Reduced viability with age of K5-Tert telomerase transgenic mice compared to wild-type littermates.
Transgenic mice overexpressing the murine telomerase reverse transcriptase gene (Tert) under the control of the bovine keratin 5 (K5) promoter, K5-Tert, were previously described by us (9
). K5 promoter targets Tert expression to basal keratinocytes of stratified epithelia in the mouse (21
; this paper). We have previously shown that adult K5-Tert mice show histologically normal stratified epithelia; however, upon chemical carcinogenesis, the transgenic epithelia showed a higher susceptibility to developing papillomas than those of the corresponding wild-type littermates (9
). Furthermore, the skin of K5-Tert transgenics showed a faster rate of wound healing and a higher proliferation rate upon mitogenic stimuli than that of the corresponding wild-type controls (9
To study the impact of telomerase overexpression on spontaneous cancer incidence in aging of K5-Tert tissues, we maintained five independent K5-Tert founder lines (T1, T3, T4, T5, and T8) for more than 2 years (9
). Lines T1 and T8 showed the highest transgenic telomerase activity expression, lines T4 and T5 showed high to intermediate levels, and line T3 had low levels of transgenic telomerase expression (9
). Only a reduced number of mice of lines T3, T4, and T5 were included in the study. In the case of lines T1 and T8, with the highest telomerase expression, and the wild-type controls, we generated large mouse colonies of 79, 46, and 71 mice, respectively (Fig. ).
FIG. 1. (A) Decreased life span of K5-Tert transgenic mice compared to the wild-type controls. Survival curves for wild type (71 mice), T1 K5-Tert (79 mice), and T8 K5-Tert (46 mice). (B) Percentage of mice of each genotype that presented tumors at the time of (more ...)
Transgenic K5-Tert mice from both the T1 and T8 lines showed increased mortality with age compared with the corresponding wild-type cohorts in the same genetic background. As shown in Fig. , 20% of T8 mice and 15% of T1 mice died in the first 80 weeks of life, whereas only 4% of wild-type mice died during this time. Differences in survival were even more evident after week 110 of life, when only 45% and 78% of T8 and T1 mice survived, respectively, compared to 90% of the wild-type counterparts (Fig. ). The differences in life span between the T1 and T8 K5-Tert transgenic lines could be due to (i) small differences in Tert levels (9
) or (ii) positional effects due to transgene integration in the genome.
Increased incidence of spontaneous preneoplastic and neoplastic lesions in aged K5-Tert mice compared to wild-type controls.
A full-body histopathological analysis was performed on each of the moribund wild-type, T1, and T8 K5-Tert mice. We classified the neoplastic and preneoplastic lesions according to the cell type of origin. Table shows the number of mice that presented the indicated lesions at the time of death out of the total number of mice analyzed. See below for calculations of global tumor incidence per genotype, as well as for statistical analyses.
Preneoplastic lesions in stratified epithelia.
Table shows the number of mice of each genotype that showed preneoplastic lesions (hyperkeratosis or both hyperplasia and hyperkeratosis) in the indicated stratified epithelia (Table ). To compare the incidence of preneoplastic lesions per genotype, we determined the number of mice of each genotype that developed preneoplastic lesions in any of the stratified epithelia at the time of death (Table ). In addition to the moribund wild-type and T1 and T8 K5-Tert mice, we also included T3, T4, and T5 K5-Tert mice that were sacrificed in parallel (>75 weeks old) (Table ).
Telomerase activity levels and incidence of pre-neoplastic lesions in stratified epithelia (skin, esophagus, and forestomach) of aged (>75 weeks old) wild-type and K5-Tert mice from the indicated founder linesa
As shown in Table , there was a higher incidence of preneoplastic lesions in the moribund T1 and T8 mice than in the wild-type controls, 50% and 34.7% compared to 13.3%, respectively. Representative images of preneoplastic lesions in the stratified epithelia of moribund K5-Tert mice are shown in Fig. to c. Histopathological analysis of the T3, T4, and T5 K5-Tert mice revealed hyperplasias and hyperkeratosis only in T4 and T5 K5-Tert lines, which have intermediate and high levels of telomerase activity, respectively, but not in the T3 K5-Tert line, with low levels of telomerase activity (see Table ). These results suggest that the incidence of spontaneous preneoplastic lesions in the stratified epithelia of K5-Tert mice is a direct consequence of telomerase activity reconstitution in the different transgenic lines.
FIG. 2. Examples of preneoplastic and neoplastic lesions in K5-Tert transgenics. Magnifications, ×20. (a) A skin section showing areas of hyperplasia and hyperkeratosis. (b) An esophagus section showing areas of hyperplasia and hyperkeratosis. (c) Forestomach (more ...) Preneoplastic and neoplastic lesions in nonstratified epithelial tissues.
