VEGF,VEGFR3,and PDGFRB Protein Expression Is Influenced by RAS Mutations in Medullary Thyroid Carcinoma

Abstract

Background:

Tyrosine kinase inhibitors (TKIs) have achieved remarkable clinical results in medullary thyroid carcinoma (MTC) patients. However, the considerable variability in patient response to treatment with TKIs remains largely unexplained. There is evidence that it could be due, at least in part, to alterations in genes associated with the disease via their effect on the expression of TKI targets. The objective of this study was to evaluate the influence of RAS mutations on the expression levels in MTC tumors of eight key TKI target proteins.

Methods:

We assessed by immunohistochemistry the expression of EGFR, KIT, MET, PDGFRB, VEGF, VEGFR1, VEGFR2, and VEGFR3 in a series of 84 primary MTC tumors that had previously been molecularly characterized, including 14 RAS-positive, 18 RET ^M918T-positive, and 24 RET ^C634-positive tumors, as well as 15 wild-type tumors with no mutations in the RET or RAS genes.

Results:

In contrast to RET-positive tumors, RAS-positive tumors expressed neither PDGFRB nor MET (p=0.0060 and 0.047, respectively). Similarly, fewer RAS-positive than RET-related tumors expressed VEGFR3 (p=0.00062). Finally, wild-type tumors expressed VEGF more often than both RAS- and RET-positive tumors (p=0.0082 and 0.011, respectively).

Conclusions:

This is the first study identifying that the expression of TKI targets differs according to the presence of RAS mutations in MTC. This information could potentially be used to select the most beneficial TKI treatment for these patients.

Introduction

Upon neoplastic transformation, parafollicular cells of the thyroid gland give rise to medullary thyroid carcinoma (MTC), a rare malignancy that accounts for approximately 2–5% of all thyroid neoplasias (1). Around 75% of MTCs are sporadic, while the remaining cases arise as a manifestation of the hereditary multiple endocrine neoplasia type 2 syndrome (MEN2). Interestingly, in both scenarios, different activating point mutations in the “rearranged during transfection” (RET) proto-oncogene have been shown to lead to carcinogenesis (2,3). It has recently been described that a substantial proportion of non-RET-mutated sporadic MTCs are caused by mutations in RAS genes (4,5). Both types of molecular alterations have been shown to have an impact on the evolution of MTC: while RAS gene mutations have been associated with better outcomes, RET mutations give rise to a more aggressive phenotype with a worse prognosis (4).

The clinical outcome of MTC depends greatly on when in the disease process the patient is diagnosed. The 10-year survival rate of patients with MTC diagnosed at an advanced stage is less than 20% (6), mainly because treatment with cytotoxic drugs and/or standard radiotherapy has been proven to be ineffective (7). In this regard, because RET is a tyrosine kinase receptor, MTC patients with metastatic or locally advanced disease have more recently been treated with small-molecule tyrosine kinase inhibitors (TKIs), with promising results (8 –13). Remarkably, it has already been shown that the expression of the key target proteins of these drugs varies in MTC according to the specific RET mutation present (14), a finding that could undoubtedly have an important impact on clinical practice. However, little is known about how RAS-positive MTC patients respond to TKI drugs.

On the basis of the evidence that expression of TKI target proteins is associated with the presence of particular genetic mutations and, more importantly, since RAS mutations have also been associated with resistance to TKI therapy for other tumor types (15), we assessed the relationship between the expression of key TKI targets in a series of 84 molecularly characterized primary MTC tumors. We observed differences in the frequency of expression of VEGFR3, MET, and PDGFRB between RAS- and RET-mutated tumors. Wild-type (WT) tumors, those with no mutations in these two genes, more frequently expressed VEGF. These results could be of clinical importance when enrolling these patients in clinical trials of TKI treatments.

Materials and Methods

Human MTC samples

Eighty-four formalin-fixed paraffin-embedded MTC primary tumor samples were collected at the Spanish National Cancer Research Center (CNIO) in collaboration with the CNIO Tumor Bank. Written informed consent was obtained from all study participants, and the study was approved by the institutional review board (Comité de Bioética y Bienestar Animal) of the Instituto de Salud Carlos III. The tumor samples corresponded in most cases to patients diagnosed as sporadic or familial, based on the analysis of the RET proto-oncogene in peripheral blood samples. The mutational status of exons 10, 11, 13, 14, 15, and 16 of RET was assessed in genomic germline DNA using standard PCR conditions, primers, and automated sequencing, as previously described (16). When no RET mutation was found in peripheral blood, the same RET screening was performed in the corresponding tumor. Twenty-nine RET-negative samples were subsequently screened for somatic alterations in H-, N-, and K-RAS mutation hotspots: codons 12 and 13 in exon 2, and codon 61 in exon 3, as previously described (5) (see Table 1 for details).

