Increased telomere fragility and fusions resulting from TRF1 deficiency lead to degenerative pathologies and increased cancer in mice

Abstract

The telomere repeat-binding factor 1 (TERF1, referred to hereafter as TRF1) is a component of mammalian telomeres whose role in telomere biology and disease has remained elusive. Here, we report on cells and mice conditionally deleted for TRF1. TRF1-deleted mouse embryonic fibroblasts (MEFs) show rapid induction of senescence, which is concomitant with abundant telomeric gamma-H2AX foci and activation of the ATM/ATR downstream checkpoint kinases CHK1 and CHK2. DNA damage foci are rescued by both ATM and ATM/ATR inhibitors, further indicating that both signaling pathways are activated upon TRF1 deletion. Abrogation of the p53 and RB pathways bypasses senescence but leads to chromosomal instability including sister chromatid fusions, chromosome concatenation, and occurrence of multitelomeric signals (MTS). MTS are also elevated in ATR-deficient MEFs or upon treatment with aphidicolin, two conditions known to induce breakage at fragile sites, suggesting that TRF1-depleted telomeres are prone to breakage. To address the impact of these molecular defects in the organism, we deleted TRF1 in stratified epithelia of TRF1(Delta/Delta)K5-Cre mice. These mice die perinatally and show skin hyperpigmentation and epithelial dysplasia, which are associated with induction of telomere-instigated DNA damage, activation of the p53/p21 and p16 pathways, and cell cycle arrest in vivo. p53 deficiency rescues mouse survival but leads to development of squamous cell carcinomas, demonstrating that TRF1 suppresses tumorigenesis. Together, these results demonstrate that dysfunction of a telomere-binding protein is sufficient to produce severe telomeric damage in the absence of telomere shortening, resulting in premature tissue degeneration and development of neoplastic lesions.

Figure 1.

Activation of a DDR in TRF1-deficient MEF. (A) TRF1 deletion in TRF1^flox/flox MEFs upon retroviral infection with Cre recombinase was confirmed by PCR amplification of TRF1 alleles. Alleles are schematically depicted. (B) Knockdown of p53 expression in MEFs infected with a vector containing shRNA-p53 (shp53). (Left) Western blot showing p53 knockdown. (Right) Growth curves of TRF1^+/+ and TRF^Δ/Δ cells doubly infected with Cre and shp53 and of TRF^Δ/Δ cells only infected with Cre. n = independent cultures used per genotype and condition. Error bars indicate standard error. (C, left) Quantification of senescence-associated β-gal-positive cells 4 d post-infection. Error bars indicate standard error. Student's t-test was used for statistical analysis and P-values are indicated. (Right) Representative images of senescence-associated β-gal staining in the indicated genotypes. Black arrows indicate senescent cells. (D, left) Western blot of TRF1 protein and its quantification in the indicated MEFs. Values were normalized to wild-type levels. n = independent MEFs per genotype. Error bars indicate standard error. The Student's t-test was used for statistical analysis and P-values are indicated. (Right) Growth curves of the indicated MEFs. n = independent cultures used per genotype and condition. Error bars indicate standard error. Statistical differences were calculated using the Student's t-test and P-value is indicated. (E) Western blot detection of mouse TRF1 in TRF1^Δ/Δ MEFs treated with empty pBabe vector (ve) and H&R Cre (Cre). Extracts from cells collected 4 and 6 d post-selection were immunoblotted as indicated. Recombinant human TRF1-His₆, migrating more quickly due the smaller size of the human TRF1 protein, served as a control. Histone H3 was used as loading control. (F) Immunofluorescence detection of mouse TRF1 in TRF1^Δ/Δ MEFs treated with empty vector and H&R Cre. Metaphase chromosome spreads were stained with anti-TRF1 rabbit polyclonal antibody (red). DNA was counterstained with DAPI (blue). (G) Immunofluorescence detection of γH2AX (green) combined with FISH staining of the telomeres (red) in TRF1^Δ/Δ MEFs treated with empty vector and H&R Cre. (H) The percentage of metaphase nuclei exhibiting 10–20 or >20 γH2AX foci was determined for at least 50 metaphases prepared as in G and collected 4 and 6 d post-selection. (I) Western blot detection of DDR factors in TRF1^Δ/Δ MEFs treated with empty pBabe vector (ve) and H&R Cre (Cre). Extracts were collected 4 and 6 d post-selection as indicated. MEFs treated with 0 or 10 Gy of IR served as a control for checkpoint activation. Tubulin was used as loading control.

