Ethical clearance from the Human Ethics Committee of All India Institute of Medical Sciences was obtained before commencement of the study (IESC T-244/2012), and informed consent from all patients or their legally acceptable representatives was taken before their inclusion. All patients included in this study underwent treatment at the Department of Gastrointestinal Surgery (GIS) at the All India Institute of Medical Sciences (AIIMS), New Delhi.

In the present study, surgically resected gallbladder tumor tissues were obtained from GBC patients (n = 43) who underwent a presumed curative surgical resection. All cases were staged clinically, according to the American Joint Committee on Cancer (AJCC) guidelines. Tissue samples obtained from patients undergoing cholecystectomy for gallstone disease and from patients with periampullary carcinoma where the normal gallbladder was removed as a part of pancreaticoduodenectomy were also collected and served as controls (Total Controls, n = 69). These gallbladders were histologically proven chronic cholecystitis with no evidence of any malignancy. A section of all resected gallbladder tissues (both tumor and controls) was immediately snap-frozen in liquid nitrogen and stored at −80°C to be used for enzymatic assays and RNA isolation. Another portion of the resected tissue was fixed in 10% neutral buffered formalin solution for immunohistochemical analysis.

Further, preoperative serum samples were obtained from cytologically proven cases of GBC (n = 66, median age 54 years, males = 23 and females = 43) including both resectable and locally advanced/metastatic GBC, attending the Outpatient Department of GIS. For controls, cholecystitis patients with gallstone gallbladder disease (GSGB) who underwent cholecystectomy (n = 34) and healthy individuals (n = 20) with no active inflammation, gallstones, or malignancy were recruited in the present clinical setup. All blood samples were processed for serum isolation and stored at −80°C until used for further analysis. The schematic representation of sample collection and workflow for GBC patients and controls is outlined in Figure 1.

A total of 5–10 mg of frozen gallbladder tissue was lysed in Tris-HCl buffer (50 mM Tris-HCl, pH6.8; 150 mM NaCl; 10% Glycerol; 1% Nonidet P-40). After two cycles of freeze-thaw, the homogenate was centrifuged at 10,000 × g at 4°C for 15 min to remove cell debris. Subsequently, total protein in the supernatant was estimated by BCA protein estimation. An equal amount of 100 μg protein was used to assay the combined CTSL+B activity in the supernatant using CBZ-Phe-Arg-NMec (Sigma-Aldrich, U.S.A), a synthetic fluorogenic substrate. Simultaneously, the same assay was performed in the presence of 5 μM CA074 Me (Calbiochem, Germany), a specific CTSB inhibitor to measure CTSL activity. Values obtained from CTSL activity were subtracted from total CTSL+B activity to calculate CTSB enzyme activity. The enzymatic activities were expressed as Relative Fluorescence Units (RFU)/min/mg protein.

Immunohistochemistry was performed on 4 μm thick paraffin-embedded tissue sections of control and carcinoma gallbladder using mouse monoclonal anti CTSL (1:500, ab6314, Abcam, USA) and CTSB antibody (1:200, ab58802, Abcam, USA) as described previously (16). Briefly tissue sections were mounted on glass slides and deparaffinized in xylene, xylene: alcohol, and alcohol gradients, followed by antigen retrieval in citrate buffer (0.01 M, pH 6:0) using a microwave oven. The slides were allowed to cool at room temperature and incubated with 3% serum for 30 min to preclude non-specific binding. These sections were then incubated with the primary antibody in a humidified chamber at 4°C overnight. On the following day, slides were washed thrice with Tris buffer saline (TBS), and sections were treated with hydrogen peroxide (0.3% v/v in ethanol) for 20 min to quench the endogenous peroxidase activity. Primary antibody was detected using VECTASTAIN® Elite® ABC peroxidase kit as per the manufacturer's protocol using diaminobenzidine (DAB) as the chromogen. Finally, the sections were counterstained with Mayer's hematoxylin and mounted with D.P.X. Tissue sections were microscopically examined.

