Alternative splicing may be responsible for heterogeneity of thyroglobulin structure

The evolution of biomarkers in thyroid cancer-from mass screening to a personalized biosignature

Thyroid cancer is the most common malignancy of the endocrine system. The diagnosis of thyroid nodules, made by neck examination and ultrasonography, is a common event occurring in over 50% of the patient population over the age of 50. Yet, only 5% of these patients will be diagnosed with cancer. Fine needle aspiration biopsy is the gold standard for diagnosing thyroid nodules. However, 10–15% of these biopsies are inconclusive, ultimately requiring a diagnostic thyroid lobectomy. Consequently, research in thyroid biomarkers has become an area of active interest. In the 40 years since calcitonin was first described as the biomarker for medullary thyroid cancer, new biomarkers in thyroid cancer have been discovered. Advances in genomic and proteomic technologies have defined many of these novel thyroid biomarkers. The purpose of this article is to provide a comprehensive literature review of how these biomarkers have evolved from simple screening tests into a complex array of multiple markers to help predict the malignant potential and genetic signature of thyroid neoplasms.

Circulating thyroid cancer markers

PURPOSE OF REVIEW

To describe the progress in the field of circulating markers of thyroid cancer.

RECENT FINDINGS

Thyroid cancer cells in the circulation can be detected by measuring the mRNA of thyroid-specific genes. Among these, thyroglobulin, and more recently thyroid-stimulating hormone receptor mRNAs' provide high diagnostic sensitivity and specificity for thyroid cancer detection. These markers can be used in synergy with current diagnostic modalities, i.e. fine-needle aspiration and ultrasound, for preoperative diagnosis and serum thyroglobulin measurement for monitoring.

SUMMARY

For the detection of recurrent/residual thyroid cancer, serum thyroglobulin remains the sole circulating marker, but lacks sensitivity and is unreliable in the presence of antithyroglobulin antibodies. The measurement of thyroid-specific mRNA in blood may provide sensitive/specific markers, but significant variability exists among various studies for thyroglobulin mRNA in particular, questioning the validity of this marker. Recent studies have demonstrated the high sensitivity and specificity of thyroid-stimulating hormone receptor mRNA in detecting recurrent/residual disease even in the presence of thyroglobulin antibodies. Fine-needle aspiration biopsy is currently the sole method for evaluating thyroid nodules. Indeterminate fine-needle aspiration cytology is found in approximately 15-30% of specimens. Thyroid-stimulating hormone receptor mRNA measurement in patients with indeterminate fine-needle aspiration may enhance cancer detection and save unnecessary surgeries.

Molecular diagnostic methods in the diagnosis and follow-up of well-differentiated thyroid carcinoma

Thyroid cancers are the most common endocrine malignancies and are being diagnosed with increased frequency in modern clinical practice. Among other diagnostic modalities, fine-needle aspiration (FNA) biopsy of clinically suspicious thyroid nodules is becoming increasingly popular. Preliminary investigations have suggested that molecular diagnostic assays using tumor-specific markers may improve the sensitivity and accuracy of FNA and so may be expected to reduce the frequency of open surgical procedures by identifying those patients with demonstrably benign lesions who do not require definitive surgical excision of their lesions for diagnosis. At the same time, thyroid-specific mRNA assays (especially thyroglobulin mRNA testing) have been used by investigators in the postoperative follow-up of patients with thyroid cancer as a potential means of detecting tumor recurrence in the peripheral blood. Although these studies have not all reported unqualified successes--indeed, some problems based on both technical and biologic limitations have been identified-these assays still hold out the possibility that potentially important new advances in the management of patients with well-differentiated thyroid cancer may be offered by these and other molecular diagnostic methods.

