LINE-1 transposition into murine Thyroglobulin results in congenital thyroid dysplasia

Abstract

A spontaneous mutation in the wild type C57BL/6NTac mouse was discovered that is associated with early-onset histopathologic sequalae typical of thyroid dysplasia. The spontaneous mutation resulted from insertion of a L1 long interspersed nuclear element (LINE-1) into an intron within the Thyroglobulin (Tg) gene. The mouse genome contains a significant amount of retrotransposon DNA, and these mobile genetic elements routinely change genomic location through retrotransposition, including in germ cells. Analysis of the thyroid transcriptome suggested that the presence of the LINE-1 interferes with the Tg gene splicing, resulting in exclusion of exon 26 from most Tg transcripts in animals homozygous (HOM) for the insertion. The LINE-1 insertion allele of the Tg gene has been designated Tgtdys-Tac. The resulting phenotype is inherited in an autosomal dominant manner with affected mice exhibiting thyroid follicular cell dysplasia that progresses to thyroid adenoma by 9 months of age with complete penetrance in homozygotes. Serum thyroid hormone measurements revealed a decrease in triiodothyronine (T3) levels in homozygotes at 12 months of age, as well as a decrease in tetraiodothyronine (T4) levels at 6-9 months and at 12 months of age in both heterozygotes and homozygotes. In addition, serum Thyroid Stimulating Hormone (TSH) level was strongly increased in homozygotes at 6-8 months of age, consistent with hypothyroidism. Computational molecular modeling showed that omission of the 64 amino acids from the TG protein arm domain, which is the consequence of exon 26-skipping in the Tg transcript, results in decreased local stability. This result in combination with the observed up-regulation in unfolded protein response (UPR) pathways in the thyroids of affected animals, identifies the arm domain of TG as important for its proper cellular distribution. This report describes a spontaneous retrotransposon insertion causatively linked to dysregulated physiological phenotypes in a widely used inbred mouse strain.

Fig 1. RNA analysis reveals a ~6-fold reduced coverage of exon 26 in the murine Tg gene.

(A) Tg transcriptional coverage in B6NCrl (CRL) and B6NTac (Taconic) thyroid samples determined by RNAseq. Shown are exons 24 through 31 of the murine Tg gene. Coverage of exon 26 is ~ 6x decreased in B6NTac affected versus unaffected B6NTac and B6Crl samples. (B) Percentage of RNAseq reads continuing downstream of Tg exon 25 (indicated by purple arrow). A custom python script was used to count the 10 nucleotides found downstream of the 3’ end of exon 25. Three types of transcripts were observed continuing from the 3’ end of exon25: 1) intron 25, indicating unspliced RNA (blue arrow), 2) exon 26, indicating normal splicing (green arrow) or, 3) exon 27, indicating aberrant splicing that skipped exon 26 (red arrow). The five affected B6NTac mice all had the majority of reads (73.4-76.3%) continuing with exon 27, whereas the unaffected mice all had less than 0.25% of reads in this category.

Fig 2. Whole genome sequencing map of the intron 25-exon 26 transition of the murine Tg locus.

A variant cluster indicated by the red square is present in affected (lower track) but not unaffected (upper track) animals in intron 25, near the 5’ end of exon 26. Variant cluster reads were used for the identification of a LINE-1 element insertion in the intron between exon 25 and 26.

Fig 3. Genotype-based analysis of the LINE-1.

(A) Pedigree analysis indicating that thyroid dysplasia is strongly associated with the LINE-1 and is inherited in a dominant fashion. Affected animals indicated in red font and filled symbols, unaffected animals indicated in green font and open symbols. (B) Exon 26 inclusion in the Tg transcript in WT, HET and HOM thyroid samples. Graph plots the ratio of the Ct value for the exon 25-26 and exon 26-27 junctions over the Ct value for the exon 24-25 junction, relative to WT. The exon 25-26 and 26-27 junctions are affected by the skipping of exon 26. The data were analyzed by One-way ANOVA with Bonferroni correction. Significance is indicated by **** at adjusted p < 0.0001.

