COS-7 cells (ATCC, Manassas, VA, USA) were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, as previously described (29). The cells were transiently transfected with polyethylenimine-PEI (Polysciences, Inc., Warrington, PA, USA). Plasmid cDNA encoding full-length, Flag-tagged LRIG1 has been described previously (30). The expression vector encoding RET2A (C634R) was previously described (31). The cDNA expression vector encoding RET2B (M918T) was kindly provided by Dr M. Santoro at University Federico II (Naples, Italy) (32).

The cells were lysed at 4°C in TNE-buffer (50 mM Tris pH 7.5, 150 mM NaCl and 2 mM EDTA) supplemented with 0.5% Triton X-100, 1% octyl-β-glucoside, phosphatase inhibitors (50 mM NaF, 2 mM Na₃VO₄), and complete EDTA-free protease inhibitors (Roche). Cell lysates were clarified by centrifugation and analyzed by immunoprecipitation and western blotting using previously described methodologies (27). Ligand-independent RET phosphorylation was examined at 48 h after transfection in COS-7 cells expressing RET2A or RET2B constructs together with either an empty vector or Flag-LRIG1. The blots were scanned in a Storm 845 PhosphorImager and quantifications were conducted with ImageQuant software (both from GE Healthcare Life Sciences, Capital Federal, Argentina). The antibodies used were anti-phosphotyrosine (p-Tyr, clone PY99, catalogue no. SC7020) at a 1:10,000 dilution, anti-Ret (goat antibodies C20 and T20, catalogue no. SC-1290) (both from Santa Cruz Biotechnology, Santa Cruz, CA, USA) at a 1:1,000 dilution, and anti-Flag M2 mouse monoclonal antibody (catalogue no. F1804; Sigma-Aldrich, St. Louis, MO, USA) at a 1:1,500 dilution.

Three different patient cohorts were investigated. The first cohort (Cohort1-TCGA) comprised the thyroid carcinoma cases available at The Cancer Genome Atlas (TCGA), which included 502 PTC and 58 normal thyroid gland tissues (http://cancergenome.nih.gov/; December 15, 2016). The RNAseq count data and patient survival data were downloaded from TCGA (http://cancergenome.nih.gov/; December 15, 2016). The samples were upper quartile normalized FPKM (UQ-FPKM) and the samples were log-transformed to follow a normal distribution.

The second cohort (Cohort2-MTC) has been described previously (33) and included 39 MTC patient samples and 12 normal tissue controls collected with informed consent and local ethical approval at the Karolinska University Hospital. The clinical data, survival and mutation status of the RET and RAS genes were as reported in a previous study (33).

The third cohort (Cohort3-TMA) was collected with the aim of including all patients diagnosed with thyroid carcinoma in the county of Dalarna, Sweden, between the years 1980 and 2008. The study was approved by the Regional Ethics Review Board in Uppsala (dnr 2010/351), and the Swedish Cancer Registry was used to identify the patients. Only patients who provided written informed consent were included in the study. Paraformaldehyde-fixed and paraffin-embedded tissues were retrieved from hospital archives and used to assemble a tissue microarray (TMA). The diagnostic groups PTC, FTC, MTC, PDTC and ATC were represented, and FTCs with or without oncocytic features (so called Hurthle cell or oxyphilic) were separately studied. All patients gave written informed consent prior to their inclusion in the study. The clinical data were collected from patient files, including histology, TNM stage, time to relapse, and date and cause of death.

The RNA from Cohort2-MTC, which had been prepared previously (33), was treated with TURBO DNA-free kit (Life Technologies) according to the manufacturer's instructions. Human LRIG1 and the reference gene RN18S were analyzed in triplicate samples of 20 ng total RNA through quantitative reverse transcription (RT) polymerase chain reaction (PCR), essentially as previously described (9). However, in contrast to what was previously described (9), qScript 1-Step qRT-PCR Kit (Quanta Biosciences, Gaithersburg, MD, USA) and a Bio-Rad CFX96 apparatus (Bio-Rad Laboratories AB) were used. The RN18S-normalized LRIG1 levels were divided by the corresponding level in the QPCR Human Reference Total RNA (Agilent Technologies, Santa Clara, CA, USA). Thus, the normalized LRIG1 level in the reference RNA was set to the value of 1.

