Inactivation of ribosomal protein L22 promotes transformation by induction of the stemness factor, Lin28B

Abstract

Ribosomal protein (RP) mutations in diseases such as 5q- syndrome both disrupt hematopoiesis and increase the risk of developing hematologic malignancy. However, the mechanism by which RP mutations increase cancer risk has remained an important unanswered question. We show here that monoallelic, germline inactivation of the ribosomal protein L22 (Rpl22) predisposes T-lineage progenitors to transformation. Indeed, RPL22 was found to be inactivated in ∼ 10% of human T-acute lymphoblastic leukemias. Moreover, monoallelic loss of Rpl22 accelerates development of thymic lymphoma in both a mouse model of T-cell malignancy and in acute transformation assays in vitro. We show that Rpl22 inactivation enhances transformation potential through induction of the stemness factor, Lin28B. Our finding that Rpl22 inactivation promotes transformation by inducing expression of Lin28B provides the first insight into the mechanistic basis by which mutations in Rpl22, and perhaps some other RP genes, increases cancer risk.

Figure 1

Deletions encompassing the RPL22 locus are observed in approximately 10% of primary T-ALL samples. (A) aCGH copy number analysis of primary T-ALL samples. Genomic DNA from primary T-ALL samples was subjected to copy number analysis as described. Vertical dark blue bars denote the position of the deletion. Adjacent genes and their orientation relative to the RPL22 locus are indicated on the right. (B) Sequence analysis of the RPL22 alleles of T-ALL relapse patient samples and cell lines. Representative DNA sequencing chromatograms from T-ALL cell lines with wild-type (left) or mutant (right) RPL22 alleles are depicted. The loss of a single A nucleotide in the stretch of 8 consecutive A on the mutant allele causes a shift in the translational reading frame that truncates Rpl22 after 18 amino acids.

Figure 2

Rpl22 haploinsufficiency accelerates the development of cancer in a mouse model of T-cell malignancy. (A) Kaplan-Meier curves of mice of the indicated genotypes, which were killed on manifestation of outward signs of disease. Myr-Akt2;Rpl22^+/+, n = 10; Myr-Akt2;Rpl22^+/− n = 14; (B) Distribution of CD4/8 subpopulations in thymi of mice with the indicated genotypes. Single-cell suspensions of thymocytes from young adult mice (4-6 weeks) were stained with antibodies reactive CD4 and CD8. Absolute numbers of thymocytes were determined and the mean ± SD are depicted graphically to the right. Analysis was performed on a minimum of 3 mice per group and is representative of 3 experiments performed. (C) Proliferation of explanted thymocytes from MyrAkt2 Tg mice measured by BrdU incorporation. Proliferation of the indicated populations was assessed flow cytometrically by determining the extent of BrdU incorporation after a 4-hour pulse. The mean ± SD of the fraction of BrdU⁺ cells for a representative experiment is depicted graphically. Each bar represents an individual experiment involving at least 3 mice. Three experiments were performed. *P < .05. (D) Assessment of the extent of proliferation of Rpl22^+/+ and ^+/− thymic lymphoma cells by Ki-67 staining in situ. Thymic sections from the indicated mice were either stained with hematoxylin and eosin (H&E) or with anti-Ki67 antibodies to detect the number of proliferating cells. The micrograph was generated using the ×20 objective (×200 total magnification) of a Nikon Eclipse 50i microscope and a Digital Sight DS-Fi1 camera. Mean ± SEM of the thymic organ weight relative to body weight from Rpl22^+/+ (n = 6) and Rpl22^+/− (n = 9) mice at the time of sacrifice is depicted graphically below. *P < .05. Representative thymi are shown on the left. (E) Evaluation of the rate of protein synthesis in thymocytes measured by metabolic labeling. Thymocyte suspensions from mice of the indicated genotypes were metabolic labeling for 30 minutes with [³⁵S]methionine after which the counts incorporated were quantified by TCA precipitation of aliquots of the detergent lysates. Data were derived from triplicate values from 2 independent experiments. In addition, extracts were resolved directly by SDS-PAGE and visualized by fluorography (right).

