Carbonic anhydrase I as a new plasma biomarker for prostate cancer

Abstract

Serum prostate-specific antigen (PSA) levels ranging from 4 to 10 ng/mL is considered a diagnostic gray zone for detecting prostate cancer because biopsies reveal no evidence of cancer in 75% of these subjects. Our goal was to discover a new highly specific biomarker for prostate cancer by analyzing plasma proteins using a proteomic technique. Enriched plasma proteins from 25 prostate cancer patients and 15 healthy controls were analyzed using a label-free quantitative shotgun proteomics platform called 2DICAL (2-dimensional image converted analysis of liquid chromatography and mass spectrometry) and candidate biomarkers were searched. Among the 40,678 identified mass spectrum (MS) peaks, 117 peaks significantly differed between prostate cancer patients and healthy controls. Ten peaks matched carbonic anhydrase I (CAI) by tandem MS. Independent immunological assays revealed that plasma CAI levels in 54 prostate cancer patients were significantly higher than those in 60 healthy controls (P = 0.022, Mann-Whitney U test). In the PSA gray-zone group, the discrimination rate of prostate cancer patients increased by considering plasma CAI levels. CAI can potentially serve as a valuable plasma biomarker and the combination of PSA and CAI may have great advantages for diagnosing prostate cancer in patients with gray-zone PSA level.

Figure 1

(a) Two-dimensional display of all MS peaks. The 117 MS peaks whose mean intensities significantly differed between prostate cancer patients and healthy controls (P < 0.05, Welch's t-test) are highlighted in red. RT: retention time. (b) 2DICAL images of peak ID 396 in a representative prostate cancer patient (left) and a healthy control (right).

Figure 2

(a) MS peaks of peak ID 396 in triplicate LC-MS runs (25 with prostate cancer (left) and 15 with healthy controls (right)) aligned along LC RT. Columns represent the mean intensities of triplicate runs. (b) Immunoblotting of CAI and complement C3b-α (loading control) for 7 prostate cancer patients selected from the most intense MS peaks and 7 healthy controls from the least intense peaks.

Figure 3

(a) Plasma CAI levels in patients with prostate cancer (n = 54), BPH (n = 22), prostatitis (n = 6), and RCC (n = 20) and healthy controls (n = 60) were 1.43 ± 0.69, 1.03 ± 0.51, 0.98 ± 0.46, 0.99 ± 0.36, and 1.09 ± 1.82 μg/mL (mean ± SD), respectively. There was a significant difference between prostate cancer patients and healthy controls (P = 0.020, Mann-Whitney U test). Horizontal lines represented the average levels. (b) Scatter plot correlating PSA and CAI levels in prostate cancer patients (red) and healthy controls (black). The dotted lines were added to indicate the PSA gray zone. Serum PSA levels were measured using the Tandem-R kit before the first treatment in each patient. (c) ROC curve of PSA plus CAI (red line) and PSA alone (blue line) confined to the cases with PSA levels in the gray zones. ROC curves were created using a composite index of the 2 markers generated from the results of multivariate logistic regression analysis. As a reference, ROC curves of CAI (black line) for all cases were included.

Figure 4

CAI expression in cell lines. (a) Immunofluorescence detection of CAI in normal prostate epithelial cells and prostate cancer cells. The cells were fixed for 1 h, permeabilized, and stained with antibodies to detect CAI. Alexa Fluor 488- or Alexa Fluor 568-phalloidin was used to visualize normal prostate epithelial cells (upper panels) and prostate cancer cells of 22Rv1 (lower panels). (b) Western blot analysis of media conditioned by the normal prostate epithelial cells and prostate cancer cells probed using a CAI antibody. C3b-α was used as a loading control.

Figure 5

Immunohistochemical staining of CAI. CAI was strongly stained at prostate cancer (magnification ×100).