Stem cell expansion during carcinogenesis in stem cell-depleted conditional telomeric repeat factor 2 null mutant mice

Abstract

To examine the role of telomeric repeat-binding factor 2 (TRF2) in epithelial tumorigenesis, we characterized conditional loss of TRF2 expression in the basal layer of mouse epidermis. These mice exhibit some characteristics of dyskeratosis congenita, a human stem cell depletion syndrome caused by telomere dysfunction. The epidermis in conditional TRF2 null mice exhibited DNA damage response and apoptosis, which correlated with stem cell depletion. The stem cell population in conditional TRF2 null epidermis exhibited shorter telomeres than those in control mice. Squamous cell carcinomas induced in conditional TRF2 null mice developed with increased latency and slower growth due to reduced numbers of proliferating cells as the result of increased apoptosis. TRF2 null epidermal stem cells were found in both primary and metastatic tumors. Despite the low-grade phenotype of the conditional TRF2 null primary tumors, the number of metastatic lesions was similar to control cancers. Basal cells from TRF2 null tumors demonstrated extreme telomere shortening and dramatically increased numbers of telomeric signals by fluorescence in situ hybridization due to increased genomic instability and aneuploidy in these cancers. DNA damage response signals were detected at telomeres in TRF2 null tumor cells from these mice. The increased genomic instability in these tumors correlated with eightfold expansion of the transformed stem cell population compared with that in control cancers. We concluded that genomic instability resulting from loss of TRF2 expression provides biological advantages to the cancer stem cell population.

Fig. 1

TRF2 null mice exhibit some stem cell depletion phenotypes consistent with mouse models of dyskeratosis congenita (DKC). (A) Mouse tail from K14-Cre;TRF2+/+ mouse. (B) Crinkled tail from K14-Cre;TRF2f/f mouse. (C) Nail dystrophy in K14-Cre;TRF2f/f mouse. (D) Genotyping of sorted CD34+/K15+ and CD34−/K15− cells from epidermis of K14-Cre;TRF2f/f and K14-Cre;TRF2+/+ mice was performed by PCR using The Jackson Laboratory protocol (upper panel). TRF2 null, TRF2 wild type (wt), and Cre PCR products are shown. TRF2 mRNA expression in sorted CD34+/K15+ and CD34−/K15− cells from epidermis of K14-Cre;TRF2f/f and K14-Cre;TRF2+/+ mice by was performed by RT-PCR (lower panel). β-actin expression is shown as the internal control. Representative gels are shown. (E) Quantitative PCR of genotyping and TRF2 expression shown in (D). Error bars represent SEM.

Fig. 2

TRF2 deficiency results in telomeric DNA damage response and apoptosis of CD34+/K15+ stem cells. TUNEL analysis of sorted CD34+/K15+ cells from K14-Cre;TRF2+/+ epidermis. DAPI (A) and FITC (B) fluorescence is shown. (C) Combined 53BP1 immunofluorescence (FITC) and telomere FISH (Cy3) in sorted CD34+/K15+ cells from K14-Cre;TRF2+/+ epidermis. Cells were counterstained with DAPI. TUNEL analysis of sorted CD34+/K15+ cells from K14-Cre;TRF2f/f epidermis. DAPI (D) and FITC (E) fluorescence is shown. (F) Combined 53BP1 immunofluorescence (FITC) and telomere FISH (Cy3) in sorted CD34+/K15+ cells from K14-Cre;TRF2f/f epidermis. Arrows indicate 53BP1 localization at telomeres (yellow foci). Cells were counterstained with DAPI.

Fig. 3

TRF2 deficiency in basal layer of mouse epidermis induces DNA damage response, apoptosis, and stem cell depletion. (A) Western blot analysis demonstrating DNA damage response in K14-Cre;TRF2+/+ (wt) and K14-Cre;TRF2f/f (TRF) mice. Blots were incubated with antibodies indicated at left using independent protein samples. Epidermis from K14-Cre;TRF2+/+ (B) and K14-Cre;TRF2f/f mice (G). Stem cells (arrow), hair follicles (f), and sebaceous glands (s) are shown. Scale bar = 10 μm. TRF2 expression in skin from K14-Cre;TRF2+/+ (C) and K14-Cre;TRF2f/f mice (H) is shown by immunohistochemistry. Apoptosis in epidermis as determined by TUNEL analysis in K14-Cre;TRF2+/+ (D, DAPI; E, FITC) and K14-Cre;TRF2f/f (I, DAPI; J, FITC) mice. Cell cycle analysis of dissociated epidermal keratinocytes in K14-Cre;TRF2+/+ (F) and K14-Cre;TRF2f/f (K) mice. FACS of dissociated mouse epidermal keratinocytes incubated with control IgG (L), epidermal keratinocytes from K14-Cre;TRF2+/+ (M) or K14-Cre;TRF2f/f (N) mice incubated with phycoerythrin conjugated CD34 antibody. Hematocrit (Hct; O) and white blood cell (WBC) counts (P) in K14-Cre;TRF2+/+ and K14-Cre;TRF2f/f mice.

