Conceived and designed the experiments: RR PGW. Performed the experiments: AKCS. Analyzed the data: JK IK JA. Contributed reagents/materials/analysis tools: CM PGW. Wrote the paper: JK IK RR. Retrieved and collected clinicopathological information for creating and organizing database: IK JA.
The authors have declared that no competing interests exist.
Biotinidase was identified in secretome analysis of thyroid cancer cell lines using proteomics. The goal of the current study was to analyze the expression of biotinidase in thyroid cancer tissues and fine needle aspiration (FNA) samples to evaluate its diagnostic and prognostic potential in thyroid cancer. Immunohistochemical analysis of biotinidase was carried out in 129 papillary thyroid cancer (PTC, 34 benign thyroid tissues and 43 FNA samples and correlated with patients’ prognosis. Overall biotinidase expression was decreased in PTC compared to benign nodules (p = 0.001). Comparison of aggressive and non-aggressive PTC showed decrease in overall biotinidase expression in the former (p = 0.001). Loss of overall biotinidase expression was associated with poor disease free survival (p = 0.019, Hazards ratio (HR) = 3.1). We examined the effect of subcellular compartmentalization of nuclear and cytoplasmic biotinidase on patient survival. Decreased nuclear expression of biotinidase was observed in PTC as compared to benign tissues (p<0.001). Upon stratification within PTC, nuclear expression was reduced in aggressive as compared to non-aggressive tumors (p<0.001). Kaplan-Meier survival analysis showed significant association of loss of nuclear biotinidase expression with reduced disease free survival (p = 0.014, HR = 5.4). Cytoplasmic biotinidase expression was reduced in aggressive thyroid cancers in comparison with non-aggressive tumors (p = 0.002, Odds ratio (OR) = 0.29) which was evident by its significant association with advanced T stage (p = 0.003, OR = 0.28), nodal metastasis (p<0.001, OR = 0.16), advanced TNM stage (p<0.001, OR = 0.21) and extrathyroidal extension (p = 0.001, OR = 0.23). However, in multivariate analysis extrathyroidal extension emerged as the most significant prognostic marker for aggressive thyroid carcinomas (p = 0.015, HR = 12.8). In conclusion, loss of overall biotinidase expression is a novel marker for thyroid cancer aggressiveness.
Thyroid cancer is the most common malignant endocrine tumor and accounts for >90% of cancers of the endocrine glands, with an estimated annual incidence of 122,803 cases worldwide
At present, fine-needle aspiration (FNA) is the most commonly used pre-operative technique for diagnosis of thyroid nodules >1 cm in size. However, even the use of ultrasound-guided FNA is often beset with inconclusive biopsy results (10–20% of all cases)
In search of new cancer biomarkers for this malignancy, we analyzed the secretome from thyroid cancer cell lines to identify cancer-relevant secreted proteins that can serve as potential biomarkers
This study was approved by the Research Ethics Board of Mount Sinai Hospital, Toronto, Ontario, Canada. Archived formalin-fixed paraffin-embedded tissue blocks from the Mount Sinai Hospital Tumor Bank were retrieved, reviewed by the pathologist (CM), and cut to provide 5 µm-thick sections for immunohistochemical staining. Diagnoses were derived from histopathological analyses and clinical examination. Benign cases included multinodular goiters, hyperplastic nodules, and follicular adenomas, whereas all non-aggressive and aggressive tumors examined here were well-differentiated papillary thyroid carcinomas which included the following variants: classic, follicular, oncocytic, diffuse sclerosing, and tall cell. Defining features of tumor aggressiveness were TNM stage IV classification, distant metastasis, perineural invasion, and disease recurrence or persistence. Additional consideration was given to the following potential risk factors of aggressiveness including: TNM stage III, vascular invasion, extrathyroidal extension, lymph node metastasis, and non-classical papillary thyroid cancer variant type (especially the presence of three or more of these factors in the cases examined).
Based on these criteria, the 163 tissue samples examined in this study were categorized as 34 benign (median age: 51 years; range: 16 to 76 years), 81 non-aggressive (median age: 44 years; range: 23 to 71 years), and 48 aggressive (median age: 52 years; range: 21 to 86 years) tumors.
Formalin-fixed paraffin-embedded cell blocks from fine-needle aspiration preparations from 43 patients were similarly obtained from the Mount Sinai Hospital Tumor Bank and each sample diagnosis was confirmed by histological examination of the respective thyroidectomy specimen (CM). Accordingly, the FNA samples were categorized as 23 benign, 13 non-aggressive, and 7 aggressive cases.
The patient follow up data were retrieved from the clinical database to correlate the protein expression in tumors with clinical outcome for evaluation of the prognostic relevance of biotinidase. The patients were followed up for a maximum period of 19.5 years.