Two pathologists independently performed immunohistochemistry scoring. A semi-quantitative H-score was used for evaluating CTSL and CTSB immunostaining pattern. It was calculated by the addition of staining intensity and percentage distribution scores evaluated in epithelial cells, tumor cells, and endothelial cells. The staining intensity was scored from 0 to 3 as follows: 0 (no staining), 1 (weak), 2 (moderate), and 3 (strong). Percentage distribution was scored as follows: No staining = 0, < 40% = 1, 40–80% = 2, > 80% = 3. Thus, the H-score ranged from 0 to 6. Based on this score, immunoexpression of CTSL and CTSB was divided into two subgroups: Low (H score, 1–4) and high (H score 5–6) expression groups.

Total cellular RNA was extracted from snapped frozen gallbladder tissues using TRI Reagent BD™ (Sigma-Aldrich, U.S.A) as per the manufacturer's instructions from snapped frozen tissues. The procedure followed for cDNA preparation, and qPCR has been described previously (17). Transcripts were normalized using 18S ribosomal RNA. After normalization, 2^−ΔCt values obtained were plotted in a dot plot. The nucleotide sequences of primer sets used for human CTSL, CTSB, and 18S are given in Supplementary Table 1.

Levels of serum CTSL and CTSB were assayed using a commercially available sandwich ELISA kits (RayBiotech, Norcross, GA and Bosterbio, Pleasanton, CA, U.S.A.) as per the manufacturer's instructions as described in detail elsewhere (18).

STATA 13.1 software (Lakeway Drive College Station, Texas USA) was used for all statistical evaluation. The skewed distributed variables were analyzed by the Mann-Whitney U test and were presented as median (interquartile range, IQR). The cut-off values of CTSL and CTSB levels (in tissues and serum) were determined using receiver operating characteristic (ROC) curves analysis along with their associated diagnostic measures. Chi-square test assessed the association between two categorical variables in semi-quantitative immunohistochemistry scoring. Of the 43 patients who were operated, only 35 underwent curative surgery (Radical cholecystectomy); as a result, these 35 patients were enrolled in the follow-up protocol to assess the association of CTSL and CTSB over expression with survival after presumed curative resection. The additional 8 patients could not undergo a radical cholecystectomy in view of metastatic or locally advanced unresectable disease that was detected during surgery. The overall survival time of the patients (in months) was calculated from the date of their surgery to the date of the last follow-up. To verify whether CTSL+B, CTSL, or CTSB expression correlated with survival, these variables were dichotomized according to their median values, and Kaplan–Meier analysis was performed using log-rank tests. Spearman rank correlation coefficient (ρ) analysis was used to assess the correlations between the activity and mRNA levels of cysteine cathepsins. All tests were two-sided, and ^*p ≤ 0.05, ^**p ≤ 0.01, and ^***p ≤ 0.001 were considered to be statistically significant.

All subjects enrolled in the present study were ethnic South Asians residing in the northern part of India. The median age group of GBC patients who underwent curative surgical resection in the study group was 56 years. Out of 43 patients, 11 were males and 32 females. Based on the TNM stage classification, 17 GBC patients were stage I ± II and 26 were stage III ± IV. The other details including tumor size, histological type, grading, the involvement of lymph node and liver in these patients are given in Table 1.