Molecular advances in thyroglobulin disorders

Synthesis of tri-iodothyronine (T(3)) and thyroxine (T(4)) follows a metabolic pathway that depends on the integrity of the thyroglobulin structure. This large glycoprotein is a homodimer of 660 kDa synthesized and secreted by the thyroid cells into the lumen of thyroid follicle. In humans it is coded by a single copy gene, 270 kb long, that maps on chromosome 8q24 and contains an 8.5 kb coding sequence divided into 48 exons. The preprotein monomer is composed of a 19-amino acid signal peptide followed by a 2749-amino acid polypeptide. In the last decade, several mutations in the thyroglobulin gene were reported. In animals, four of them have been observed in Afrikander cattle (p.R697X), Dutch goats (p.Y296X), cog/cog mouse (p.L2263P) and rdw rats (p.G2300R). Mutations in the human thyroglobulin gene are associated with congenital goiter or endemic and nonendemic simple goiter. Thirty-five inactivating mutations have been identified and characterized in the human thyroglobulin gene: 20 missense mutations (p.C175G, p.Q310P, p.Q851H, p.S971I, p.R989C, p.P993L, p.C1058R, p.C1245R, p.S1447N, p.C1588F, p.C1878Y, p.I1912V, p.C1977S, p.C1987Y, p.C2135Y, p.R2223H, p.G2300D, p.R2317Q, p.G2355V, p.G2356R), 8 splice site mutations (g.IVS3-3C>G, g.IVS5+1G>A, g.IVS10-1G>A, g.IVS24+1G>C, g.IVS30+1G>T, g.IVS30+1G>A, g.IVS34-1G>C, g.IVS45+2T>A) 5 nonsense mutations (p.R277X, p.Q692X, p.W1418X, p.R1511X, p.Q2638X) and 2 single nucleotide deletions (p.G362fsX382, p.D1494fsX1547). The thyroglobulin gene has been also identified as the major susceptibility gene for familial autoimmune thyroid diseases (AITD) by linkage analysis using highly informative polymorphic markers. In conclusion the identification of mutations in the thyrogobulin gene has provided important insights into structure-function relationships.

Genomic organization of the 3' region of the human thyroglobulin gene

The genomic organization of the 3' end of the human Thyroglobulin (Tg) gene has not previously been characterized. We isolated and characterized seventeen lambda phage clones from a human genomic library that included nucleotides 6263 to 8410 of the Tg mRNA, encompassing the last thirteen 3' exons of the Tg gene. The region contained exons ranging in size from 94 to 222 nucleotides, split by introns of 1 to 64 kb. We estimate a total of 48 exons in the Tg gene. All the intron-exon boundaries were sequenced. We found that the splicing sequences diverged considerably from the 3' and 5' consensus. However, the GT-AG rule was perfectly respected in all the exons. A total of 5788 intronic bases and most of the sequences contained in the 13 exons were analyzed (1846 bases). One sequence variation, TT to CC at positions 8377-8378, was found in the 3' untranslated segment. The three tyrosine residues involved in thyroid hormones synthesis (amino acids 2554, 2568, and 2747) at the carbosyl termini of Tg, are encoded by exons 44, 45, and 48. The knowledge of the precise organization of the Tg gene should help to direct studies of Tg gene mutations in families in which a defect in the synthesis of Tg occurs.

Two novel cysteine substitutions (C1263R and C1995S) of thyroglobulin cause a defect in intracellular transport of thyroglobulin in patients with congenital goiter and the variant type of adenomatous goiter

We analyzed the thyroglobulin (Tg) gene of 2 unrelated patients with congenital goiter and the Tg gene of 2 siblings with the variant type of adenomatous goiter. The clinical characteristics of the patients with congenital goiter and the variant type of adenomatous goiter were very similar, except for serum Tg levels, which were less than 15 pmol/L in the patients with congenital goiter, but 117-181 pmol/L in the patients with the variant type of adenomatous goiter (normal, 15-50 pmol/L). The tissue content of Tg in the thyroid glands of all 4 patients was reduced at 0.9-3.8% of total protein (normal, 19-40%). The missense mutation C1263R was detected in the 2 unrelated patients with congenital goiter; the pedigree study showed an autosomal recessive pattern of inheritance. In the 2 siblings with the variant type of adenomatous goiter, the missense mutation C1995S was homozygously detected. In the Tg complementary DNA of 110 normal subjects, the allelic frequencies of the C1263R and C1995S mutations were each less than 0.5%. Also in the normal subjects were detected 35 nucleotide polymorphisms, the insertion of 3 nucleotides, and 1 alternative splicing, each of which was not associated with any specific thyroid disease. From these data, the molecular mechanism of the C1263R and C1995S mutations was elucidated. We first analyzed the carbohydrate residues of C1263R Tg and C1995S Tg. Sensitivity to treatment by endoglycosidase H suggests that C1263R Tg and C1995S Tg were retained in the endoplasmic reticulum (ER). Also, the presence of endoglycosidase H-resistant Tg as well as endoglycosidase H-sensitive Tg in the patients with the variant type of adenomatous goiter suggests that a fraction of C1995S Tg was transported to the Golgi and associated with the mildly increased serum Tg levels. Native PAGE and Western blot analysis with anti-Tg antibody showed that C1263R Tg and C1995S Tg form high mol wt aggregates in the ER. Our results suggest that missense mutations that replace cysteine with either arginine or serine cause an abnormal three-dimensional structure of Tg. Such misfolded Tg polypeptides are retained in the ER as high mol wt aggregates.

