ORIGINAL ARTICLE
Expression of LIN28 and its regulatory microRNAs in adult adrenocortical cancer
André M. Faria*, Silviu Sbierat, Tamaya C. Ribeiro*, Ibere C. Soares+ 11, Beatriz M.P. Mariani*, Daniel S. Freire##, Gabriela R.V. de Sousa*, Antônio M. Lerario*##, Cristina L. Ronchi§§, Timo Deutschbeint, Alda Wakamatsu+, Venancio A.F. Alvest, Maria Claudia N. Zerbini, Berenice B. Mendonca*, Maria Candida B.V. Fragoso*##, Ana Claudia Latronico*, Martin Fassnachtt’§§ and Madson Q. Almeida*##
*Unidade de Suprarrenal & Unidade de Endocrinologia do Desenvolvimento, Laboratório de Hormônios e Genética Molecular LIM42, Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo, São Paulo, Brasil, ¡Endocrine and Diabetes Unit, Department of Internal Medicine I, University Hospital, University of Wuerzburg, Wuerzburg, Germany, ¿Divisão de Anatomia Patológica, Laboratório de Patologia Hepática LIM14, Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo, São Paulo, Brasil, 1Hospital do Câncer de Barretos - Unidade Porto Velho, Porto Velho, ¿¿ Instituto do Câncer do Estado de São Paulo (ICESP) da Faculdade de Medicina da Universidade de São Paulo, São Paulo, Brasil and §§Comprehensive Cancer Center Mainfranken, University of Wuerzburg, Wuerzburg, Germany
Summary
Objective LIN28 control cells reprogramming and pluripotency mainly through miRNA regulation and has been overexpressed in many advanced cancers. In this study, we evaluated the prog- nostic role of LIN28 and its regulatory miRNAs in a large cohort of adrenocortical tumours (ACTs).
Patients and methods LIN28 protein expression was assessed in 266 adults ACTs (78 adenomas and 188 carcinomas) from Brazil and Germany. LIN28A and LIN28B gene expression was analysed in 59 ACTs (31 adenomas and 28 carcinomas) and copy number variation in 39 ACTs. In addition, we determined the expression of let-7 family, mir-9, mir-30 and mir-125 in 28 carcinomas.
Results LIN28A gene was overexpressed in aggressive ACCs when compared with adenomas and nonaggressive ACCs, but no LIN28A copy number variation was found in ACTs. Unex- pectedly, weak LIN28 protein expression was significantly associ- ated with reduced disease-free survival in ACC patients (P = 0-01), but for overall survival only a trend was detectable (P = 0-117). In the multivariate analysis, only Ki67 index ≥10% (HR 4.6, P = 0.000) and weak LIN28 protein expression (HR 2.0, P = 0.03) were independent predictors of recurrence in ACC patients. Interestingly, mir-9 expression, a negative LIN28A/B regulator, was significantly higher in aggressive than in nonaggressive ACCs [2076 (from 36 to 9307) vs 133-4 (from
2.4 to 5193); P = 0-011] and was highly associated with reduced overall (P = 0-01) and disease-free survival (P = 0-01). However, mir-9 prognostic role should be further evaluated in a larger cohort.
Conclusion Weak LIN28 protein expression was associated with recurrence in ACCs. Additionally, overexpression of mir-9, a negative LIN28A regulator, was associated with poor outcome.
(Received 10 July 2014; returned for revision 30 July 2014; finally revised 25 August 2014; accepted 29 August 2014)
Introduction
Adrenocortical cancer (ACC) is a rare malignancy with incom- pletely understood pathogenesis and poor prognosis.1-3 To date, systemic treatment of advanced ACC has showed unsatisfactory overall response.4,5 Over the past two decades, considerable advances have been made into the understanding of molecular mechanisms associated with tumour progression and metastasis.1 Overexpression of IGF2/IGFIR and constitutive activation of ß- catenin were identified as key factors involved in the develop- ment of ACC.6-8
MicroRNAs (miRNA) play a key role in the regulation of gene expression through post-transcriptional repression by interacting with complementary sequences of target mRNAs. miRNA expression in human tumours has been characterized by an overall downregulation.9 Recently, several studies demonstrated the potential of miRNA profiling in differentiating between adrenocortical adenomas and carcinomas, risk stratification and prognosis.1º Although these studies have produced some diver- gent data, mir-483 overexpression and mir-195 downregulation in adult ACCs seem consistent.11,12
Nevertheless, little is known about post-transcriptional regula- tion of miRNAs. LIN28 (also known as LIN28A) is a highly con- served RNA-binding protein that has emerged as a modulator of the processing of let-7, an important family of miRNAs widely known for its tumour-suppressive effects.13 LIN28B is the only homologue of LIN28 in humans, sharing structure and certain function.14 Besides from let-7, LIN28 has also shown to regulate and be regulated by mir-9, mir-30 and mir-125.14,15 By repress- ing let-7, LIN28 consequently activates many cancer driver genes, including promitogenic factors and cell cycle progression genes.13 Both LIN28 and LIN28B were overexpressed in a variety of advanced human malignancies.16 Importantly, IGF1R is an important target of the let-7 family and has been demonstrated to be upregulated in tumours that overexpress LIN28B.17,18 Moreover, LIN28A and B have also miRNA-independent func- tions.19,20 Polesskaya et al. 21 demonstrated that LIN28 directly enhances translation efficiency of IGF2 mRNA during murine muscle cell differentiation, a finding with potential implication to ACC.