Figure 3

Rpl22 haploinsufficiency and deficiency promote growth and transformation in cell models in vitro. (A) Effect of Rpl22 inactivation on growth of primary MEFs. Primary Rpl22^+/+, +/−, and −/− MEFs were seeded in triplicate, cultured in 3% O₂ at 37°C, and counted at the indicated intervals for 8 days. Mean cell number ± SD at each time point is represented graphically. Results are representative of 3 independent experiments. (B) Effect of Rpl22 inactivation on transformation of primary MEFs. Primary Rpl22^+/+, +/−, and −/− MEFs were transduced with oncogenes E1A and H-RasV12, followed by drug selection for 1 week, and plating in 0.7% agar. After 3 weeks, colonies were stained with crystal violet and enumerated. Images of representative wells were captured using an EPSON Perfection V700 Photo scanner and are depicted in the top panels. Mean colony number per well ± SD for each genotype is represented graphically in the bottom panel. **P < .005. Data are representative of 3 independent experiments performed in triplicate. (C) Knockdown of Rpl22 expression in immortalized MEFs. Immortalized Rpl22^+/+ MEFs were transduced with control or Rpl22 shRNA constructs, after which the effect on Rpl22 mRNA and protein expression was evaluated by real-time PCR (top) and immunoblotting (bottom). (D-E) Evaluation of immortalized MEF growth and transformation after Rpl22 knockdown. MEFs stably expressing control or 2 Rpl22 shRNA constructs were transformed by oncogenic H-RasV12, after which their growth rate was assessed by counting (panel D; *P > .05 vs control shRNA) and their transformation by colony formation in soft agar (E) as in panel B.

Figure 4

Increased transformation potential associated with of Rpl22 loss or inactivation is accompanied by induction of Lin28B. Effect of Rpl22 knockdown on Lin28 expression in immortalized MEF. MEF lines stably expressing control or 2 different Rpl22 shRNA constructs were harvested for RNA and protein, after which Lin28A and Lin28B mRNA levels were evaluated by real-time PCR (A). **P < .005 for Lin28B expression in Rpl22 shRNA relative to controls. Protein levels were measured by blotting (B). Data are representative of 2 independent experiments. (C) Let-7 miRNA levels in immortalized MEFs where Rpl22 expression was suppressed by shRNA. Expression of Let-7 family miRNA was evaluated by real-time PCR in immortalized MEFs stably expressing control or Rpl22 shRNA constructs. Expression levels were normalized to sno202 RNA and to the expression level in cells transduced with control shRNA. Mean expression levels of triplicate measurements ± SD are represented graphically. Data are representative of 3 independent experiments. P < .05 for Let-7 miRNA levels in Rpl22 shRNA compared with control shRNA. (D) Effect of Rpl22 knockdown on expression of Let-7 targets. Expression of Ras and Myc was assessed by immunoblotting in control or Rpl22 knockdown MEFs. GAPDH served as a loading control. Expression of Lin28B in primary MEFs. Lin28B protein and mRNA levels were measured in primary MEFs of the indicated genotypes by immunoblotting (E) and real-time PCR (F), respectively. Mean ± SD of Lin28B mRNA expression levels are depicted graphically. Results are representative of at least 3 experiments performed. *P < .05. (G) Lin28B expression in primary thymocytes. The expression of Lin28B in thymocytes from mice with the indicated genotypes was evaluated by immunoblotting. GAPDH served as a loading control. Lin28B expression in MEFs after knockdown of Rpl11 and Rpl24. Immortalized MEFs were transduced with 2 different shRNA constructs targeting Rpl24 or Rpl11, after which protein levels were evaluated by immunoblotting (H). GAPDH served as loading control. (I) The level of Lin28B mRNA expressed by these cells was measured by real time PCR as in panel F. **P < .005.

