Molecular and genetic evidence for the role of AMBRA1 in suppressing S-phase entry and tumorigenesis

Abstract

AMBRA1, which was initially reported to be essential for nervous system development via autophagy and cell proliferation control, also functions as a tumor suppressor by regulating the ubiquitination of D-type cyclins through interaction with DDB1-Cullin4A/4 B E3 ligase. We had identified a missense mutation in AMBRA1 through exome analysis of a family with Cowden syndrome. The patient-type mutant showed reduced DDB1 binding and impaired cyclin D degradation. To investigate the physiological role of AMBRA1, we generated Ambra1 flox mice crossed with Rosa-Cre-ERT2-Tg mice. These inducible Ambra1 conditional knock out mice exhibited increased body weight, organ size, and enhanced S phase entry, with elevated cyclin D expression in a cell lineage- or differentiation-specific manner. Notably, their susceptibility to spontaneous, radiation-, and chemically induced malignancies was significantly higher. These findings support the role of AMBRA1 as a tumor suppressor that regulates cyclin Ds, although other targets may also contribute.

Keywords: Cancer; Cell biology.

Graphical abstract

Figure 1

AMBRA1 is involved in controlling the cell cycle and the stability of cyclin Ds (A) AMBRA1 with non-synonymous alterations of A89G:Q30R and G3585C:R1195S appeared exclusively in the symptomatic family members. Sequence alignment of AMBRA1 of Homo sapiens, Pan troglodytes, Macaca mulatta, Canis lupus, Bos taurus, Mus musculus, Rattus norvegicus, Gallus gallus, Danio rerio, and Xenopus tropicalis at the N- and C-terminus. Substitutions found in the patient are indicated. (B) Five days after the culture of MEFs with the indicated genotype with 4-OHT, the cells were observed with a phase contrast inverted microscope (Axio Vert A1, Carl Zeiss). While active RAS-expressing MEFs showed an extended shape, presumably because of the effects on cytoskeletal organization, AMBRA1-deficient cells did not show such a morphological change. Scale bars, 100 μm. (C) The growth curves of MEFs of the indicated genotype after adding 4-OHT at day 0. The values are presented as mean ± standard deviations (SD); n = 3 in each group (∗∗p < 0.01 compared with the Ambra1+ active RAS- group, ^##p < 0.01 compared with the Ambra1+ active RAS+ group, ^¶¶p < 0.01 compared with the Ambra1 KO active RAS- group, analyzed by two-way repeated measures ANOVA, followed by Tukey’s post hoc test). Three independent experiments showed similar results. (D) Time-course analysis of cyclin D expression in MEFs of the indicated genotype cultured with 4-OHT. (E) Three days after culture with 4-OHT, MEFs with active RAS expression or Ambra1 deficiency were cultured with cycloheximide (CHX) for the indicated periods, and extracted protein samples were subjected to immunoblot analysis for cyclin D1, D2, D3, and tubulin. (F) Densitometric analysis of the data in (E) is shown. Closed squares are Ambra1 KO MEFs, and open triangles are active RAS-expressing MEFs. The expression of cyclin D1, D2, and D3 was higher and more stable in Ambra1 KO MEFs. Three independent experiments showed similar results. The values are presented as mean ± SD (∗p < 0.05 in two-way repeated measures ANOVA, followed by Tukey’s post hoc test).

