Abstract
Mutation in nucleophosmin (NPM1) causes relocalization of this normally nucleolar protein to the cytoplasm (NPM1c+). Despite NPM1 mutation being the most common driver mutation in cytogenetically normal adult acute myeloid leukemia (AML), the mechanisms of NPM1c+-induced leukemogenesis remain unclear. Caspase-2 is a proapoptotic protein activated by NPM1 in the nucleolus. Here, we show that caspase-2 is also activated by NPM1c+ in the cytoplasm and DNA damage–induced apoptosis is caspase-2 dependent in NPM1c+ but not in NPM1wt AML cells. Strikingly, in NPM1c+ cells, caspase-2 loss results in profound cell cycle arrest, differentiation, and down-regulation of stem cell pathways that regulate pluripotency including impairment of the AKT/mTORC1 pathways, and inhibition of Rictor cleavage. In contrast, there were minimal differences in proliferation, differentiation, or the transcriptional profile of NPM1wt cells lacking caspase-2. Our results show that caspase-2 is essential for proliferation and self-renewal of AML cells expressing mutated NPM1. This study demonstrates that caspase-2 is a major effector of NPM1c+ function.
Mutated NPM1 (NPM1c+) activates caspase-2 to maintain AML pluripotency through regulation of the mTOR pathway.
INTRODUCTION
Accounting for 80% of adult acute leukemia, acute myeloid leukemia (AML) is an aggressive malignant blood cancer that impairs hematopoiesis resulting in the accumulation of immature blood cells (1). Among the known driver mutations for AML, mutation of nucleophosmin (NPM1) is the most common, occurring in 60% of adult cytogenetically normal AML (2). NPM1 is a ubiquitous, multi-functional phosphoprotein that localizes in the nucleolus and primarily functions in ribosomal biogenesis (3). The most common mutation of NPM1 (type A) results from a 4–base pair insertion in the C terminus that harbors a nucleolar localization signal (NuLS) (2). This insertion disrupts the NuLS and instead generates a leucine-rich strong nuclear export motif (NES), resulting in constitutive nuclear export of NPM1. Thus, the mutant version of NPM1 is referred to as NPM1 cytoplasmic positive or NPM1c+ (2).
Although mutation in NPM1 is a leukemia-generating event, NPM1c+ AML has a favorable prognostic index compared to NPM1wt AML. As such, the World Health Organization has defined it as a distinct leukemia entity (4). Compared to NPM1wt AML, NPM1c+ AML is associated with high rates of complete molecular remission following standard induction therapy (2). The improved response to treatment is thought to be a result of increased sensitivity of NPM1c+ cells to apoptosis induced by chemotherapeutic drugs (5). However, the mechanisms by which NPM1c+ can both drive leukemogenesis and improve treatment responses are unclear.
We previously reported that NPM1 is an upstream activator of the proapoptotic protein caspase-2 in the nucleolus (6). Caspase-2 is an initiator caspase that is activated by proximity-induced dimerization of inactive monomers following recruitment to specific high molecular weight protein complexes known as activation platforms (7). The activation platform for caspase-2 is the PIDDosome, consisting of PIDD1 (p53-induced protein with a death domain) and RAIDD (receptor-interacting protein–associated interleukin-1β–converting enzyme homolog 1/cell-death determining 3 homologous protein with a death domain) (8). In response to DNA damage, we showed that NPM1 provides a scaffold for PIDDosome formation and caspase-2 activation in the nucleolus (6). Cytoplasmic activation of caspase-2 did not require PIDD1 or NPM1 but it was RAIDD dependent, indicating that two separate activation platforms are assembled in response to DNA damage: one in the cytoplasm and one in the nucleolus. The two different subcellular localizations of caspase-2 activation suggest that, depending on where it is activated in the cell, caspase-2 may have access to different substrates leading to distinct downstream functions. Therefore, it is possible that the subcellular localization of caspase-2 activation is a major determinant of its downstream functions, and caspase-2 may even regulate nonapoptotic functions from the nucleolus in an NPM1-dependent manner. Because NPM1c+ changes the cellular localization of NPM1 (2), this could have a marked effect on the downstream activation of caspase-2 and resulting functional outcomes.
While caspase-2 is considered a proapoptotic protein, it has been shown to have nonapoptotic roles. For example, caspase-2–deficient cells proliferate at higher rates and have increased replication stress (9, 10). As a result, caspase-2–deficient cells and tissues show many features of persistent DNA damage and genomic instability (11–13). In caspase-2–deficient murine models, this genomic instability manifests as accelerated aging or, in certain tumor models, accelerated tumorigenesis (9, 12–15). In aged mice, loss of caspase-2 leads to increased DNA damage in hematopoietic stem cells and increased hematopoiesis (16). Despite the fact that caspase-2 is a proven tumor suppressor in the Eμ-Myc lymphoma mouse model, and increased caspase-2 expression is associated with somewhat improved survival and drug sensitivity in AML (16, 17), a role for caspase-2 in regulating hematopoiesis in leukemia has not been studied.
NPM1c+ AML is characterized by an increased stem cell molecular signature (18). NPM1c+ is associated with increased expression of several homeobox (HOX) genes, which are highly expressed in hematopoietic stem cells (18). This HOX gene signature is specific to NPM1c+ AML and has been proposed to be important for leukemia development (19). This idea is supported by studies showing that enforced relocalization of NPM1c+ back to the nucleus results in the down-regulation of the HOX/MEIS1 (myeloid ecotropic viral integration site 1) network, cell cycle arrest, and terminal differentiation of AML cells (20).
Given the ability of NPM1 to activate caspase-2, and proposed roles for caspase-2 in cell cycle regulation and hematopoiesis, we set out to determine the downstream consequences of NPM1-dependent caspase-2 activation in NPM1wt and NPM1c+ AML cells. On the basis of the known roles of caspase-2 in promoting apoptosis and restricting proliferation, our initial prediction was that loss of caspase-2 would increase proliferation and/or enhance resistance to apoptosis in one or both of these cell types. Instead, we found minimal effects of caspase-2 loss in NPM1wt cells. In contrast, in NPM1c+ cells, loss of caspase-2 reduced apoptosis. Loss of caspase-2 also had profound effects on cell division in NPM1c+ cells, but these were opposite to our initial hypothesis that blocking caspase-2 would increase proliferation. Rather, we made the surprising discovery that, in NPM1c+ AML, caspase-2 is essential for proliferation and self-renewal.
RESULTS
Loss of caspase-2 inhibits apoptosis when NPM1 is localized to the cytoplasm
To investigate how mutation in NPM1 and the subsequent localization changes affect cellular function, we used OCI-AML-2 and OCI-AML-3 cells. These cells express NPM1wt (nucleolus) and NPM1c+ (cytoplasm), respectively (Fig. 1A). For simplicity, we will refer to the OCI-AML-2 cells as NPM1wt and the OCI-AML-3 cells as NPM1c+ throughout this manuscript. To demonstrate that NPM1c+ increases sensitivity to apoptosis in AML cells, we incubated NPM1wt and NPM1c+ cells with the DNA-damaging agents doxorubicin, daunorubicin, and etoposide (Fig. 1, B to D). These drugs are commonly used in AML chemotherapy. In each case, the NPM1c+ cells were significantly more sensitive to apoptosis over a range of concentrations of each drug.
Fig. 1. Increased sensitivity of NPM1c+ AML cells to DNA damage–induced apoptosis is caspase-2 dependent.
(A) Schematic representation showing the protein domains of NPM1 and NPM1c+. The NuLS sequence is indicated in NPM1, which is mutated to a nuclear exclusion signal (NES) in NPM1c+. The localization of NPM1 in NPM1wt cells and in NPM1c+ cells is shown in the nucleolus or cytoplasm, respectively, in blue. (B to D) OCI-AML-2 (NPM1wt) and OCI-AML-3 (NPM1c+) cells were treated with the indicated doses of doxorubicin (B), daunorubicin (C), and etoposide (D) for 16 hours. Apoptosis was assessed by flow cytometry for annexin V binding. Results are the mean of three to four independent experiments ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 (two-way ANOVA with Sidak’s multiple comparison test). (E) NPM1wt and NPM1c+ cells and their respective CRISPR/Cas9-generated caspase-2–deficient clones (ΔC2) were treated with and without daunorubicin (Dnr, 0.5 μM) for 16 hours. The percentage of cells undergoing apoptosis was measured by flow cytometry for annexin V binding. Results are the mean of four independent experiments ± SD. ****P < 0.0001 (two-way ANOVA with Sidak’s multiple comparison test). (F) Parental NPM1wt cells, NPM1c+ cells, and the ΔC2 clones of each were treated with daunorubicin for 16 hours (0.5 μM). Lysates were immunoblotted for caspase-3 and caspase-2. Pro-C3 and Pro-C2, the pro-form of caspase-3 and caspase-2; p17, the large catalytic subunit. Actin was used as a loading control. Results are representative of three independent experiments.
To examine the role of caspase-2 in apoptosis in NPM1wt and NPM1c+ cells, we used CRISPR/Cas9 to generate caspase-2–deficient NPM1wt and NPM1c+ cells via direct electroporation of the Cas9 ribonucleic protein (RNP) complex. We used two single-guide RNAs (sgRNAs) to delete a large portion of the gene, to ensure gene disruption. We identified two single-cell clones of each line that showed complete deletion of caspase-2. Following exposure to daunorubicin, loss of caspase-2 in NPM1wt cells had little effect on the susceptibility of the cells to apoptosis (Fig. 1E). Strikingly, in NPM1c+ cells, loss of caspase-2 resulted in significantly decreased apoptosis compared to the parental line (Fig. 1E). We noted a similar effect in the bulk cell populations following doxorubicin and daunorubicin treatment, and in independently generated cell lines using a single sgRNA (fig. S1). Apoptosis of NPM1c+ cells was not completely blocked in the absence of caspase-2; rather, it decreased to the level induced in the NPM1wt cells. This result suggests that the increased sensitivity of NPM1c+ cells to apoptosis compared to NPM1wt cells is caspase-2 dependent.
To confirm that the death induced by caspase-2 in NPM1wt and NPM1c+ AML cells was apoptosis, we measured the protein levels of caspase-3. In NPM1wt parental cells, we observed caspase-3 cleavage following daunorubicin treatment. Loss of caspase-2 resulted in a slight increase in the amount of the cleaved p17 caspase-3 subunit after daunorubicin treatment (Fig. 1F). Consistent with the higher rates of apoptosis in NPM1c+ parental cells, caspase-3 cleavage was increased compared to the NPM1wt cells. We also observed the disappearance of the full-length caspase-2 band in NPM1c+ cells with treatment, suggesting caspase-2 cleavage and activation. In the absence of caspase-2, the level of caspase-3 cleavage was decreased, as measured by lower levels of the p17 subunit and higher levels of the full-length caspase. Caspase-2 deficiency in NPM1c+ cells led to a reduction of caspase-3 cleavage to the level of that observed in the NPM1wt cells. These results show that caspase-2–dependent death in NPM1c+ cells is apoptosis.
Caspase-2 is activated in the same cellular compartment as NPM1
We previously reported that caspase-2 is activated in the nucleolus by NPM1-mediated PIDDosome assembly (6). We showed that NPM1 binds to PIDD1 through its central disordered region (6). Since the NPM1c+ mutation is in the C terminus of NPM1, this binding pattern would suggest that NPM1c+ retains the ability to activate caspase-2. Despite this prediction, we showed that caspase-2 cleavage was impaired in NPM1c+ cells treated with the combination of irradiation and a Chk1 inhibitor (6). However, this finding is not consistent with the reduced apoptosis we observed in caspase-2–deficient NPM1c+ cells (Fig. 1E), which suggests a functional role for caspase-2 in these cells. Given that caspase-2 cleavage is not the most reliable indicator of activation (21), we revisited the question of whether NPM1c+ can activate caspase-2 in the cytoplasm. We first determined the subcellular localization of caspase-2 in both cell types. We fractionated NPM1wt and NPM1c+ cells treated with or without daunorubicin into cytoplasmic, nuclear, and nucleolar fractions (Fig. 2A). The purity of the nucleolar fraction was confirmed by probing for the nucleolar protein fibrillarin. In NPM1wt cells, NPM1 was concentrated in the nucleolus and was also detected in the cytoplasm. This finding reflects the known normal nucleus-cytoplasm shuttling activity of NPM1 (22). Caspase-2 was detected in the nucleolus and nucleus of NPM1wt cells. In NPM1c+ cells, NPM1 was concentrated in the cytoplasmic and nuclear fraction as expected, but had minimal presence in the nucleolar fraction. Caspase-2 was also concentrated in the cytoplasmic and nuclear fractions but not in the nucleolus of NPM1c+ cells. This suggests that NPM1 and caspase-2 are localized in the same subcellular fractions in NPM1wt and NPM1c+ AML cells (Fig. 2A). In the cells treated with daunorubicin, we only detected reduction of full-length caspase-2 in the nucleolar fraction for NPM1wt cells and in the cytoplasmic and nuclear fractions for NPM1c+ cells. In the nuclear NPM1c+ fraction, minimal reduction of caspase-2 was observed. This result is consistent with the data shown in Fig. 1F where caspase-2 cleavage appears increased in the NPM1c+ cells, and suggests that caspase-2 is cleaved primarily in the cytoplasm of NPM1c+ cells and the nucleolus of NPM1wt cells. We also ran caspase-2–deficient clones of each cell type. The NPM1c+ ΔC2 clone #3 was a single-cell clone from the same batch of cells from which clones #1 and #2 were derived. This NPM1c+ ΔC2 clone #3 had reduced but not full knockout of caspase-2 expression. Loss or reduction of caspase-2 appeared to disrupt the expression or stability of NPM1 in both cell types (Fig. 2A). This expression pattern is different to what we observed in other cell types (6), but the significance of this is unclear.