Table 1.

Molecular Characteristics of the 84 Medullary Thyroid Carcinoma Primary Tumors Included in the Study

Tumor group	Total=84, N (%)	Specific mutation	Other features
RAS gene mutation	14 (17%)	H-RAS exon 2 (n=3; 4%)	Sporadic (n=3)
		H-RAS exon 3 (n=7; 8%)	Sporadic (n=7)
		K-RAS exon 2 (n=2; 2%)	Sporadic (n=2)
		K-RAS exon 3 (n=2; 2%)	Sporadic (n=2)
RET gene mutation	55 (65%)	RET M918T (n=18; 22%)	Sporadic (n=18)
		RET C634 (n=24; 29%)	Sporadic (n=4)
			Familial (n=20)
		RET 618 (n=6; 7%)	Sporadic (n=1)
			Familial (n=5)
		RET 611, 619, 620, 804, 883, or 891 (n=6; 7%)	Sporadic (n=4)
			Familial (n=2)
		Other (n=1; 1%)	Sporadic (n=1)
WT	15 (18%)

H-RAS accession number: ENSG0000017477; K-RAS accession number: ENSG00000133703.

Of the 84 formalin-fixed paraffin-embedded samples, 79 were distributed across 3 tissue microarrays as previously described (14), and 5 were evaluated as complete sections. The tissue microarrays were constructed with 2 cores of 1 mm from each tumor. Sections of each tissue microarray and individual tumors were immunostained using antibodies specific for the EGFR, KIT, MET, PDGFRB, VEGF, VEGFR1, VEGFR2, and VEGFR3 proteins. The immunohistochemistry (IHC) protocols used, immunostaining, and the scoring applied are detailed in Supplementary Table S1 (Supplementary Data are available online at www.liebertpub.com/thy) and have been previously described (14). Briefly, for EGFR, VEGFR2, and PDGFRB, tumors with moderate/strong staining were considered positive (17,18); for KIT, both the intensity and the extent of the staining were evaluated, and cases with aggregate scores >3 were regarded as positive (19). The MET protein was considered positive when the expression was positive for 30% of tumor cells with moderate/strong staining (20); for VEGF, the intensity of the staining was estimated on the 4-tiered scale (0, absent; 1, weak; 2, moderate; 3, strong), and immunopositivity was defined by strong staining (21); for VEGFR1, a tumor was considered positive if cytoplasmic expression was detected; for VEGFR3, tumors with a percentage of cells with positive staining greater than the observed median (50%) were considered positive.

Statistical analysis

All statistical analyses were performed using SPSS version 17.0. The χ ²-test or Fisher's exact test was used to assess associations between mutation status and IHC expression of each protein. RET-mutated tumors were classified into three groups: RET-mutated group as a whole (including all RET-related tumors, regardless the mutation), RET ^C634 tumors, and RET ^M918T tumors. The RET ^C634-mutated group included tumors from familial and sporadic cases with germline and somatic mutations, respectively; the RET ^M918T-mutated group comprised exclusively tumors from sporadic cases. Tumors were classified as RAS-mutated regardless of the particular RAS gene involved, although we also carried out an analysis stratified by gene. The WT group consisted exclusively of tumors from sporadic cases in which both germline and somatic RET mutations and RAS somatic mutations were not found. Two-sided p-values <0.05 were considered statistically significant.

Results

The results of mutational screening are summarized in Table 1. In our series, 55 (65%) tumors harbored a mutation in RET proto-oncogene (27 familial and 28 sporadic), with codons 634 and 918 being the most frequently affected (24 and 18 cases, respectively). Other, less frequent RET mutations were present in 13 tumors. A RAS gene mutation was found in 14 of the 29 RET-negative primary tumors (48%), which represented 17% of the entire collection of primary tumors. The majority of the mutations were located in H-RAS, while four were in K-RAS.

There were clear differences in the immunohistochemical expression of the TKI receptors (Fig. 1). Table 2 summarizes significant results of the analysis of IHC status with the underlying mutated gene. Briefly, we observed differences in the frequency of expression of four key TKI target genes (PDGFRB, VEGFR3, MET, and VEGFR1) between RAS gene-mutated samples and RET-mutated tumors, and one, VEGF, when comparing RAS with WT.

FIG. 1.

Examples of immunohistochemical staining of the selected proteins. Representative cases are shown with low and high protein expression of KIT (A), EGFR (B), MET (C), PDGFRB (D), VEGF (E), VEGFR2 (F), VEGFR1 (G), and VEGFR3 (H).

