RET mutated C-cells proliferate more rapidly than non-mutated neoplastic cells

Introduction

Medullary Thyroid Carcinoma (MTC) accounts for about 5–7% of all thyroid cancer and can occur in a hereditary (25%) or a sporadic form (75%) (1).

According to the Next Generation Studies published in the last years (2, 3, 4, 5, 6), somatic RET mutations are the most frequent alterations found in sporadic MTC. Although a large spectrum of RET mutations are described, the commonest alteration is the p.Met918Thr mutation in exon 16 of the RET gene (1). In addition to somatic RET mutations, sporadic MTC shows the presence of somatic RAS mutations that have been mostly reported in RET negative tumours and are almost always mutually exclusive with RET mutations (7, 8). RET somatic mutations have been reported to be a factor of a bad prognosis and are significantly associated with a more aggressive biological behavior and a reduced survival (9, 10, 11). Although less investigated, RAS mutations seem to predict a better outcome when compared to somatic RET mutations (7, 9).

We previously demonstrated that the presence of a somatic RET p.Met918Thr mutation correlated with larger tumor size while it was significantly lower in tumors smaller than 2 cm (12). We hypothesized that p.Met918Thr mutation might not be an early event or that it could be present from the beginning but only in a subpopulation of cells not detectable with the conventional sequencing analysis because of its low sensitivity. This latter hypothesis has been supported by the evidence of RET mutation heterogeneity in about 20% of MTC (13, 14). However, due to the low sensitivity of the Sanger method and to the fact that only RET exon 11 and 16 have been investigated, the problem of false negative cases should be considered as a bias in the former studies.

In the present study, we correlated the prevalence of any type of RET and RAS somatic mutations, as obtained with a Next Generation Sequencing (NGS) approach, with the MTC tumor size. Moreover, the correlation of their allelic frequency (AF) and the size of the tumor was evaluated to better understand their driver role in the tumoral transformation of the original C cells. Finally, the correlation with Ki67 positivity, that is an index of cell proliferation, with the size and the RET mutations was also analyzed.

Patients and methods

Our study group was represented by 149 MTC patients belonging to a larger series of 201 MTC patients submitted to total thyroidectomy and central neck dissection at our hospital whose tumoral tissues were analyzed by NGS for many gene alterations as previously reported (15). To the purpose of this study we included only the 149 cases analyzed on the primary tumor. All patients had no history of familial disease, were negative for the presence of other endocrine neoplasia and no germline RET mutations were found.

Clinical, biochemical and pathological data, with particular regard to tumor size, were collected from a computerized database. Cases were classified according to the size of the tumor as follows: group 1, ≤1 cm; group 2, >1 and ≤2 cm; group 3, >2 and ≤3 cm; and group 4, >3 cm.

Data about the presence or absence of RET and RAS somatic mutations were retrieved from the row data of the previous study (15), including the allelic frequency of the driver mutations.

Ki67 proliferative index was evaluated by immunohistochemistry. Five micrometer sections were cut from formalin-fixed paraffin-embedded (FFPE) blocks for 130 out of 149 cases. Ki67 immunostaining was performed automatically by the Ventana Benchmark immunostaining system (Ventana Medical Systems, Inc., Tucson, AZ) using a rabbit monoclonal primary antibody (immunoglobulin (Ig)G) directed against the C-terminal portion of Ki67 (CONFIRM anti-Ki67, clone 30-9; Ventana). Neoplastic cells were considered positive when the nuclei showed an immunoreactivity variable from weak to strong; neoplastic cells without nuclear immunoreactivity were considered as negative. Ki67 score was independently evaluated by two pathologists (L T and F B) who were blinded to the clinicopathologic data. Immunostaining was evaluated using a standard Leica DM4000 microscope with the ‘hot spots’ method: the field under high power magnification (original magnification, ×40) with the highest apparent Ki67 index was selected. The Ki67 score was defined as the percentage of positive cells among a minimum of 100 neoplastic cells. The results were scored according to the number of Ki67 positive cells: <3% as low, 3–20% as intermediate and >20% as high (16).

An informed consent form for RET genetic screening and other clinical procedures was signed by all patients. The present study was approved by the Institutional Review Board and by the ‘Comitato Etico Regionale per la Sperimentazione Clinica della Regione Toscana’ Prot n 6714, 05/02/2019.