The aim of our study was to investigate the expression of LIN28 protein in a large cohort of adult adrenocortical tumours (ACTs). Additionally, we analysed LIN28A and LIN28B expres- sion and the presence of copy number variation in subgroups of this cohort. Expression of the miRNAs (let-7 family, mir-9, mir- 30 and mir-125) involved with their post-transcriptional regula- tion was also investigated. To further characterize their role in tumour progression and prognosis, we correlated these findings with clinicopathological parameters and known molecular mark- ers, such as IGF1R, IGF2, ß-catenin and Ki67 index.
Material and methods
The study was approved by the Ethics Committees of the Hospi- tal das Clínicas, University of São Paulo and of the University of Wuerzburg, and informed written consent was obtained from all patients. The Weiss criteria were used to classify adenomas and carcinomas (Weiss score <3 and ≥3, respectively). LIN28 protein expression was assessed in a total of 266 adrenocortical tumours (ACTs, 78 adenomas and 188 carcinomas) and Ki67 protein expression was evaluated in 158 ACCs. Among them, 59 ACTs (31 adenomas and 28 carcinomas) were used to analyse LIN28A and LIN28B gene expression, 39 ACTs to study copy number variation study and 28 ACCs to determine let-7 family, mir-9, mir-30 and mir-125 expression (Figure S1). Further evaluation of mir-9 expression was also done in 13 adenomas.
The ACT cohort includes 156 tumours (78 adenomas and 78 carcinomas) from the University of São Paulo and 110 carcino- mas from the German ACC Registry. All tumours samples derived from primary surgery. Clinical parameters, such as sex, age at diagnosis, date of surgery, tumour size, pathological clas- sification and hormone analysis were collected from patient records. Tumour stage was classified according to the European Network for the Study of Adrenal Tumours (ENSAT) classifica- tion. Only patients with at least 12 months of follow-up were included in this study. The presence of distant metastases or
recurrence was evaluated at the time of diagnosis and during follow-up visits by computerized tomography of chest and abdo- men every 3-6 months. ACCs were further classified as aggres- sive or nonaggressive according to the presence or absence of locoregional recurrence and/or distant metastasis in the initial 12 months after first surgery.
Tissue microarray (TMA) and immunohistochemical analysis
Representative areas of the ACTs (viable tumour tissue without necrosis) were identified on haematoxylin- and eosin-stained slides and marked on paraffin donor blocks. The spotted areas of donor blocks were punched (1-0 mm punch) and mounted into 3 recipient paraffin blocks using a precision microarray instrument (Beecher Instruments, Sun Prairie, WI, USA). One set of three slides was selected (one slide from each of the 3 TMA paraffin blocks of the triplicate) for staining with anti- LIN28 rabbit polyclonal antibody (titre 1:100; 11724-1-AP, Pro- teinTech Group Inc, Chicago, IL, USA), which recognizes both LIN28A and LIN28B. An immunoperoxidase immunohisto- chemical modified method with humid heat antigen retrieval was used as previously described.22 LIN28 immunostaining was blindly evaluated by two independent observers (I.C.S. and S.S.) in both Brazilian and German cohorts, and the mean of the two evaluations was taken for statistical analysis. The interobserver reproducibility for LIN28 was 0-71. TMA samples were included in the analysis only if two or more evaluable cores were available after the staining procedure. Cytoplasmic staining was evaluated according intensity as negative (0), low (1), medium (2), or strong (3). The percentage of positive tumour cells was visually scored as follow: 0 if 0% of tumour cells were positive; 0-1 if 1- 9%; 0.5 if 10-49% and 1 if ≥50%. A semiquantitative H-score was then calculated by multiplying the staining intensity score by the proportion score, with a final score ranging from 0 to 3.23 The median H-score value from each ACC cohort was a pri- ori chosen as the cut-off point for separating tumours with low and strong LIN28 staining (Table 1).