The enzymatic activity of CTSL and CTSB was assayed in tumor and control gallbladder tissues using a synthetic fluorogenic substrate CBZ-PheArg-NMec and expressed as RFU/min/mg protein. The median CTSL+B activity in GBC patients was 109.6 × 10³ RFU/min/mg protein (IQR 44.6–177.4) as compared to 22.3 × 10³ RFU/min/mg protein (IQR 8.2–73.1) in control gallbladder. Thus, the median CTSL+B activity was significantly higher (5-fold) in GBC with respect to controls (p < 0.0001, Figure 2A). Similarly, the median CTSL activity in the same set of patients was 6.0 × 10³ RFU/min/mg protein (IQR 2.4–11.4), ~2.4-fold more than that in controls (2.6 × 10³ RFU/min/mg protein, IQR 0.7–6.6, p = 0.0004, Figure 2B). Interestingly, as compared to controls (17.7 × 10³ RFU/min/mg protein, IQR 7.1–65.7,) a drastic increase of around 5-fold in CTSB activity was also observed in GBC patients (98.9 × 10³ RFU/min/mg protein; IQR 47.2–170.1, p < 0.0001, Figure 2C). Furthermore, we observed a strong correlation between the activities of these proteases and clinicopathological parameters of GBC patients. The elevated activity of CTSL+B, as well as CTSB, displayed a significant association with the T stage and lymph node involvement (Table 1, p < 0.050).

Having established increased activities of both CTSL and CTSB in GBC as compared to control gallbladder with chronic cholecystitis, we then sought to examine the diagnostic potential of these cathepsins in GBC using ROC analysis (Figure 3A and Table 2).

Our analysis revealed an optimal cut-off value of 68 × 10³ RFU/min/mg protein for CTSL+B in GBC with 72% diagnostic sensitivity and an AUC value of 80%. Thus, 72% of GBC patients (31/43) displayed higher activity of CTSL+B than the assigned cut-off value. For CTSL, the identified cut off was 4.5 × 10³ RFU/min/mg protein, with 62% diagnostic sensitivity and an AUC value of 69%. This again translated to a reasonable diagnostic accuracy with 62% (27/43) of GBC patients having higher values than the calculated cutoff for CTSL. Similarly, the identified cut-off value for CTSB was 63.1 × 10³ RFU/min/mg protein, with 72% sensitivity and an AUC value of 80%, again distinguishing over 72% of GBC patients with increased CTSB activity as compared to only 28% (12/43) of cases with values lower than the cut-off. These results imply the excellent diagnostic performance of combined CTSL+B and CTSB activity values in differentiating GBC from chronic cholecystitis in gallbladder tissues.

Further, we aimed to determine if the higher activity of these cathepsins also correlates with poor survival in GBC patients who underwent curative surgical resection (n = 35). For this, the median CTSL+B activity values were taken as a cut-off to divide GBC patients into two subgroups, i.e., one with the higher median enzymatic activity and the other with the lower values than the cut-off. Kaplan-Meier analysis revealed significantly reduced overall survival in patients with higher CTSL+B activity (median value >107.5 × 10³ RFU/min/mg protein) as compared to patients with lower values (p = 0.049, log-rank test, hazard ratio 2.93, 95% CI = 1.02–8.39, Figure 3B). These results indicate that a combined increase in the activities of CTSL+B not only correlates with the active disease but also portends a poor prognosis in GBC.

Next, we sought to study the protein expression and localization of CTSL and CTSB in gallbladder tissues. For this immunohistochemical analysis was performed on paraffin-embedded tissue sections of gallbladder obtained from GBC patients and controls. A weak cytoplasmic immunopositivity was noted for both CTSL and CTSB at the apical surface of the columnar epithelium and endothelial cells of control gallbladder. However, staining for both these proteases was very strong in tumor cell populations, tumor-associated macrophages, and tumor endothelial cells of histologically proven GBC tissues (Figures 4A–I). Among total cases of GBC, 77% (31/40) displayed increased expression of CTSL (H-score>4) in tumor cells as compared to only 16% (10/60) of control gallbladder tissues (Figures 4A,B). Similarly, intense staining of CTSL in tumor endothelial cells was also noted in 55% (22/40) of GBC patients as compared to only 20% (12/60) in endothelial cells of controls (Figures 4D,E), indicating a remarkable increase in CTSL levels in GBC (p < 0.0001, Table 3). Interestingly, immunohistochemical staining of CTSB was markedly higher (H-score >4) in 67% (27/40) of the GBC patients compared to only 10% (6/60) in control gallbladder sections (Figures 4F,G). Likewise, 75% (30/40) of GBC cases also exhibited higher CTSB expression (H-score >4) in tumor endothelial cells as compared to only 11% (7/60) of controls, implying significant overexpression of CTSB protein (p < 0.0001, Figures 4H,I, Table 3) in this malignancy.