Results

We distinguished 149 MTC cases according to the size of the primary tumor. As shown in Table 1, in this subgroup of primary tissues, we found 81 (54%) cases with a somatic RET mutation, 40 (27%) cases with RAS mutations and 28 (19%) cases that were negative. Among the RAS positive cases, 27 were positive for HRAS and 13 for KRAS respectively. Data on the presence of the somatic mutations were retrieved by a larger study performed in our center and already published (15).

As shown in Fig. 1 panel A, RET somatic mutations were the most prevalent in each group. However, a statistically significant different mutation profile (P = 0.02) was observed in the four groups with the highest prevalence (20/29, 69%) of the RET mutations found in MTC cases with the largest size (group 3 and 4) and the lowest prevalence in group 2 (28/51, 11.8%) and 1 (13/36, 36%). At variance, RAS positive cases and RET/RAS negative cases were found to be more frequent in smaller tumors (1 and 2). The mutation profile was also compared between tumors smaller and larger than 2 cm. As shown in Fig. 1 panel B, the prevalence of RET positive cases was significantly higher in tumors larger than 2 cm while the prevalence of RAS positive and RET/RAS negative cases was lower in larger tumors.

View Full Size

Prevalence of somatic RET and RAS mutations according to tumor size. A statistically significant difference in the mutations profile was observed both when 4 different groups were considered (group 1, ≤1 cm; group 2, >1 and ≤2 cm; group 3, >2 and ≤3 cm; and group 4, >3 cm (P = 0.02)) (panel A) and when grouped into 2 bigger groups (1 + 2, ≤2 cm and 3 + 4, >2 cm) (P = 0.01) (panel B).

Citation: Endocrine Connections 10, 2; 10.1530/EC-20-0589

• Download Figure
• Download figure as PowerPoint slide

We then focused on the different types of RET mutations according to the American thyroid association risk classification (17) as highest, high and moderate. As shown in Fig. 2 panel A, a gradual increase of the highest risk RET mutation (i.e. p.Met918Thr) has been observed with the increase of tumor size and a simultaneous gradual decrease was observed for high and moderate RET mutations (P = 0.0027). When considering the distribution of the different RET mutations in tumors either smaller (group 1 and 2) or larger (group 3 and 4) than 2 cm, we observed that in the group 1 + 2 RET mutations were uniformly distributed (highest 29.3%, high 36.6% and moderate 34%) while in the group 3 + 4, a significant greater prevalence of the highest RET mutations (29/40, 72.5%) was found with respect to the high (7/40, 17.5%) and the moderate (4/40, 10%) RET mutations.

View Full Size

Prevalence of somatic highest, high and moderate RET mutations (17) according to tumor size. Four different groups were considered (group 1, ≤1 cm; group 2, >1 and ≤ 2 cm; group 3, >2 and ≤3 cm; and group 4, >3 cm) (panel A); groups A and B vs groups C and D have been considered (panel B). A statistically higher prevalence of the most aggressive RET mutations was observed (P= 0.0027 and P = 0.0004, respectively).

Citation: Endocrine Connections 10, 2; 10.1530/EC-20-0589

• Download Figure
• Download figure as PowerPoint slide

As shown in Fig. 3 panels A and B, the AF of RET mutations was significantly higher in larger tumors both when comparing the four different groups (group 1, mean AF 23.42 ± 10.38; group 2, mean AF 29.49 ± 11.16; group 3, mean AF 39.27 ± 9.3, group 4 mean AF 42.48 ± 11.16) and when comparing group 1 + 2 and group 3 + 4 (group 1 + 2, mean AF 27.51± 10.38; group 3 + 4 mean AF 40.91 ± 12.63). In addition, the RET AF was higher in tumors of group 3 + 4 than in group 1 + 2 for every type of RET mutations (moderate: 46.72% vs 29.72%, high: 39.48% vs 25.52%, highest: 41.52% vs 27.7%, respectively). In particular, when we focused on the p.Met918Thr, we observed a statistically significant increase in the AF of this mutation with the increase of tumor size (group 1, mean AF 18.16 ± 8.6; group 2, mean AF 28.9 ± 13.8; group 3, mean AF 39.8 ± 13.1, group 4 mean AF 44.3 ± 16.3).