Histopathological analysis of moribund T1 and T8 K5-Tert mice showed an increased frequency of preneoplastic and neoplastic lesions in a wide range of nonstratified epithelial tissues compared to moribund wild-type mice (Table ). The tissues analyzed included the salivary glands, the glandular stomach, the intestine, the liver, the lung, and the male and female genital organs. Figure shows a severe hyperplasia in the glandular stomach of a K5-Tert mouse; this type of lesion was never found in the moribund wild-type cohorts (Table ). Similarly, the K5-Tert male genital organs, including the prostate, prepucial glands, seminal vesicles, and testes, showed preneoplastic and neoplastic lesions that were not found in any of the wild-type cohorts (Table ). Representative images of some of the lesions are shown in Fig. and .
We also detected a higher incidence of tumors in the liver and lung of K5-Tert mice compared to similarly aged wild-type mice. Liver hepatocellular carcinomas and liver lymphomas were present in the K5-Tert mice but were never found in the wild-type cohorts (Table ) (Fig. and ). Other liver tumors such as histiocytic sarcomas were also detected in wild-type mice at a similar frequency. In addition, carcinomas of the mammary gland were detected in T1 and T8 K5-Tert mice but never in the wild-type cohorts. It is interesting that T8 K5-Tert mice showed a higher incidence of sarcomas in various tissues than the T1 K5-Tert line. Again, this difference in tumor susceptibility between T1 and T8 K5-Tert lines could be due to (i) small differences in Tert levels and/or telomerase activity levels or (ii) positional effects due to transgene integration in the genome.
Tumors in lymphoid tissues.
Moribund K5-Tert mice showed elevated frequencies of hyperplasias, lymphomas, and sarcomas in the lymph nodes, as well as hyperplasias and lymphomas in the spleen, compared to similarly aged moribund wild-type mice (Table ).
Tumor incidence by genotype and statistical significance.
The results presented above indicate that two independent K5-Tert transgenic lines, T1 and T8 K5-Tert, showed a modest but consistent increased incidence of spontaneous preneoplastic and neoplastic lesions as they aged compared to the wild-type cohorts. In particular, we determined that 69% and 77% of the moribund T1 and T8 transgenic mice, respectively, showed tumors, compared with only 40% of the wild-type mice (Fig. ; the preneoplastic lesions are not included in this calculation). Furthermore, 22% and 30% of the T1 and T8 transgenic mice with tumors, respectively, showed more than one tumor, compared to 13% of the wild-type mice (Fig. ).
To increase the number of mice studied, we sacrificed wild-type and T1 and T8 transgenic mice (between 91 and 124 weeks old) and determined the percentage of these mice that showed spontaneous tumors at the time of sacrifice (Fig. ). In this case, 79% and 67% of the sacrificed T1 and T8 transgenic mice, respectively, showed tumors, compared to 50% of the wild-type controls (Fig. ). Similarly, 21% and 22% of the T1 and T8 transgenic mice with tumors, respectively, contained more than one tumor, compared to only 17% of the wild-type controls (Fig. ).
To calculate the statistical significance of the differences between genotypes in the incidence and severity of the preneoplastic and neoplastic lesions, we assigned a tumor severity value to each mouse analyzed (see Materials and Methods). Student's t test analysis indicates that moribund T1 and T8 K5-Tert mice showed a significantly increased tumor susceptibility compared to the wild-type controls, P = 0.044 and P = 0.033, respectively (see Materials and Methods).
K5-driven Tert mRNA expression in the mouse is not restricted to stratified epithelia.
As mentioned above, the known expression pattern of keratin 5 in the mouse includes all the stratified epithelia (21
). In agreement with this, we observed an elevated frequency of lesions in the stratified epithelia of aged T1 and T8 K5-Tert mice compared to the age-matched littermate controls (Tables and ). However, the fact that we also observed an elevated frequency of spontaneous neoplasias in nonstratified epithelial tissues as well as in nonepithelial tissues from the K5-Tert mice suggested that transgene expression was not restricted to the stratified epithelia.