Fig. 2. Caspase-2 is activated in the same cellular compartment as NPM1.
(A) OCI-AML-2 (NPM1wt), OCI-AML-3 (NPM1c+) [parental (P)], and CRISPR/Cas9-generated caspase-2–deficient clones (ΔC2) were fractionated into the cytosol, nucleoplasm, and nucleolus and immunoblotted for caspase-2, NPM1, and fibrillarin (nucleolus). (B) Schematic of the caspase-2 BiFC reporter construct (top) and the readout, where recruitment of the C2-Pro VN or C2-Pro VC monomers to the PIDDosome results in induced proximity and refolding of Venus (bottom). (C) NPM1wt and NPM1c+ cells stably expressing C2-Pro VC-2A-C2-Pro VN-2A-mCherry (C2-Pro BiFC) were treated with daunorubicin (0.5 μM) and imaged by confocal microscopy for 16 hours. Confocal time-lapse images show a representative NPM1wt cell (red) with induction of caspase-2 BiFC (yellow) in the nucleolus (dashed outline) and a representative NPM1c+ cell with caspase-2 BiFC induced in the cytoplasm (at the periphery). Scale bar, 5 μm. (D) Quantification of the imaging shown in (C). Percentage of cells per field positive for caspase-2 BiFC in the cytoplasm (C) or the nucleolus (No) at the start of the time lapse (resting) or that became positive during the time lapse (induced) was calculated from 20 fields of view with 10 to 20 cells per field. Error bars represent SEM. ****P < 0.0001 (one-way ANOVA with Tukey’s multiple comparison test). Results are representative of multiple independent experiments.
In the fractionation experiments, we were unable to detect the cleaved version of caspase-2, possibly due to the extensive processing required for this type of fractionation. Because of this complication, and the fact that cleavage is not a direct measure of caspase-2 activation, we used a more specific measure of caspase-2 activation, caspase-2 bimolecular fluorescence complementation (BiFC). BiFC uses nonfluorescent fragments of the yellow fluorescent protein Venus that can associate to reform the fluorescent complex when fused to interacting proteins (23). Caspase-2 is recruited to the PIDDosome via the CARD (caspase recruitment domain) in its prodomain. Upon recruitment to the PIDDosome, caspase-2 undergoes induced proximity, which facilitates caspase-2 dimerization, which is essential for caspase-2 activation (21). For BiFC, we use the prodomain of caspase-2 (C2-Pro) fused to each half of split Venus (VC or VN) so that when caspase-2 is recruited to its activation platform, the subsequent induced proximity of the Venus fragments leads to refolding of the intact fluorescent molecule (Fig. 2B). Thus, Venus fluorescence acts as a readout for caspase-2 induced proximity, the proximal step in its activation (24). We previously described a lentiviral bicistronic construct where the C2 Pro-VC and C2 Pro-VN are expressed in a single vector separated by the viral 2A self-cleaving peptide that is also linked to an mCherry reporter of expression (6). We used this construct to create OCI-AML-2 and OCI-AML-3 cell lines stably expressing the caspase-2 BiFC reporter.
We treated the NPM1wt and NPM1c+ C2-Pro BiFC cells with doxorubicin or daunorubicin. Caspase-2 was activated after treatment, as measured by increased Venus fluorescence (Fig. 2, C and D, and fig. S2). As measured by BiFC, activation of caspase-2 was minimal in NPM1wt cells and high in NPM1c+ cells (Fig. 2D). In these cells, the mCherry fluorescence is excluded from the nucleolus, allowing us to precisely assess nucleolar localization of the BiFC. Using imaging flow cytometry, we observed that, in NPM1wt cells, caspase-2 BiFC was detected in the nucleolus, while caspase-2 BiFC in NPM1c+ cells was only detected outside of the nucleolus at the periphery of the cells (fig. S2A). We confirmed these localizations using time-lapse confocal microscopy (Fig. 2C and movies S1 and S2). Because AML cells are suspension cells, they are difficult to track by imaging. To circumvent this problem, we plated them in a custom-designed microscaffold to keep them restricted to small square-shaped microwells with an area of 50 × 50 μm (fig. S2B). Thus, we were able to track the cells over time by microscopy. Although there was very little induction of caspase-2 BiFC in NPM1wt cells, we only saw nucleolar localization of caspase-2 BiFC in this cell type. An example of an NPM1wt cell with nucleolar caspase-2 activation is shown in Fig. 2C. In contrast, in NPM1c+ cells, we did not detect caspase-2 BiFC in the nucleolus at any time in the cells assessed in the time lapse. Together, these results suggest that caspase-2 is activated in different cellular compartments in a manner that is dictated by the subcellular localization of NPM1. The distinct localization pattern of caspase-2 activation may explain the different caspase-2 dependencies of apoptosis in NPM1wt and NPM1c+ cells. We noted a high level of basal caspase-2 activation in the NPM1c+ cells, suggesting some level of constitutive caspase-2 activation when NPM1 is restricted to the cytoplasm.
Loss of caspase-2 induces cell cycle arrest in NPM1c+ AML cells
As we cultured the cells, we noted that we could not maintain the NPM1c+ caspase-2–deficient cells in culture for longer than 3 to 4 weeks. This observation was surprising because caspase-2–deficient cells are known to proliferate faster (10). To determine if this growth pattern represented a consequence of NPM1c+-dependent caspase-2 activation, we investigated the impact of the loss of caspase-2 on cell cycle progression in NPM1wt and NPM1c+ cells. Using 5-bromo-2′-deoxyuridine (BrdU)/7-amino-actinomycin D (7-AAD) staining and flow cytometry, we profiled each cell type to quantitate the proportion of cells in each phase of the cell cycle (fig. S3A). Parental NPM1wt cells exhibited normal cell cycle progression with the expected horseshoe pattern of BrdU/7-AAD staining showing the three distinct cell cycle phases, G1, S, and G2 (Fig. 3A). Loss of caspase-2 in NPM1wt cells had a minimal impact on the cell cycle profile compared to the parental cells (Fig. 3, A and B). In stark contrast, caspase-2 loss in NPM1c+ cells resulted in a profound increase in cells in G1 phase (Fig. 3, A and B). These data suggest that caspase-2 is required for normal cell cycle progression of NPM1c+ AML cells. To investigate further, we measured cell proliferation over 3 weeks in freshly thawed cells using an MTT [3-(4, 5-dimethylthiazolyl-2)-2, 5-diphenyltetrazolium bromide] assay. While NPM1wt parental, NPM1wt caspase-2–deficient, and NPM1c+ parental cells proliferated as expected (Fig. 3, C and D), caspase-2–deficient NPM1c+ cells showed no growth and a complete decline in viability (Fig. 3D). This result provides strong evidence that caspase-2 is required for the sustained proliferation of NPM1c+ cells.
Fig. 3. Loss of caspase-2 induces growth arrest in NPM1c+ cells.
(A) OCI-AML-2 (NPM1wt) and OCI-AML-3 (NPM1c+) parental or ΔC2 cells were harvested following a 30-min BrdU (10 μM) pulse. The proportion of cells in S, G1, and G2 phase was determined by flow cytometry. Representative flow plots gated on live cells are shown. (B) The percent of cells in each phase of the cell cycle was determined for each cell line. Results are the average of three independent experiments ± SD (two-way ANOVA with Sidak’s multiple comparison test). (C and D) Cells of each genotype were seeded at 1 × 104 cells per well, and cell viability was measured by MTT at the indicated times for NPM1wt parental and ΔC2 cells (C) and NPM1c+ parental and ΔC2 cells (D). Results are the average of three independent experiments ± SD (unpaired t test with Holm-Sidak’s multiple comparison test). (E) THP-1 (NPM1wt) and IMS-M2 (NPM1c+) cells (2 × 105) were electroporated with Cas9RNP with a scramble sgRNA or sgRNAs targeting caspase-2 or RAIDD. Representative bright-field images taken 1 week later are shown. Scale bar, 400 μm. (F) Viable cells were counted 1 week after electroporation of the Cas9 RNPs. Results represent the average cell counts resulting from three independent Cas9RNP electroporations ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 (two-way ANOVA with Sidak’s multiple comparison test). (G) NPM1wt and NPM1wtΔC2 OCI-AML-2 or THP-1 cells transduced with NPM1c+ linked to an mCherry reporter (pRRL.NPM1c+-2A-mCherry) were imaged 2 weeks after antibiotic selection. The viable cells are shown in red. Scale bar, 100 μm.
To confirm our results in a different cell line, we introduced the Cas9RNP targeting caspase-2 into additional AML cell lines. We used THP-1 cells, which are NPM1wt, and IMS-M2 cells, which are NPM1c+. Shortly after the electroporation of the Cas9RNP, THP-1 cells loaded with the caspase-2 sgRNA grew out as expected. In stark contrast, deletion of caspase-2 in IMS-M2 cells prevented the cells from growing out (Fig. 3, E and F). To ensure that this result was not an electroporation artifact leading to cell death, we repeated the Cas9RNP electroporation on two more separate occasions and observed the same results of decreased proliferation in NPM1c+ IMS-M2 cells.
To determine if this proliferative effect of caspase-2 was dependent on the canonical PIDDosome, we used CRISPR/Cas9 to delete the other components of the complex, PIDD1 and RAIDD. We confirmed reduced expression of RAIDD in both NPM1wt (OCI-AML-2) and NPM1c+ (OCI-AML-3) cells. We could only detect reduced expression of PIDD1 in the NPM1wt and not the NPM1c+ cells (fig. S3B). Therefore, we proceeded to focus on characterizing the RAIDD-deficient cells. The absence of RAIDD reduced cell viability in the NPM1c+ cells to a similar degree as caspase-2 loss in IMS-M2 and OCI-AML-3 cells (Fig. 3, E and F, and fig. S3, C and D). The absence of caspase-2 in THP-1 cells led to a significant increase in cell number compared to the CRISPR scramble control, while RAIDD deletion reduced growth in the THP-1 cells but not in the OCI-AML-2 NPM1wt cells (Fig. 3F and fig. S3, C and D). This result may suggest a possible caspase-2–independent function for RAIDD in certain cell types. Together, these results indicate that RAIDD is required for NPM1c+-associated viability of AML cells.
To test if the loss of viability was specific to cells harboring the NPM1c+ mutation, we reconstituted NPM1c+ expression and function in NPM1wt cells. NPM1c+ is always a heterozygous mutation, and upon expression, it hetero-oligomerizes with the wild-type copy of NPM1 and translocates it from the nucleolus to the cytoplasm (25). Hence, we expected that exogenously expressed NPM1c+ would export endogenous NPM1 to the cytoplasm, functionally mimicking NPM1c+ cells. As a proof of principle, we first tested this hypothesis in Hela cells that normally express wild-type NPM1. We used this cell line because, unlike the AML suspension cells, they are much easier to image and identify subcellular compartments. We transiently expressed NPM1 with a TCTG insertion mutation in exon 12 (NPM1c+) in HeLa cells. We immunostained the cells with a C-terminal NPM1 antibody that only detects wild-type NPM1 and not the mutant (fig. S3, E and F). As expected, in control cells, wild-type NPM1 was detected in the nucleolus, forming a ring around the nucleolar protein fibrillarin, and in the nucleus, colocalizing with the transfected nuclear marker H2B mCherry (fig. S3E). In cells transfected with NPM1c+, we detected wild-type NPM1 in the cytoplasm. The NPM1 staining was nonnuclear and overlapped with phalloidin, which stains actin as a marker of the cytoplasm (fig. S3F). This finding indicates that exogenously expressed NPM1c+ induces cytoplasmic translocation of wild-type NPM1. Transient transfection of NPM1c+ into Hela cells induced cell death, which confirms that the NPM1c+ overexpression also functionally converts the cells to an NPM1c+ phenotype. However, this death was not blocked by caspase-2 deficiency and was only partially inhibited by the pan-caspase inhibitor qVD-OPh, which suggests that other forms of cell death than apoptosis were induced by NPM1c+ overexpression in HeLa cells (fig. S3G).
