S-adenosylmethionine decarboxylase overexpression inhibits mouse skin tumor promotion

Abstract

Neoplastic growth is associated with increased polyamine biosynthetic activity and content. Tumor promoter treatment induces the rate-limiting enzymes in polyamine biosynthesis, ornithine decarboxylase (ODC), and S-adenosylmethionine decarboxylase (AdoMetDC), and targeted ODC overexpression is sufficient for tumor promotion in initiated mouse skin. We generated a mouse model with doxycycline (Dox)-regulated AdoMetDC expression to determine the impact of this second rate-limiting enzyme on epithelial carcinogenesis. TetO-AdoMetDC (TAMD) transgenic founders were crossed with transgenic mice (K5-tTA) that express the tetracycline-regulated transcriptional activator within basal keratinocytes of the skin. Transgene expression in TAMD/K5-tTA mice was restricted to keratin 5 (K5) target tissues and silenced upon Dox treatment. AdoMetDC activity and its product, decarboxylated AdoMet, both increased approximately 8-fold in the skin. This enabled a redistribution of the polyamines that led to reduced putrescine, increased spermine, and an elevated spermine:spermidine ratio. Given the positive association between polyamine biosynthetic capacity and neoplastic growth, it was somewhat surprising to find that TAMD/K5-tTA mice developed significantly fewer tumors than controls in response to 7,12-dimethylbenz[a]anthracene/12-O-tetradecanoylphorbol-13-acetate chemical carcinogenesis. Importantly, tumor counts in TAMD/K5-tTA mice rebounded to nearly equal the levels in the control group upon Dox-mediated transgene silencing at a late stage of tumor promotion, which indicates that latent viable initiated cells remain in AdoMetDC-expressing skin. These results underscore the complexity of polyamine modulation of tumor development and emphasize the critical role of putrescine in tumor promotion. AdoMetDC-expressing mice will enable more refined spatial and temporal manipulation of polyamine biosynthesis during tumorigenesis and in other models of human disease.

Fig. 1.

Transgene construct for regulated AdoMetDC expression. (A) Human AdoMetDC cDNA was modified by PCR to add specific restriction sites and replace the C-terminal six residues with an HA epitope. The modified cDNA was then ligated into a vector containing an IRES-luciferase cassette to generate the TAMD construct. A 5.6-Kb fragment released by Not I digestion was purified and used for microinjection. (B) TAMD mice were bred with K5-tTA mice to generate bitransgenic animals with regulated AdoMetDC expression in the skin. In this system, AdoMetDC and luciferase are translated from the same mRNA in the absence of Dox, and transgene expression is silenced by Dox treatment. (C) Genotyping by PCR with primers P1 and P2 in (A) yielded a 330-bp product only from the tail DNA of transgenic animals (lane 2). Primers that amplify a 520-bp product from the mouse AZ gene (Oaz1) were included in all reactions as a positive control.

Fig. 3.

Tumor development in DMBA/TPA-treated TAMD/K5-tTA mice. Mice were initiated at 1 day of age with DMBA (200 nmol) and promoted twice weekly with TPA (6.8 nmol) beginning 8 weeks later and continuing for 25 weeks. (A) Tumor incidence is presented as the percentage of mice in each group containing at least one tumor. Group sizes were wild type, n = 23; TAMD, n = 30; K5-tTA, n = 31; and TAMD/K5-tTA, n = 22. TAMD/K5-tTA versus wild type, P = 0.001; TAMD/K5-tTA versus K5-tTA, P = 0.06. (B) Tumors 2mm and larger were counted weekly and multiplicity is presented as the mean for each genotype. Error bars are not included in order to improve the clarity of the graph, and group sizes are the same as in (A). TAMD/K5-tTA versus wild type and K5-tTA, P < 0.0001. (C) Tumor volume (mm³) was calculated and size range (mm in maximum dimension) was classified for all tumors after 25 weeks of tumor promotion. Total tumor counts were wild type, n = 105; TAMD, n = 151; K5-tTA, n = 75; and TAMD/K5-tTA, n = 31. (D) Following 25 weeks of tumor promotion, a cohort of the mice was switched to Dox-containing diet (2g/kg) to silence transgene expression, and tumor counts (group average) were monitored for an additional 10 weeks, while TPA promotion was continued. K5-tTA, n = 13 and TAMD/K5-tTA, n = 11.

Fig.2.

Hair follicle analysis and polyamine content in newborn TAMD/K5-tTA mice. (A) The hair follicle morphogenesis stage was determined in skin sections harvested from newborn mice (1 day old). Complete longitudinal follicles (n = 60) were scored for 3–7 mice per genotype. (B) Hair follicle density was determined in the same sections as in (A). The average number of follicles connected to the interfollicular epidermis was determined in eight high-power fields for 3–7 mice per genotype. Bars represent mean ± standard deviation (SD). (C) Polyamine levels (left y-axis) and the spermine:spermidine ratio (right y-axis) were determined in whole skin samples harvested from the dorsal surface of newborn mice (1 day old). Bars represent mean ± SD, n = 3 to 5. ^*P < 0.05, ^**P < 0.005 versus wild type; ^***P < 0.05 versus K5-tTA.