Table 2.

Proportion of Tumors with Proteins Differentially Expressed

	RAS vs. RET			WT vs. RET			WT vs. RAS
Proteins ^a	RAS	RET	p	WT	RET	p	WT	RAS	p
VEGF	2/12 (17%)	15/52 (29%)	0.49	9/13 (69%)	15/52 (29%)	0.011	9/13 (69%)	2/12 (17%)	0.0082 ^b
PDGFRB	0/13 (0%)	21/53 (40%)	0.0060 ^b	2/12 (17%)	21/53 (40%)	0.19	2/12 (17%)	0/13 (0%)	0.22
VEGFR1	6/10 (60%)	13/37 (35%)	0.28	6/13 (46%)	13/37 (35%)	0.52	6/13 (46%)	6/10 (60%)	0.68
VEGFR2	4/10 (40%)	22/39 (56%)	0.48	10/13 (77%)	22/39 (56%)	0.19	10/13 (77%)	4/10 (40%)	0.10
VEGFR3	1/14 (7%)	31/53 (59%)	0.00062 ^b,c	5/12 (42%)	31/53 (59%)	0.35	5/12 (42%)	1/14 (7%)	0.065
MET	0/12 (0%)	12/40 (30%)	0.047	4/14 (29%)	12/40 (30%)	0.99	4/14 (29%)	0/12 (0%)	0.10

	RAS vs. RET^C634			RAS vs. RET^M918T
Proteins ^a	RAS	RET^C634	p	RAS	RET^M918T	p
VEGF	2/12 (17%)	6/24 (25%)	0.69	2/12 (17%)	6/16 (38%)	0.40
PDGFRB	0/13 (0%)	11/24 (46%)	0.0032 ^b	0/13 (0%)	6/17 (35%)	0.024 ^b
VEGFR1	6/10 (60%)	3/17 (18%)	0.039 ^c	6/10 (60%)	5/10 (50%)	0.99
VEGFR2	4/10 (40%)	10/17 (59%)	0.44	4/10 (40%)	6/12 (50%)	0.69
VEGFR3	1/14 (7%)	16/24 (67%)	0.00037 ^b,c	1/14 (7%)	10/17 (59%)	0.0067 ^b,c
MET	0/12 (0%)	4/17 (24%)	0.12	0/12 (0%)	5/14 (36%)	0.042

Only proteins with a p-value ≤0.10 in at least one of the comparisons are shown.

Remained statistically significant when considering only tumors with mutations in H-RAS. H-RAS accession number: ENSG0000017477.

Remained statistically significant when considering only tumors with mutations in K-RAS. K-RAS accession number: ENSG00000133703.

p-Values <0.05 highlighted in boldface.

Notably, all RAS-related samples lacked PDGFRB expression, while RET-related MTC frequently stained positive (0% vs. 40%, p=0.0060). In order to assess whether this association was because of one RET mutation in particular, the analysis was repeated considering only the two more prevalent RET mutations. PDGFRB expression was more strongly associated with RET ^C634-mutated (46%) than RET ^M918T-mutated cases (35%, p=0.0032 and 0.024, respectively). RAS-related tumors less often expressed VEGFR3 than RET-related tumors (p=0.00062), seemingly more associated with harboring a RET ^C634 mutation (p=3.7×10⁻⁴) than the RET ^M918T change (p=0.0067). Although VEGFR1 expression in RAS-mutated tumors was not significantly different from that in RET-mutated tumors as a whole, it was compared with the RET ^C634-mutated group (60% vs. 18%, p=0.039). Finally, RAS-related tumors less often expressed MET than RET tumors, mainly because of RET ^M918T samples (0% vs. 36%, p=0.042).

In order to identify a characteristic staining profile for WT cases, we compared this group to RET and RAS samples separately. In both comparisons, the only significant difference observed was in the frequency of VEGF expression, which was more common in the WT group (p=0.011 and 0.0082, respectively).

Discussion

The appropriate clinical management of familial MTC is already well established, including prophylactic thyroidectomy at an early age, determined according to the particular germline mutation detected (22). It is the management of sporadic and de novo patients with germline RET mutations that presents a clinical challenge. These patients are often diagnosed at advanced stage with local or distant metastases (23), for which the standard therapeutic options are not effective (7).