To efficiently convert NPM1wt AML cells to NPM1c+ cells, we expressed NPM1c+ linked to an mCherry reporter via lentiviral transduction in NPM1wt cells. Parental NPM1wt cells expressing NPM1c+ (NPM1wt[c+]) were viable and showed normal proliferation. However, when we expressed NPM1c+ in NPM1wt caspase-2–deficient OCI-AML-2 o r THP-1 cells (NPM1wt[c+] ΔC2), we noted a rapid loss of viability and we could not establish a stable line (Fig. 3G). These results demonstrate that the loss of viability in cells lacking caspase-2 is due to the presence of NPM1c+ and not any other genetic difference between the NPM1wt and NPM1c+ cell lines. This result confirms that NPM1c+ cells require caspase-2 for growth and proliferation.
Loss of caspase-2 induces NPM1c+-dependent terminal differentiation
Upon depletion from the cytoplasm of AML cells, NPM1c+ has been shown to induce cell cycle arrest in G1 and terminal differentiation (20). Given that deletion of caspase-2 in NPM1c+ cells induces extended G1 arrest (Fig. 3B), we investigated if these cells undergo differentiation in the absence of caspase-2. We measured the expression of the cell surface markers CD34, to measure immature hematopoietic precursors, and CD14, to measure myeloid differentiation (fig. S4A). CD34 expression was minimal in the NPM1wt and NPM1c+ cells, as previously reported (26, 27), and reduced slightly in the absence of caspase-2 (fig. S4B). In contrast, CD14 surface expression increased more than 10-fold in caspase-2–deficient NPM1c+ cells (Fig. 4, A and B). These data indicate that loss of caspase-2 induces differentiation of NPM1c+ cells into a macrophage-like state and that these cells can no longer proliferate.
Fig. 4. Loss of caspase-2 induces NPM1c+-dependent differentiation.
(A) Cell surface expression of CD34 and CD14 was measured in OCI-AML-2 (NPM1wt) and OCI-AML-3 (NPM1c+) parental or ΔC2 cells cultured for 2 weeks by flow cytometry. Representative flow plots are shown. (B) Percentage of cells expressing CD14. Results are the average of three independent experiments ± SD. ***P < 0.001, ****P < 0.0001 (two-way ANOVA with Sidak’s multiple comparison test). (C) NPM1wtΔC2 or NPM1wtΔC2[c+] cells expressing a Tet-repressible caspase-2 (TetOFF-C2WT) or a Tet-repressible catalytically inactive caspase-2 (TetOFF-C2C/A) were treated with doxycycline (2 μg/ml) for 16 hours to repress caspase-2 expression. Lysates were immunoblotted for caspase-2 and actin as a loading control. (D) NPM1wtΔC2-TetOFF-C2WT or NPM1wtΔC2[c+]-TetOFF-C2WT cells were grown in DMSO (C2ON) or doxycycline (2 μg/ml, C2OFF) for 3 weeks. Cell surface CD14 and CD34 expression was measured once a week by flow cytometry. Results represent the average of three independent experiments ± SD. **P < 0.01, ****P < 0.0001 (two-way ANOVA with Sidak’s multiple comparison test). (E) Representative flow plots from (D) are shown (see also fig. S4D). (F) NPM1wtΔC2 or NPM1wtΔC2[c+] cells expressing a TetOFF-C2WT or TetOFF-C2C/A were grown in DMSO or doxycycline (2 μg/ml) for 1 week. Cell surface expression of CD14 was measured by flow cytometry. Results are the average of three independent experiments ± SD. ***P < 0.001 (two-way ANOVA with Sidak’s multiple comparison test). (G) IMS-M2 cells were incubated with the indicated selective caspase-2 pentapeptide inhibitors (20 μM). Cell surface expression of CD14 was measured by flow cytometry 24 hours later. Results are the average of three independent experiments ± SD. *P < 0.05, **P < 0.01 (one-way ANOVA with Dunnett’s multiple comparison test).
To demonstrate that caspase-2 is required for this process and no other genetic difference between the cell lines, we reintroduced caspase-2. We used a TET-OFF–based expression system where we could both reconstitute caspase-2 expression and repress caspase-2 transcription in the same cells. We also introduced a catalytically inactive version of caspase-2 where we mutated the catalytic cysteine to an alanine in the TET-repressible construct. Because of the reduced viability of caspase-2–deficient NPM1c+ cells, they did not survive long enough to establish a stable cell line. To circumvent this problem, we stably expressed the repressible caspase-2 constructs in NPM1wt caspase-2–deficient cells (NPM1wt ΔC2) and then converted them to NPM1c+ by expressing exogenous NPM1c+ as shown in Fig. 3G. The addition of doxycycline to the NPM1wtΔC2 [c+]TETOFFC2WT and NPM1wtΔC2 [c+]TETOFFC2C/A cells efficiently inhibited expression of caspase-2 (Fig. 4C and fig. S3C). We assessed the repressible caspase-2 cells for CD34 and CD14 surface expression under caspase-2 ON (grown in the absence of doxycycline) and caspase-2 OFF (grown in the presence of doxycycline) conditions over a 3-week period. As expected, neither restoration of caspase-2 (C2ON) nor depletion of caspase-2 after doxycycline treatment (C2OFF) affected the percentage of CD14-positive cells in the NPM1wt lines (Fig. 4D and fig. S3D).
In NPM1wt[c+] cells with cytoplasmic expression of NPM1, NPM1wt[c+]C2TET cells grown in dimethyl sulfoxide (DMSO) (C2ON) did not differentiate to express CD14 over the 3 weeks (Fig. 4, D and E). This result confirms that the reintroduction of caspase-2 to the cells restores them to the NPM1c+ parental phenotype. Depletion of caspase-2 with doxycycline (C2OFF) in the NPM1wt[c+]C2TET cells resulted in a time-dependent increase in CD14 surface expression, peaking at week 3 (Fig. 4, D and E). Unlike the NPM1wt[c+]C2TET cells, caspase-2(C/A) was unable to suppress differentiation in the NPM1wt[c+]ΔC2 cells. The inability of the catalytically inactive version of caspase-2 to inhibit differentiation like the wild-type caspase-2 indicates that this function is due to the catalytic activity of caspase-2. Following doxycycline treatment to reduce expression of caspase-2(C/A), the level of differentiation was unchanged, indicating that caspase-2(C/A) expression phenocopies loss of caspase-2 (Fig. 4F). Together, these results confirm that caspase-2 is required to sustain cellular viability, at least in part, by preventing the monocytic differentiation of NPM1c+ cells and this function requires its catalytic activity.
The commercially available caspase-2 inhibitor VDVAD.fmk has overlapping specificities and also inhibits caspase-3 (28). Therefore, to test if we could use a chemical inhibitor to induce differentiation of NPM1c+ cells, we used NH-23-C2. NH-23-C2 inhibits caspase-2 with high efficiency without inhibiting caspase-3 or caspase-8 (29). We tested the original published version of the compound NH-Idc-hGlu-Thr(Bzl)-Ser-Asp-AOMK (C2i1) along with three derivatives: NH-Idc-hGlu-Thr(Bzl)-Ser-Asp(Me)-AOMK (C2i2), which has a methylated Asp at the P1 position to increase probe cellular permeability; NH-Idc-hGlu(tBu)-Thr(Bzl)-Ser(tBu)-Asp-AOMK (C2i3), where the hGlu in the P4 position and the Ser in the P2 position are protected with a tert-Butyl group, tBu; and NH-Idc-hGlu(tBu)-Thr(Bzl)-Ser(tBu)-Asp(Me)-AOMK (C2i4), which has the protected hGlu, Ser, and a methylated Asp (fig. S5A). We tested each of these four inhibitors for their ability to block caspase-2 cleavage in our panel of AML cell lines. After treatment with the Aurora B kinase inhibitor ZM447439 (ZM), a strong inducer of caspase-2 activation, we could only detect robust cleavage of caspase-2 in IMS-M2 cells, which are NPM1c+ (fig. S5B). This cleavage was effectively blocked with three of four of the inhibitors, with only the methylated version of the original compound not working. Because caspase-2 cleavage is not the most faithful readout of caspase activation, we also probed for cleavage of the known caspase-2 substrates, MDM2 (mouse and human double minute 2) and BID (BH3-interacting domain death agonist). BID cleavage was observed in IMS-M2 cells as a disappearance of the full-length protein. Full-length BID was restored in cells treated with C2i1, C2i3, and C2i4 to similar efficiencies as the cleavage of caspase-2. ZM only induced MDM2 cleavage in OCI-AML-2 (NPM1c+) cells. MDM2 cleavage was measured by the appearance of a 60-kDa fragment. This cleavage was blocked by all of the inhibitors. There was some background cleavage of BID in OCI-AML-3 cells and of MDM2 in IMS-M2 cells, but this was not further increased by ZM treatment. On the basis of these results, we used IMS-M2 cells for further investigation with the inhibitors. We then treated the IMS-M2 cells with each of the three inhibitors and measured differentiation. After 24 hours, we noted a significant increase in differentiation of the IMS-M2 cells with C2i1 and C2i3 and a trend toward increased differentiation with C2i4 (Fig. 4G). Together, these results show that inhibition of caspase-2 induces differentiation of NPM1c+ AML cells.
Caspase-2 loss down-regulates genes regulating stem cell pluripotency
To understand further the molecular mechanisms by which caspase-2 sustains NPM1c+ cell proliferation and prevents differentiation, we performed transcriptomic analysis. We carried out total RNA sequencing (RNAseq) on parental NPM1wt, NPM1c+ cells, and their respective caspase-2–deficient counterparts (Fig. 5A). By comparing the differentially expressed genes (DEGs) with greater than fivefold expression changes across the genotypes, we noted that caspase-2 deficiency had an extensive impact on the transcriptomic profile of NPM1c+ cells but not NPM1wt cells (Fig. 5B). Given this large difference in expression between NPM1c+ cells with and without caspase-2, we focused on the differences between these cell lines. We used gene enrichment analysis to first determine cell types associated with the changes in gene expression after loss of caspase-2 in NPM1c+ cells. We found that the changes in expression profile were associated with an increased identity with CD14+ monocytes according to the Human Gene Atlas (fig. S6A). There were significant increases in the macrophage markers CD14, CD86, and CYBB (fig. S6B). These results confirm our observations that loss of caspase-2 induces monocytic differentiation of NPM1c+ cells. Next, we carried out KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway analysis. We found that the most enriched pathways between NPM1c+ parental and NPM1c+ ΔC2 cells were “Pathways in Cancer,” “Hippo Signaling Pathway,” and “Signaling Pathways Regulating Pluripotency of Stem Cells” (Fig. 5C). The most significantly down-regulated group was Signaling Pathways Regulating Pluripotency of Stem Cells (Table 1). In particular, the FGF (fibroblast growth factor) pathway (FGF2, FGFR, AKT3), the IGFR (insulin growth factor receptor) pathway (IGF1R, AKT3), the Wnt (wingless-related integration site) pathway (WNT3, WNT5B, WNT7B, WNT11, FZD1 to FZD4, TCF7), the LIF (leukemia initiating factor) pathway (LIF, LIFR, JAK3), and one of the core members of the self-renewal transcriptional network, OCT4 ([organic cation transporter 4], POU5F1B) were all down-regulated in the absence of caspase-2 in NPM1c+ cells (Fig. 5D). Together, these data provide strong evidence that caspase-2 is required to maintain NPM1c+ AML self-renewal.
Fig. 5. Loss of caspase-2 in NPM1c+ cells down-regulates pathways regulating pluripotency.