Additionally, for the Brazilian cohort, another set of three slides was stained with mouse monoclonal anti-human Ki67 antigen (titre 1:40, clone MIB-1, code M7240, Dako, Glostrup, Denmark). Nuclear staining for Ki67 was blindly evaluated by two independent observers (I.C.S. and M.C.N.Z.), and the mean of the two evaluations taken for statistical analysis. The interob- server reproducibility for Ki67 analysis was 0.9. Each tumour sample was categorized by subjective assessment based on visual estimation of percentage of positive nuclei, as follows: 0-9%, 10-19%, 20-29%, 30-39%, 40-49%, 50-59%, 60-69%, 70-79%, 80-89% and ≥90%. In the German cohort, Ki67 analysis was performed in full slides of each patient which were stained with monoclonal antibody anti-human Ki67 antigen (titre 1:100, clone MIB-1, Dosseheim, Germany). Nuclear staining was evalu- ated by the reference pathologist of the German ACC Registry (Dr. Wolfgang Saeger), with all positive nuclei counted and reported to the number of high-power field.
| Weak LIN28 protein expression | Strong LIN28 protein expression | P | |
|---|---|---|---|
| n (%) | 120 (64%) | 68 (36%) | |
| H-score, median | 0-83 | 1.73 | |
| Age, median (year) | 45-8 | 41-5 | 0-13 |
| Sex [male, n (%)] | 37 (31) | 19 (28) | 0.68 |
| Tumour size, median (cm) | 12.0 | 10-2 | 0-05 |
| Tumour stage (ENSAT), n (%) | |||
| Known | 117 | 67 | 0-23 |
| 1-2 | 52 (59) | 36 (41) | |
| 3-4 | 65 (68) | 31 (32) | |
| Hormonal status, n (%) | |||
| Known | 83 | 53 | |
| Cushing | 37 (67) | 18 (33) | 0-22 |
| Noncushing | 46 (57) | 35 (43) | |
| Weiss score, median | 5 | 5 | 0-22 |
| Ki 67 index (%) | |||
| Known | 87 | 56 | |
| <10 | 52 (67) | 26 (33) | 0-12 |
| ≥10 | 35 (54) | 30 (46) | |
| Recurrence, n (%) | |||
| Known | 74 | 50 | |
| Affected patients | 48 (68) | 22 (32) | 0-02 |
| Nonaffected patients | 26 (48) | 28 (52) |
ACC, adrenocortical cancer.
Quantitative real-time RT-PCR (qRT-PCR)
After surgical resection, tumour fragments were immediately fro- zen in liquid nitrogen and stored at -80 ℃ until total RNA extraction using the Trizol reagent (Invitrogen, Carlsbad, CA, USA). RNA samples were treated with DNAse using standard procedures. cDNA was generated from 1 µg of total RNA using the commercial kit Superscript III First Strand S (Invitrogen). Quantitative real-time PCR was performed in the ABI Prism 7000 sequence detector using TaqMan gene expression assays accord- ing to the manufacturer’s instructions (Applied Biosystems, Carlsbad, CA, USA). The PCR cycling conditions were as follows: 2 min at 95 ℃, 40 cycles of 95 ℃ for 15 s and 60 ℃ for 30 s and a final step at 72℃ for 30 s. The assays for target genes were LIN28A (Hs04189307_g1) and LIN28B (Hs01013729_m1). ß- actin (ACTB, 4310881E) and ß-glucoronidase (GUSB, 4310888E) were used as endogenous genes for normalization.
For measuring miRNAs expression, single-stranded cDNA was synthesized from 1 µg of total RNA using Megaplex RT Primers, Human Pool A v2.1 (PN 4399966; Applied Biosystems) and TaqMan MicroRNA Reverse Transcription Kit (PN 4366596; Applied Biosystems). PCR products were amplified using Taq- man Universal Master Mix II, no UNG (PN 4440040; Applied Biosystems). The following TaqMan MicroRNA Assays were used: let-7a (000377), let-7b (002619), let-7c (000379), let-7d (002283), let-7e (002406), let-7f (000382), let-7g (002282), mir- 98 (000577), mir-9 (000583), mir-30b (000602), mir-30c
(000419), mir-125a-3p (002199), mir-125a-5p (002198) and mir-125b (000449). RNU44 (001094) and RNU48 (001006) were used as endogenous genes for normalization. The relative expres- sion levels were analysed using the 2-AACT method. A commer- cial of pool of normal human adrenal cortex from autopsies was used as reference sample (Clontech, Palo Alto, CA, USA).
Multiplex ligation-dependent probe amplification (MLPA)
Probes were designed for each exon of LIN28A and LIN28B genes and are available upon request. SALSA MLPA KIT P300- A1 Human DNA Reference-2 was used as reference (MRC-Hol- land, Amsterdam, the Netherlands). This kit contains reference probes and control fragments specific for unique human DNA sequences. This probemix was developed to be added to ‘home- made’ synthetic MLPA probes, as previously described.24 PCR product was submitted to capillary electrophoresis on an ABI Prism 310 Genetic Analyzer (PE Applied Biosystems, The Per- kin-Elmer Corporation). Data analysis was performed with Genescan 3-7. Dosage quotient areas outside the range 0-70-1-3 were considered abnormal.