Immunohistochemistry (IHC) was performed for Cohort3-TMA. IHC was performed on 4-µm TMA sections using the LRIG1-151 antibody (18,30) at a dilution of 1:50. The antibody against human LRIG1 has been previously described. The intensity and proportion of positive cells were evaluated by an experienced pathologist (M.T.), who had previously performed a similar evaluation using the same semi-quantitative evaluation scale in another material (18). Briefly, the intensity was evaluated as absent (0), weak (1⁺), intermediate (2⁺), or strong (3⁺), and the fraction of positive cells were evaluated and scored as 0%, 1–25%, 26–50%, 51–75% or 76–100% positive cells.

All mice were housed and maintained, and all experiments were performed in accordance with the European Communities Council Directive (86/609/EEC). The experimental protocols were approved by the Regional Ethics Committee of Umeå University in Umeå, Sweden (registration nos. A80-08 and A81-11). The animals were housed under controlled conditions with a 12-h day/night cycle and fed water and standard chow pellets (cat. no. 801730; Special Diets Services, NOVA-SCB Sweden, Sollentuna, Sweden) ad libitum. The Lrig1-deficient allele has been described previously (34) and was bred onto a pure C57BL/6J genetic background for five generations. Heterozygous Lrig1 mice were crossed with a transgenic mouse strain that ectopically expresses the human RET2B oncogene under the human calcitonin promoter and develop MTC at a high frequency (35,36). The RET2B mice were of a mixed genetic background (C57BL/6J:DBA2, 1:1). The mouse genotypes were determined using tail DNA, PCR, and standard molecular biology techniques. The PCR primers that were used for the Lrig1 wild-type were: 5′-GCGAGCGTGTGTTGTGGAGAGGAT-3′ and 5′-CGATTGTCTGGTGATCAGGAGACTGC-3′, those used for the Lrig1 knockout were: 5′-ACCTCAGGGAGCGAGCGCTCTTATGGGTTAGGACG-3′ and 5′-CCCGTGATATTGCTGAAGAGCTTGGCGGCGAATGG-3′, and those used for the RET2B were: 5′-TGGAGACCCAAGACATCAAC-3′ and 5′-GGAGAAGAGGACAGCGGCT-3′. The PCR products (405 bp for the Lrig1 wild-type; 1,060 bp for the Lrig1 knockout; and 220 bp for the RET2B) were separated and visualized using analytical agarose gel electrophoresis.

The plasma calcitonin levels of the mice were measured every third month according to the methodology as previously described (36). Plasma calcitonin is an established marker of MTC and increased levels of it are usually detectable long before general signs of the disease. At 2 years of age or at signs of the disease, the mice were sacrificed and their thyroid glands were dissected, fixed in 4% phosphate-buffered formaldehyde, embedded in paraffin, and stained with routine hematoxylin and eosin staining.

To analyze differential gene expression, the tumor samples were compared with normal samples using a Student's t-test in Cohort1-TCGA and a Mann-Whitney U test in Cohort2-MTC. Associations between ordinal variables were tested using the Wilcoxon rank sum test, χ² test, one-way analysis of variance (ANOVA) not assuming equal variances, or Fisher's exact test, where appropriate. For survival analysis, the tumor samples were divided into two groups of high and low LRIG1, with the median as the cut-off, and the cumulative survival was then analyzed in the two groups and compared using the log-rank test. Survival was assessed in a Kaplan-Meier graph and a log-rank test was used for comparing different groups. Progression-free survival (PFS, the end-point date of relapse, censoring at the date of the last follow-up), disease-specific survival (DSS, the end-point date of death when thyroid cancer was specified as the cause of death, censoring at the date of the last follow-up), and overall survival (OS, the end-point date of death from any cause, censoring at the date of the last follow-up) were all addressed. The groups that were investigated included the expression of LRIG1, as well as other known prognostic factors, including clinical stage and age. All of the significance testing was performed at the 0.05 level and only two-sided P-values were presented.