(A) Experimental design for total RNA sequencing and analysis of DEGs from parental OCI-AML-2 (NPM1wt) and OCI-AML-3 (NPM1c+) and their respective caspase-2–deficient (ΔC2) cells. (B) DEGs between the parental NPM1wt and NPM1c+ and their respective ΔC2 transcriptomes are shown by heatmap. Genes with more than log5 fold changes are shown. Decreased expression is shown in blue, and increased expression is shown in red. (C) Total RNAseq data were analyzed with ENRICHR to identify genes involved in the indicated KEGG term biological processes that are most significantly differentially regulated in NPM1c+ parental and ΔC2 cells. The size of the dots represents the number of genes associated with each biological pathway. The color of the dots represents the significance (adjusted P value). (D) Mean log2 fold change of significant DEGs between NPM1c+ parental and ΔC2 cells involved in the KEGG term pathways that regulate pluripotency of stem cells [the most significantly down-regulated pathway (see Table 1)]. Gene names in bold are referred to in the text.
Table 1. Most significantly up- and down-regulated pathways in NPM1c+ cells in the absence of caspase-2.
| Term | Overlap | P | Adjusted P | Odds ratio | Combined score | |
|---|---|---|---|---|---|---|
| Down- regulated | Signaling pathways regulating pluripotency of stem cells | 21/143 | 5.93 × 10−6 | 0.001671943 | 3.414446184 | 41.09516759 |
| Hippo signaling pathway | 21/163 | 4.5 × 10−5 | 0.006338397 | 2.930434654 | 29.33332658 | |
| Pathways in cancer | 46/531 | 0.000117 | 0.008361907 | 1.894965339 | 17.15021715 | |
| Human papillomavirus infection | 32/331 | 0.000183 | 0.008361907 | 2.127629963 | 18.31472613 | |
| Cushing syndrome | 19/155 | 0.000198 | 0.008361907 | 2.763396982 | 23.5675838 | |
| Hepatocellular carcinoma | 20/168 | 0.000203 | 0.008361907 | 2.674089184 | 22.74221776 | |
| Up-regulated | Malaria | 10/56 | 9.92624 × 10−6 | 0.002729716 | 6.745091164 | 77.70566909 |
| Pathways in cancer | 39/531 | 2.41339 × 10−5 | 0.003318406 | 2.176980816 | 23.14543067 | |
| Cytokine-cytokine receptor interaction | 25/295 | 7.58836 × 10−5 | 0.006956 | 2.521357317 | 23.91837579 | |
| Viral protein interaction with cytokine and cytokine receptor | 12/100 | 0.000259819 | 0.017862524 | 3.680283851 | 30.38268317 | |
| Toll-like receptor signaling pathway | 12/104 | 0.000375033 | 0.019012637 | 3.5195377 | 27.76386473 |
We next confirmed that these changes in gene expression resulted in changes in protein levels. We did so in an unbiased manner by using reverse phase protein array (RPPA) analysis of the NPM1c+ parental and caspase-2–deficient cells. Of the protein set tested, we found that most proteins with significant differences between the two cell lines (q < 0.01) had decreased abundance in the absence of caspase-2 (Fig. 6A). Many of the tested proteins with decreased abundance in NPM1c+ caspase-2–deficient cells are known to be required for or support hematopoiesis (Fig. 6A). These include a number of epigenetic modifiers like JMJD2A (jumonji domain-containing protein), JMJD2B, and LSD1 (lysine-specific histone demethylase 1A), which have all been shown to be required for stem cell maintenance (30, 31) (Table 2). Of the proteins that were significantly changed, only one, HIF2A (hypoxia-inducible factor), was increased in the absence of caspase-2. In the largest group, 15 of the most significantly down-regulated proteins are involved in the AKT/mTORC1 (a serine/threonine protein kinase/mammalian target of rapamycin complex) pathway, and 3 are involved in the Wnt pathway, including β-catenin, the downstream effector of Wnt signaling (Table 2). These findings are consistent with our RNAseq data that showed down-regulation of pathways that converge on AKT and Wnt signaling (Fig. 5D).
Fig. 6. Caspase-2 regulates the mTORC and WNT pathways in NPM1c+ cells.
(A) RPPA analysis was carried out on OCI-AML-2 (NPM1c+) parental and ΔC2 cells. Signal intensities were normalized and filtered as described in Materials and Methods (see data file S1). The heatmap shows z scores of three biological replicates of parental NPM1c+ cells and each ΔC2 clone with q > 0.01. Up-regulated proteins are shown in red, and down-regulated proteins are shown in blue. Proteins in the mTORC1/AKT pathway are labeled in blue text, and proteins in the Wnt signaling pathway are labeled in green text. (B) OCI-AML-2 (NPM1wt) and OCI-AML-3 (NPM1c+) parental or ΔC2 cells were left untreated or treated with IL-2 for 16 hours. Lysates were harvested and immunoblotted for phosphorylated AKT (p-AKT) and total AKT. Actin was used as the loading control. The spacer lane was omitted from the p-AKT blot. The red line serves to align the remaining lanes. (C to E) The indicated cell lines were treated with or without IGF-1 (10 ng/ml) (C to E) or Wnt (5.0 nM) (E) for 16 hours. Lysates were harvested and immunoblotted for the indicated proteins with actin as a loading control. (F and G) Unstimulated OCI-AML-2 (NPM1wt) and OCI-AML-3 (NPM1c+) parental or ΔC2 cells were immunoblotted for the indicated proteins with actin or GAPDH used as a loading control. Experiments shown in (B) to (G) are each representative of two to three independent experiments.
Table 2. Significantly down-regulated proteins in NPM1c+ cells in the absence of caspase-2.
| Pathway | Proteins | Total |
|---|---|---|
| mTORC1 pathway | FoxO1, BRaf, p70S6K, p-Raptor(S792), IKKb, p-Erk1/2(T202/Y204), p-NF-kB p65 (Ser536), p-IKK a/b (Ser176/180), p-PRAS40(S183), Raptor, p-p38(T180/Y182), PTEN, PI3Kp85, p-ALK(Y1586), p-Bad(S136) | 15 |
| Histone modification | JMJD2A, JMJD2B, JARID1C, SET1A, LSD1, PHF8, HDAC6, SRC-1, HDAC4, p-HP1g (Ser83) | 10 |
| DNA repair | CHAF1A, BRCA1, SirT6, DNMT1 | 4 |
| Metabolism | LDHA, ALDH2, PKM2, AOX1, PFKFB3 (C-terminal), SCD1_R_V | 6 |
| Wnt pathway | ASH2, p-β-catenin(S33/37/T41), LRP6 | 3 |
| Transcription regulators | p-TRAP220/MED1(T1457), MED12, Brg1, Zeb1, NCOA2, p-c-Fos(S32), c-Fos, Twist, STAT2, p-Stat1(Y701), Stat5a | 11 |
| Cell cycle | Cyclin C, p-ATM(S1981), AuroraA/AIK, p-Chk2(S33/35), S100A7 | 5 |
| Growth factor pathways | VEGFR2, p-Smad2(S465/467), Notch1 | 3 |
| Other | AR-441, Lipocalin-1, CD24 | 4 |
The FGFR, IGF1R, and LIFR pathways were all down-regulated in the RNAseq dataset in the absence of caspase-2. These pathways all converge on activation of the phosphatidylinositol 3-kinase (PI3K) pathway, resulting in phosphorylation of AKT. Because the AKT pathway featured heavily in the RPPA results, we tested the impact of the loss of caspase-2 on AKT activation. We measured the effects of the loss of caspase-2 on general AKT activation using IL-2 (interleukin-2) as a stimulus. IL-2 induced phosphorylation of AKT in both the NPM1wt and NPM1c+ parental cells. Loss of caspase-2 had minimal effect on the amount of AKT phosphorylation in NPM1wt cells. In contrast, AKT phosphorylation was completely inhibited in the NPM1c+ caspase-2–deficient cells with and without IL-2 treatment (Fig. 6B). Next, we treated the cells with human IGF-1 recombinant protein. Following IGF-1 treatment, we observed slightly higher phosphorylation of AKT in the NPM1wt and NPM1c+ parental lines compared to the untreated cells but none in the caspase-2–deficient NPM1c+ cells with or without IGF-1 treatment (Fig. 6C). Total AKT levels were similar in all cell types. Therefore, the lower phosphorylation of AKT in caspase-2–deficient NPM1c+ cells cannot be explained just by lower AKT levels. We also noted an increase in the basal level of AKT phosphorylation in caspase-2–deficient NPM1wt cells (Fig. 6, B and C). These changes in AKT phosphorylation suggest that caspase-2 has opposite effects on the NPM1wt cells compared to the NPM1c+ cells. These data confirm that caspase-2 loss impairs the AKT pathway, specifically in NPM1c+ cells. In the phosphoinositide 3-kinase pathway, PTEN (phosphatase and tensin homolog) inhibits the phosphorylation of AKT and PTEN phosphorylation helps to maintain PTEN in an inactive state. Therefore, we measured PTEN phosphorylation under the same conditions. Phosphorylated PTEN was not changed under any of the conditions measured (Fig. 6D). Thus, PTEN is not aberrantly activated in caspase-2–deficient cells and is likely not responsible for the inhibition of AKT phosphorylation.
Another down-regulated pathway that was common between the RNAseq and RPPA data was the Wnt signaling pathway. Therefore, we confirmed the effect of loss of caspase-2 on Wnt signaling. The downstream effector of the canonical Wnt pathway is β-catenin. Wnt induces enhanced transcriptional activation and stabilization of β-catenin (32), and this pathway can also be indirectly activated by IGF-1 (33). We treated the AML cells with IGF-1 or with the Wnt surrogate–Fc fusion recombinant protein that stabilizes β-catenin. In untreated cells, we detected β-catenin in NPM1wt cells with and without caspase-2, and in NPM1c+ parental cells, but it was absent in the NPM1c+ ΔC2 cells. (Fig. 6E). Following treatment with IGF1 or Wnt, β-catenin was stabilized in parental NPM1wt and NPM1c+ parental cells but was not present in the absence of caspase-2. This indicates that the expression or stabilization of β-catenin is impaired in the absence of caspase-2 in both NPM1wt and NPM1c+ AML. This is consistent with the reduced expression of the upstream Frizzled receptors in the Wnt signaling pathway in NPM1c+ caspase-2–deficient AML cells (Fig. 5D).
Caspase-2 cleaves Rictor in NPM1c+ cells
In Fig. 4F, we showed that the catalytic activity of caspase-2 is required for it to repress differentiation. This result indicates that this function of caspase-2 requires cleavage of a substrate or substrates. To test this idea, we measured cleavage of known caspase-2 substrates in NPM1wt and NPM1c+ AML cells. Given that AKT activation was measured in our cell lines at baseline, we assessed substrate cleavage in untreated cells that had been in culture for around 2 weeks. We did not detect any cleavage of BID in any of the cell lines. It has been reported that the mTORC2 complex component Rictor is a caspase-2 substrate in neurons (34). Given that we measured down-regulation of the mTORC1 pathway in our RPPA data (Table 2), we decided to immunoblot for Rictor. In NPM1c+ cells, we noted a clear cleavage of Rictor producing two fragments at around 75 and 35 kDa (Fig. 6F) that was absent in caspase-2–deficient NPM1c+ cells.
In the previous report of Rictor cleavage by caspase-2, cleavage was suggested to block AKT activation (34). Their conclusion is opposite to our results where we see AKT phosphorylation in the same conditions where Rictor is cleaved. In the previous report, a direct link between caspase-2–mediated Rictor cleavage and AKT activation was not directly shown, and Rictor cleavage was only induced by overexpressed caspase-2. The antibody we used to detect Rictor was raised against an epitope around Lys1125 in the C terminus of human Rictor. Therefore, the 35-kDa cleaved product contains the C terminus of the protein. This region also contains Thr1135, which when phosphorylated mediates binding to 14-3-3 and inhibits AKT phosphorylation at Ser473 (35). We immunoblotted with an antibody that detects the phosphorylated Thr1135 site. Phosphorylation of this residue was intact in NPM1wt cells with and without caspase-2 (Fig. 6F). In the parental NPM1c+ cells, we did not detect full-length phosphorylated Rictor but we did detect it on the 35-kDa cleavage fragment (Fig. 6F). Phosphorylation of the Thr1135 residue was intact in caspase-2–deficient NPM1c+ cells. This confirms that caspase-2–dependent cleavage of Rictor in NPM1c+ cells removes the phosphorylation domain from Rictor, resulting in a protein that is constitutively able to phosphorylate AKT.