Fig. 4

Increased tumor latency and reduced proliferation in chemically induced squamous cell carcinomas from K14-Cre;TRF2f/f mice. (A) Percent of K14-Cre;TRF2+/+ and K14-Cre;TRF2f/f mice with tumors is shown. (B) Tumor volume after ten weeks tumor induction in K14-Cre;TRF2+/+ and K14-Cre;TRF2f/f mice. Error bars represent SEM. H&E stained sections of primary SCC in K14-Cre;TRF2+/+ (C) and K14-Cre;TRF2f/f (D) mice. Immunohistochemical analysis of TRF2 protein expression in SCC from K14-Cre;TRF2+/+ (E) and K14-Cre;TRF2f/f mice (F). H&E stained sections of metastatic tumors from K14Cre;TRF2+/+ (G) and K14-Cre;TRF2f/f (H) mice. CD34 and K15 immunofluorescent localization of keratinocyte stem cells in primary (I) and metastatic (J) SCC. Scale bar = 5 μm. (K) TRF2 expression in normal human epidermal keratinocytes (NHEK) and SCC lines is shown by qRT-PCR. These experiments were performed three times with independent samples. Error bars indicate SEM.

Fig. 5

Decreased proliferating cells in SCC from K14-Cre;TRF2f/f mice. Immunohistochemical analysis of PCNA positive cells from SCC in K14Cre;TRF2+/+ (A) and K14-Cre;TRF2f/f (B) mice. Scale bar = 50 μm. (C) Quantification of PCNA positive cells. Error bars indicate SEM.

Fig. 6

Increased apoptotic cells in basal cells of SCC from K14-Cre;TRF2f/f mice. TUNEL localization of apoptotic cells in SCC from K14-Cre;TRF2+/+ (A, DAPI; B, FITC) and K14-Cre;TRF2f/f (C, DAPI; D, FITC). Scale bar = 10 μm. (E) Quantification of TUNEL positive cells. Error bars indicate SEM.

Fig. 7

Increased DNA damage signaling in SCC from K14-Cre;TRF2f/f mice. Localization of p53 positive cells in SCC from K14-Cre;TRF2+/+ (A) and K14-Cre;TRF2f/f mice (B). Scale bar = 10 μm. (C) Quantification of p53 cells. Localization of 53BP1 positive cells in SCC from K14-Cre;TRF2+/+ and K14-Cre;TRF2f/f mice (E). Quantification of 53BP1 positive cells (F). bars indicate SEM.

Fig. 8

Telomere dysfunction in stem cells from epidermis leads to aneuploidy, cancer stem cell expansion, and metastasis in SCC from K14-Cre;TRF2f/f mice. (A) CD34+/K15+ stem cells were sorted from K14-Cre;TRF2+/+ and K14-Cre;TRF2f/f epidermis and SCC. Average telomere length ratios were determined in CD34+/K15+ and CD34−/K15− cells from epidermis and SCC in K14-Cre;TRF2+/+ and K14-Cre;TRF2−/− mice. Error bars indicate SEM. The number of telomere ends in tumor sections from K14-Cre;TRF2+/+ (B) and K14-Cre;TRF2f/f (C) mice was determined by FISH. Representative sections are shown with DAPI counterstain. Scale bar = 10 μm. Telomere signals (Cy3) and 53BP1 foci (FITC) in tumor cells from K14-Cre;TRF2+/+ (D) and K14-Cre;TRF2f/f (E) mice are shown with DAPI counterstain. Colocalized telomere and 53BP1 signals (yellow) in K14-Cre;TRF2f/f cells are shown at arrowheads. Scale bar = 2 μm. Telomere signals (Cy3) are shown by FISH in DAPI stained metaphase chromosomes from K14-Cre;TRF2+/+ (F) and K14-Cre;TRF2f/f (G) tumor cells. Chromosomal fusions are shown by arrowheads in K14-Cre;TRF2f/f metaphase spreads. Scale bar = 5 μm. (H) Percentage of histopathologically confirmed metastatic lymph nodes in K14-Cre;TRF2+/+ and K14-Cre;TRF2f/f mice. Error bars indicate SEM. FACS of dissociated SCC cells from K14-Cre;TRF2+/+ (I) or K14-Cre;TRF2f/f (J) mice incubated with phycoerythrin conjugated CD34 antibody. (K) Relative CXCR3 expression in K14-Cre;TRF2+/+ and K14-Cre;TRF2f/f SCC is shown by qRT-PCR. Error bars indicate SEM.