Figure 5

Lin28B is necessary for the enhancement of growth and transformation by Rpl22 inactivation. (A) Effect of knockdown of Lin28B on the growth of immortalized Rpl22^−/− MEFs. Immortalized MEFs of the indicated genotypes were transduced with control or Lin28B shRNA. The effect on expression of Lin28B was assessed by immunoblotting. Growth of triplicate wells of MEFs stably expressing control shRNA or Lin28B shRNA was determined by counting and then plotted as the mean cell number ± SD *P < .05, for Rpl22^−/− compared with Rpl22^−/− in which Lin28B was knocked down. (B) Effect on growth of reintroducing Rpl22 into Rpl22^−/− MEF. Immortalized Rpl22^−/− MEFs were transfected with Rpl22 or empty vector (EV) control, after which we assessed cell growth by counting triplicate wells and depicting the mean ± SD graphically. The expression of Lin28B, c-myc, Rpl22, and GAPDH (loading control) were evaluated by immunoblotting. (C) Dependence of soft agar colony formation on expression of Lin28B. Primary MEFs of the indicated genotypes were transduced with Lin28B shRNA followed by E1A and Ras, and then plated in triplicate in soft agar as Figure 3B. Representative images were captured with a Digital Slight DS-Fi1 camera and NIS Element AR3.0 imaging software at 1× with 3× zoom (×30 total magnification) using a Nikon SMZ1500 stereomicroscope and are shown in the top panels. The mean colony number ± SD is represented graphically beneath. **P < .005 for colonies in Rpl22^+/− and −/− relative to Rpl22^+/+. (D) Lin28B expression in Rpl22-haploinsufficient thymic lymphomas. Explanted thymic lymphomas from MyrAkt2;Rpl22^+/+ and MyrAkt2;Rpl22^−/− mice were evaluated for Lin28B and c-myc expression by immunoblotting. GAPDH served as a loading control. (E) Lin28B mRNA levels in RPL22^+/+ and Rpl22^+/− human T-ALL lines. Lin28B mRNA levels in the indicated T-ALL cell lines were quantified by real-time PCR. Data are presented as Log2 value relative to Jurkat cells (control).

Figure 6

The increased Lin28B expression that results from Rpl22 loss or inactivation is dependent on NF-κB activity. (A) Measurement of NF-κB activity in Rpl22-haploinsufficient primary MEF. EMSA analysis was performed using equal quantities of nuclear extract protein from primary MEFs of the indicated genotypes using both an intact (wt) and p65 binding mutant (mut) NF-κB probe. NF-κB activity was measured in untreated cells, cells pretreated with leptomycin B to trap NF-κB in the nucleus, and after TNFα stimulation (positive control). The composition of the NF-κB complexes was evaluated using supershift analysis using the indicated Abs. Effect of NF-κB inhibition on Lin28B expression. (B) Lin28B mRNA levels were quantified by real-time PCR on RNA extracted from immortalized MEFs stably expressing 2 different Rpl22 shRNAs, in which NF-κB activity had been pharmacologially inhibited by treatment with 1μM NF-κB inhibitor, IMD-350. **P < .01 for IMD-350 treated compared with control treated. (C) Rpl22 was knocked down by shRNA in primary MEF from p65 wild-type (p65+) or Rela^−/−, p65 knockout mice (p65⁻). Lin28B induction was blocked in p65 knockout cells in which Rpl22 was knocked down. Lin28B and Rpl22 mRNA levels were quantified by real-time PCR, and data are plotted as the average of 2 experiments. (D) Model of Rpl22 function in transformation. The model proposes that Rpl22 normally acts to restrain NF-κB activity by an unknown mechanism. However, when Rpl22 expression is diminished either by shRNA knockdown or mutation, NF-κB activity is increased, resulting in increased expression of Lin28B. Lin28B, in turn, promotes transformation at least in part by repressing Let-7 MiR processing, which results in derepression of oncogenic targets such as c-myc. Rpl22 is also likely to regulate additional targets that contribute to transformation.