Next, we measured activation of two other downstream effectors of Rictor: PKCα and NDRG1. Phosphorylation of PKCα occurs constitutively, which is required for stability of the protein (36). PKCα was down-regulated in the parental NPM1c+ cells and present in the NPM1wt cells with and without caspase-2 as well as the caspase-2–deficient NPM1c+ cells (Fig. 6G). NDRG1 is a highly specific substrate of SGK1, which is a target of mTORC2 (37). Similarly, we noted decreased phosphorylation of NDRG1 in the NPM1c+ cells expressing caspase-2 that was restored in the absence of caspase-2 (Fig. 6G). It was previously shown that disruption of phosphorylation of Rictor on Thr1135 only inhibits AKT Ser473 phosphorylation and not other Rictor targets (35). Thus, together, our data suggest that when Rictor is cleaved by caspase-2 its activity is both impaired for certain targets and enhanced for AKT phosphorylation.
DISCUSSION
NPM1c+ is the most frequent genetic alteration in AML and, as such, represents a potential therapeutic target (2). Despite this importance, how this driver mutation induces and maintains the leukemic state is not clear. The data presented here demonstrate that the cellular effects and signaling mechanisms of NPM1c+ converge on the proapoptotic protein caspase-2. We show that caspase-2 is a crucial link between NPM1c+ and changes in oncogene signature, without which the cells undergo substantial differentiation and lose their proliferative and self-renewing capacity. Moreover, our results show that caspase-2 activation has vastly different functional outcomes depending on the specific upstream signals.
Building on our previous discovery that NPM1 is an upstream activator of caspase-2 in the nucleolus (6), we show here that NPM1c+ has a similar ability to activate caspase-2 in the cytoplasm. This leads to vastly different functional consequences than would be expected. Instead of enhancing proliferation as has been previously reported (9, 10), loss of caspase-2 in NPM1c+ cells induced profound cell cycle arrest and what appears to be terminal differentiation. These findings provide strong evidence that caspase-2 is essential for survival of NPM1c+ AML cells. The marked impact on the stem cell transcriptional signature converging mainly on the AKT/mTORC1 and Wnt signaling pathways indicates that the essential role for caspase-2 in survival is not solely due to a direct regulation of the cell cycle but rather that caspase-2 plays an integral role in maintaining the self-renewing capacity of NPM1c+ AML cells.
The few studies that have investigated caspase-2 in hematopoiesis have shown that it serves to limit this process. For example, in aged mice, the absence of caspase-2 increased the number of myeloid progenitor cells and impaired the differentiation of hematopoietic stem cells (16). In addition, hematopoietic cell lineage pathways were found to be significantly up-regulated in Casp2−/− Eμ-Myc lymphoma compared to Casp2+/+ Eμ-Myc (38). When NPM1c+ is expressed, however, we found that differentiation was increased, and stem cell pathways were down-regulated in the absence of caspase-2. These results suggest that, instead of leading to inhibition of hematopoiesis, caspase-2 activation by NPM1 outside of the nucleolus leads to the promotion of pluripotency. This opposite action of NPM1-induced caspase-2 activation in the cytoplasm versus the nucleolus is supported by the small changes we see in the NPM1wt cells. For example, there is a small increase in the amount of DNA damage–induced caspase-3 cleavage in NPM1wt caspase-2–deficient cells, although this does not result in a short-term impact on apoptosis. We also observed an increase in basal AKT activation in NPM1wt caspase-2–deficient cells and increased proliferation of NPM1wt cells deficient in caspase-2. Our results support a model where caspase-2 has opposing functions depending on where it is activated in the cell.
How caspase-2 both induces apoptosis and proliferation in the same cell type is somewhat of a conundrum. The key element here appears to be the stimulus. The caspase-2 knockout NPM1c+ cells show short-term resistance to apoptosis when challenged with a DNA-damaging agent but, in the absence of an external stimulus, will arrest, differentiate, and eventually lose viability. Apoptosis is not completely inhibited in the absence of caspase-2 in NPM1c+ cells but is rather reduced to the same level as that observed in the NPM1wt cells. It is possible that the increased sensitivity to proapoptotic stimuli is a bystander effect of the enhanced self-renewal capacity of the NPM1c+ cells. Pluripotent stem cells have been shown to have increased sensitivity to apoptosis induced by genotoxic stimuli (39). This mechanism has been proposed to be a way to limit the accumulation of DNA damage and genomic instability (40). Therefore, it is possible that caspase-2 itself is not directly inducing apoptosis, but the enhanced sensitivity to cell death is an indirect consequence of the up-regulation pathways involved in pluripotency.
Mutation of NPM1 in AML is associated with better outcomes, and it is thought that this is due to enhanced chemosensitivity. Because we show that chemotherapy-induced apoptosis is caspase-2 dependent, it is tempting to speculate that caspase-2 is a major contributor to the enhanced chemosensitivity of these cells. This poses a problem when thinking how to adapt our findings to a clinically relevant therapeutic approach. Our data show that inhibiting caspase-2 induces cell cycle arrest and differentiation. However, a side effect may be increased resistance to apoptosis. In future work, it will be essential to determine how these dual functions of caspase-2 are balanced and how we can inhibit its pro-proliferative function without compromising the increased sensitivity of the NPM1c+ cancer cells.
Because caspase-2 is a protease, it is likely that the compartmentalization of caspase-2 activation into different regions of the cell provides access to different substrates. It has been notoriously difficult to identify bona fide caspase-2 substrates (7). Although a large-scale caspase substrate screen has been published identifying a long list of potential caspase-2 substrates (41), the only two that have been shown to have clear physiological consequences when cleaved by caspase-2 are BID and the p53-negative regulator MDM2 (42). More recently, Xu et al. (34) proposed Rictor as a caspase-2 substrate in neurons. Here, we confirm that Rictor is cleaved in a caspase-2–dependent manner. Rictor is an essential gene with known roles in regulating cell proliferation and the actin cytoskeleton, and is an important upstream regulator of mTORC1 via AKT activation (43, 44). As a component of the mTORC2 signaling complex, Rictor phosphorylates AKT on serine-473 (45). Additional Rictor substrates include PKCα and SGK1 (36, 37). Our data suggest that Rictor cleavage impairs its ability to activate PKCα and SGK1 but enhances its ability to phosphorylate AKT. Rictor is composed of an N-terminal ARM domain (AD), a central HEAT-like domain (HD) followed by a disordered region, and a C-terminal domain (CD) (37). We predict that cleavage of Rictor by caspase-2 removes the disordered domain and CD. The disordered domain contains several phosphorylation sites, one of which, Thr1135, has been shown to mediate 14-3-3 binding to Rictor (35). This binding is proposed to impair AKT S473 phosphorylation while leaving phosphorylation of other Rictor targets intact. Other structural studies have shown that the Rictor binding protein SIN1 changes its confirmation and binding contacts when Rictor binds SGK1, forming a salt bridge between SIN1 and Rictor/Asp1679 (46). This conformational change is not induced when Rictor complexes with AKT. Since this region of Rictor is also removed when Rictor is cleaved, this may further explain why SGK1 activation, as measured by NDG2 phosphorylation, is impaired, while AKT phosphorylation is not. The full functions and mechanisms of mTORC2 are not fully understood, and therefore, it is not clear how caspase-2 could lead to simultaneous activation and inhibition of its substrates.
The role of AKT in modulating the expression of core stem cell regulators, OCT4, SOX2, and Nanog, is described in multiple studies (47–49). Such studies advocate that activation of AKT signaling is sufficient to maintain stem cell pluripotency (50). AKT has been shown to phosphorylate OCT4 in embryonal carcinoma cells (ECCs), enhancing its ability to bind to the promotor regions of several stemness genes including POU5F1, the gene that encodes OCT4 (51). POU5F1 was down-regulated in our NPM1c+ caspase-2–deficient cells. In addition, constitutive AKT activation transforms hematopoietic stem cells into myeloproliferative diseases, including AML (52). AML cells often have hyperactivated AKT, a prerequisite to sustain the oncogenic potential of leukemia stem cells and prevent differentiation (53, 54). Consistent with our data, NPM1wt and NPM1c+ have opposing effects on AKT signaling. It has been shown that NPM1 can negatively regulate AKT and NPM1c+ antagonizes NPM1-mediated suppression of AKT signaling (55). Thus, our data support a model where caspase-2 cleaves Rictor to induce AKT activation and downstream enhancement of OCT4-mediated transcription. This pathway may be central to the transcriptional changes we see in NPM1c+ cells deficient in caspase-2.
It has been proposed that NPM1c+-mediated expression of HOX genes is required for the maintenance of the leukemic state of the AML cells (20). We did not observe large differences in HOX gene expression in NPM1c+ cells with or without caspase-2. The C terminus of NPM1 is composed of a cluster of helices (Helix 1 to 3) that bind both duplex and single-stranded DNA with no sequence specificity (56). When mutated, Helix 1 becomes denatured and unstructured, which affects the ability of NPM1 to bind DNA and protein (56). NPM1c+ is recruited to the HOX gene cluster by CRM1 via binding to the newly formed NES in its sequence (57). Disruption of the NES or depletion of NPM1c+ from the cytoplasm disrupts this interaction (20). Because NPM1c+ is thought to directly affect HOX expression, it is perhaps not surprising that it is not disrupted in caspase-2–deficient cells. Our data suggest that caspase-2 is another downstream effector of NPM1c+ that helps maintain pluripotency. The relationship between AKT and HOX function is not clear, but it has been suggested that the HOXD gene cluster induces AKT activation (58). Therefore, it is possible that caspase-2 functions in synergy with HOX genes to engage the AKT/mTORC1 pathway.
It is unlikely that AKT is the only effector of NPM1c+-mediated caspase-2 activation. Caspase-2 also engages the WNT signaling pathway in NPM1c+ cells, as evidenced by reduced expression of the Frizzled receptors and reduced β-catenin levels. Among the differentially expressed Wnt signaling genes between NPM1c+ parental and caspase-2–deficient cells, WNT5b is up-regulated in the absence of caspase-2. This ligand is antagonistic to the canonical WNT signaling pathway (59), fitting with the overall down-regulation of this pathway in the absence of caspase-2. β-Catenin is a transcriptional cofactor that regulates the transcription of numerous genes including many Wnt signaling genes (60). This may be another route of transcriptional regulation by caspase-2. There is considerable cross-talk between the AKT and Wnt pathways. Expression of Wnt increases AKT activity, and AKT can promote β-catenin nuclear localization via the regulation of glycogen synthase kinase 3β (GSK3β) phosphorylation (61).
A major outstanding question of this work is how NPM1c+ engages caspase-2 in the cytoplasm. We previously showed that NPM1 provides a scaffold for PIDDosome formation in the nucleolus, while caspase-2 can also be activated in a RAIDD-dependent, PIDD1-independent manner in the cytoplasm (6). Here, we show that viability of NPM1c+ cells is dependent on RAIDD, but we have yet to demonstrate whether PIDD1 is required. It is possible that NPM1c+ facilitates the formation of the PIDDosome in the cytoplasm, but based on the work shown here, we cannot rule out a PIDD1-independent complex being formed to activate caspase-2 in the cytoplasm. More work is needed to fully determine the upstream mechanisms of NPM1c+-mediated caspase-2 activation. We noted that in certain cell types, RAIDD knockout also led to loss of viability in certain NPM1 wild-type cell lines. While the significance of this is currently unclear, this could suggest that RAIDD can induce similar functions in NPM1wt cells in the absence of caspase-2 through recruitment of another caspase or similar protease. This could provide a possible therapeutic avenue for NPM1wt AML.
The high rate of relapse and associated mortalities in AML underscores the need to develop targeted therapies. In the absence of additional mutations, NPM1c+ AML is associated with high rates of complete molecular remission following standard induction therapy (62). However, when combined with one or more mutations like FLT3-ITD, DNMT3a, or IDH1, prognosis is worse (63), and overall relapse-free survival is the same as in patients with NPM1wt AML (64). The self-renewal properties of AML cells provide a potential reservoir for relapse. Despite our studies suggesting that the AKT pathway is a major convergence point in NPM1c+ AML pluripotency, clinical studies with AKT pathway inhibitors as monotherapy for NPM1c+ AML showed insufficient anti-leukemic activity (65). The poor efficacy of AKT inhibitors for leukemia treatment underscores the importance of understanding the full extent of the pathways engaged by NPM1c+ and caspase-2. Our work here holds the promise that targeting caspase-2 in NPM1c+ AML could be an effective method to treat the disease and prevent relapse.