View Full Size

Correlation between the allelic frequency (AF) of the driver mutations and tumor size. RET AF is higher in larger tumors. Four different groups were considered (group 1, ≤1 cm; group 2, >1 and ≤2 cm; group 3, >2 and ≤ 3 cm; and group 4, >3 cm) (panel A); groups 1 and 2 vs groups 3 and 4 have been considered. (P < 0.001) (panel B). RAS AF was not correlated to tumor size when both the four groups (panel C) and when considered groups 1 and 2 vs groups 3 and 4 (panel D) were considered.

Citation: Endocrine Connections 10, 2; 10.1530/EC-20-0589

• Download Figure
• Download figure as PowerPoint slide

As far as the AF of RAS mutations was concerned, although a trend of increase was observed, there was no a statically significant difference neither when comparing the four different groups (group 1, mean AF 29.8 ± 11.9; group 2, mean AF 33.87 ± 12.2; group 3, mean AF 36.13 ± 11.4, group 4 mean AF 46.62 ± 4.6) (Fig. 3 panel C) nor when comparing group 1 + 2 and group 3 + 4 (group 1 + 2, mean AF 32.27 ± 12.08; group 3 + 4 mean AF 39.63 ± 107) (Fig. 3 panel D).

Ki67 positive expression was high in 15/130 (11.5%), intermediate in 79/129 (60.8%), low in 34/129 (26.2%) and negative in 2/149 (1.5%). A statistically high significant correlation (P < 0.0001) was found when the Ki67 low, intermediate and high positivity was analyzed according to the tumor size (Fig. 4, panel A). A positive trend or correlation was also found when the analysis was done with the type of mutation being Ki67 expression higher in RET positive cases than in RAS positive cases and even lower in RET/RAS negative cases (Fig. 4, panel B). Although not statistically significant, we found that, among all mutations, the p.Met918Thr RET mutation showed the highest proliferation rate (Table 2). The correlation become significant when the analysis was performed between Ki67 positivity and the AF of RET positive cases (Fig. 4, panel C).

View Full Size

A statistically significant correlation was found when Ki67 low, intermediate and high positivity was correlated to the tumor size (panel A); although not statistically significant, a trend of correlation was observed when Ki67 positivity was analyzed according to the type of mutation (panel B). The correlation of Ki67 positivity was statistically significant when the analysis was performed with the AF of RET mutation in RET positive cases (panel C).

Citation: Endocrine Connections 10, 2; 10.1530/EC-20-0589

• Download Figure
• Download figure as PowerPoint slide

Data on the outcome were available in 127/149 cases: 88/127 (69.3%) patients were free of disease while 39/127 (30.7%) patients were either dead or with a persistent disease (i.e. biochemical and/or structural disease). In the group of disease free patients, the prevalence of cases smaller than 2 cm (n = 63) was significantly higher than those with larger tumors (n = 25). On the contrary in the group of patients dead or with a persistent disease, the prevalence of cases larger than 2 cm (n = 26) was significantly higher than those with smaller tumors (n = 13) (P = 0.0008).

Discussion

The sporadic form of MTC is mainly characterized by the presence of RET (about 40–50% of cases) and RAS (about 10–20% of cases) somatic mutations (Catalogue of Somatic Mutation in Cancer https://cancer.sanger.ac.uk/cosmic). The p.Met918Thr mutation in exon 16 is the most common RET somatic mutation being present in up to 90% of RET-positive cases, while RAS gene point mutations in MTC mainly occur in H- and KRAS, and they are usually mutually exclusive with RET mutations. The recent introduction of NGS techniques has largely improved the identification of the molecular alterations involved and causative of many human diseases. Several studies carried out in MTC confirmed the role of RET and RAS somatic mutations as the main drivers in MTC and only few different genetic alterations have been identified (2, 3, 15). Nevertheless, a rather large portion of cases are negative for the presence of common somatic gene alterations. In the majority of MTC cases, RET and RAS mutations are mutually exclusive indicating that RET-mediated and RAS mediated oncogenic transformations occur separately. Since both RET and RAS alterations lead to an uncontrolled activation of the MAPKinase pathway (18, 19) they are considered driver mutations in MTC.