To investigate whether LRIG1 interacts with and regulates the oncogenic RET2A and RET2B receptors, COS-7 cells were co-transfected with LRIG1 and the respective RET variant. After 48 h, cell lysates were prepared and subjected to immunoprecipitation and western blot analysis. LRIG1 physically interacted with both RET2A and RET2B, as shown by the co-immunoprecipitation of LRIG1 and the RET2A and RET2B proteins (Fig. 1A and B). Furthermore, LRIG1 co-expression induced the downregulation of the phosphorylation levels of RET2A and RET2B without significantly affecting the corresponding total protein levels (Fig. 1C–F). Thus, LRIG1 physically interacted with and inhibited the activation of both RET2A and RET2B.

To investigate the importance of LRIG1 expression in thyroid cancer, we i) analyzed the PTC expression data available at TCGA (Cohort1-TCGA), ii) performed quantitative RT-PCR analysis of a series of MTC samples (Cohort2-MTC), and iii) analyzed a thyroid cancer TMA (Cohort3-TMA) through immunohistochemistry.

Cohort1-TCGA contained data from 502 PTC samples and 58 normal thyroid gland samples. In this material, the mean LRIG1 expression level was lower in PTC than in the normal thyroid gland (P=2.2×10⁻¹⁶, Student's t-test) (Fig. 2A). There was a significant association between a higher PTC clinical stage and lower LRIG1 expression (P=0.0034, one-way ANOVA) (Fig. 2B). There was no significant association between the OS of patients and the LRIG1 expression level in the TCGA data set (high versus low LRIG1 expression level, median used as cut-off; P=0.435, log-rank test).

In Cohort2-MTC, the relative LRIG1 mRNA levels were analyzed in 39 MTC and 12 normal thyroid glands through quantitative RT-PCR. Twenty-two of the 39 (56%) MTC cases were RET mutated, whereas six (15%) showed either HRAS or KRAS mutations (33). The intragroup variability of LRIG1 expression was relatively high, as evidenced from the sample standard deviations from the means that were 179 and 70% for the normal and MTC samples, respectively. Nevertheless, the relative LRIG1 expression level was significantly lower in MTC than in normal thyroid gland (0.38, SD ±0.27 vs. 1.94, SD ±3.47; P=0.00063, Mann-Whitney U test) (Fig. 2C). Both the RET mutated and RET wild-type cases showed significantly lower LRIG1 levels than the normal thyroid glands (0.41, SD ±0.33 and 0.34, SD ±0.15, respectively vs. 1.94, SD ±3.47; P=0.0036 and P=0.0016, respectively). There was a statistically insignificant association between a higher MTC clinical stage and lower LRIG1 expression (P=0.136, Kruskal-Wallis test) (Fig. 2D). There was no significant difference in LRIG1 levels between the RET-mutated, RAS-mutated, or other MTCs.