Since the transforming event in 30–50% of MTC sporadic cases is the activation of RET by point mutations, one promising approach to extend the progression-free survival of patients with advanced disease is targeted therapy to inhibit tyrosine kinase receptors. However, the molecular basis underlying the great variability in the response of MTC patients to TKI treatment remains unknown (8). There is emerging evidence that expression profiles in MTC are driven by the underlying genetics (24) and, in consequence, that the expression of TKI target proteins could be dictated by particular mutations (14). In our series, the proportion of non-RET but RAS-positive MTCs was 48% (17% of the entire collection), with H- and then K-RAS being the most often mutated genes, and N-RAS mutations being totally absent. The reported prevalence of RAS gene mutations in non-RET MTC varies considerably, ranging from 17.6% to 81.0% (4,5,25), explaining nevertheless a substantial proportion of sporadic patients. Therefore, there is an urgent need to distinguish this subset of patients and to find out if they share a specific targetable expression pattern.

Currently, two main TKIs are available in the treatment of advanced MTC. Vandetanib (ZD6474), which targets VEGFR2, VEGFR3, RET, and EGFR, was the first TKI approved for the treatment of adults with symptomatic or progressive MTC (9). Recently, this drug has been used for treatment of pediatric patients with MTC, who harbor almost exclusively the RET ^M918T mutation, with encouraging results (10). In addition, Cabozantinib (XL184), which inhibits VEGFR2, MET, RET, KIT, VEGFR1/3, FLT3, Tie2, and AXL (11), was approved by the FDA for metastatic MTC in 2012.

Recent results from phase III clinical trials showed shorter progression-free survival in RET ^M918T-negative patients treated with Vandetanib when compared with M918T-mutated patients (26). Even though RAS gene mutation status was not assessed in this study, according to the known RAS gene mutations' prevalence, it seems reasonable to assume that a proportion of the RET ^M918T-negative patients with worse response to Vandetanib carried RAS alterations, which according to our results express less frequently some of its targets. Ciampi et al. (4) reported a higher but not significant prevalence of disease-free patients among the patients with RAS-mutated MTC. Thus, it could be expected that less RAS-positive patients were included in TKI trials, which usually require patients to have advanced metastatic disease. However, it should be noted that even RET mutations with lower transforming capacity eventually trigger advanced disease. In addition, our findings could also explain in part the tendency for longer progression-free survival observed for RET-mutated patients treated with Cabozantinib when compared with RAS-mutated patients (60 vs. 47 weeks) (13), as the latter group expresses much less frequently important targets of this drug. Additionally, differential expression of other Cabozantinib targets (e.g., Flt-3, Tie2, or AXL) could also contribute to differences in drug response.

There are other TKIs such as sorafenib, sunitinib, motesanib, and axitinib, currently being tested in clinical trials to treat aggressive MTC. Some of these drugs target PDGRFB (12), which according to our observations was not expressed by RAS-related tumors at all. On the other hand, it was particularly interesting that a major part of WT tumors expressed VEGF, suggesting that antiangiogenic therapy could be an option for these patients. In this regard, bevacizumab, a humanized monoclonal antibody that produces inhibition of angiogenesis by inhibiting vascular endothelial growth factor, has shown promising results in various cancers.

This is an exploratory study, which requires further confirmation in an independent series of samples. Moreover, we did not have access to information about the treatment that patients received or their response to treatment. Thus, it was not possible to assess the impact of the differential expression of TKI targets on the treatment outcomes of these patients.

To conclude, this is the first report evaluating the expression of key TKI target proteins in RAS-related MTC tumors. RAS-related MTCs do not express MET and PDGFRB, and stain less frequently for VEGFR3. VEGF was notably more frequently expressed in WT MTCs. These findings could have an important impact on treatment decisions for MTC patients based on the likelihood of benefiting from a particular therapy, and therefore constitute a first step toward personalized medicine for these patients.

Footnotes

Acknowledgments

This work was supported by Fondo de Investigaciones Sanitarias (FIS project PI11/01359 to M.R.) and the project from the Spanish Ministry of Science and Innovation (SAF2012-35779 to C.R.-A.). L.I.-P. is supported by CIBERER. V.M., A.A.D.C., and M.A.-R. are predoctoral fellows of the “la Caixa”/CNIO international PhD program.

We would like to thank Manuel Morente and María Jesús Artiga of the Spanish National Tumor Bank Network (CNIO) for their hard work collecting tumor samples used in this study.

Author Disclosure Statement

No competing financial interests exist.

References

Randolph

, Maniar

. 2000. Medullary carcinoma of the thyroid. Cancer Control, 7:253–261.

Hofstra

, et al. 1994. A mutation in the RET proto-oncogene associated with multiple endocrine neoplasia type 2B and sporadic medullary thyroid carcinoma. Nature, 367:375–376.

Wells

Jr. , et al. 2013. Multiple endocrine neoplasia type 2 and familial medullary thyroid carcinoma: an update. J Clin Endocrinol Metab, 98:3149–3164.