Slides were immunostained as described previously
Immunopositive staining was evaluated in five areas of the tissue sections as described
Of the 129 thyroid cancer cases, follow-up data were available for 116 (90%) patients, while 13 patients (10%) were lost to follow-up. Thyroid cancer patients were monitored for a maximum period of 19.5 years (Range 2–234 months; mean 43 months and median 29 months). Recurrence with or without metastases was observed in 19 of 116 (16.4%) patients monitored during the follow-up. Ninety seven patients who did not show recurrence were alive until the end of the follow-up period. Disease-free survival was expressed as the number of months from the date of surgery to recurrence or till the last possible follow up in case of patients who are disease free.
Clinico-pathologic parameters | Total N | Cyt Positiven (%) | Cyt P value | Cyt OR (95% C.I) | Nuc Positiven (%) | Nuc P value | Nuc OR (95% C.I) | Overall Positiven (%) | overall P value | Overall OR (95% C.I) |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
||||||
|
59 | 38 (64.4) | 0.573 | 1.29 (0.62–2.69) | 24 (40.7) | 0.266 | 0.63 (0.30–1.29) | 40 (67.8) | 0.703 | 0.84 (0.40–1.79) |
|
70 | 49 (70) | 21 (30) | 50 (71.4) | ||||||
|
105 | 72 (68.6) | 0.631 | 0.76 (0.30–1.92) | 37 (35.2) | 1.000 | 0.92 (0.36–2.35) | 74 (70.5) | 0.806 | 1.19 (0.46–3.08) |
|
24 | 15 (62.5) | 8 (33.3) | 16 (66.7) | ||||||
|
96 | 72 (75) |
|
0.28 (0.12–0.63) | 42 (43.8) |
|
0.13 (0.04–0.45) | 75 (78.1) |
|
4.29 (1.85–9.92) |
|
33 | 15 (45.5) | 3 (9.1) | 15 (45.5) | ||||||
|
80 | 66 (82.5) |
|
0.16 (0.07–0.36) | 40 (50) |
|
0.11 (0.04–0.32) | 68 (85) |
|
6.96 (3.02–15.99) |
|
49 | 21 (42.9) | 5 (10.2) | 22 (44.9) | ||||||
|
97 | 74 (76.3) |
|
0.21 (0.09–0.50) | 41 (42.3) |
|
0.20 (0.06–0.60) | 76(78.4) |
|
4.65 (1.99–10.88) |
|
32 | 13 (40.6) | 4 (12.5) | 14(43.8) | ||||||
|
81 | 63 (77.8) |
|
0.29 (0.13–0.62) | 38 (46.9) |
|
0.19 (0.08–0.48) | 65 (80.2) |
|
3.74 (1.70–8.21) |
|
48 | 24 (50) | 7 (14.6) | 25 (52.1) | ||||||
|
41 | 31 (75.6) | 0.227 | 0.57 (0.25–1.30) | 18 (43.9) | 0.167 | 0.57 (0.26–1.22) | 31 (75.6) | 0.411 | 1.52 (0.66–3.53) |
|
88 | 56 (63.6) | 27 (30.7) | 59 (67) | ||||||
|
43 | 25 (58.1) | 0.117 | 1.86 (0.86–4.01) | 14 (32.6) | 0.845 | 1.17 (0.54–2.53) | 26(60.5) | 0.110 | 0.53 (0.24–1.15) |
|
86 | 62 (72.1) | 31 (36) | 64 (74.4) | ||||||
|
64 | 40 (62.5) | 0.263 | 1.57 (0.75–3.29) | 17 (26.6) | 0.065 | 2.09 (1–4.39) | 41 (64.1) | 0.183 | 0.58 (0.27–1.25) |
|
65 | 47 (72.3) | 28 (43.1) | 49 (75.4) | ||||||
|
126 | 84 (66.7) | 0.550 | 0.67 (0.59–0.75) | 44 (34.9) | 1.000 | 0.93 (0.08–10.57) | 87 (69) | 0.553 | 0.69 (0.61–.078) |
|
3 | 3 (100) | 1 (33.3) | 3 (100) | ||||||
|
94 | 72 (76.6) |
|
0.23 (0.1–0.52) | 40 (42.6) |
|
0.23 (0.08–0.63) | 75 (79.8) |
|
5.27 (2.28–2.16) |
|
35 | 15 (42.9) | 5 (14.3) |
|
15 (42.9) |
N = number of cases, Cyt = Cytoplasmic staining, Nuc = Nuclear staining, OR = Odds ratio,
C.I = Confidence interval, Non-agg = Non-aggressive, Agg = Aggressive
Paraffin embedded sections of benign thyroid nodules and malignant tumors were stained using anti-biotinidase polyclonal antibody as described in the Methods section: a) Benign tissue section showing nuclear and overall biotinidase immunostaining; b) Papillary non-aggressive thyroid cancer section illustrating reduction in nuclear staining and increase of cytoplasmic biotinidase immunostaining in the tumor cells; c) Papillary aggressive thyroid cancer section showing reduced overall (nuclear and cytoplasmic) biotinidase immunostaining; d) Thyroid cancer section used as a negative control, showing no immunoreactivity in cells (a–d, original magnification x 400).