Statistical analysis
Statistical analysis was performed using spss Software (PASW version 19.0; SPSS Inc., Chicago, IL, USA). Continuous data are expressed as median values (from minimum to maximum). Dif- ferences in expression levels between two groups were analysed by means of the two-tailed Mann-Whitney U-test. One-way ANO- VA model followed by Bonferroni post hoc test was employed when comparing three groups. The Fisher’s exact test or the x2 test was used to investigate dichotomous variables. Overall sur- vival was defined as the time from the date of primary diagnosis to death related to ACC or last follow-up. Disease-free survival was defined as the time from the date of complete tumour resection to the first radiological evidence of disease relapse or death. All the survival curves were obtained by Kaplan-Meier estimates, and the differences between survival curves were assessed by the log-rank (Mantel-Cox) test. In this context, the mRNA or miRNA expression was considered as a categorical value (cut-off value for this data set was the median value of all samples for each mRNA or miRNA). Predictive factors of prog- nosis were identified by means of Cox proportional hazards regression models, which were used to estimate hazard ratios (HR) and their 95% confidence intervals in univariate and mul- tivariate analysis. P < 0-05 was considered significant.
Results
LIN28A and LIN28B gene expression
LIN28A was overexpressed in aggressive ACCs when compared with adenomas [7-0 (from 0 to 174.3) fold change vs. 3.6 (from 0 to 18.3); P = 0.006, respectively] and a trend towards greater expression when compared with non-aggressive carcinomas [7-0
(from 0 to 174.3) vs. 7.1 (from 0 to 17.1); P = 0-092]. Most of the adenomas (54%, 17 of 31) had null expression of LIN28A. However, LIN28A at the mRNA level was not associated with overall and disease-free survival. LIN28B expression was negative in 92% of the ACTs. Thus, although the LIN28 antibody employed at the immunohistochemical analysis recognizes both LIN28A and LIN28B, we can assume that LIN28 staining in ACCs represents mainly the protein product of LIN28A gene.
Because LIN28/LIN28B has shown to have an important role in regulating the IGF pathway both directly by promoting the trans- lation of the IGF2 mRNA21 and indirectly by derepressing IGF1R,17,18 we evaluated the association of LIN28A expression with our previously published data.6 We did not find correlation between LIN28A and IGF2 expression. As the Wnt/ß-catenin path- way has been shown to induce LIN28 upregulation, we also looked at association between LIN28A/LIN28B levels and CTNNB1 muta- tions or abnormal ß-catenin accumulation from our previous published data,7 but no significant association was found.
Among 39 ACTs studied by MLPA, no LIN28A copy number variation was detected. Only a metastatic ACC from a 44-year- old woman showed deletion of LIN28B gene and concordant negative LIN28B expression.
LIN28 protein expression
The frequency of strong LIN28 protein expression was not sta- tistically different between nonaggressive ACCs (37 of 77, 48%) and adenomas (26 of 78, 33%; x2 = 2.9; P = 0-11). Unexpect- edly, a weak LIN28 protein expression was significantly more frequent in aggressive (75 of 105, 71%) than in nonaggressive ACCs (40 of 77, 51%; x2 = 7.6; P = 0-02).
Interestingly, weak LIN28 protein expression was significantly associated with reduced disease-free survival in ACC patients (P = 0-01), but for overall survival only a trend was detectable (P = 0-117; Fig. la-d). Tumour size and recurrence were signifi- cantly associated with weak LIN28 protein expression (P = 0-05 and P = 0-02, respectively) (Table 1).
Ki67 staining was assessed in 158 ACCs. A Ki67 index ≥10% was significantly correlated with reduced disease-free and overall survival (P = 0-0001 and P = 0-0001, respectively) (Fig. 2). In the univariate analysis, tumour size ≥8 cm (P = 0-009), ENSAT 4 stage (P = 0-0001), Weiss score ≥6 (P=0-001) and Ki67 index ≥10% (P = 0-0001) were associated with reduced overall survival (Table 2). In the multivariate analysis, only ENSAT 4 stage (HR 3.3, 95% CI 1.9-5-7; P = 0.0001) and Ki67 index ≥10% (HR 3-0, 95% CI 1-8-5-0; P = 0-0001) remained as pre- dictors of reduced overall survival.
Regarding disease-free survival, male sex (P = 0-01), tumour size ≥8 cm (P = 0-001), Weiss score ≥6 (P = 0-003), Ki67 index ≥10% (P = 0-0001) and weak LIN28 protein expression (P = 0-01) were significantly associated with recurrence (local or metastasis) in the univariate analysis (Table 2). In the multivari- ate analysis, only Ki67 index ≥10% (HR 5-7, 95% CI 3-0-10-8; P = 0-0001) and weak LIN28 protein expression (HR 2-3, 95% CI 1.2-4-4; P = 0.008) were independent predictors of recur- rence of ACC patients.
