Dicer1 functions as a haploinsufficient tumor suppressor

Abstract

While the global down-regulation of microRNAs (miRNAs) is a common feature of human tumors, its genetic basis is largely undefined. To explore this question, we analyzed the consequences of conditional Dicer1 mutation (Dicer1 "floxed" or Dicer1(fl)) on several mouse models of cancer. Here we show Dicer1 functions as a haploinsufficient tumor suppressor gene. Deletion of a single copy of Dicer1 in tumors from Dicer1(fl/+) animals led to reduced survival compared with controls. These tumors exhibited impaired miRNA processing but failed to lose the wild-type Dicer1 allele. Moreover, tumors from Dicer1(fl/fl) animals always maintained one functional Dicer1 allele. Consistent with selection against full loss of Dicer1 expression, enforced Dicer1 deletion caused inhibition of tumorigenesis. Analysis of human cancer genome copy number data reveals frequent deletion of DICER1. Importantly, however, the gene has not been reported to undergo homozygous deletion, suggesting that DICER1 is haploinsufficient in human cancer. These findings suggest Dicer1 may be an important haploinsufficient tumor suppressor gene and, furthermore, that other factors controlling miRNA biogenesis may also function in this manner.

Figure 1.

Dicer1 mutation reduces post-infection survival in a genetically engineered mouse model of K-Rasdriven lung cancer. Kras^LSL-G12D mice, either wild type or heterozygous or homozygous conditional for Dicer1 (KD^+/+, KD^fl/+, and KD^fl/fl, respectively), were intranasally infected with Ad-Cre, and survival was assessed. Median survival was 194 d for KD^+/+ mice, 108 d for KD^fl/+ mice, and 143 d for KD^fl/fl mice. Statistical significance was assessed by the log-rank test.

Figure 2.

Dicer1 undergoes hemizygous loss in lung tumors. (A) DNA was prepared from Kras^LSL-G12D mouse embryonic fibroblasts (MEFs) either heterozygous conditional or mutant for Dicer1 (KD^fl/+ and KD^+/−) and lung cancer cell lines from KPD^fl/+ mice, and the Dicer1 locus was examined by PCR. (B) DNA was prepared from Kras^LSL-G12D mouse embryonic fibroblasts either homozygous conditional or heterozygous mutant for Dicer1 (KD^fl/fl and KD^+/−) and lung cancer cell lines from KPD^fl/fl mice, and the Dicer1 locus was examined by PCR.

Figure 3.

Hemizygous deletion of Dicer1 causes a global reduction in steady-state miRNA levels. (A) miRNA profiling and hierarchical clustering were performed on lung cancer cell lines either wild type (KPD^+/+) or heterozygous for Dicer1 (KPD^+/− and KPD^fl/−). (B) Small RNA Northern blotting analysis of miRNAs and glutamine tRNA was performed on lung cancer cell lines wild type or heterozygous for Dicer1 as above.

Figure 4.

Complete deletion of Dicer1 inhibits sarcoma development. (A) KPD^fl/+ and KPD^fl/− sarcoma cell lines were infected with MSCV.CreER^T2.puro. Cre-mediated recombination was induced by treatment with 4-OHT for defined time points. DNA was prepared, and recombination of the Dicer1 locus was assessed by PCR. (B) Two independent KPD^fl/− sarcoma cell lines infected with MSCV.CreER^T2.puro were injected subcutaneously into C57Bl6/129SV F1 animals. Animals were treated with or without tamoxifen by intraperitoneal injection, and tumor growth was measured over time. Values are mean ± SEM (n = 8 each). (C) Tumors were isolated from KPD^fl/− sarcoma cell line transplants treated with or without tamoxifen. DNA was prepared, and recombination of the Dicer1 locus was assessed by PCR. Asterisk corresponds to the wild-type Dicer1 locus from the C57Bl6/129SV F1 host animals.