Figure 2

Attenuated function for the control of the cell cycle and cyclin D expression in Cowden syndrome (A) Cell cycle analysis of murine CD4⁺CD8⁺ immature cell line OVA53 (Parent) and Ambra1-disrupted OVA53 cells transfected with either empty vector (+Empty), AMBRA1 wild-type (+WT), or patient-type AMBRA1 mutant (+CS mut), before (TCR-) and 48 h after TCR/CD3 crosslinking (TCR+). Using the Fucci (SA) system, cells at the G1 phase were visualized to identify the mCherry-positive cells. The percentages of the G1 phase cells are indicated and shown in bar graphs. Means ± SD were calculated using three independent clones per group (∗∗∗p < 0.005 in one-way ANOVA followed by Tukey’s post hoc test). (B) Immunoblot analysis of OVA53 cells (Parent), Ambra1 KO OVA53 cells expressing empty vector (+Empty), AMBRA1 WT (+WT), and patient-type AMBRA1 mutant (+CS Mut). Immunoblotting was performed using antibodies against cyclin D3, AMBRA1, and Actin. (C) Cells used in (B) were cultured in the absence (left) and presence (right) of proteasome inhibitor MG132 at 10 μmol/L for 6 h, lysed, and subjected to immunoblot analysis using antibodies against AMBRA1, cyclin D3, and Tubulin. (D) Following treatment with 25 μg/mL of cycloheximide (CHX) for the indicated time, cell lysates were prepared and subjected to immunoblot analysis using antibodies against cyclin D3 and Tubulin. (E) The relative expressions to the level at 0 min were measured using the ImageJ software and represented as mean ± SD; number of samples: n = 3. The p values were calculated by two-way repeated measures ANOVA, followed by Tukey’s post hoc test. ∗p < 0.05 and ∗∗p < 0.01 compared with the WT group; #p < 0.05 and ##p < 0.01 compared with the CS mut group.

Figure 3

AMBRA1 mutant had a reduced ability to bind to DDB1 (A) FLAG-tagged DDB1 was introduced in Ambra1-disrupted OVA53 cells which had been transfected with either empty vector (+Empty), AMBRA1 wild-type (+WT), or patient-type AMBRA1 mutant (+CS mut). Cell lysates were subjected to immunoprecipitation (IP) using an anti-AMBRA1 antibody. The IP products were analyzed via Immunoblot to detect FLAG (DDB1). (B) Cell lysates were subjected to IP using an anti-FLAG antibody. The IP products were analyzed via Immunoblot to detect AMBRA1 WT and CS mutant.

Figure 4

Marked increase in the body size and sizes of organs in Ambra1 cKO mice (A) Ambra1^flox/flox mice were crossed with Rosa-Cre-ERT2-Tg mice. Tamoxifen (Tam) was administered to the mice at age 8 weeks to delete Ambra1 exon 4 when the mice possessed Cre-ERT2 transgene (Cre-ER+). PCR primers for amplifying deleted/undeleted segments are indicated as arrows shown in Figure S2A (5′F and 3′R). DNA from various organs was subjected to PCR analysis, and the amplified undeleted segment and deleted allele of 1097 bp and 576 bp, respectively, were detected. (B) Ubiquitously expressed AMBRA1 was diminished in Ambra1 cKO mice. Lysates of the thymus, lung, liver, spleen, kidney, colon, and uterus of Ambra1^flox/floxCre-ERT2 (+) and Ambra1^flox/floxCre-ERT2 (−) mice (Cre-ER+ and Cre-ER-, respectively), which were administered with Tam were subjected to immunoblot analysis. (C) Body size of TAM administered-Ambra1 cKO mice and TAM administered-littermates (WT) at 31 weeks of age. (D) Body weights of Ambra1 cKO (Cre-ER(+), red lines) and control mice (Cre-ER(−), blue lines) after Tam administration at 7–9 weeks of age (Top: male, middle: female). Body weights of male Ambra1^+/+ mice with and without Rosa- Cre-ERT2 are shown as controls (bottom). Individual value (left) and mean value ±standard error (SE) of body weights relative to that at the start point (week 0) are indicated (right). (Ambra1^flox/flox male Cre-ER (+): n = 10, Ambra1^flox/flox male Cre-ER(−): n = 8, Ambra1^flox/flox female Cre-ER(+): n = 8, Ambra1^flox/flox female Cre-ER(−): n = 13, Ambra1^+/+ male Cre-ER(+): n = 8, and Ambra1^+/+ male Cre-ER(−): n = 4) (∗p < 0.05 and ∗∗p < 0.01 in the Student’s t test). (E) The liver and kidneys were larger in TAM administered-Ambra1 cKO mice compared to those in TAM administered-littermate control (female littermates, 12 months old). Scale bars, 1 cm. (F) The weights of the heart, lungs, liver, pancreas, kidney, spleen, and uterus of the mice shown in (D). (TAM administered-Ambra1 cKO (cKO, n = 8 and 9, female and male, respectively) and TAM administered-control mice (WT, n = 13 and 8, female and male, respectively)). Data are presented as medians with 25^th–75^th percentiles. (∗p < 0.05, ∗∗p < 0.01, and n.s.: not significant in the Mann-Whitney U-test). (G) The daily average food intake of Ambra1 cKO (cKO, n = 6 and 8, female and male, respectively) and control mice (WT, n = 7 and 8, female and male, respectively). Food intake was measured before Tamoxifen injection (Tamoxifen 0 weeks) and after 5weeks Tamoxifen injection (Tamoxifen 5 weeks). Data are presented as medians with 25^th–75^th percentiles. (∗p < 0.05 and n.s.: not significant in the Mann-Whitney U-test).