LIN28 regulatory miRNA expression
From the 14 miRNAs studied in a subgroup of 28 ACCs, only mir-9 showed significant differential expression between aggres- sive and nonaggressive ACCs. Expression levels of let-7a, let-7b,
(a)
(b)
(c)
(d)
1.0
Adult ACCs (n = 124)
1.0
Adult ACCs (n = 188)
Disease-free survival
0-8
0-8
Overall survival
0-6
Strong LIN28 staining (n = 50)
0-6
Strong LIN28 staining (n = 68)
0-4
0-4
Weak LIN28 staining (n = 74)
0-2
0-2
Weak LIN28 staining (n = 120)
p = 0-01
p = 0.11
0-0
0-0
0-0
100
200
300
400
0-0
100
200
300
400
Time (m)
Time (m)
(a)
(b)
(c)
(d)
1.0
Adult ACCs (n = 107)
1.0
Adult ACCs (n = 158)
Disease-free survival
0-8
0.8
Ki67 < 10% (n=65)
Overall survival
Ki67 <10% (n=82)
0-6
0.6
0-4
p < 0.0001
0.4
p < 0.0001
0-2
Ki67 ≥ 10% (n=42)
0.2
Ki67 ≥ 10%(n=76)
0-0
0.0
0-0
100
200
300
400
0-0
100
200
300
400
Time (m)
Time (m)
| Univariate analysis | Multivariate analysis | |||||
|---|---|---|---|---|---|---|
| HR | 95% CI | P | HR | 95% CI | P | |
| Overall survival | ||||||
| Male sex | 1.2 | 0-8-1-8 | 0-35 | |||
| Tumour size ≥8 cm | 6.5 | 1.6-26-4 | 0-009 | 5.5 | 0-7-40 | 0-09 |
| Tumour stage (ENSAT 4) | 3-4 | 2-4-5-0 | 0-0001 | 3.3 | 1.9-5-7 | 0-0001 |
| Cushing | 1-5 | 0.9-2-2 | 0-09 | |||
| Weiss score ≥6 | 1.9 | 1-3-2-7 | 0-001 | 1.3 | 0-8-2.0 | 0-21 |
| Ki67 index ≥10% | 4.1 | 2.5-6-6 | 0-0001 | 3.0 | 1-8-5-0 | 0-0001 |
| Weak LIN28 protein expression | 1-4 | 0.9-1-1 | 0-12 | |||
| Disease-free survival | ||||||
| Male sex | 1-7 | 1.1-2-7 | 0-016 | 1.0 | 0-6-1.9 | 0.91 |
| Tumour size ≥8 cm | 2.7 | 1-5-5-0 | 0-001 | 1.7 | 0-7-4-2 | 0-26 |
| Cushing | 1-3 | 0-8-2-2 | 0-32 | |||
| Weiss score ≥6 | 1.9 | 1-2-3-0 | 0-003 | 1.2 | 0-6-2-1 | 0-63 |
| Ki67 index ≥10% | 5.1 | 2.9-9-0 | 0-0001 | 5-7 | 3.0-10-8 | 0-0001 |
| Weak LIN28 protein expression | 1-8 | 1.1-3-1 | 0-017 | 2.3 | 1.2-4-4 | 0-008 |
ACC, adrenocortical cancer.
let-7c, let-7d, let-7e, let-7f, let-7 g, mir-98, mir-30b, mir-30c, mir- 125a-3p, mir-125a-5p and mir-125b were not significantly different between nonaggressive and aggressive ACCs. LIN28 staining and mir-9 expression did not correlate, possibly because of small sample size.
Patients with aggressive ACC had significantly higher mir-9 levels in comparison with both nonaggressive ACC and
adenomas [2076 (from 36 to 9307) vs 133-4 (from 2-4 to 5193) and 313 (from 2.8 to 2129); P = 0-024 and P = 0-01, respectively). In addition, ACCs with mir-9 overexpression also presented a trend towards a higher Weiss score (P = 0-07) and a Ki67 index ≥10% (P = 0-11) (Table 3). Overexpression of mir-9 was also highly associated with reduced overall and disease-free survival (P= 0-012 and P = 0-010, respectively; Fig. 3).
| Low mir-9 | High mir-9 | P | |
|---|---|---|---|
| n (%) | 14 (50) | 14 (50) | |
| Age, median (year) | 38.7 | 26.5 | 0-13 |
| Sex [male, n (%)] | 4 (28-6) | 1 (7-1) | 0-14 |
| Tumour size, median (cm) | 8.7 | 11-5 | 0-22 |
| Tumour stage (ENSAT), n (%) | |||
| Known | 14 | 14 | |
| 1 | 3 (75) | 1 (25) | 0-19 |
| 2 | 6 (67) | 3 (33) | |
| 3 | 1 (17) | 5 (83) | |
| 4 | 4 (44) | 5 (56) | |
| Hormonal status, n (%) | |||
| Known | 14 | 13 | |
| Cushing only | 4 (67) | 2 (33) | 0-16 |
| Virilization only | 3 (100) | 0 (0) | |
| Mixed syndrome | 4 (31) | 9 (69) | |
| Nonfunctioning | 2 (50) | 2 (50) | |
| Oestrogen secreting | 1 (100) | 0 (0) | |
| Weiss score, median | 4 | 6.5 | 0-07 |
| Ki67 index (%) | |||
| Known | 13 | 11 | |
| <10 | 10 (67) | 5 (33) | 0-11 |
| ≥10 | 3 (33) | 6 (67) | |
| LIN28 staining | |||
| Known | 13 | 11 | |
| Low expression | 8 (57) | 6 (43) | 0-73 |
| High expression | 5 (50) | 5 (50) | |
| Distant metastasis or local recurrence, n (%) | |||
| Known | 11 | 9 | |
| Affected patients | 2 (25) | 6 (75) | 0-028 |
| Nonaffected patients | 9 (75) | 3 (25) | |
ACC, adrenocortical cancer.