The immunohistochemical data were subjected to statistical analysis using SPSS 17.0 software (SPSS Inc., Chicago, IL). Scatter plots were used to determine the distribution of cytoplasmic, nuclear and overall biotinidase expression in benign, non-aggressive and aggressive thyroid cancers. Sensitivity and specificity were calculated and quantified using receiver operating characteristic (ROC) analysis. Based on the optimal sensitivity and specificity as revealed by ROC analysis, cut-offs were generated for biotinidase protein expression. For overall biotinidase expression, a cut-off value of ≥3.6 was defined as positive immunoreactivity for statistical analysis. A cut-off value of ≥2 and ≥4 was defined as positive criterion for individual nuclear and cytoplasmic biotinidase immunopositivity respectively for detailed statistical examination. The relationships between biotinidase expression and clinicopathological parameters were tested using Chi-Square and Fischer’s exact test. Two-sided p values were calculated and p<0.05 was considered to be significant. Prognostic significance of overall, nuclear and cytoplasmic biotinidase expression was assessed by Kaplan-Meier survival and multivariate Cox-proportional Hazards regression analysis.
Scatter plots showing distribution of total immunostaining scores determined by IHC of tissue sections of benign (n = 34), non-aggressive (n = 81) and aggressive (n = 48) thyroid cancer tissues. The vertical axis gives the total immunohistochemical score as described in the Methods. a) Decreased nuclear accumulation of biotinidase was observed in aggressive and non-aggressive thyroid cancers analyzed. Scatter plot shows the reduction in nuclear biotinidase immunostaining with increasing tumor aggressiveness; b) Increased cytoplasmic accumulation of biotinidase was observed in the thyroid cancer analyzed with reduction in expression in aggressive compared to non-aggressive PTC; c) Scatter plot showed reduced overall biotinidase immunostaining with increasing tumor aggressiveness.
Benign vs Malignant Thyroid | ||||||
Biotinidase | AUC | Sensitivity (%) | Specificity (%) | NPV (%) | PPV (%) | Asymptotic significance |
Cytoplasmic | 0.662 | 66.9 | 74.0 | 36.2 | 90.8 | <0.001 |
Nuclear loss | 0.972 | 97.1 | 65.0 | 1.1 | 58.2 | <0.001 |
Overall loss | 0.816 | 79.4 | 69.2 | 5.3 | 59.7 | <0.001 |
|
||||||
|
|
|
|
|
|
|
Cytoplasmic | 0.347 | 76.2 | 49.0 | 71.9 | 54.4 | 0.003 |
Nuclear loss | 0.696 | 46.4 | 85.7 | 84.8 | 48.3 | <0.001 |
Overall loss | 0.715 | 42.9 | 89.8 | 71.7 | 56.1 | <0.001 |
AUC, area under the curve; NPV, negative predictive value; PPV, positive predictive value.
The vertical axis of each curve indicates sensitivity and the horizontal axis indicates the 1-specificity. The sensitivity, specificity, and AUC values are summarized in
Of the 34 benign tissues analyzed, 33 tissues (97.1%) showed overall as well as nuclear accumulation of biotinidase protein (
FNA specimens from benign (Panel a), non-aggressive papillary thyroid cancer (Panel b), and aggressive papillary thyroid cancer (Panel c) cases were immunostained with 1∶100 α-biotinidase K-17 rabbit polyclonal antibody. FNA specimen used as a negative control shows no immunoreactivity in cells (Panel d). Photomicrographs show a pronounced decrease in nuclear biotinidase expression in more aggressive thyroid cancer cases and are presented at 400× original magnification.
The reduction in overall biotinidase expression significantly correlated with advanced T stage (p = 0.001, OR = 4.29), nodal metastasis (p<0.001, OR = 6.96), Stage III+IV tumors (p = 0.001, OR = 4.65), and extrathyroidal extension (p<0.001, OR = 5.27). Furthermore, loss of cytoplasmic and nuclear biotinidase individually also showed significant association with advanced T stage (p = 0.003, p<0.001 respectively), nodal metastasis (p<0.001, p<0.001 respectively), stage III+IV tumors (p<0.001, p = 0.002 respectively), and extrathyroidal extension (p = 0.001, p = 0.003 respectively;
Disease free survival curves showing biotinidase expression in (a) nucleus [median disease-free survival 116 months]; (b) cytoplasm [median disease-free survival 111 months]; (c) and overall (nucleus and cytoplasm) [median disease-free survival 109 months]. Disease free survival curves showing (d) extrathyroidal extension [median disease-free survival 36 months], and (e) nodal status [median disease-free survival 54 months].
