Report

Dyskeratosis Congenita and Cancer in Mice Deficient in Ribosomal RNA Modification

See allHide authors and affiliations

Science  10 Jan 2003:
Vol. 299, Issue 5604, pp. 259-262
DOI: 10.1126/science.1079447

Abstract

Mutations in DKC1 cause dyskeratosis congenita (DC), a disease characterized by premature aging and increased tumor susceptibility. The DKC1 protein binds to the box H + ACA small nucleolar RNAs and the RNA component of telomerase. Here we show that hypomorphic Dkc1 mutant (Dkc1m ) mice recapitulate in the first and second generations (G1 and G2) the clinical features of DC.Dkc1m cells from G1 and G2 mice were impaired in ribosomal RNA pseudouridylation before the onset of disease. Reductions of telomere length in Dkc1m mice became evident only in later generations. These results suggest that deregulated ribosome function is important in the initiation of DC, whereas telomere shortening may modify and/or exacerbate DC.

Dyskeratosis congenita (DC) is a rare X-linked recessive disease caused by point mutations in theDKC1 gene (1). Individuals with DC display features of premature aging, as well as nail dystrophy, mucosal leukoplakia, interstitial fibrosis of the lung, and increased susceptibility to cancer (2, 3). The tissues consistently affected by DC, such as the bone marrow and skin, are characterized by a high turnover of their progenitor cells.

DKC1 codes for dyskerin, a putative pseudouridine synthase, which is in a complex with box H + ACA small nuclear RNAs and mediates posttranscriptional modification of ribosomal RNA (rRNA) through conversion of uridine (U) to pseudouridine (Ψ) (4–6). Point mutations in the yeast homolog of DKC1 and loss of DKC1 expression inDrosophila reduce pseudouridylation and processing of rRNA (7–10). Mutations in the catalytic domain of dyskerin required for rRNA pseudouridylation have been implicated in Hoyeraal-Hreidarsson syndrome, which is characterized by a more severe spectrum of pathologies than that of DC, including immunodeficiency, growth retardation, and microcephaly (11, 12).

Dyskerin is also physically associated with the RNA component of human telomerase (hTR), which contains a H + ACA RNA motif (13). Cell lines derived from individuals with DC have reduced telomerase activity and shorter telomeres, as compared with those from individuals without DC (13). Primary cells from the peripheral blood of individuals with DC also have shorter telomeres than those of individuals without DC, but they have normal levels of telomerase activity (14). hTR is mutated in an autosomal dominant form of DC, which has a phenotype that is similar to, but markedly milder than, that of X-linked DC (15). Therefore, an outstanding question is whether the disease state associated with mutations in DKC1 is the result of a defect in rRNA modification, telomere maintenance, or both. The mouse is an ideal model organism to address this issue genetically because defects in telomerase function resulting in telomere attrition would only be manifest in late generations (16, 17).

To investigate the role of DKC1 in disease pathogenesis, we generated hypomorphic Dkc1 mutant mice (Dkc1m ) (Fig. S1, A and B and supporting online text) (18). Dkc1m mice were born at normal Mendelian ratios with no overt developmental defects. Cells from hemizygous male and heterozygous female Dkc1m mice displayed a decreased level of Dkc1 expression (four- and twofold, respectively) (Fig. S1C).

One of the most consistent features of human DC is bone marrow failure (BMF). Starting from 6 months of age, ∼60% ofDkc1m mice in the first and second generations (G1 and G2, respectively) presented with severe anemia [mean hemoglobin at 8 months was 10.6 ± 4 g/dl, versus 15.1 ± 1.7 g/dl in wild-type (WT) controls (P < 0.001,n = 16 in each group)] and lymphopenia [mean absolute number of lymphocytes was 2428 ± 1978 per μl, versus 5352 ± 2945 per μl in WT controls (P = 0.01)]. Because of the lymphopenia, the Dkc1m mice became mildly leukopenic [mean white blood cell count was 4526 ± 3311 per μl, versus 8215 ± 3795 per μl in WT controls (P = 0.07)]. Some of the Dkc1m mice also had reduced platelet counts, as well as a specific decrease in B lymphocytes (19). Analysis of bone marrow (BM) from anemic Dkc1m mice revealed a decrease in cellularity, with a relative increase in fat and fibrous tissue (Fig. 1A) [mean BM cellularity at 8 months was 67.8 ± 7.1%, versus 87.3 ± 5.7% in WT controls (P < 0.0001, n = 14 per group)]. A decrease in erythroid and lymphoid cells (Fig. 1B) and, in certain cases, a decrease in myeloid cells (19) were observed in BM cytospins from Dkc1m mice. Flow cytometric analysis confirmed the reduction in erythroid cells and B lymphocytes (Fig. 1C).