To demonstrate this, we developed a real-time PCR-based method that allows quantification of both endogenous and transgenic Tert expression at the mRNA level. The primers used were designed against sequences flanking an intron of the Tert gene to avoid detection of a putative contamination with endogenous Tert genomic DNA (see A in Fig. ; Table ). As expected, all stratified epithelia, including skin, oral mucosa, tongue, esophagus, forestomach, trachea, and vagina, showed a 10- to 100-fold increase in Tert mRNA expression compared to the corresponding wild-type tissues (see asterisk in Table ). In addition, nonstratified epithelial tissues, in which we did not expect transgene expression, such as the prepucial glands, penis, seminal glands, testes, uterus, glandular stomach, thymus, lymph nodes, and lung, also showed an increase of between 10- and 100-fold in Tert mRNA levels compared to the wild-type tissues (Table ). The colon, liver and spleen showed lower but still significantly elevated Tert mRNA expression compared to the corresponding wild-type tissues (Table ). Tert mRNA expression was not elevated in the small intestine, kidney, ovary, heart, and brain of the K5-Tert mice compared to the corresponding wild-type tissues (Table ). These results indicate that, although highly expressed in the stratified epithelia, K5-Tert is also expressed in a wide spectrum of adult tissues in the mouse. This broad expression pattern of the K5-Tert transgene is the most likely explanation for the wide tumor spectrum detected in aged T1 and T8 K5-Tert mice.
Specific K5-Tert mRNA expression in K5-Tert tumors.
Telomerase activity has been described to be upregulated in murine tumors despite the fact that mice have very long telomeres (1
). Similarly, quantitative real-time RT-PCR with Tert-specific primers (primers A in Fig. ) showed that Tert is upregulated in tumors derived from both wild-type and K5-Tert mice (not shown). With this set of primers, however, we could not distinguish between endogenous and transgenic Tert in the tumors derived from K5-Tert mice.
To specifically address whether cells expressing the K5-Tert mRNA contributed to the formation of spontaneous tumors in aged moribund K5-Tert mice, we designed a second RT-PCR strategy that distinguishes endogenous Tert from transgenic K5-Tert expression. For this, we took advantage of the fact that transgenic K5-Tert mRNA contains a 5′ UTR region that is not present in the endogenous Tert transcript (see B in Fig. ). In particular, we designed a forward primer directed against the 5′ UTR region of keratin 5 and a reverse primer directed against the Tert coding sequence (see B in Fig. ). This pair of primers cannot amplify the endogenous Tert mRNA (Fig. ). These primers also distinguish between specific K5-Tert mRNA amplification and a possible contamination of K5-Tert genomic DNA. The expected size for the specific K5-Tert mRNA amplified product is 522 bp, and in the case of genomic DNA contamination, it would be 1,096 bp.
We cloned and sequenced the 522-bp band and demonstrated that it specifically amplifies the K5-Tert mRNA (not shown). With this strategy, we were able to detect K5-Tert mRNA in 9 of 15 K5-Tert tumors analyzed (60%) (Fig. shows examples). This does not rule out that, in addition to K5-Tert, the endogenous Tert is also upregulated in these tumors or in the K5-Tert tumors that are negative for transgene expression. The K5-Tert tumors with detectable specific K5-Tert mRNA expression included skin papillomas, liver hemangiosarcomas, liver histiocytic sarcomas, hepatocellular carcinomas, spleen hyperplasias, and spleen lymphomas. Figure shows specific K5-Tert expression in two K5-Tert liver hemangiosarcomas, one K5-Tert spleen lymphoma, and one K5-Tert skin papilloma (9
). As a positive control, a K5-Tert skin sample was used (Fig. ). As expected, none of the wild-type tumors, zero of four tumors tested, showed transgene expression. As an example, Fig. shows undetectable K5-Tert-specific expression in a hemangiosarcoma from a moribund wild-type mouse.
Aggravated loss of viability and spontaneous tumor incidence of K5-Tert mice in p53 mutant backgrounds.
To test whether forced telomerase expression in adult tissues cooperates with p53 loss to promote cancer development during aging, we generated large colonies of T1 and T8 K5-Tert mice as well as of the corresponding wild-type controls in both p53−/−
genetic backgrounds. Deletion of the p53 tumor suppressor in mice, p53−/−
, results in a dramatic decrease in the life span of these mice coincident with an increased incidence of tumors, particularly, lymphomas (7
; this paper). Heterozygous p53+/−
mice also show a reduced life span and an increased tumor incidence compared to the wild-type controls, although the appearance of tumors is delayed in time with respect to the p53−/−
; this paper).