MATERIALS AND METHODS
Chemicals and antibodies
The following antibodies were used: anti–caspase-2 (clone 11B4; EMD Millipore), anti–caspase-3 (9662; Cell Signaling Technology), anti–caspase-3 cleaved (9661; Cell Signaling Technology), anti-actin (clone 4; Fisher Scientific), anti-NPM1 (clone FC82291; Sigma-Aldrich), anti-fibrillarin (C13C3; Cell Signaling Technology), anti–phospho-AKT (4060; Cell Signaling Technology), anti-AKT (4691; Cell Signaling Technology), anti–phospho-PTEN (9551, Cell Signaling Technology), anti–β-catenin (8814; Cell Signaling Technology), anti-Rictor (2114; Cell Signaling Technology), anti–phospho-Rictor (Thr1135) (3806; Cell Signaling Technology), anti-MDM2 (sc-812; Santa Cruz Biotechnology), anti-BID (Alexa Fluor 860; R&D Systems), anti-PKCα (2056; Cell Signaling Technology), anti–phospho-NDRG1(Thr346) (5482; Cell Signaling Technology), anti-NDRG1 (5196; Cell Signaling Technology), anti–glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (sc-265062; Santa Cruz Biotechnology), anti-B23 (172-6195; Fisher Scientific), allophycocyanin (APC)/Cy7 anti-human CD34 (clone 581; BioLegend), and APC anti-human CD14 (clone 63D3; BioLegend). Alexa Fluor 647 Phalloidin was purchased from Cell Signaling Technology. All cell culture medium reagents were purchased from Thermo Fisher Scientific. Unless otherwise indicated, all other reagents were purchased from Sigma-Aldrich (St. Louis, MO, USA). Caspase-2 inhibitors were synthesized and characterized by liquid chromatography–mass spectrometry (LC-MS) as previously described by Poreba et al. (29).
Plasmids
The caspase-2 BiFC reporter pRRL-C2 Pro-VC-2A-C2 Pro-VN 2A-mCherry-2A-Puro was described previously (6). The pcDNA3.1.NPM1c+ plasmid was generated by site-directed mutagenesis of pcDNA3.1-NPM1 using the primer sequences: 5′-CTATTCAAGATCTCTGTCTGGCAGTGGAGGAAGGCG-3′ (forward) and 5′-CTATTCAAGATCTCTGTCTGGCAGTGGAGGAAGGCG-3′ (reverse) to insert the TCTG nucleotide in exon 12 of the NPM1 gene between sites 1864 and 1868. NPM1c+ was cloned into pRRL-MND-MCS-2A-mCherry-2A-Puro by standard polymerase chain reaction (PCR) strategies with primers designed to incorporate restriction enzymes, Xho I and Not I. pLVX-Tet-Off Advanced and pLVX Tight-Puro-GOI vectors were purchased from Takara Bio (catalog no. 632163) as a part of the Lenti-X Tet-Off Advanced Inducible Expression system. Caspase-2 or caspase-2(C320A) was cloned into the pLVX Tight-Puro vector by the standard PCR strategies. Histone H2B-mCherry was purchased from Addgene [#20972 (66)]. Each construct was verified by sequencing.
Cell culture and generation of cell lines
OCI-AML-2, THP-1 (NPM1wt), OCI-AML-3, and IMS-M2 (NPM1c+) cells were grown in RPMI containing fetal bovine serum (FBS) (10%, v/v), l-glutamine (2 mM), and penicillin (50 IU/ml)/streptomycin (50 μg/ml). Human embryonic kidney (HEK) 293T and HeLa cells were grown in DMEM containing FBS (10%, v/v), l-glutamine (2 mM), and penicillin (50 IU/ml)/streptomycin (50 μg/ml). Stable cell lines were generated by lentiviral transduction. HEK 293T cells were transiently transfected with pRRL.C2 Pro-VC-2A-C2 Pro-VN-2A-mCherry along with the packaging vector pSPAX-2 and the envelope vector pVSV-G using Lipofectamine 2000 transfection reagent (Thermo Fisher Scientific) according to the manufacturer’s instructions. Additional lentiviral vectors were packaged in HEK 293T cells using Lenti-X Packaging Single Shots (VSV-G) (Takara Bio USA) per the manufacturer’s protocol. After 48 hours, virus-containing supernatants were cleared by centrifugation and incubated with OCI-AML-2 or OCI-AML-3 cell lines with polybrene (5 μg/ml). For the Tet-repressible cell lines, cell lines were cotransduced with pLVX-Tet-Off Advanced and pLVX Tight-Puro-caspase-2 vectors in a 1:1 ratio. Cells were selected in puromycin (1 to 5 μg/ml) or G418 (400 μg/ml), followed by bulk fluorescence-activated cell sorting (FACSAria II, BD) when a linked fluorescent gene was expressed.
CRISPR/Cas9 gene editing
Caspase-2 was deleted from OCI-AML-2 and OCI-AML-3 cells by electroporation of Cas9RNP, as adapted from the protocol described in (67). Protospacer sequences for caspase-2 were identified using the CRISPRscan scoring algorithm [www.crisprscan.org (68)]. DNA templates for sgRNAs were made by PCR using a pX459 plasmid containing the sgRNA scaffold sequence and the following primers:
Scramble: ttaatacgactcactataGGCGCGATAGCGCGAATATATTgttttagagctagaaatagc
ΔCASP2(76): ttaatacgactcactataGGCGTGGGCAGTCTCATCTTgttttagagctagaaatagc ΔCASP2(73): ttaatacgactcactataGGTGTGGAGGGCGCCATCTAgttttagagctagaaatagc
ΔPIDD1(97): ttaatacgactcactataGGCTTGGACCTGTACCCCGGgttttagagctagaaatagc
ΔPIDD1(76): ttaatacgactcactataGGCGGTTGTGTGTCACTGTGgttttagagctagaaatagc
ΔRAIDD(57): ttaatacgactcactataGGCCCAGGGAAACTCCTGTAgttttagagctagaaatagc
ΔRAIDD(76): ttaatacgactcactataGGAGGAGCATTGTTTTCCGGgttttagagctagaaatagc
Universal reverse primer: AGCACCGACTCGGTGCCACT
sgRNAs were generated by in vitro transcription using the HiScribe T7 high-yield RNA synthesis kit (New England Biolabs). Purified sgRNA (0.5 μg) was incubated with Cas9 protein (1 μg, PNA Bio) for 10 min at room temperature. OCI-AML-2 and OCI-AML-3 cells were electroporated with sgRNA/Cas9 complex using the Neon transfection system (Thermo Fisher Scientific) under the condition of 1350 V, 35-ms pulse width, and one pulse. Single-cell clones were isolated, and knockout was confirmed by PCR and Western blotting. The single-cell clones were validated by sequencing. The single-cell OCI-AML-2 and OCI-AML-3 clones #1 and #2 were validated to have a full knockout of caspase-2. OCI-AML-3 clone #3, when sequenced, had a mixed population of wild-type and deleted sequence and hence had lower caspase-2 expression (Fig. 2A). Because of the fast onset of viability loss in the caspase-2–deficient cells, all experiments were initiated within 2 weeks of thawing.
Flow cytometry
All flow experiments were performed using an LSR Fortessa Flow Cytometer (BD, San Jose, CA, USA). Data were analyzed using FlowJo Software (BD). For annexin V binding, cells were treated as indicated in the figure legends. Cells were harvested by centrifugation at 300g for 5 min, washed with 1× phosphate-buffered saline (PBS), and resuspended in 200 μl of annexin V staining buffer (10 mM Hepes, 150 mM NaCl, 5 mM KCl, 1 mM MgCl2, and 1.8 mM CaCl2) containing 2 μl of annexin V–APC (Thermo Fisher Scientific). After 15 min of incubation at room temperature, annexin V–positive cells were quantified by flow cytometry. For cell cycle analysis, cell medium was exchanged for medium with 10 μM BrdU. After 30 min, cells were harvested by centrifugation at 300g for 5 min. Cells were fixed and permeabilized with BD Cytofix/Cytoperm buffer for 15 min at room temperature and then with the secondary permeabilization buffer BD Cytoperm Permeabilization Buffer Plus for 10 min on ice and then were fixed and permeabilized again with BD Cytofix/Cytoperm buffer for 5 min at room temperature. Cells were washed in BD Perm/Wash buffer with FBS and centrifuged at 300g between each step. The cells were treated with deoxyribonuclease (DNase) (30 μg/1 × 106 cells) for 1 hour at 37°C to uncover the BrdU epitope. Cells were then washed in BD Perm/Wash buffer and centrifuged at 300g. The cell pellets were incubated in a 1:50 dilution of fluorescein isothiocyanate (FITC)–labeled anti-BrdU antibody in BD Perm/Wash buffer for 20 min at room temperature. The cells were then resuspended in stain buffer (3% FBS in PBS) containing 10 μl of 7-AAD/M cells. BrdU- and 7-AAD–positive cells were quantitated by flow cytometry. For quantitation of cell surface markers, cells were centrifuged at 1400 rpm for 5 min. Cell pellets were suspended in 100 μl of fresh medium with the appropriate antibodies to cell surface markers and incubated for 45 min at 4°C. The cells were washed in 2 ml of BD Stain Buffer and centrifuged at 1400 rpm for 5 min between each step. The cells were fixed and permeabilized with BD Cytofix/Cytoperm buffer for 15 min. Cells were resuspended in 300 μl of stain buffer and quantified for cell surface marker expression by flow cytometry. Live cells were selected using the LIVE/DEAD Fixable Dead Cell Stain Kit (Thermo Fisher Scientific).
Immunoblotting
Cells were lysed in radioimmunoprecipitation assay (RIPA) buffer [150 mM NaCl, 50 mM tris-HCl (pH 7), 0.1% SDS, 0.5% sodium deoxycholate, 1% NP-40 (IGEPAL) plus complete protease inhibitors-EDTA (1 mini tablet/10 ml)]. Cleared protein lysates (40 μg) were resolved by SDS-PAGE on a 4 to 12% gradient gel (Thermo Fisher Scientific). The proteins were transferred to nitrocellulose membrane (Bio-Rad Laboratories) and immunodetected using relevant primary and peroxidase-conjugated secondary antibodies: donkey anti-rabbit immunoglobulin G (IgG) (catalog no. 84-854, Prometheus), goat anti-rat IgG (catalog no. 20-307, Prometheus), sheep anti-mouse IgG (catalog no. 84-848, Prometheus), and donkey anti-goat IgG (sc-2020; Santa Cruz). Proteins were visualized with West Dura and West Pico chemiluminescence substrate (Thermo Fisher Scientific).
Nucleolar isolation
Nucleoli were isolated as per the protocol adapted from (69). OCI-AML-2 and OCI-AML-3 cells (1 × 106) were washed in 10 ml of PBS and centrifuged at 400g for 5 min. Cell pellets were resuspended in five pellet volumes of nucleolar isolation buffer (NIB: 10 mM tris, 2 mM MgCl2, 0.5 mM EDTA) containing complete protease inhibitor (Roche). Samples were incubated in NIB buffer for 2 min at room temperature and then on ice for 10 min. Plasma membranes were lysed by the addition of 10% (v/v) IGEPAL to a final concentration of 1% and light vortexing. Crude nuclei pellets were isolated by centrifugation at 500g for 3 min. The supernatant was removed and stored as the cytoplasmic fraction. The nuclear pellets were washed in 15 pellet volumes of NIB, 1% IGEPAL. Pellets were then resuspended in 10 pellet volumes of NIB and sonicated at 20% power for five cycles of 1 s on followed by 5 s off on an Misonix XL 2020 sonicator fitted with a microtip probe (Misonix, Farmingdale, NY, USA). The sonication was repeated until the intact nucleoli were visible. The samples were centrifuged, and the supernatant was removed and stored as the nucleoplasmic fraction. The nucleoli pellets were washed once more with five pellet volumes of NIB and finally resuspended in 50 μl of NIB.
Microscopy
Cells were imaged using a spinning disk confocal microscope (Carl Zeiss MicroImaging, Thornwood, NY, USA), equipped with a CSU-X1A 5000 spinning disk unit (Yokogawa Electric Corporation, Japan), multilaser module with wavelengths of 458, 488, 514, 561, and 647 nm, and an Axio Observer Z1 motorized inverted microscope equipped with a precision motorized XY stage (Zeiss). Images were acquired with a Zeiss Plan-Neofluar 40× 1.3 numerical aperture (NA) or 63× 1.4 NA objective on an Orca R2 charge-coupled device (CCD) camera using Zen 2.5 software (Zeiss). HeLa cells were plated on dishes containing glass coverslips coated with fibronectin (Mattek Corp., Ashland, MA, USA) 24 hours before treatment. For time-lapse experiments, OCI-AML cells in medium supplemented with Hepes (20 mM) and 2-mercaptoethanol (55 μM) were loaded onto an ethylcellulose microscaffold to restrict movement. Cells were allowed to equilibrate before focusing on the cells in a humidified incubation chamber set at 37°C with 5% CO2.