mRNA encoding human thyroglobulin's C-terminus is heterogeneous

Thyroglobulin (Tg) provides the peptide backbone for synthesis for thyroid hormones. Because previous studies by various techniques have raised the possibility of heterogeneity in Tg's message and translated protein, we have applied a highly sensitive ribonuclease protection assay (RPA) to examine the mRNA species translating part of Tg's C-terminal region, an area containing three of Tg's hormonogenic sites. Tissue samples were obtained from 18 normal and diseased human thyroids at surgery. Three probes spanning part or all of the nucleotide segment containing bases 7808-8086 in the cDNA sequence, detected full-length mRNAs as the dominant transcripts but also showed the consistent presence of at least seven discrete smaller mRNA species in the thyroid samples. The amounts of these smaller mRNAs varied among tissue samples without a discernible relationship to the underlying clinical thyroid condition. We conclude that the mRNA for this region of Tg is quite heterogeneous and offers potential opportunities for translation of different peptide sequences that might affect hormonogenesis in the C-terminal region of the protein.

Thyrotropin regulates thyroglobulin mRNA splicing and differential processing

Rat thyroid tissue and cultured rat thyrocyte lines contain two thyroglobulin (Tg) mRNAs: a 9 kb rTg-1 mRNA encoding the 330, kDa Tg monomer and a recently described 0.95 kb rTg-2 mRNA. These transcripts have identical 5' coding sequences (641 nucleotides); however, the 3' end of rTg-2 is comprised of coding and non-coding sequences not present in rTg-1. To determine if a single Tg gene encoded both mRNA species, a genomic clone was isolated which spanned the full-length rTg-2 cDNA sequence. The promoter sequence and restriction map were the same as for the previously characterized rTg-1 gene, indicating that rTg-1 and rTg-2 mRNAs are splicing variants derived from the same Tg gene. The unique 3' end of rTg-2 mRNA comprised a single exon which was intronic with respect to rTg-1 mRNA formation. The level of rTg-2 in cultured rat thyrocytes was more sensitive to thyrotropin (TSH) regulation than was rTg-1. rTg-2 mRNA was rapidly (and reversibly) depleted to nearly undetectable levels after TSH removal, unlike rTg-1. Conversely, TSH rapidly restored control levels of rTg-2 mRNA in such depleted cells. The data thus support a model of TSH-induced splicing and regulation of the two Tg mRNAs in the rat.

Identification of a minor Tg mRNA transcript in RNA from normal and goitrous thyroids

Polymerase chain reaction (PCR) amplification of nt 4502 to nt 5184 of the thyroglobulin (Tg) mRNA from several patients, with or without elevated serum thyrotropin (TSH), showed a predominant fragment of the expected size (683 bp) and a minor fragment of 512 bp. The sequence of this minor fragment revealed that 171 bp were missing between position 4567 and 4737. It is highly probable that the deleted sequence corresponds to a complete exon, suggesting an alternative splicing as mechanism for the generation of the minor transcript.

A 3' splice site mutation in the thyroglobulin gene responsible for congenital goiter with hypothyroidism

A case of congenital goiter with defective thyroglobulin synthesis has been studied in molecular terms. The patient is the fifth of a kindred of six, three of which have a goiter. The parents are first cousins. Segregation of thyroglobulin alleles in the family was studied by Southern blotting with a probe revealing a diallelic restriction fragment length polymorphism (RFLP). The results demonstrated that the three affected siblings were homozygous for the RFLP. Northern blotting analysis of the goiter RNA with a thyroglobulin probe suggested that thyroglobulin mRNA size was slightly reduced. Polymerase chain reaction amplification of the 8.5-kb thyroglobulin mRNA as overlapping cDNA fragments demonstrated that a 200-bp segment was missing from the 5' region of the goiter mRNA. Subcloning and sequencing of the cDNA fragments, and of the patient genomic DNA amplified from this region, revealed that exon 4 is missing from the major thyroglobulin transcript in the goiter, and that this aberrant splicing is due to a C to G transversion at position minus 3 in the acceptor splice site of intron 3. The presence in exon 4 of a putative donor tyrosine residue (Tyrosine nr 130) involved in thyroid hormone formation provides a coherent explanation to the hypothyroid status of the patient.

