 0.006). The score for each group was achieved with pre-processed NuQ® ELISA data from four cf-nucleosome biomarkers: histone modifications H2AK119Ub, H3K9Ac, H3K27Ac, and the global level of nucleosome. Fisher’s linear discriminant model was used to calculate the score. p values were determined by Mann-Whitney rank-sum test. The box plot shows the median and the 25th and 75th percentiles; the whiskers indicate the 5th and 95th percentiles. A.U. arbitrary unit

Fig. 2

ROC curves for discrimination of cancer vs. healthy controls. The area under the curve (AUC) for the established tumor marker CEA (0.66) was improved by using the best combination of four cf-nucleosome biomarkers (0.87); or the panel of four-cf-nucleosome biomarker panel with CEA (0.87). The AUC was further increased by using the age-adjusted four-cf-nucleosome panel (0.97); the gray line indicates random chance

Fig. 3

Discrimination of four NuQ® assay panels in an age-adjusted algorithm CRC, polyps, and healthy controls. The box plot shows significantly higher scores in patients with a CRC (n = 23) compared with healthy controls (n = 19) (p < 0.001) and between CRC patients (n = 23) and patients with polyps (n = 16) (p < 0.001). The score for each group was performed with pre-processed NuQ® ELISA data from four cf-nucleosome biomarkers: histone modifications H2AK119Ub, H3K9Ac, H3K27Ac, and the global level of nucleosome and age. Fisher’s linear discriminant model was used to calculate the score. p values were determined by Mann-Whitney rank-sum test. The box plot shows the median and the 25th and 75th percentiles; the whiskers indicate the 5th and the 95th percentiles. A.U. arbitrary unit

Fig. 4

Combination of four cf-nucleosome biomarkers in polyps vs. healthy controls. The box plot shows significantly higher scores in patients with a polyp (n = 16) compared with healthy controls (n = 19) (p = 0.025). Improved discrimination between the polyp and the healthy control groups was achieved using a second algorithm. The score for each group was achieved with pre-processed NuQ® ELISA data from four cf-nucleosome biomarkers: histone modifications H2AK119Ub, H3K9Ac, H4K20Me3, and the global level of nucleosome. Fisher’s linear discriminant model was used to calculate the score. p values were determined by Mann-Whitney rank-sum test. The box plot shows the median and the 25th and 75th percentiles; the whiskers indicate the 5th and the 95th percentiles. A.U. arbitrary unit

Fig. 5

ROC curve for the combination of four biomarkers distinguishing between patients with polyps and healthy controls. Area under the curve (AUC) 0.72. The gray line indicates random chance