Recurrence developed in 6 out of 9 (75%) patients with mir-9 overexpression and in only 2 out of 9 (22%, x2 = 4.8; P = 0.028; Table 3) with mir-9 under expression.
Discussion
Overexpression of LIN28A and B has been demonstrated in a variety of different human malignancies, specifically in a subset
of advanced and poorly differentiated cancers.14,16 Indeed, it has been shown to have a key role in promoting cancer progression and metastasis.13 In the present study, LIN28A expression was significantly higher in carcinomas when compared to adenomas. Unexpectedly, most of the patients with aggressive ACC had weak LIN28 protein expression. On the other hand, we have also demonstrated that mir-9, a miRNA known to negatively regulate LIN28, was highly expressed in the subgroup of aggressive when compared to nonaggressive ACC. mir-9 overexpression was strongly associated with both reduced overall and disease-free survival in ACCs. Importantly, we demonstrated the association of low LIN28 protein expression with reduced disease-free sur- vival in a significant cohort of ACCs.
Despite their high degree of homology, LIN28A and LIN28B function through distinct mechanisms and many tumours usu- ally express either one or the other.25 In our cohort, LIN28B was not expressed in most of the adrenocortical adenomas and carcinomas, whereas LIN28A was overexpressed in ACCs. How- ever, the majority of aggressive ACC have a very weak LIN28 expression at the protein level. We speculate here if the negative post-transcriptional regulation of LIN28 by mir-9 decreased LIN28 protein expression in aggressive ACCs. Nevertheless, this finding does not rule out the clinical relevance of weak LIN28 protein expression as a prognostic marker of recurrence in adult ACCs.
Ki67 index has been proposed as an important marker to pre- dict poor prognosis in ACC patients.1 A Ki67 index ≥10% has been employed to identify high-risk ACC patients who would be benefited from adjuvant approaches to avoid recurrence.2 More recently, Duregon et al. 26 demonstrated that Ki67 is the best prognostic indicator of overall survival, being superior to the mitotic index. In this study, Ki67 index ≥10% was significantly associated with reduced overall and disease-free survival in the multivariate analysis. Then, we found here in a TMA-based analysis from the Brazilian cohort in conjunction with data from full section analysis of the German ACC Registry the same cut- off of Ki67 index previously associated with poor prognosis.
A genomewide expression data of 317 cancer cell lines reveal that NCIH69 (small cell lung cancer), SNB19 (glioblastoma) and NCIH295R (ACC) are among the cancer cell lines that most express mir-9 in order of frequency, respectively (http://www. ebi.ac.uk/gxa/experiment/E-MTAB-37/ENSG00000207570#).27
(a)
(b)
1.0
1.0
Low mir-9 (n = 11)
0-8
Disease-free survival
0-8
Overall survival
0-6
0-6
Low mir-9 (n = 14)
0-4
High mir-9 (n = 9)
0-4
High mir-9 (n = 14)
0-2
0-2
p = 0.01
p = 0-01
0-0
0-0
0-0
100
200
300
400
0-0
100
200
300
400
Time (m)
Time (m)
Additionally, transcriptome profiling data obtained from the Michigan University ACC cohort confirmed that mir-9 is overexpressed in ACCs when compared to adenomas and the normal adrenal cortex (http://www.ebi.ac.uk/gxa/experiment/ E-GEOD-10927/ENSG00000207570)28.
Aberrant expression of mir-9 has been reported in a wide range of solid tumour and blood cancers. Expression of miRNAs and their role in cancer are known to be tissue specific and tumour specific. mir-9 may act as an oncomir in some cancers, such as breast cancer29 and hepatocellular carcinoma,30 but can also have tumour-suppressive activity in others, such as ovarian serous cell carcinoma31 and gastric cancer.32 For instance, mir-9 has shown to downregulate E-cadherin in human breast carci- noma cells leading to increased cell motility and invasiveness and consequent activation of ß-catenin signalling and elevated expression of VEGFA.29 However, E-cadherin is not significantly expressed in ACCs,33 indicating that different players should be involved in the mir-9 pathway in ACC tumourigenesis.