For the current analysis, large colonies of mice of the following genotypes were obtained and studied: wt/p53+/+ and wt/p53−/− (wild-type for the transgene and null for p53) and T8 K5-Tert/p53−/− (heterozygous for the transgene and null for p53), as well as wt/p53+/− (wild-type for the transgene and heterozygous for p53), T1 K5-Tert/p53+/− (heterozygous for the transgene and for p53), T8 K5-Tert/p53+/− (heterozygous for the transgene and for p53), and T8 K5-Tert/p53+/+ (heterozygous for the transgene and wild-type for p53). We found no significant differences in life span between wt/p53+/+ and T8 K5-Tert/p53+/+ controls at week 80 after birth (Fig. ). This is in agreement with the fact that the negative impact of high telomerase on life span is only seen at older ages, as shown above. T8 K5-Tert/p53−/− and wt/p53−/− cohorts showed a reduced life span compared to that of mice with wild-type p53 (Fig. ). In particular, all of the T8 K5-Tert/p53−/− and wt/p53−/− mice died 40 weeks after birth (Fig. ). Therefore, there were no significant differences in life span between T8 K5-Tert/p53−/− and wt/p53−/− mice, indicating that high telomerase expression in adult tissues does not significantly affect viability in a p53 null background, most likely because these mice die very rapidly of tumors (Table ).
FIG. 5. (A) Survival curves of wild-type and K5-Tert mice in the p53−/− and p53+/− genetic backgrounds. The number of mice of each genotype used in the analysis is indicated. (B) Percentage of mice of each genotype that presented (more ...)
Preneoplastic and neoplastic lesions in 18 moribund wild-type and 18 T8 K5-Tert mice in a p53−/− genetic background after a full-body histopathological analysisa
Interestingly, the moribund K5-Tert/p53−/− cohorts had a reproducibly higher tumor incidence than the control wt/p53−/− mice. In particular, 100% of the K5-Tert/p53−/− mice died of tumors, compared to 83% of the wild-type cohorts (Fig. ). In addition, K5-Tert/p53−/− mice showed a broader spectrum of preneoplastic and neoplastic lesions than the wt/p53−/− mice (Table ). Indeed, most of the lesions, except for lymphomas and sarcomas, were detected only in the K5-Tert/p53−/− mice but not in the wt/p53−/− cohorts (Table ). In addition, the incidence of lymphomas in the thymus and in the lymph nodes was also increased in the K5-Tert/p53−/− mice compared to the wt/p53−/− cohorts (Table ). These results suggest that transgenic telomerase expression in adult tissues favors tumorigenesis, in this case cooperating with p53 deficiency.
This fact was more evident when wt/p53+/−, T1 K5-Tert/p53+/−, and T8 K5-Tert/p53+/−mice were analyzed. In this case, T1 K5-Tert/p53+/− and T8 K5-Tert/p53+/− mice showed a reduced life span compared to the wt/p53+/− controls (Fig. ). In particular, 70 weeks after birth, most of the T1 K5-Tert/p53+/− and T8 K5-Tert/p53+/− mice had died, whereas 60% of the wt/p53+/− had survived (Fig. ). Extensive histopathological analysis of large numbers of moribund mice revealed a significantly higher tumor incidence and a wider tumor spectrum in the T1 K5-Tert/p53+/− and T8 K5-Tert/p53+/− mice than in the wt/p53+/− littermates (Fig. ; Table ). In particular, 90% and 91% of the T8 K5-Tert/p53+/− and T1 K5-Tert/p53+/− mice, respectively, showed tumors at the time of death, whereas only 73% of the wt/p53+/− controls showed tumors (Fig. ). In addition, T1 K5-Tert/p53+/− and T8 K5-Tert/p53+/− mice showed a higher incidence of preneoplastic lesions in stratified epithelia (skin, esophagus, oral cavity), as well as tumors in other tissues that were not affected in the moribund wt/p53+/− mice studied (prepucial glands, vagina, salivary glands, lymphoid tissues; Table ). The differences in the incidence and severity of the preneoplastic and neoplastic lesions between wild-type and T1 K5-Tert/p53+/− or T8 K5-Tert/p53+/− mice were statistically significant (P < 0.05) (Materials and Methods).
Preneoplastic and neoplastic lesions in moribund mice of the indicated genotypea
As described for the single K5-Tert transgenics, a significant proportion of tumors from K5-Tert transgenics in both the p53−/− and p53+/− genetic backgrounds showed K5-Tert mRNA expression by RT-PCR (Fig. , B primers). In particular, 8 of 25 K5-Tert/p53+/− tumors analyzed showed detectable transgenic Tert mRNA expression. The K5-Tert/p53+/− tumors studied included thymic lymphomas, keratoacanthomas, papillomas, and tumors of the salivary glands and the prepucial glands (not shown). As expected, none of the tumors that appeared in the wt/p53+/− mice showed K5-Tert mRNA expression (0 of 10 tumors analyzed).
Altogether, these results indicate that transgenic telomerase expression in adult mouse tissue cooperates with p53 loss in inducing tumorigenesis.