Figure 5

Histological manifestations of Ambra1 cKO mice (A) Hematoxylin and eosin (H&E) stained sections of the lung (top) and colon (bottom). Papillary growth of the bronchus (upper right) and hyperplasia of the colon epithelium (lower right) in Ambra1 cKO mice are presented. Scale bars, 100 μm (lung) and 200 μm (colon). (B) H&E-stained sections of the liver (upper). Scale bars, 200 μm. The numbers of cell nuclei per microscopic field (lower, WT: n = 5, cKO1: n = 8, and cKO2: n = 8) were calculated by ImageJ software. The signals from the nuclei were divided into two groups: <15.5 μm² (small nucleus) and >15.5 μm² (large nucleus). Small nuclei mainly corresponded to non-parenchymal cells, such as blood-derived cells, endothelial cells, and stellate cells, while large nuclei corresponded to hepatocytes. Data are presented as medians with 25^th–75^th percentiles. (∗p < 0.05 in the Steel–Dwass test). (C and D) Hematoxylin and eosin (H&E)-stained sections of the liver (C) and lung (D) of WT and Ambra1 cKO mice. Infiltration of hematopoietic cells was frequently observed (C, D, arrowheads). Scale bars, 200 μm (C), 100 μm [(D) upper], and 50 μm [(D) bottom]. (E and F) FACS analysis of hematopoietic cells in the liver (E) and lung (F). Eight months after Tam administration, female mice were used for the experiment. After collagenase treatment and Percoll gradient centrifugation, fractions enriched in hematopoietic cells were analyzed by FACSverse. The number of CD45⁺, CD11b⁺, Ly6G⁺Ly6C^low, Ly6G⁻Ly6C^hi, and NK1.1⁺ cells in the liver (E, upper) and lungs (F, upper) is shown. The proportions of CD11b⁺ cells in CD45⁺ cells, Ly6G⁺Ly6C^low cells in CD11b⁺ cells, Ly6G⁻Ly6C^hi cells in CD11b⁺ cells, and NK1.1⁺ cells in CD45⁺ cells are shown (E and F, bottom). Data are presented as medians with 25^th–75^th percentiles. (∗p < 0.05, ∗∗p < 0.01, and n.s.: not significant in the Mann-Whitney U-test).