Figure 2.

TRF1 deficiency leads to telomere fusions and increased telomere fragility in the presence of normal binding of other shelterin components. (A,C) Frequency of aberrations in metaphase spreads from the indicated genotypes and conditions. n = metaphases used for the analysis from a total of two MEFs per genotype. Error bars indicate standard error. Statistical comparisons using the χ² test are shown and P-values are indicated. (B,D) Representative images of metaphases from the indicated genotypes. In B, red arrows indicate MTS, white arrows indicate end-to-end fusions and yellow arrows indicate sister telomere fusions. In D, arrows are described in the figure. (E,F) Chromosomal aberrations in untreated MEFs or upon treatment with aphidicolin of the indicated MEFs. (G,H) Telomere fluorescence (Q-FISH) distribution in metaphase spreads from the indicated genotypes. The fused telomeres were quantified as single telomeres in G or divided by a factor of 2 in H. Mean telomere fluorescence and standard deviation values are shown for each genotype. n = number of telomeres used for the analysis. The Student's t-test was used for statistical calculations and P-values are indicated. (I, left) Subcellular fractionation of TRF1^Δ/Δ MEFs treated with empty pBabe vector (ve) and H&R Cre (Cre). Tubulin was used as loading control for the cytoplasmic fraction and histone H3 was used for the chromatin-bound fraction. (Right) Quantification of Western blots probed with TRF2, POT1, TIN2, and RAP1 antibodies is shown. Intensity was calculated relative to the histone H3, which served as a loading control for the chromatin-bound fraction, and was expressed relative to the wild type.

Figure 3.

TRF1 deficiency in mouse tissues leads to massive induction of TIFs. (A) Schematic outline of wild-type (TRF1⁺), floxed (TRF1^flox-neo), and deleted (TRF1^Δ) loci. (B, left) Breeding strategy to generate TRF1^Δ/ΔK5-Cre mice. (Right) PCR amplification of TRF1 alleles in whole skin (dermis and epidermins, D + E), dermis only (D), and epidermis only (E) of TRF1^Δ/ΔK5-Cre mice using the F and R primers depicted in A. (C, left) Quantification of TRF1-positive cells in the dermis and epidermis of the indicated genotypes. n = mice analyzed per genotype. The total number of cells scored per genotype is shown. Error bars indicate standard error. Student's t-test was used for statistical analysis and P-values are indicated. (Right) Representative images of TRF1 immunofluorescence. (D, left) Quantification of TRF1 protein in the epidermis of the indicated genotypes. n = newborn mice used per genotype. Error bars indicate standard error. The Student's t-test was used for statistical analysis and P-values are indicated. (Right) Representative Western blot of TRF1 protein in newborn primary keratinocytes from the indicated genotypes. Actin was used as a loading control. #1, #2, #3 correspond to individual mice. (E, left) Percentage of cells showing γ-H2AX foci, 53BP1 foci, and doubly γ-H2AX–53BP1-positive foci in back skin sections of mice of the indicated genotype. n = mice analyzed per genotype. The total number of cells scored for each genotype is shown. Error bars indicate standard error. Student's t-test was used for statistical analysis and P-values are indicated. (Right) Representative images. Arrows indicate cells with DNA damage foci. (F, left) Percentage of cells with colocalization of 53BP1 foci to telomeres. n = mice analyzed per genotype. The total number of cells scored per genotype is shown. Error bars indicate standard error. Student's t-test was used for statistical analysis and P-values are indicated. (Right) Representative images of 53BP1 foci and of telomeres, and of the combined images. White arrows indicate colocalization of DNA damage foci to telomeres.

Figure 4.

TRF1 deficiency leads to increased p53, p21, and p16 levels and severe proliferative defects in vivo. (A–C) Percentage of p53-positive cells (A), p21-positive cells (B), and p16-positive cells (C) in back skin sections from mice of the indicated genotypes. Representative images are also shown. n = mice analyzed per genotype. The total number of cells scored per genotype is shown in brackets. Error bars indicate standard error. Student's t-test was used for statistical analysis and P-values are indicated. (D, left) Representative Western blot showing TRF1, p53, and p21 protein levels in primary keratinocytes isolated the indicated genotypes. Actin was used as a loading control. #1, #2, #3 correspond to individual mice. (Right) Quantification of p53 and p21 protein levels after normalization to wild-type levels. n = independent mice used per genotype. Error bars indicate standard error. The Student's t-test was used for statistical analysis and P-values are indicated. (E) Quantification of percentage of cells at different phases of the cell cycle by FACS analysis of newborn keratinocytes. Cells were stained with propidium iodide. n = independent newborn mice used per genotype. Error bars indicate standard error. Student's t-test was used for statistical analysis and P-values are indicated.