Recently, Zhuang et al.34 cocultured endothelial cells with tis- sue-matched tumour cells and have found significant upregula- tion of a group of miRNAs in five endothelial cell lines. From the group of upregulated miRNAs, mir-9 had the greatest effect on human umbilical vein endothelial cell migration. Taken up by endothelial cells, the released mir-9 targeted SOCS5, an inhibitor of the JAK-STAT signalling pathway, to promote endothelial migration. Furthermore, mir-9 levels were increased in the plasma of mice implanted with colorectal or lung tumour cells, and a mir-9 antagomir slowed tumour growth and angiogenesis.34 Interestingly, specific serum circulating miRNAs were detected in ACC patients and associated with both shorter overall and disease-free survival.35 Additionally, a recent phase 2 study in patients with hepatitis C had first eval- uated a miRNA-based targeted therapy.36 As miRNA-based tar- get therapy has now been used in clinical trials, we could consider mir-9 as a potential target for ACC therapy. In addi- tion, serum mir-9 detection can function as a precocious mar- ker of tumour recurrence (Figure S2). Although promising, mir-9 prognostic role should be further evaluated in a larger cohort of ACCs.
In conclusion, weak LIN28 protein expression was associated with recurrence in adult ACCs. LIN28A gene was overexpressed in aggressive ACCs, suggesting a negative post-transcriptional regulation of LIN28. Additionally, mir-9, a negative LIN28A reg- ulator, was higher expressed in aggressive ACCs.
Declaration of interest
We declare no duality of financial interest or direct or indirect conflict of interest on the part of any author of this manuscript.
Funding
FAPESP Grant number 2011/09092-0 (to A.M.F. and M.Q.A.) and 2012/21272-6 (to G.R.V.S. and M.Q.A.); CNPq Grant num- ber 470428/2013-9 (to MQA), 470631/2012-0 (to MQA) and 302825/2011-8 (to A.C.L.).
References
1 Fassnacht, M., Libe, R., Kroiss, M. et al. (2011) Adrenocortical carcinoma: a clinician’s update. Nature Reviews Endocrinology, 7, 323-335.
2 Fassnacht, M., Kroiss, M. & Allolio, B. (2013) Update in adreno- cortical carcinoma. The Journal of Clinical Endocrinology and Metabolism, 98, 4551-4564.
3 Else, T., Kim, A., Sabolch, A. et al. (2014) Adrenocortical carci- noma. Endocrine Reviews, 35, 282-326.
4 Fassnacht, M., Terzolo, M., Allolio, B. et al. (2012) Combination chemotherapy in advanced adrenocortical carcinoma. The New England Journal of Medicine, 366, 2189-2197.
5 Berruti, A., Baudin, E., Gelderblom, H. et al. (2012) Adrenal can- cer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Annals of Oncology, 23(Suppl 7), vii131-vii138.
6 Almeida, M.Q., Fragoso, M.C., Lotfi, C.F. et al. (2008) Expres- sion of insulin-like growth factor-II and its receptor in pediatric and adult adrenocortical tumors. The Journal of Clinical Endocri- nology and Metabolism, 93, 3524-3531.
7 Heaton, J.H., Wood, M.A., Kim, A.C. et al. (2012) Progression to adrenocortical tumorigenesis in mice and humans through insulin-like growth factor 2 and beta-catenin. The American Jour- nal of Pathology, 181, 1017-1033.
8 Faria, A.M. & Almeida, M.Q. (2012) Differences in the molecular mechanisms of adrenocortical tumorigenesis between children and adults. Molecular and Cellular Endocrinology, 351, 52-57.
9 Calin, G.A. & Croce, C.M. (2006) MicroRNA signatures in human cancers. Nature Reviews Cancer, 6, 857-866.
10 Singh, P., Soon, P.S., Feige, J.J. et al. (2012) Dysregulation of microRNAs in adrenocortical tumors. Molecular and Cellular Endocrinology, 351, 118-128.
11 Patterson, E.E., Holloway, A.K., Weng, J. et al. (2011) MicroR- NA profiling of adrenocortical tumors reveals miR-483 as a mar- ker of malignancy. Cancer, 117, 1630-1639.
12 Soon, P.S., Tacon, L.J., Gill, A.J. et al. (2009) miR-195 and miR-483-5p identified as predictors of poor prognosis in adrenocortical cancer. Clinical Cancer Research: An Official Journal of the American Association for Cancer Research, 15, 7684-7692.
13 Viswanathan, S.R. & Daley, G.Q. (2010) Lin28: a microRNA reg- ulator with a macro role. Cell, 140, 445-449.
14 Zhou, J., Ng, S.B. & Chng, W.J. (2013) LIN28/LIN28B: an emerging oncogenic driver in cancer stem cells. The International Journal of Biochemistry & Cell Biology, 45, 973-978.
15 Zhong, X., Li, N., Liang, S. et al. (2010) Identification of mi- croRNAs regulating reprogramming factor LIN28 in embryonic stem cells and cancer cells. The Journal of Biological Chemistry, 285, 41961-41971.
16 Viswanathan, S.R., Powers, J.T., Einhorn, W. et al. (2009) Lin28 promotes transformation and is associated with advanced human malignancies. Nature Genetics, 41, 843-848.
17 Wang, Y.C., Chen, Y.L., Yuan, R.H. et al. (2010) Lin-28B expres- sion promotes transformation and invasion in human hepatocel- lular carcinoma. Carcinogenesis, 31, 1516-1522.