Figure 1

Bone marrow (BM) failure inDkc1m mice. (A) BM sections stained with hematoxylin and eosin, showing decreased BM cellularity inDkc1m mice, as compared with WT controls. Arrows indicate adipose tissue. Scale bar, 50 μm. (B) Cytomorphological analysis of Dkc1m and WT BM cytospin preparations. Bars represent the myeloid-to-lymphoid (left) and the myeloid-to-erythroid ratios (right) of BM cells. Each bar represents the mean ± SE (error bars) from 14 mice. (C) Flow cytometric analysis ofDkc1m BM. The dot plot cytograms show the decrease in B lym phocytes (CD3 and B220 positive cells, top row, top left quadrant) and erythroid cells (CD71 and TER119 double-positive cells, bottom row, top right quadrant) inDkc1m BM. (D) BM colony-forming assay in WT and Dkc1m mice. BM cultures were scored for the number of erythroid and B lymphocyte colony-forming units (CFU-E and CFU-preB, respectively), and the data are expressed as the percentage of colony numbers present in the cultures. Each value represents the mean ± SE (error bars) from triplicates of one of three independent experiments performed on healthyDkc1m mice before the onset of disease.

We next performed BM colony-forming assays to assess the clonogenic and differentiation potential of hematopoietic precursors from WT and Dkc1m mice before the onset of disease. The Dkc1m mice exhibited a reduced number of erythroid and B colony-forming units (Fig. 1D). [3H]thymidine incorporation assays in BM liquid cultures revealed no difference in the proliferation rate ofDkc1m and WT cells (19). Most analyzed mice were G2 hemizygous males. However, features of BMF were also observed at a lower penetrance in heterozygous G1 and G2 female Dkc1m mice (19), consistent with the observation that some human female carriers of DKC1mutations display features of X-linked DC (20,21). Thus, BMF (one of the distinctive features of DC) is highly penetrant in G1 and G2 Dkc1m mice.

Dkc1m mice starting from the G1 generation showed additional pathological features, including dyskeratosis of the skin (Figs. 2A and S3, A and B), extramedullary hematopoieis associated with splenomegaly (Fig. S2), and abnormalities in the lungs and kidneys (Fig. S3, C through F). A subset of individuals with DC develop malignant tumors of various histological origin (2).Dkc1m mice were highly prone to tumors and developed a variety of them (Fig. 2B), the most common being lung and mammary gland tumors, as well as one case of renal cell carcinoma (Figs. 2B and S4, A through F) (19). Strikingly, 50% ofDkc1m mice developed tumors during their life-span, suggesting that Dkc1 is an important tumor-suppressor gene in vivo. All of the defects observed in theDkc1m mice have been previously documented in human DC (2, 3). Thus, the hypomorphicDkc1m mouse is a faithful model of the human disease.

Figure 2

Dyskeratosis of the skin and increased tumor susceptibility in Dkc1m mice. (A) Histopathological analysis of Dkc1m skin. Photomicrographs of skin from the palm of a WT mouse and aDkc1m mouse. The epidermis of 7 out of 25Dkc1m mice analyzed shows hyperplastic changes produced by an expansion of the nucleated cell layers (square bracket). Nucleoli of keratinocytes can be prominently seen in the spinous layer (with the area noted by the bracket), and there is an increase in the granular layer stained with antibody to loricrin (arrows). Scale bar, 200 μm. (B) Tumor incidence is expressed as the percentage of mice that developed tumors in WT andDkc1m mouse cohorts over time (n= 50). Pheocro., pheocromocytoma.

Figure 3

Analysis of telomere status inDkc1m mice. (A) Relative telomere length of primary B lymphocytes derived from WT andDkc1m mice obtained by flow-FISH analysis. Fluorescence intensity values of the selected group (30%) of WT mice that showed the shortest telomere are represented as 100%. The relative intensity of Dkc1m cells was calculated in cells from 5- to 6-month-old and 10- to 12-month-old mice. Average values ± SD (error bars) of the relative fluorescence intensity from three experiments for each age group are shown. The values forDkc1m cells are the average of the selected group (30%) of mice, which showed a telomere reduction. (B) Telomere length analysis obtained by Q-FISH. Telomere length is shown as the frequency distribution of telomere signal intensities, in metaphase spreads prepared from primary B lymphocytes of early (G2) and late (G4) WT and Dkc1m mice. Signal intensities from three mice in each genotype are shown. Data are expressed in telomere fluorescence units (TFU), and the average telomere signal intensities are shown in the top left of each frequency distribution. (C) TRAP assay for telomerase activity in primary B lymphocytes purified from the spleen of WT andDkc1m mice. Different amounts of proteins representing three consecutive dilutions were subjected to the TRAP assay, performed with a [γ-32P]adenosine 5′-triphosphate–labeled primer. R indicates samples treated with ribonuclease before the experimental reaction. The arrow indicates the 36–base pair internal control (IC) for polymerase chain reaction amplification. A representative example of WT andDkc1m samples is shown. (D) The average values ± SD (error bars) of the relative telomerase activity from four independent TRAP experiments on four different WT and mutant extracts are shown. (E) Northern blot analysis for the RNA component of the telomerase complex (mTR) in purified B lymphocytes from WT and Dkc1m mice. The RNA was normalized with a glyceraldehyde phosphate dehydrogenase (GAPDH) probe.