Figure 5.

TRF1 deficiency results in perinatal mortality and severe epithelial abnormalities. (A) TRF1^Δ/ΔK5-Cre mice were born at the expected Mendelian ratios. (B) Early perinatal lethality of TRF1^Δ/ΔK5-Cre mice. The χ² test was used to determine statistical significance. P-values are indicated. n = number of mice. (C) TRF1^Δ/ΔK5-Cre mice are born with low body weight and do not gain weight after birth. The Student's t-test was used for statistical calculations, and standard error and P-values are shown.(D, left panel) Note the lack of milk in the stomach of a 1-d-old TRF1^Δ/ΔK5-Cre mouse compared with an age-matched wild type. (Middle and right panels) Note that the skin of the 1-d-old TRF1^Δ/ΔK5-Cre mouse shows a shiny appearance and it is hyperpigmented, a phenotype that is aggravated at day 3. The mean weight of each genotype is indicated. Arrows indicate severe hyperpigmentation of the paws. (E,F) Representative images of back skin sections stained for p63 at P1 and P3 (E), as well as stained for involucrin, loricrin, cytokeratin 10, cytokeratin 14, and cytokeratin 6 (F). (G) Transepidermal water loss (TEWL) normalized to wild-type levels (set to 1). n = mice analyzed per genotype. Error bars indicate standard error. Student's t-test was used for statistical analysis and P-values are indicated.

Figure 6.

TRF1 deficiency leads to severe epidermal stem cell defects. (A) Percentage of Sox9-positive cells at interfollicular epidermis and at hair follicles of back skin sections from mice of the indicated genotypes. (Right) Representative images of Sox9 staining in E18.5 embryos and P6 mice. Arrows indicate Sox9 expression. (B) Percentage of cytokeratine 15 (K15)-positive cells in back skin sections from mice of the indicated genotypes. (Right) Representative images of K15-positive staining in the skin of P2 mice. Arrows indicate K15 expression. (C) Percentage of BrdU-positive cells in the interfollicular epidermis and in the hair follicles of back skin sections from mice of the indicated genotypes. (Right) Representative images of BrdU-positive staining in the skin of P1 mice. (D) Quantification of size and number of clones obtained. (Right) Representative examples of clones obtained from newborn mice (P1–P2) at the indicated dilutions. Clones are visualized by rhodamine stainning. n = mice analyzed per genotype. The total number of cells scored per genotype is indicated. Error bars indicate standard error. Students's t-test was used for statistical analysis and P-values are shown.

Figure 7.

p53 deficiency rescues proliferative defects and hyperpigmentation of TRF1^Δ/Δ K5-Cre mice but results in increased skin cancer. (A, left) TRF1^Δ/ΔK5-Cre p53^−/− mice grow hair. (Right) TRF1^Δ/ΔK5-Cre p53^−/− mice do not show skin hyperpigmentation. (B) Kaplan-Meyer survival curves of the different mouse cohorts. n = mice of each genotype included in the analysis. P-value for statistical comparisons between genotypes was calculated using log rank test. (C) Weight gain of the different mouse cohorts. n = mice of each genotype used for the analysis. (D) TRF1^Δ/ΔK5-Cre p53^−/− mice develop nail atrophy and dysplastic epithelium at the proximal and ventral nail fold. (E) TRF1^Δ/ΔK5-Cre p53^−/− mice develop oral leukoplakia characterized by severe epithelia hyperplasia, hyperkeratosis, and dermis invasion by basal keratinocytes (arrows). (F) SCCs in tail and ear skin of TRF1^Δ/ΔK5-Cre p53^−/− mice. (Top panel) Macroscopic appearance of tail and ear skin in a TRF1^Δ/ΔK5-Cre p53^−/− P42 mouse. (Bottom panel) Histopathological analysis of the SCC shows high mitotic index, pleomorphic cells with bizarre nuclei, and aberrant mitosis (arrowhead). Nest of epithelial tumoral cells invade the ear dermis (arrows). (G) Multinucleated giant cells and anaphase bridges are observed in SCC. n = mice or SCC analyzed. The total number of scored cells is indicated. The P-value is calculated by the Student's t-test. (H) Representative images of multinucleated giant cells and anaphase bridges.