Cyclin E phosphorylation regulates cell proliferation in hematopoietic and epithelial lineages in vivo

Abstract

Phosphorylations within N- and C-terminal degrons independently control the binding of cyclin E to the SCF(Fbw7) and thus its ubiquitination and proteasomal degradation. We have now determined the physiologic significance of cyclin E degradation by this pathway. We describe the construction of a knockin mouse in which both degrons were mutated by threonine to alanine substitutions (cyclin E(T74A T393A)) and report that ablation of both degrons abolished regulation of cyclin E by Fbw7. The cyclin E(T74A T393A) mutation disrupted cyclin E periodicity and caused cyclin E to continuously accumulate as cells reentered the cell cycle from quiescence. In vivo, the cyclin E(T74A T393A) mutation greatly increased cyclin E activity and caused proliferative anomalies. Cyclin E(T74A T393A) mice exhibited abnormal erythropoiesis characterized by a large expansion of abnormally proliferating progenitors, impaired differentiation, dysplasia, and anemia. This syndrome recapitulates many features of early stage human refractory anemia/myelodysplastic syndrome, including ineffective erythropoiesis. Epithelial cells also proliferated abnormally in cyclin E knockin mice, and the cyclin E(T74A T393A) mutation delayed mammary gland involution, implicating cyclin E degradation in this anti-mitogenic response. Hyperproliferative mammary epithelia contained increased apoptotic cells, suggesting that apoptosis contributes to tissue homeostasis in the setting of cyclin E deregulation. Overall these data show the critical role of both degrons in regulating cyclin E activity and reveal that complete loss of Fbw7-mediated cyclin E degradation causes spontaneous and cell type-specific proliferative anomalies.

Figures

Figure 1.

Analysis of cyclin ET74A T393A-induced hyperproliferation in hematopoietic tissues. (A) Age- and sex-matched animals of the indicated cyclin E genotypes were sacrificed between 20 and 30 wk of age, and spleen lengths and weights were measured. A representative comparison is shown (left panel, ruler increments are in millimeters). Mean weights and standard deviations were calculated from a cohort of 19 mice (right panel). (B) Lysates were prepared from snap-frozen tissues from the indicated genotypes and immunoblotted for cyclin E and Grb2 (loading control) and immunoprecipitated with an anti-cyclin E antibody to assay kinase activity. (C) Representative stainings are shown of formalin-fixed, paraffin-embedded (FFPE) spleens from cyclin Ewt and cyclin ET74A T393A mice, using haematoxylin and eosin (H&E), anti-cyclin E, anti-bromodeoxyuridine (BrdU), and anti-Ki67 antibodies. Displayed are 2× (H&E, Ki67) micrographs to show overall splenic architecture and 10× (cyclin E, BrdU) micrographs. Data from quantitative analysis of BrdU incorporation in splenocytes are shown with standard deviations (SD). (D) Whole spleens and bone marrow were obtained from age- and sex-matched animals (six mice per genotype) and single-cell suspensions prepared. (Top panel) S-phase fractions were determined by DAPI incorporation. (Bottom panel) Nucleated cell counts were manually performed in triplicate. Error bars indicate standard deviations.

Figure 2.

Massively expanded erythroid progenitors and abnormal erythroid maturation in cyclin ET74A T393A mice. (A) Splenocytes were harvested from age- and sex-matched adult mice (aged 20–30 wk) and stained with anti-CD71 (transferrin receptor) and anti-Ter119 (glycophorin A-associated protein) to enumerate erythroid progenitor cells. CD71, Ter119 double-positive cells (top right quadrant) were quantified and are expressed as a ratio to total splenocytes. (B) Erythroid maturation was measured in wild-type (wt), cyclin ET393A (T393A), and cyclin ET74A T393A (T74A T393A) bone marrow. (Left panel) Morphologically distinct Ter119-positive populations are identified on the basis of CD71 expression, which was confirmed by sorting and staining for morphologic analysis. (Right panels) CD71, Ter119 profiles of bone marrow cells show abnormal erythroid maturation in cyclin ET74A T393A mice. Numbers within each sector indicate ratio to total bone marrow cells. (C) Absolute numbers of high CD71-, Ter119-positive and low CD71-, Ter119-positive erythroid progenitors are displayed as mean values from two separate experiments (10 mice total). (D) Representative 40× micrographs of Wright-Giemsa stained cytospins from cyclin Ewt and cyclin ET74A T393A mice, with arrowheads indicating normal-appearing basophilic erythroblasts in the left panel (cyclin Ewt bone marrow) and dysplastic erythroids found in cyclin ET74A T393A bone marrow cytospins (right panel). (E) Cell cycle profiles were obtained from Ter119-positive, high CD71-expression cells isolated from bone marrow. Cells in S-phase and with subgenomic DNA content were measured, and mean values with standard deviations are shown with representative histograms. (F) Quantitative analyses of BrdU incorporation in total bone marrow and Ter119+ cell fractions (top panel) and of anti-Annexin V-staining in Ter119+ cells (bottom panel) are shown. (G) Representative Wright-Giemsa stained peripheral blood smears (40×) from cyclin Ewt and cyclin ET74A T393A mice showing abnormally small (*), large (+), and target (t) cells. Complete blood counts were obtained from six mice from each genotype, and mean red blood cell counts (RBC) and red blood cell distribution widths (RDW) are shown with standard deviations (SD).