18 Alajez, N.M., Shi, W., Wong, D. et al. (2012) Lin28b promotes head and neck cancer progression via modulation of the insulin- like growth factor survival pathway. Oncotarget, 3, 1641-1652.
19 Wilbert, M.L., Huelga, S.C., Kapeli, K. et al. (2012) LIN28 binds messenger RNAs at GGAGA motifs and regulates splicing factor abundance. Molecular Cell, 48, 195-206.
20 Huang, Y. (2012) A mirror of two faces: Lin28 as a master regu- lator of both miRNA and mRNA. Wiley Interdisciplinary Reviews RNA, 3, 483-494.
21 Polesskaya, A., Cuvellier, S., Naguibneva, I. et al. (2007) Lin-28 binds IGF-2 mRNA and participates in skeletal myogenesis by increasing translation efficiency. Genes & Development, 21, 1125- 1138.
22 Shi, S.R., Key, M.E. & Kalra, K.L. (1991) Antigen retrieval in for- malin-fixed, paraffin-embedded tissues: an enhancement method for immunohistochemical staining based on microwave oven heating of tissue sections. The Journal of Histochemistry and Cytochemistry: Official Journal of the Histochemistry Society, 39, 741-748.
23 Olaussen, K.A., Dunant, A., Fouret, P. et al. (2006) DNA repair by ERCC1 in non-small-cell lung cancer and cisplatin-based adjuvant chemotherapy. The New England Journal of Medicine, 355, 983-991.
24 Almeida, M.Q., Soares, I.C., Ribeiro, T.C. et al. (2010) Steroido- genic factor 1 overexpression and gene amplification are more frequent in adrenocortical tumors from children than from adults. The Journal of Clinical Endocrinology and Metabolism, 95, 1458-1462.
25 Piskounova, E., Polytarchou, C., Thornton, J.E. et al. (2011) Lin28A and Lin28B inhibit let-7 microRNA biogenesis by dis- tinct mechanisms. Cell, 147, 1066-1079.
26 Duregon, E., Molinaro, L., Volante, M. et al. (2014) Compara- tive Diagnostic and Prognostic Performances of the Hematoxy- lin-Eosin and Phospho-histone H3 Mitotic Count and Ki-67 Index in Adrenocortical Carcinoma. Modern pathology: an offi- cial journal of the United States and Canadian Academy of Pathology, Inc., 27, 1246-1254.
27 Krupp, M., Itzel, T., Maass, T. et al. (2013) Cell LineNavigator: a workbench for cancer cell line analysis. Nucleic Acids Research, 41(Database issue), D942-D948.
28 Giordano, T.J., Kuick, R., Else, T. et al. (2009) Molecular classifi- cation and prognostication of adrenocortical tumors by tran- scriptome profiling. Clinical Cancer Research, 15, 668-676.
29 Ma, L., Young, J., Prabhala, H. et al. (2010) miR-9, a MYC/ MYCN-activated microRNA, regulates E-cadherin and cancer metastasis. Nature Cell Biology, 12, 247-256.
30 Sun, Z., Han, Q., Zhou, N. et al. (2013) MicroRNA-9 enhances migration and invasion through KLF17 in hepatocellular carci- noma. Molecular Oncology, 7, 884-894.
31 Tang, H., Yao, L., Tao, X. et al. (2013) miR-9 functions as a tumor suppressor in ovarian serous carcinoma by targeting TLN1. International Journal of Molecular Medicine, 32, 381- 388.
32 Zheng, L., Qi, T., Yang, D. et al. (2013) microRNA-9 sup- presses the proliferation, invasion and metastasis of gastric can- cer cells through targeting cyclin D1 and Ets1. PLoS ONE, 8, e55719.
33 Gupta, D., Shidham, V., Holden, J. et al. (2001) Value of topo- isomerase II alpha, MIB-1, p53, E-cadherin, retinoblastoma gene protein product, and HER-2/neu immunohistochemical expres- sion for the prediction of biologic behavior in adrenocortical neoplasms. Applied Immunohistochemistry & Molecular Morphol- ogy: AIMM / Official Publication of the Society for Applied Immu- nohistochemistry, 9, 215-221.
34 Zhuang, G., Wu, X., Jiang, Z. et al. (2012) Tumour-secreted miR-9 promotes endothelial cell migration and angiogenesis by activating the JAK-STAT pathway. The EMBO Journal, 31, 3513- 3523.
35 Chabre, O., Libe, R., Assie, G. et al. (2013) Serum miR-483-5p and miR-195 are predictive of recurrence risk in adrenocortical cancer patients. Endocrine-Related Cancer, 20, 579-594.
36 Janssen, H.L., Reesink, H.W., Lawitz, E.J. et al. (2013) Treatment of HCV infection by targeting microRNA. The New England Journal of Medicine, 368, 1685-1694.
Supporting Information
Additional supporting information may be found in the online version of this article at the publisher’s web site.