Figure 6

Enhanced cell proliferation with increased cyclin D expression in Ambra1 cKO mice (A) Two weeks after Tam administration, proliferating cells in the liver (top), lung (right), and colon (bottom) were detected by Ki-67 expression. Ki-67 positive cells were counted, and signal counts per field were calculated using the ImageJ program. Scale bar, 100μm. Data are presented as medians with 25^th–75^th percentiles. (∗∗p < 0.01 in the Mann-Whitney U-test). (B) The lung, liver, and colon of Ambra1 cKO (cKO) and littermate control (WT) mice, which had been intraperitoneally injected with BrdU 24 h before analysis and subjected to immunohistochemical analysis of BrdU (green) and cyclin D1 (magenta) using fluorescence microscopy. BrdU⁺ cells per microscopic field (upper right) and BrdU-stained nuclei per total nuclei (lower right) were calculated using the ImageJ program. Scale bar, 100 μm. Data are presented as medians with 25^th–75^th percentiles. (∗∗p < 0.01, and n.s.: not significant in the Mann-Whitney U-test). (C) Expression of cyclin D proteins extracted from the thymus, lung, liver, spleen, kidney, colon, and uterus of Ambra1^flox/flox-Rosa-Cre-ERT2-Tg (−) (Cre-ER-) and Ambra1^flox/flox-Rosa-Cre-ERT2-Tg (+) (Cre-ER+) mice subcutaneously injected with TAM. AMBRA1 shown in Figure 4B and cyclin D1 shown in (C) were detected simultaneously on the same membrane, while cyclin D2 and cyclin D3 was detected after reprobing. The same Actin blot is shown in both figures as the loading control.

Figure 7

Increased susceptibility of Ambra1 cKO mice to tumorigenesis (A) Kaplan-Meier survival curve of Ambra1 cKO and control littermates (WT) after Tam administration. p value was assessed using the log rank test. (B and C) Spontaneous tumors in Ambra1 cKO mice. (B) Thymus hypertrophy (arrowhead) because of CD4⁺CD8⁺ T lymphoma, leading to lethal respiratory distress. (C) Hypertrophies of the submandibular and inguinal lymph nodes (left, arrowheads). Hematoxylin and eosin (H&E)-stained section of the submandibular lymph node showed signet ring-like cells (right, arrowheads), reflecting the secretory function of immunoglobulin. Scale bar, 50 μm. (D–H) Irradiation-induced tumors in Ambra1 cKO mice transferred with Ambra1-sufficient hematopoietic cells. (D) Experimental schedule: Two weeks after tamoxifen administration, Ambra1^flox/flox mice with or without Rosa-Cre-ERT2-Tg were irradiated with 8.5 Gy X-ray, and 1×10⁶ WT bone marrow cells were transferred. The mice were examined for carcinogenesis after 24 weeks. (E) Kaplan-Meier Survival curves of Ambra1 cKO and control littermates (WT) after BMT. p value was assessed using the log rank test. (F) The percentage of cancer incidence in irradiated Ambra1 cKO and control mice. (∗∗p < 0.01 by Chi-square method). (G) A flat polyp without a stalk in the colon of an Ambra1 cKO mouse with WT blood cells. The proliferation of atypical columnar epithelium having a large nucleus with tubular and cribriform structure indicated adenocarcinoma. (H) A round tumor (7 mm) with a smooth surface in the thoracic space of an Ambra1 cKO mouse with WT blood cells. Histologically, a diffused proliferation of tumor cells having an oval or spindle nucleus with abundant collagen fibers was observed. The lymphocytes were intermingled. (I–N) AOM-induced cancer in Ambra1 cKO mice. (I) Experimental schedule: Ambra1 cKO and control mice were intraperitoneally injected with AOM, and carcinogenesis was examined 25 weeks later in the surviving mice (cKO: 11 and WT: 20). (J) Kaplan-Meier survival curves of Ambra1 cKO and control littermates (WT) after AOM injection. p value was assessed using the log rank test. (K) The percentage of cancer incidence in Ambra1 cKO and control mice injected with AOM. (∗∗p < 0.01 by Chi-square method). AOM-induced tumor in the liver (L) and colon (M) of Ambra1 cKO mice. (N) AOM-induced angiosarcoma in Ambra1 cKO mice.