Figure 3.

Analysis of cyclin ET74A T393A expression and activity in epithelial tissues. Micrographs at 40× are shown of sections from FFPE colon tissues obtained from age-matched males (A) and mammary glands obtained from females sacrificed during timed breedings (B) and stained with an affinity purified cyclin E antibody. (C) Epithelial cell proliferation was studied in situ in cyclin Ewt, cyclin ET393A, and cyclin ET74A T393A mammary glands by staining tissues for Ki67 proliferation antigen expression and BrdU incorporation. Three age- and gestation time-matched mice per genotype were BrdU pulsed, and positive cells were counted per 40× field (right panel). (Left panels) Micrographs at 20× are shown for cyclin Ewt and cyclin ET74A T393A mammary glands obtained from pregnant mice at D7.5 pc. At least 40 fields were counted per gland. Bars indicate standard error. (D) Apoptosis was detected in mammary tissues from age- and gestation time-matched mice by staining with anti-cleaved caspase 3 antibody, and numbers of positive cells per high power field areshown.

Figure 4.

Delayed mammary involution in cyclin ET74A T393A mice. (A) Mammary acini (indicated by arrowheads) were identified in proximity to ducts (D) and numbers of acinar units were counted per 40× field. Representative 40× micrographs are shown from H&E stained FFPE sections of cyclin Ewt and cyclin ET74A T393A mammary glands (#4) obtained at involution day 16. Five mice per genotype were compared, with approximately equal-sized litters and matched to involution day number. (B) Average numbers of acinar units per field were calculated from 40 fields per mammary gland. (C) Whole mounts were prepared from involuting #4 mammary glands contralateral to those studied in A, and the size of the terminal duct lobules (TDLs) were compared. The maximal cross-sectional widths of TDLs were measured from 10× images taken of whole mounts using NIH ImageJ software. Thirty to 40 TDLs per mammary gland were measured. (D) Tissue sections from FFPE involuting mammary glands from A were stained for Ki67, and positive-staining epithelial cells per 40× field were counted. Average numbers of Ki67-positive cells per field were calculated. (E) Shown are 40× micrographs from anti-activated caspase 3 stained, involuting mammary glands day 21. (F) Average numbers of caspase-positive cells per field were calculated.

Figure 5.

Analysis of cyclin ET74A T393A stability and activity in mouse embryo fibroblasts. (A) Western blot analyses of cyclin E abundance in passage 2 MEFs using an affinity purified anti-cyclin E antibody and anti-Grb2 as loading control are shown. Kinase activity using histone H1 as a substrate was measured from cyclin E-immunoprecipitated complexes. (B) Cyclin E protein stability was measured by treating MEFs with cycloheximide and immunoblotting lysates collected at the indicated time points. (C) Freshly isolated MEFs were expanded in culture and transduced with retroviral vectors expressing a short hairpin RNA sequence targeting murine Fbw7 expression (sh-Fb) or a scrambled, control hairpin sequence (sh-c). Following selection, mRNA was harvested from cells to verify Fbw7 knockdown (Supplemental Fig. 5) and cyclin E protein abundance was measured by immunoblotting lysates (left panel). Cyclin E-associated kinase activity was measured and the relative increase (right panel) with Fbw7 knockdown for each set of MEFs determined. Error bars indicate standard deviations calculated from data from duplicate experiments.

Figure 6.

Cell cycle regulation of cyclin ET74A T393A expression and activity. (A) MEFs were synchronized by serum starvation and then released by serum addition and harvested at the indicated times. Cyclin E abundance and kinase activity are shown. Flow cytometry was performed from cells fixed at the indicated time points and indicated similar cell cycle kinetics among the three genotypes. (B) Kinase activity from A was quantified and plotted relative to starting kinase activity in wild-type MEFs. (C) Aphidicolin- and nocodazole-synchronized MEFs were harvested and subjected to cell cycle, immunoblot, and kinase activity analyses as in A. (D) Kinase activity was quantified and normalized to activity measured in the aphidicolin-synchronized wild-type sample. (E) Spontaneously immortalized wild-type and cyclin ET74A T393A MEFs were synchronized similarly as in A and harvested for immunoblot and cell cycle analyses to compare cyclin E abundance at the indicated times. (*) Background band.