The quantitative real-time PCR analysis was used to assess the mRNA expression of these cysteine cathepsins in both tumor and control gallbladder tissues. For this purpose, the total RNA was isolated from flash-frozen gallbladder tissue obtained from both GBC patients and controls. However, qPCR could be performed only in 34 GBC patient tissues and 54 controls. In the remaining cases, either the tissue sample was limited, or the quality of isolated RNA was not up to the mark. The median 2^−ΔCt values of CTSL mRNA in GBC patients was 0.19 × 10⁻⁴ AU (range 0.01–0.99) as compared to 0.03 × 10⁻⁴ AU (IQR 0.01–0.22) in controls. Thus, a significant increase in median CTSL mRNA levels was observed in GBC (p = 0.023, Figure 5A). A similar increase of around 5-fold in median 2^−ΔCt values of CTSB mRNA (1.15 × 10⁻⁴ AU, IQR 0.09–4.39) was also noticed in GBC patients with respect to the controls (0.24 × 10⁻⁴ AU, IQR 0.04–1.02; p = 0.015, Figure 5B).

To ascertain the degree of association (if any) between enzymatic activities and mRNA levels of these proteases, Spearman rank correlation analysis was performed. As evident from this analysis, we observed a positive correlation (ρ = 0.519, p = 0.001, Figure 5C) between CTSL enzyme activity and its mRNA levels. Similarly, a strong positive correlation between CTSB enzyme activity and its mRNA expression was also evident in GBC patients (ρ = 0.473, p = 0.004, Figure 5D). Therefore, these findings indicate that the transcriptional upregulation of these proteases in GBC may be responsible for the higher protein levels and increased enzymatic activities of both CTSL and CTSB in this malignancy.

Having established the diagnostic and prognostic significance of CTSL and CTSB overexpression in the tumor tissues of GBC, it was of interest to investigate if the levels of these cathepsins in serum could accurately discern GBC patients from controls. To test this hypothesis, ELISA was employed to estimate the expression of CTSL and CTSB in the serum of GBC patients and controls. The median CTSL value in GBC patients was found to be 0.90 ng/ml (IQR 0.28–2.45). This value was significantly higher as compared to the serum levels of CTSL in both GSGB (0.16 ng/ml, IQR 0.05–0.95) and healthy controls (0.06 ng/ml, IQR 0–0.60, p < 0.0001, Figure 6A). Similarly, the median CTSB antigenic levels in GBC were found to be elevated (17.8 ng/ml, IQR 13.3–46.7) with respect to its level in both GSGB (10.20 ng/ml, IQR 6.80–17.40) and healthy controls (5.43 ng/ml, IQR 1.07–7.8, p < 0.0001, Figure 6B).

Serum antigenic expression of CTSL and CTSB obtained from GBC patients and controls was used to assess the diagnostic potential of these molecules. The ROC-based assessment revealed an optimal cut-off value of 0.30 ng/ml for CTSL, with 72% diagnostic sensitivity and an AUC value of 77% in distinguishing GBC from controls (Figure 6C, Table 4). Correspondingly, for CTSB, the chosen cutoff value was 11.6 ng/ml, with an excellent diagnostic sensitivity of 83% and an AUC value of 82% (Figure 6C, Table 4). Hence, our results for the first time demonstrate the diagnostic significance of both CTSL and CTSB as novel blood-based biomarkers in GBC.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.