Fig 4. Thyroid dysplasia in inbred C57BL/6NTac mice WT, HET or HOM for the LINE-1.

(A-C) Representative light microscopic morphology of the thyroid follicular epithelial cells of B6NTac WT, HET and HOM at 12 wks of age. Thyroid dysplasia most developed in the HOM animals and to a lesser degree in HET animals. (D-E) Immunohistochemistry staining of thyroid follicular epithelial cells from B6NCrl unaffected and B6NTac affected animals using anti-Thyroglobulin antibody. Accumulation of TG is visible in the cytoplasm of thyroid follicular epithelial cells and the associated decrease in TG staining in the follicular colloid of affected mice (B6NTac) but not in unaffected mice (B6NCrl). (F-G) Representative transmission electron microscopy images of thyroid follicular epithelial cells. The images illustrate the ultrastructural dilation of the endoplasmic reticulum in B6NTac affected mice and not in B6NCrl unaffected mice. (H-K) Representative light-microscopic images of the adenohypophysis of the pituitary in B6NTac affected and B6NCrl unaffected mice. The images from H&E stained tissues (H-I) illustrate an increase in large lightly basophilic cells in the adenohypophysis of the pituitary in affected mice as compared with unaffected mice. The immunohistochemistry staining (J-K) using anti-Thyroid Stimulating Hormone (TSH) shows that the large lightly basophilic cells stain lightly for TSH. (L-M) Representative light microscopic morphology of the thyroid follicular epithelial cells of B6NCrl unaffected and B6NTac affected mice at 12 months of age. The images illustrate the presence of proliferative, inflammatory, and neoplastic changes in older affected mice, but not unaffected mice.

Fig 5. Serum hormone levels in WT, HET and HOM animals.

(A-D) T3 levels are shown for animals at 4 wks (A), 6 wks (B), 9 mo (C) and 1 yr (D) of age. (E-H) T4 levels are shown for animals at 4 wks (E), 6 wks (F), 9mo (G) and 1 yr (H) of age. (I) TSH levels are shown for animals of 6-9mo of age. (J) Matching T4 levels for the same animals as in (I). All panels show individual measurements with the average indicated by a horizontal line. The data were analyzed by One-way ANOVA with Bonferroni correction. Significance is indicated with * at adjusted p < 0.05, ** at adjusted p < 0.01, *** at adjusted p < 0.001, and **** at adjusted p < 0.0001.

Fig 6. TG protein modeling.

(A) AlphaFold2 predicted Arm segment of mouse TG with or without the 64 amino acid omission due to exon 26 skipping is shown, with each residue colored by its pLDDT score from structure prediction. High pLDDT residues are colored as blue, and low pLDDT residues are colored as red. The position of exon 26 encoded sequence is boxed in green dash line. (B) Position-specific pLDDT comparison of AlphaFold 2 predicted mouse TG Arm with (red line) or without (blue line) 64 amino acid omission. Positions of exon 26 encoded region, extended stretch of low pLDDT segments towards the N (1638-1680) and C (1745-1856) termini of exon 26 encoded sequence, and regions with significant drop in pLDDT upon exon 26 skipping (yellow bars) are respectively high-lighted. (C) Cavity analysis of WT and exon 26 skipped structural models. Exon 26 encoded sequence is colored red, and cavities near its position in each structure are shown as grey meshed surfaces. (D) Root-mean squared fluctuation (RMSF) analysis of molecular dynamics simulation outputs for WT (blue line) and exon 26 skipped (red line) mouse TG structures initially modeled by AlphaFold2. Regions with significant increase in fluctuation upon exon 26 skipping are highlighted in yellow bars. Exon 26 sequence and two long stretches of segments with low pLDDT scores (highlighted in (C)) on its N and C termini are also annotated for reference. (E) Root-mean squared deviation of MD plotted for WT (blue line) and exon 26 skipped mouse TG structures.