95% CI for HR | |||||
Univariate | Parameter | P value | Hazards ratio (HR) | Lower | Upper |
Age | 0.595 | 0.8 | 0.3 | 1.9 | |
Gender | 0.136 | 2.1 | 0.8 | 5.5 | |
Tumor stage |
|
2.2 | 1.5 | 3.2 | |
Overall stage | 0.140 | 1.3 | 0.9 | 1.8 | |
Histology type | 0.676 | 1.1 | 0.8 | 1.4 | |
Histology grade | 0.374 | 23.9 | 0.02 | 26100.0 | |
Nodal status |
|
12.2 | 2.8 | 53.8 | |
Extrathyroidal extension |
|
9.0 | 3.1 | 26.4 | |
Biotinidase Nuclear loss |
|
5.4 | 1.2 | 24.1 | |
Biotinidase Cytoplasmic loss | 0.022 | 2.8 | 1.1 | 7.2 | |
Biotinidase Overall loss |
|
3.1 | 1.2 | 7.8 | |
|
|||||
Extrathyroidal extension |
|
4.1 | 1.3 | 12.8 | |
Nodal status |
|
5.6 | 1.1 | 27.8 |
Receiver operating characteristic curve analysis was used to determine the potential of biotinidase expression as a biomarker to distinguish benign nodules and malignant tumors (
We analyzed the expression of biotinidase in 23 benign and 20 malignant thyroid FNA samples (
The follow-up data of 116 thyroid cancer patients for up to 19.5 years were used to assess the prognostic relevance of biotinidase for predicting recurrence in these patients after completion of primary treatment. Kaplan-Meier survival analysis showed significantly reduced disease free survival in patients with decreased expression of biotinidase in nucleus (p = 0.014, HR = 5.4, 95% C.I. = 1.2–24.1; median survival 116 months,
There is an unmet need to identify novel biomarker(s) that can not only help distinguish benign thyroid from malignant tumors but also have the ability to discriminate between aggressive and non-aggressive thyroid cancers. The findings of our study suggest that biotinidase satisfies both these criteria. In addition, our study demonstrates biotinidase levels are decreased in aggressive thyroid carcinomas and suggests its potential to serve as a marker for tumor aggressiveness. This novel use of biotinidase underscores its potential to serve as a tool to identify aggressive thyroid cancers in early stages for more focused therapy. Recently Kang
FNA is an essential tool for the management of thyroid nodules and FNA of all thyroid nodules >1 cm has been recommended
Classically age, gender, tumor stage, extrathyroidal extension (spread outside the thyroid capsule), nodal status, histology type, histology grade all have prognostic significance and were observed to be associated with biotinidase expression in our study. To determine independent prognostic significance for biotinidase, all these conventional markers of poor prognosis for thyroid cancer were incorporated into a multivariate model and the additional significance of biotinidase was assessed. However, in multivariate analysis extrathyroidal extension and nodal status emerged more significant than biotinidase as markers for poor prognosis (p = 0.015, HR = 4.1, 95% C.I. = 1.3−12.8, and p = 0.035, HR = 5.6, 95% C.I. = 1.1−27.8 respectively). Given the association of biotinidase with conventional markers of poor prognosis, biotinidase did not have independent prognostic significance in this study cohort. Nevertheless, our findings are of significance in view of the limited studies on biotinidase in human cancers; importantly these few reports corroborate and support our observations.
The role biotinidase plays in cancer aggressiveness remains to be elucidated. A probable hypothesis would center on the role of biotin as a co-factor for a plethora of enzymes responsible for chromatin structure and stability. Biotinidase cleaves biocytin thereby making free biotin readily available. Loss of this enzyme would subsequently cause a biotin deficient state which would in turn affect histone biotinylation in chromatin remodeling. It is known that biotinylation of K12 in histone H4 is important for repair of DNA and heterochromatin structures as well as repression of genes and transposons to maintain genomic stability and reduce cancer risk in human cells and
BRAF(V600E) is considered a negative prognostic marker in PTC and might have been a confounding prognostic factor in our analysis. One of the limitations of our study is the non-availability of BRAF(V600E) mutation data in our cohort of thyroid cancer patients. The majority of PTCs are initiated by genetic events involving mutation of BRAF or RAS and translocations producing RET/PTC oncogenes