Ciampi

, et al. 2013. Evidence of a low prevalence of RAS mutations in a large medullary thyroid cancer series. Thyroid, 23:50–57.

Moura

, et al. 2011. High prevalence of RAS mutations in RET-negative sporadic medullary thyroid carcinomas. J Clin Endocrinol Metab, 96:E863–E868.

Leboulleux

, et al. 2004. Medullary thyroid carcinoma. Clin Endocrinol (Oxf), 61:299–310.

Orlandi

, et al. 2001. Treatment of medullary thyroid carcinoma: an update. Endocr Relat Cancer, 8:135–147.

Almeida

, Hoff

. 2012. Recent advances in the molecular pathogenesis and targeted therapies of medullary thyroid carcinoma. Curr Opin Oncol, 24:229–234.

Durante

, et al. 2013. Vandetanib: opening a new treatment practice in advanced medullary thyroid carcinoma. Endocrine, 44:334–342.

10.

Fox

, et al. 2013. Vandetanib in children and adolescents with multiple endocrine neoplasia type 2B associated medullary thyroid carcinoma. Clin Cancer Res, 19:4239–4248.

11.

Kurzrock

, et al. 2011. Activity of XL184 (Cabozantinib), an oral tyrosine kinase inhibitor, in patients with medullary thyroid cancer. J Clin Oncol, 29:2660–2666.

12.

Schlumberger

, et al. 2009. Phase II study of safety and efficacy of motesanib in patients with progressive or symptomatic, advanced or metastatic medullary thyroid cancer. J Clin Oncol, 27:3794–3801.

13.

Sherman

, et al. 2013. Efficacy of cabozantinib (Cabo) in medullary thyroid cancer (MTC) patients with RAS or RET mutations: results from a phase III study. J Clin Oncol, 31(suppl):363s, abstr 6000.

14.

Rodriguez-Antona

, et al. 2013. Influence of RET mutations on the expression of tyrosine kinases in medullary thyroid carcinoma. Endocr Relat Cancer, 20:611–619.

15.

Linardou

, et al. 2008. Assessment of somatic k-RAS mutations as a mechanism associated with resistance to EGFR-targeted agents: a systematic review and meta-analysis of studies in advanced non-small-cell lung cancer and metastatic colorectal cancer. Lancet Oncol, 9:962–972.

16.

Ceccherini

, et al. 1993. Exon structure and flanking intronic sequences of the human RET proto-oncogene. Biochem Biophys Res Commun, 196:1288–1295.

17.

Maderna

, et al. 2007. Nestin, PDGFRbeta, CXCL12 and VEGF in glioma patients: different profiles of (pro-angiogenic) molecule expression are related with tumor grade and may provide prognostic information. Cancer Biol Ther, 6:1018–1024.

18.

Rodriguez-Antona

, et al. 2010. Overexpression and activation of EGFR and VEGFR2 in medullary thyroid carcinomas is related to metastasis. Endocr Relat Cancer, 17:7–16.

19.

Miliaras

, et al. 2004. KIT expression in fetal, normal adult, and neoplastic renal tissues. J Clin Pathol, 57:463–466.

20.

Lee

, et al. 2010. Prognostic implication of MET overexpression in myxofibrosarcomas: an integrative array comparative genomic hybridization, real-time quantitative PCR, immunoblotting, and immunohistochemical analysis. Mod Pathol, 23:1379–1392.

21.

Duncan

, et al. 2008. Vascular endothelial growth factor expression in ovarian cancer: a model for targeted use of novel therapies?. Clin Cancer Res, 14:3030–3035.

22.

Hoff

, Hoff

. 2007. Medullary thyroid carcinoma. Hematol Oncol Clin North Am, 21:475–488; viii.

23.

Treglia

, et al. 2012. Detection rate of recurrent medullary thyroid carcinoma using fluorine-18 fluorodeoxyglucose positron emission tomography: a meta-analysis. Endocrine, 42:535–545.

24.

Maliszewska

, et al. 2013. Differential gene expression of medullary thyroid carcinoma reveals specific markers associated with genetic conditions. Am J Pathol, 182:350–362.

25.

Boichard

, et al. 2012. Somatic RAS mutations occur in a large proportion of sporadic RET-negative medullary thyroid carcinomas and extend to a previously unidentified exon. J Clin Endocrinol Metab, 97:E2031–E2035.

26.

Wells

Jr. , et al. 2012. Vandetanib in patients with locally advanced or metastatic medullary thyroid cancer: a randomized, double-blind phase III trial. J Clin Oncol, 30:134–141.