Microscaffold preparation and assembly
The microscaffold was fabricated as a thin film by soft lithography. Briefly, ethylcellulose solution (Sigma-Aldrich; 7.5%, w/v, in ethanol) was transferred onto a template containing an array of square-shaped posts (50 × 50 μm; 30 μm height) and allowed to dry at 70°C for 3 hours. The microscaffold containing square-shaped wells was gently peeled away from the template and kept in a sealed container. The integrity of the wells was confirmed by light microscope. The internal glass surface of 35-mm glass bottom dishes was cleaned using a Q-tip cotton swab dipped in acetone and briefly washed three times with Milli-Q water. Next, a solution of 1 M NaOH was incubated overnight before being finally rinsed off three times with Milli-Q water and allowed to dry under a biosafety cabinet airflow. A polydimethylsiloxane (PDMS) solution (SYLGARD 184 Silicone, Dow Corning) was prepared using the manufacturer’s instructions. Ten microliters was deposited in the center of the glass bottom dish and spread into a thin film. A square of the microscaffold was applied to the PDMS while carefully avoiding trapping air bubbles underneath, and then incubated at 37°C overnight or until the PDMS was fully polymerized. The dishes were ultraviolet (UV)–sterilized by a handheld germicidal UV light (Genesee), followed by sonication in sterile water for 5 min to dislodge air bubbles. The dishes were washed with sterile PBS, before seeding the cells.
Image stream analysis
OCI-AML-2 and OCI-AML-3 cells stably expressing the C2-Pro BiFC reporter were analyzed using an Amnis ImageStream Mark II (405-, 488-, 561-, and 642-nm excitation lasers; ×40 magnification) and INSPIRE software (Cytek Biosciences). Using the bright-field channel, a cell area–versus–cell aspect ratio plot was used to gate single cells from beads and cellular debris. Subsequently, out-of-focus objects were gated out from the dataset using the gradient root-mean-square (RMS) measurement (<45%) on the same bright-field channel. Nontransduced cells were used as fluorescence-minus-one (FMO) controls for the gating strategy. Data analysis was performed using IDEAS software (version 6.2).
Transient transfection and immunofluorescence
HeLa cells (1 × 105) were plated on dishes containing glass coverslips coated with fibronectin (1 mg/ml). Cells were transfected with relevant expression plasmids as described in the figure legends using Lipofectamine 2000 transfection reagent (Thermo Fisher Scientific) according to the manufacturer’s instructions. After 24 hours, cells were washed in 3 × 2 ml PBS and fixed in 2% (w/v) paraformaldehyde in PBS (pH 7.2) for 15 min. Cells were washed with PBS followed by permeabilization with 0.15% (v/v) Triton X-100 for 15 min. Cells were blocked in 2% (w/v) bovine serum albumin (BSA) for 30 min. Cells were stained with relevant antibodies in 1:100 dilutions in PBS with 2% (w/v) BSA for 1 hour. After washing in PBS with 2% (w/v) BSA, the cells were incubated with Alexa Fluor 488– or Alexa Fluor 647–conjugated secondary antibodies (Thermo Fisher Scientific) at a 1:250 dilution in 2% (w/v) BSA for 45 min. Cells were washed in PBS before imaging.
Viability assay
Parental and caspase-2–deficient NPM1wt and NPM1c+ cells (1 × 104) were plated per well in 96-well plates. The plates were incubated for 1, 2, 3, 4, and 5 days at 37°C. After each respective day of incubation, the cells were gently centrifuged at 1000 rpm for 2 min to collect the suspension cells. The medium from each well was removed and replaced with fresh medium containing 10 μl of 12 mM MTT as per the manufacturer’s instruction (Roche). After 4 hours of incubation, MTT medium was replaced by 100 μl of DMSO followed by orbital plate shaking. The plates were then incubated for an additional 15 min. The absorbance was read at 540- and 680-nm wavelengths.
RNAseq analysis
Total RNA was extracted from the cells using RNeasy kit with QIAshredder (Qiagen) according to the manufacturer’s protocol. RNA was resuspended in diethylpyrocarbonate (DEPC)–treated water and quantified using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific). Quality testing and RNAseq were carried out at the BGI Americas facility (San Jose, CA). RNAseq libraries were sequenced using the BGISEQ-500 platform using the DNA nanoball technology [https://bgiamericas.com/technologies/complete-genomics/ (Shenzhen, China)]. Base-calling was performed by the BGISEQ-500 software version 0.3.8.1111. Bioinformatics analysis was carried out using Dr.Tom (BGI) and ENRICHR pathway analysis tools (https://maayanlab.cloud/Enrichr/). Pathway analysis was carried out on the complete dataset (13,043 genes were included for analysis). Significantly DEGs shared between both caspase-2–deficient clones (813 genes) were selected to generate the heatmap. The heatmap was generated using R version 4.3.0 and RStudio version 2022.12.0.
RPPA analysis
RPPA assays for antibodies to proteins or phosphorylated proteins in different functional pathways were carried out as described previously (70, 71). Specifically, protein lysates were prepared from cultured cells with modified Tissue Protein Extraction Reagent (TPER) (Thermo Fisher Scientific) and a cocktail of protease and phosphatase inhibitors (Roche, Pleasanton, CA). Three technical replicates were spotted for each of the four independent biological replicates for each cell line. The lysates were diluted into 0.5 mg/ml in SDS sample buffer and denatured on the same day. The Quanterix 2470 Arrayer (Quanterix, Billerica, MA) with a 40-pin (185-μm) configuration was used to spot samples and control lysates onto nitrocellulose-coated slides (Grace Bio-labs, Bend, OR) using an array format of 960 lysates per slide (2880 spots per slide). The slides were processed as described and probed with a set of 258 antibodies against total proteins and phosphoproteins using an automated slide stainer Autolink 48 (Dako, Santa Clara, CA). Each slide was incubated with one specific primary antibody, and a negative control slide was incubated with antibody diluent without any primary antibody. Primary antibody binding was detected using a biotinylated secondary antibody followed by streptavidin-conjugated IRDye680 fluorophore (LI-COR Biosciences, Lincoln, NE). Total protein content of each spotted lysate was assessed by fluorescent staining with Sypro Ruby Protein Blot Stain according to the manufacturer’s instructions (Molecular Probes, Eugene, OR). Fluorescence-labeled slides were scanned on a GenePix 4400 AL scanner, along with accompanying negative control slides, at an appropriate PMT (photomultiplier tube) to obtain optimal signal for this specific set of samples. The images were analyzed with GenePix Pro 7.0 (Molecular Devices, Silicon Valley, CA). Total fluorescence signal intensities of each spot were obtained after subtraction of the local background signal for each slide and were then normalized for variation in total protein, background, and nonspecific labeling using a group-based normalization method as described (70). For each spot on the array, the background-subtracted foreground signal intensity was subtracted by the corresponding signal intensity of the negative control slide (omission of primary antibody) and then normalized to the corresponding signal intensity of total protein for that spot. Each image, along with its normalized data, was evaluated for quality through manual inspection and control samples. Antibody slides that failed the quality inspection were either repeated at the end of the staining runs or removed before data reporting. A total of 258 antibodies remained in the list. Antibodies were filtered to remove samples with >25% CV (coefficient of variation) across the technical replicates and less than 200 signals. A total of 208 antibodies were used for subsequent data analyses. Multiple t tests were used to determine significance with a cutoff of q < 0.01. Z-prime scores and heatmaps were generated using Heatmapper (http://www.heatmapper.ca).
Statistical analysis
Statistical significance was assessed by using two-tailed Student’s t test or two-way analysis of variance (ANOVA) for multiple comparisons. All statistical analyses were performed using GraphPad Prism 9 for Windows (San Diego, CA, and USA RRID: SCR_002798).
Supplementary Material
Acknowledgments
We would like to thank the entire Bouchier-Hayes laboratory for helpful discussion. We would like to acknowledge the Texas Children’s Hospital William T. Shearer Center for Human Immunobiology for their support for this research and the expert assistance of D. Carrasco Di Lallo. We thank X. Wang and Z. Shi from the Antibody-based Proteomics Core/Shared Resource for their excellent technical assistance in performing RPPA experiments. We thank C. Coarfa and S. L. Grimm for RPPA data processing and normalization. Graphics were created using BioRender.
Funding: This work was supported by NIH grants R01GM121389 (to L.B.-H.), R21CA256606 (to L.B.-H.), T32DK060445 (to S.S.), and S10OD028648 (to S.H.); Houston Junior Women’s Club gift (to L.B.-H.); Cancer Prevention Institute of Texas CPRIT RP210027 (to D.S.); Cancer Prevention & Research Institute of Texas Proteomics & Metabolomics Core Facility Support Award RP210227 [Antibody-based Proteomics Core/Shared Resource (to S.H.)]; NCI Cancer Center Support Grant P30CA125123 [Antibody-based Proteomics Core/Shared Resource (to S.H.)]; and National Science Centre in Poland, OPUS, UMO-2022/47/B/NZ5/02405 (to M.P.).
Author contributions: Conceptualization: D.S., A.N.B.-S., and L.B.-H. Methodology: D.S., A.N.B.-S., S.S., K.E.L., S.H., L.E.C., D.J.R., B.L., C.S.S., A.V.-H., A.F.C., M.P., J.M.F., and L.B.-H. Investigation: D.S., A.N.B.-S., S.S., K.E.L., F.K., S.H., L.E.C., K.M.S., C.I.C., K.P.D., D.J.R., J.M.F., M.P., and L.B.-H. Visualization: D.S. and L.B.-H. Supervision: L.B.-H. Writing—original draft: D.S. and L.B.-H. Writing—review and editing: D.S. and L.B.-H.
Competing interests: D.S. is now employed by Lonza. All other authors declare that they have no competing interests.
Data and materials availability: RNAseq data have been deposited in NCBI GEO (accession number GSE249545). All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials.
Supplementary Materials
This PDF file includes:
Legends for movies S1 and S2
Legend for data file S1
Figs. S1 to S6
Other Supplementary Material for this manuscript includes the following:
Movies S1 and S2
Data file S1
REFERENCES AND NOTES
- 1.A. M. Noone, N. Howlader, M. Krapcho, D. Miller, A. Brest, M. Yu, J. Ruhl, Z. Tatalovich, A. Mariotto, D. R. Lewis, H. S. Chen, E. J. Feuer, K. A. Cronin, Eds. SEER Cancer Statistics Review, 1975-2015 (National Cancer Institute, 2018); https://seer.cancer.gov/csr/1975_2015/.