Cohort3-TMA was analyzed by immunohistochemistry (IHC) using a TMA with 118 thyroid carcinomas (Table I). In this cohort, 125 patients gave informed consent to participate (Table I) and it was possible to perform IHC on samples from 118 patients, whereas for 7 patients, it was not possible to perform this analysis because of the limited amount of available tissue. Examples of cases with different LRIG1 staining intensities and proportions of positive cells are shown in Fig. 3. Detailed data on the clinical parameters and results of IHC for Cohort3-TMA are presented in Table I. The possible associations between LRIG1 immunoreactivity (proportion of positive cells or intensity of staining) and histology, clinical stage, age, sex, and clinical outcome were investigated. A significant difference was found when histology was compared with the percentage of LRIG1-positive cells (P=0.026, χ² test) and intensity of LRIG1 staining (P=0.033, χ² test). This showed that the distribution of LRIG1 immunoreactivities (percentage or intensity) was different between the different histological groups. However, because of the small sample sizes, it was not possible to draw firm conclusions about the distributions within individual groups. Both the fraction of LRIG1-positive cells and the intensity of LRIG1 staining were associated with whether the patients were dead or alive at the latest follow-up (P=0.036 and P=0.007, respectively, χ² test). The patients who died during the follow-up period were older at the time of diagnosis than those still alive (P<0.001, Mann-Whitney U test), but the age distribution was similar across all categories for the fraction of LRIG1-positive cells and all of the intensities (P=0.15 and P=0.12, respectively, Kruskal-Wallis test). No other significant associations were found between LRIG1 immunoreactivity and the clinical parameters. The same analyses were performed including only PTC and FTC, since these entities have a more favorable clinical course. Both the percentage of positive cells and intensity of staining were still associated with whether the patients were dead or alive at the latest follow-up (P=0.014 and P=0.002, respectively, χ² test). The median follow-up time was 86 months for PFS, 96 months for DSS, and 96 months for OS. Survival analyses were performed for clinical stage, the percentage of LRIG1-positive cells and the intensity of LRIG1 staining. As expected, the clinical stage was a strong prognostic factor for OS, DSS and PFS, both when comparing all of the stages and when comparing stages I and II with stages III and IV (data not shown). No significant associations were found between LRIG1 immunoreactivity and the cumulative survival rates in Cohort3-TMA, both when analysing the whole cohort as well as when analysing PTC only (n=72). Performing additional sub-group analysis for smaller groups was not relevant due to the limited size of each group.

To experimentally address the role of LRIG1 in mutated RET-driven MTC, a mouse strain carrying a deficient Lrig1 allele was crossed with a transgenic mouse strain that expresses human RET2B in thyroid parafollicular cells (C-cells) and is known to develop MTC or C-cell hyperplasia within 2-years and at high frequencies (13 and 77%, respectively) (34,35). The resulting offspring that were heterozygous for both the Lrig1 and the RET2B alleles were crossed with heterozygous Lrig1 mice to yield Lrig1 wild-type, heterozygous, or knockout mice with or without the RET2B transgene (Fig. 4A). Only mice that were positive for the RET2B transgene were included in the study. In total, 125 mice were included, of which 46 (37%) were Lrig1 wild-type, 49 (39%) were Lrig1 heterozygous, and 30 (24%) were Lrig1-deficient. The mice were followed up for up to 2 years. Blood samples were obtained and calcitonin levels were measured every third month beginning at 6 months of age. Surprisingly, the plasma calcitonin levels were below the detection limit of the assay in all of the samples that were analyzed, except for four samples, which were calcitonin positive and represented three different mice. Of the three mice with elevated calcitonin levels, one was Lrig1 wild-type and two were heterozygous for the Lrig1 allele. Eleven of the 46 wild-type mice (24%), 8 of the 49 heterozygous mice (16%), and 7 of the 30 knockout mice (23%) died prematurely, that is, before the planned termination at 24 months. For most of the prematurely dead mice, the cause of death was undetermined; therefore, they were excluded from further analysis. At 24 months of age, the remaining mice were sacrificed and their thyroid glands were collected, formaldehyde-fixed, paraffin embedded, sectioned and routinely stained. Pre-invasive lesions were found in 13 of the 35 wild-type glands (37%), 12 of the 41 heterozygous glands (29%), and in 11 of the 23 Lrig1-deficient glands (48%) (Fig. 4B). The remaining 63 thyroid glands showed no evidence of pathological changes. The difference in the incidence of pre-invasive lesions between the wild-type and Lrig1-deficient mice was not statistically significant (P=0.42, χ² test). Furthermore, no overt cancers were observed.