In the present study, we found that the prevalence of RET somatic mutations, any type, was significantly higher in larger than in smaller MTC cases and in particular that the RET ATA highest mutation (i.e. p.Met918Thr) was the most represented in the larger tumors. With these results, we confirmed our previous study (12) that was partially affected by the low sensitivity of the method of sequencing used with the risk to have lost some positive cases especially in the smallest tumors. The prevalence of RET mutations found with NGS can be considered reliable and we can now confirm that the different prevalence of RET in smaller and bigger tumors is true. At variance, a higher proportion of tumors positive for RAS somatic mutation or negative for any mutation was observed in smaller tumors. According to these results, we can hypothesize that three types of MTC tumors exist: those with RAS or noRAS/noRET mutations. At the time of their development they are similar in prevalence but then RAS or noRAS/noRET cases remain smaller likely because of growing slowly. While those with RET, especially p.Met918Thr, mutation becomes bigger likely because of rapidly growing.

A second important observation of the present work is that larger tumors are characterized by a higher AF of RET mutations, and in particular of the p.Met918Thr, with a mean AF of 41.54 vs 26.93 in the smaller. This finding indicates that a bigger number of tumoral cells, almost the totality, are mutated in the bigger tumors but not in the smaller tumors. This was not observed for the AF of RAS mutations that were present in about 30% of tumoral cells without differences related to the tumor size. These findings support the previous observations that a genetic intra- and inter-tumor heterogeneity and non-clonal origin of some MTC cases exists (13, 14). Considering that RET p.Met918Thr mutation has been demonstrated to have the highest transforming ability (20), we can postulate that cells carrying this mutation can duplicate more rapidly and take a growth advantage respect to the non-mutated concomitant cells. The fact that bigger tumors have the RET p.Met918Thr mutation more frequently at a higher AF is in line with this hypothesis.

The higher prevalence of RET somatic mutations, and in particular of the p.Met918Thr, in larger tumors is in keeping with the evidence that larger tumors have a worse outcome (21, 22) as shown also in the present series. On this regard, it is useful to recall that RET p.Met918Thr is also the germline mutation of the MEN 2B, whose MTC is the most aggressive and rapidly growing tumor among all genotype of MEN 2 (23). These findings are in keeping with the hypothesis that RET mutations, particularly p.Met918Thr, identify a subgroup of tumors rapidly growing for the presence of a highly transforming mutation and with a more aggressive behavior with respect to those with RAS mutations or without any mutations. The great correlation of Ki67 positivity, that is a well-recognized index of tumor proliferation, with the greater size of the tumor and with a higher percentage of AF of RET in RET positive cases can be considered the proof of the fact that RET positive cells take a growth advantage with respect to the negative cells (i.e. higher percentage of AF of the mutation), divide and reproduce more rapidly (i.e. higher Ki67 positivity) thus determining a rapid increase of the tumor respect to RET negative, either RAS positive or negative, cases.

Simultaneously, we can consider that, at least cases with RAS mutations, are more clonal and less aggressive similarly to RAS positive follicular adenomas and carcinomas (24). This behavior has been also reported in colon cancer (25, 26) in which both KRAS and NRAS mutations are usually present in the majority of neoplastic cells and have been considered a clonal event in the tumoral transformation. It has been proposed (27, 28) that, at variance with PIK3CA and NOTCH1 mutations that instead seem to have a biological behavior similar to that of RET mutations (13, 14), RAS mutations would likely accumulate at the beginning of tumor formation thus representing the main driver alteration.

In conclusion, this study highlights the predominant role of RET somatic mutations in MTC tumorigenesis and demonstrates a high prevalence of these alterations in all size categories. We demonstrate that RET mutation prevalence and RET mutation AF are significantly higher in larger than in smaller tumors indicating that RET alterations are clonal event in cases larger than 2 cm and subclonal event in small MTC cases. At variance, both the prevalence and the AF of RAS mutations are similar in MTC size categories. Finally, bigger tumors, that have higher prevalence of RET mutations as well as higher level of Ki67 positivity, showed a more aggressive behavior and a worse outcome.

Declaration of interest

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.

Funding

This study has been supported by grants to R E from Associazione Italiana per la Ricerca sul Cancro (AIRC, Investigator grant 2018, project code 21790), Agenzia Italiana del Farmaco (AIFA, project code AIFA-2016-02365049), and Progetto di Ricerca di Ateneo (PRA_2018_27) from University of Pisa.

Author contribution statement

C R designed the study, performed the in silico analysis and prepared the manuscript. T R and C M performed all DNA preparation and the MLPA experiments. A P, V C and L L contributed to patients selection and collection of clinical data. L T and F B reviewed histological slides and gave the final diagnosis of MTC. R C performed all NGS experiments. R E is the team leader, she contributed to the preparation of the manuscript.