Figure 4

Impaired rRNA pseudouridylation and ribosome function in Dkc1m mice. (Athrough D) Ψ formation in rRNA from G2 WT andDkc1m mutant B lymphocytes. Representative TLC analysis of labeled 28S (A) and 18S rRNA (C) from WT cells. There is a decrease in Ψp formation relative to Up in both 28S (B) and 18S rRNA (D) fromDkc1m cells. Ψp/Up fromDkc1m was then normalized against that for the WT, adjusted to 100%. The averages for 100% of the animals studied (n = 8) are [Ψp/Up 28S: WT, 100%;Dkc1m , 78 ± 9.3 (P < 0.05)] [Ψp/Up 18S: WT, 100%; Dkc1m , 76 ± 10 (P < 0.05)]. Spots corresponding to Ap (adenosine 3′-monophosphate), Cp (eytidine 3′-monophosphate), Gp (guanosine 3′-monophosphate), Up, and Ψp are indicated. (E) Pulse-chase labeling experiments of rRNA from G2 WT andDkc1m B lymphocytes. There is an accumulation of 45S, 41S, and 32S rRNA species with a concomitant decrease in the mature 18S and 28S rRNA fromDkc1m cells. A schematic representation of all intermediate and mature rRNA species is illustrated. (F) G2Dkc1m and WT mouse embryonic fibroblasts (MEFs) were treated with anisomycin (20 μg/ml), cyclohexamide (20 μg/ml), and puromycin (40 μg/ml). Apoptosis was measured by subdiploid peak fluorescence-activated cell sorting analysis. The average of three independent experiments ± SD (error bars) from early passage MEFs is shown. Cells of the same passage from Dkc1m and WT mice were studied.

To determine the molecular basis for DC in theDkc1m mice, we examined their telomere status. As assessed by flow–fluorescence in situ hybridization (flow-FISH), a reduction in telomere length was first evident in 30% of early fourth generation (G4) Dkc1m mice, as compared to WT mice (n = 20 in each group) (Fig. 3A). There were no detectable changes in telomere length in Dkc1m mice before G4 (Fig. S5A). Furthermore, using quantitative-FISH (Q-FISH) analysis (22) to quantitate the relative number of telomere repeats on individual chromosome ends, we did not detect any differences between early generation Dkc1m mice and WT mice, whereas G4 Dkc1m mice showed a loss of telomere repeats, which is consistent with the flow-FISH data (Fig. 3B).

We next investigated whether the decreased telomere length in early G4 mice was accompanied by a decrease in telomerase activity. A telomeric repeat amplification protocol (TRAP) assay revealed a 40% decrease in telomerase activity inDkc1m cells, as compared with WT controls, and this was associated with a 1.2-fold reduction in mouse telomerase RNA (mTR) levels (Fig. 3, C through E). These data indicate that although dyskerin may be an integral component of the telomerase complex, the DC phenotype in the early generation Dkc1m mice is likely to be independent of its role in maintaining telomere length.

We next examined rRNA pseudouridylation in early generation Dkc1m mice. The nucleotide composition of 28S and 18S rRNA was analyzed with two-dimensional thin-layer chromatography (TLC), and the ratio of modified (pseudouridine 3′-monophosphate) to unmodified uridines (uridine 3′-monophosphate) (Ψp/Up) was calculated. In six out of eight Dkc1m samples analyzed, there was a 10 to 40% reduction in Ψp/Up for both the 28S and 18S rRNAs (Fig. 4, A through D). Analysis of newly synthesized rRNA revealed thatDkc1m cells contained immature intermediates of 45S, 41S, and 32S rRNA at time points at which these rRNA species in WT controls were fully processed (Fig. 4E). Finally, to assess ribosome function, we measured the sensitivity of WT and Dkc1m cells to drugs that inhibit translation (23, 24).Dkc1m cells were hypersensitive to these drugs, undergoing apoptosis at a markedly higher rate than WT cells (Fig. 4F). These results indicate that dyskerin acts as a pseudouridine synthase in mammalian cells and that impairment of its activity disrupts ribosome function.

Dkc1m mice recapitulate the major features of DC, including an increased susceptibility to tumor formation. G1 and G2 Dkc1m mice presented with DC and showed alterations in rRNA modification, whereas defects in telomere length were not evident until G4 mice. This is in agreement with the analysis of mTR knockout mice, which show an aging phenotype and tumor susceptibility only in late generations (fourth through sixth) (16). The fact thatDkc1m mice develop the full spectrum of DC features in G1 and G2 strongly suggests that deregulated ribosome function is important for the initiation of DC and that impairment in telomerase activity in Dkc1m mice may modify and/or exacerbate the disease in later generations. Although we cannot exclude additional unrecognized functions of dyskerin, our results establish a role for deregulated rRNA modification in tumor formation and disease pathogenesis.

Supporting Online Material

www.sciencemag.org/cgi/content/full/299/5604/259/DC1

Materials and Methods

SOM Text

Figs. S1 to S4

References

  • * To whom correspondence should be addressed. E-mail: p-pandolfi{at}ski.mskcc.org

REFERENCES AND NOTES

View Abstract

Navigate This Article