Demonstration of a heterogeneous transcription pattern of thyroglobulin mRNA in human thyroid tissues

Previous reports on human thyroglobulin (hTg) modifications in thyroid carcinomas prompted us to study hTg mRNA in thyroid adenomas and carcinomas. The quantification of hTg mRNA showed a decrease in its levels of expression in both pathological conditions which differed by a factor of 2 between adenomas and carcinomas. Furthermore, PCR was used to analyse the characteristics of hTg mRNA by amplifying 4 regions of the hTg mRNA. When applied to 2 normal, 17 benign and 13 malignant pathological tissue specimens, PCR showed no modification in the size of Tg mRNA. However, abnormal sized cDNAs appeared in all tissues with no distinction between the pathologies; the Restriction Fragment Length Polymorphism study of these cDNAs suggests the existence of alternate splicing patterns in thyroglobulin mRNAs.

Structure of rabbit butyrylcholinesterase gene deduced from genomic clones and from cDNA with introns

1. Three clones were isolated from a rabbit genomic library. They covered the entire coding sequence of the rabbit BChE gene. The positions of splice sites between exons 2, 3, and 4 are identical to those found in the human gene (Arpagaus et al., 1990). Exon 2 covers 83% of the coding sequence. This contrasts with the small size of exon 3 (167 bp) and large size of introns 2 and 3 (greater than 20 kb each). The active-site serine at position 198 is found in a highly conserved region. Aspartic acids in positions 91 and 170 are conserved in human and rabbit, and one of them could be involved in the calytic triad. Aspartic acid 70, present in the anionic site of human BChE, is also conserved in rabbit BChE. The coding sequences of human and rabbit BChE are 89% identical over 744 bp around the active-site serine. 2. In addition to the genomic clones, one cDNA clone (BNY1) was isolated. This cDNA was unusual in that it contained intronic sequences. The insert of 1 kb contained 167 coding bases homologous to the nucleotide sequence 1434 to 1600 of human cDNA and corresponded to exon 3 of the BChE gene. On each side of this coding region, consensus sequences of intron-exon boundaries were found. 3. The presence of large-size transcripts in Northern blots and the existence of a cDNA copy of unprocessed mRNA found in the BNY1 clone suggest a slow processing of transcripts. A genomic sequence unspliced in a cDNA of Torpedo AChE could give a transmembrane domain (Sikorav et al., 1988); the corresponding sequence in rabbit BChE gene, also found in a cDNA, had no homology with Torpedo AChE but could be translated in a hydrophobic C-terminal domain if maintained in mature mRNA.

Thyroglobulin structure and function: recent advances

Thyroglobulin is a large-size iodoglycoprotein specific to thyroid tissue and is the substrate for the synthesis of thyroid hormones, thyroxine and 3,5,3'-triiodothyronine. Recent studies, which greatly benefited from recombinant DNA methodologies, improved the knowledge of several structural features of this dimeric protein and permitted insights into some structure-function relationships. Analysis-function of the primary structure of the human thyroglobulin monomer revealed several main characteristics: 1) 3 types of internal homologies; 2) extensive homology with the bovine thyroglobulin monomer and known partial sequences in the thyroglobulins of other mammalian species; 3) significant homologies with 2 other non-thyroid proteins (acetylcholinesterase and the invariant chain of the Ia class II histocompatibility antigen); 4) a terminal localization of the hormonogenic sites at both ends of the monomer. Current studies aim at determining conformational characteristics, understanding the molecular mechanisms of thyroid hormone formation and unraveling those interactions which in the thyroid cell and the thyroid follicle will permit this large pro-hormone to synthesize and release a few small thyroid hormone molecules. A more precise knowledge of this molecule in higher vertebrates and during evolution would impart valuable information concerning thyroid pathology, since thyroglobulin has been implicated in some genetic and in autoimmune thyroid diseases.