- 2.Falini B., Mecucci C., Tiacci E., Alcalay M., Rosati R., Pasqualucci L., la Starza R., Diverio D., Colombo E., Santucci A., Bigerna B., Pacini R., Pucciarini A., Liso A., Vignetti M., Fazi P., Meani N., Pettirossi V., Saglio G., Mandelli F., Lo-Coco F., Pelicci P. G., Martelli M. F., Cytoplasmic nucleophosmin in acute myelogenous leukemia with a normal karyotype. N. Engl. J. Med. 352, 254–266 (2005). [DOI] [PubMed] [Google Scholar]
- 3.Grisendi S., Mecucci C., Falini B., Pandolfi P. P., Nucleophosmin and cancer. Nat. Rev. Cancer 6, 493–505 (2006). [DOI] [PubMed] [Google Scholar]
- 4.Arber D. A., Orazi A., Hasserjian R. P., Borowitz M. J., Calvo K. R., Kvasnicka H. M., Wang S. A., Bagg A., Barbui T., Branford S., Bueso-Ramos C. E., Cortes J. E., Dal Cin P., DiNardo C. D., Dombret H., Duncavage E. J., Ebert B. L., Estey E. H., Facchetti F., Foucar K., Gangat N., Gianelli U., Godley L. A., Gökbuget N., Gotlib J., Hellström-Lindberg E., Hobbs G. S., Hoffman R., Jabbour E. J., Kiladjian J. J., Larson R. A., le Beau M. M., Loh M. L. C., Löwenberg B., Macintyre E., Malcovati L., Mullighan C. G., Niemeyer C., Odenike O. M., Ogawa S., Orfao A., Papaemmanuil E., Passamonti F., Porkka K., Pui C. H., Radich J. P., Reiter A., Rozman M., Rudelius M., Savona M. R., Schiffer C. A., Schmitt-Graeff A., Shimamura A., Sierra J., Stock W. A., Stone R. M., Tallman M. S., Thiele J., Tien H. F., Tzankov A., Vannucchi A. M., Vyas P., Wei A. H., Weinberg O. K., Wierzbowska A., Cazzola M., Döhner H., Tefferi A., International Consensus Classification of Myeloid Neoplasms and Acute Leukemias: Integrating morphologic, clinical, and genomic data. Blood 140, 1200–1228 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Martelli M. P., Gionfriddo I., Mezzasoma F., Milano F., Pierangeli S., Mulas F., Pacini R., Tabarrini A., Pettirossi V., Rossi R., Vetro C., Brunetti L., Sportoletti P., Tiacci E., di Raimondo F., Falini B., Arsenic trioxide and all-trans retinoic acid target NPM1 mutant oncoprotein levels and induce apoptosis in NPM1-mutated AML cells. Blood 125, 3455–3465 (2015). [DOI] [PubMed] [Google Scholar]
- 6.Ando K., Parsons M. J., Shah R. B., Charendoff C. I., Paris S. L., Liu P. H., Fassio S. R., Rohrman B. A., Thompson R., Oberst A., Sidi S., Bouchier-Hayes L., NPM1 directs PIDDosome-dependent caspase-2 activation in the nucleolus. J. Cell Biol. 216, 1795–1810 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Brown-Suedel A. N., Bouchier-Hayes L., Caspase-2 substrates: To apoptosis, cell cycle control, and beyond. Front. Cell Dev. Biol. 8, 610022 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Tinel A., Tschopp J., The PIDDosome, a protein complex implicated in activation of caspase-2 in response to genotoxic stress. Science 304, 843–846 (2004). [DOI] [PubMed] [Google Scholar]
- 9.Ho L. H., Taylor R., Dorstyn L., Cakouros D., Bouillet P., Kumar S., A tumor suppressor function for caspase-2. Proc. Natl. Acad. Sci. U.S.A. 106, 5336–5341 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Boice A. G., Lopez K. E., Pandita R. K., Parsons M. J., Charendoff C. I., Charaka V., Carisey A. F., Pandita T. K., Bouchier-Hayes L., Caspase-2 regulates S-phase cell cycle events to protect from DNA damage accumulation independent of apoptosis. Oncogene 41, 204–219 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Dorstyn L., Puccini J., Wilson C. H., Shalini S., Nicola M., Moore S., Kumar S., Caspase-2 deficiency promotes aberrant DNA-damage response and genetic instability. Cell Death Differ. 19, 1288–1298 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Puccini J., Shalini S., Voss A. K., Gatei M., Wilson C. H., Hiwase D. K., Lavin M. F., Dorstyn L., Kumar S., Loss of caspase-2 augments lymphomagenesis and enhances genomic instability in Atm-deficient mice. Proc. Natl. Acad. Sci. U.S.A. 110, 19920–19925 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Parsons M. J., McCormick L., Janke L., Howard A., Bouchier-Hayes L., Green D. R., Genetic deletion of caspase-2 accelerates MMTV/c-neu-driven mammary carcinogenesis in mice. Cell Death Differ. 20, 1174–1182 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Terry M. R., Arya R., Mukhopadhyay A., Berrett K. C., Clair P. M., Witt B., Salama M. E., Bhutkar A., Oliver T. G., Caspase-2 impacts lung tumorigenesis and chemotherapy response in vivo. Cell Death Differ. 22, 719–730 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Zhang Y., Padalecki S. S., Chaudhuri A. R., De Waal E., Goins B. A., Grubbs B., Ikeno Y., Richardson A., Mundy G. R., Herman B., Caspase-2 deficiency enhances aging-related traits in mice. Mech. Ageing Dev. 128, 213–221 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Dawar S., Shahrin N. H., Sladojevic N., D’Andrea R. J., Dorstyn L., Hiwase D. K., Kumar S., Impaired haematopoietic stem cell differentiation and enhanced skewing towards myeloid progenitors in aged caspase-2-deficient mice. Cell Death Dis. 7, e2509 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Holleman A., den Boer M. L., Kazemier K. M., Beverloo H. B., von Bergh A. R. M., Janka-Schaub G. E., Pieters R., Decreased PARP and procaspase-2 protein levels are associated with cellular drug resistance in childhood acute lymphoblastic leukemia. Blood 106, 1817–1823 (2005). [DOI] [PubMed] [Google Scholar]
- 18.Alcalay M., Tiacci E., Bergomas R., Bigerna B., Venturini E., Minardi S. P., Meani N., Diverio D., Bernard L., Tizzoni L., Volorio S., Luzi L., Colombo E., Lo Coco F., Mecucci C., Falini B., Pelicci P. G.; for the Gruppo Italiano Malattie Ematologiche Maligne dell’Adulto (GIMEMA) Acute Leukemia Working Party , Acute myeloid leukemia bearing cytoplasmic nucleophosmin (NPMc+ AML) shows a distinct gene expression profile characterized by up-regulation of genes involved in stem-cell maintenance. Blood 106, 899–902 (2005). [DOI] [PubMed] [Google Scholar]
- 19.Eklund E., The role of Hox proteins in leukemogenesis: Insights into key regulatory events in hematopoiesis. Crit. Rev. Oncog. 16, 65–76 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Brunetti L., Gundry M. C., Sorcini D., Guzman A. G., Huang Y. H., Ramabadran R., Gionfriddo I., Mezzasoma F., Milano F., Nabet B., Buckley D. L., Kornblau S. M., Lin C. Y., Sportoletti P., Martelli M. P., Falini B., Goodell M. A., Mutant NPM1 maintains the leukemic state through HOX expression. Cancer Cell 34, 499–512.e9 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Baliga B. C., Read S. H., Kumar S., The biochemical mechanism of caspase-2 activation. Cell Death Differ. 11, 1234–1241 (2004). [DOI] [PubMed] [Google Scholar]
- 22.Borer R. A., Lehner C. F., Eppenberger H. M., Nigg E. A., Major nucleolar proteins shuttle between nucleus and cytoplasm. Cell 56, 379–390 (1989). [DOI] [PubMed] [Google Scholar]
- 23.Shyu Y. J., Liu H., Deng X., Hu C. D., Identification of new fluorescent protein fragments for bimolecular fluorescence complementation analysis under physiological conditions. Biotechniques 40, 61–66 (2006). [DOI] [PubMed] [Google Scholar]
- 24.Bouchier-Hayes L., Oberst A., McStay G. P., Connell S., Tait S. W. G., Dillon C. P., Flanagan J. M., Beere H. M., Green D. R., Characterization of cytoplasmic caspase-2 activation by induced proximity. Mol. Cell 35, 830–840 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Brodska B., Kracmarova M., Holoubek A., Kuzelova K., Localization of AML-related nucleophosmin mutant depends on its subtype and is highly affected by its interaction with wild-type NPM. PLOS ONE 12, e0175175 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Pianigiani G., Rocchio F., Peruzzi S., Andresen V., Bigerna B., Sorcini D., Capurro M., Gjertsen B. T., Sportoletti P., di Ianni M., Martelli M. P., Brunetti L., Falini B., The absent/low expression of CD34 in NPM1-mutated AML is not related to cytoplasmic dislocation of NPM1 mutant protein. Leukemia 36, 1931–1934 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Hayun M., Zaatra M., Itzkovich C., Sahar D., Rosenberg D., Filatova M., Ringelstein-Harlev S., Baris H., Moustafa-Hawash N., Louria-Hayon I., Ofran Y., ERK activity in immature leukemic cells drives clonal selection during induction therapy for acute myeloid leukemia. Sci. Rep. 10, 8349 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.McStay G. P., Salvesen G. S., Green D. R., Overlapping cleavage motif selectivity of caspases: Implications for analysis of apoptotic pathways. Cell Death Differ. 15, 322–331 (2008). [DOI] [PubMed] [Google Scholar]
- 29.Poreba M., Rut W., Groborz K., Snipas S. J., Salvesen G. S., Drag M., Potent and selective caspase-2 inhibitor prevents MDM-2 cleavage in reversine-treated colon cancer cells. Cell Death Differ. 26, 2695–2709 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Agger K., Nishimura K., Miyagi S., Messling J. E., Rasmussen K. D., Helin K., The KDM4/JMJD2 histone demethylases are required for hematopoietic stem cell maintenance. Blood 134, 1154–1158 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Hu X., Li X., Valverde K., Fu X., Noguchi C., Qiu Y., Huang S., LSD1-mediated epigenetic modification is required for TAL1 function and hematopoiesis. Proc. Natl. Acad. Sci. U.S.A. 106, 10141–10146 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Clevers H., Nusse R., Wnt/β-Catenin signaling and disease. Cell 149, 1192–1205 (2012). [DOI] [PubMed] [Google Scholar]
- 33.Desbois-Mouthon C., Cadoret A., Blivet-van Eggelpoël M. J., Bertrand F., Cherqui G., Perret C., Capeau J., Insulin and IGF-1 stimulate the beta-catenin pathway through two signalling cascades involving GSK-3beta inhibition and Ras activation. Oncogene 20, 252–259 (2001). [DOI] [PubMed] [Google Scholar]
- 34.Xu Z. X., Tan J. W., Xu H., Hill C. J., Ostrovskaya O., Martemyanov K. A., Xu B., Caspase-2 promotes AMPA receptor internalization and cognitive flexibility via mTORC2-AKT-GSK3β signaling. Nat. Commun. 10, 3622 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Dibble C. C., Asara J. M., Manning B. D., Characterization of Rictor phosphorylation sites reveals direct regulation of mTOR complex 2 by S6K1. Mol. Cell. Biol. 29, 5657–5670 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Facchinetti V., Ouyang W., Wei H., Soto N., Lazorchak A., Gould C., Lowry C., Newton A. C., Mao Y., Miao R. Q., Sessa W. C., Qin J., Zhang P., Su B., Jacinto E., The mammalian target of rapamycin complex 2 controls folding and stability of Akt and protein kinase C. EMBO J. 27, 1932–1943 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Garcia-Martinez J. M., Alessi D. R., mTOR complex 2 (mTORC2) controls hydrophobic motif phosphorylation and activation of serum- and glucocorticoid-induced protein kinase 1 (SGK1). Biochem. J. 416, 375–385 (2008). [DOI] [PubMed] [Google Scholar]
- 38.Dorstyn L., Hackett-Jones E., Nikolic A., Norris M. D., Lim Y., Toubia J., Haber M., Kumar S., Transcriptome profiling of caspase-2 deficient EμMyc and Th-MYCN mouse tumors identifies distinct putative roles for caspase-2 in neuronal differentiation and immune signaling. Cell Death Dis. 10, 56 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Smith A. J., Nelson N. G., Oommen S., Hartjes K. A., Folmes C. D., Terzic A., Nelson T. J., Apoptotic susceptibility to DNA damage of pluripotent stem cells facilitates pharmacologic purging of teratoma risk. Stem Cells Transl. Med. 1, 709–718 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Liu J. C., Lerou P. H., Lahav G., Stem cells: Balancing resistance and sensitivity to DNA damage. Trends Cell Biol. 24, 268–274 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Julien O., Zhuang M., Wiita A. P., O’Donoghue A. J., Knudsen G. M., Craik C. S., Wells J. A., Quantitative MS-based enzymology of caspases reveals distinct protein substrate specificities, hierarchies, and cellular roles. Proc. Natl. Acad. Sci. U.S.A. 113, E2001–E2010 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Oliver T. G., Meylan E., Chang G. P., Xue W., Burke J. R., Humpton T. J., Hubbard D., Bhutkar A., Jacks T., Caspase-2-mediated cleavage of Mdm2 creates a p53-induced positive feedback loop. Mol. Cell 43, 57–71 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Shiota C., Woo J. T., Lindner J., Shelton K. D., Magnuson M. A., Multiallelic disruption of the rictor gene in mice reveals that mTOR complex 2 is essential for fetal growth and viability. Dev. Cell 11, 583–589 (2006). [DOI] [PubMed] [Google Scholar]
- 44.Sarbassov D. D., Ali S. M., Kim D.-H., Guertin D. A., Latek R. R., Erdjument-Bromage H., Tempst P., Sabatini D. M., Rictor, a novel binding partner of mTOR, defines a rapamycin-insensitive and raptor-independent pathway that regulates the cytoskeleton. Curr. Biol. 14, 1296–1302 (2004). [DOI] [PubMed] [Google Scholar]
- 45.Sarbassov D. D., Guertin D. A., Ali S. M., Sabatini D. M., Phosphorylation and regulation of Akt/PKB by the rictor-mTOR complex. Science 307, 1098–1101 (2005). [DOI] [PubMed] [Google Scholar]
- 46.Yu Z., Chen J., Takagi E., Wang F., Saha B., Liu X., Joubert L. M., Gleason C. E., Jin M., Li C., Nowotny C., Agard D., Cheng Y., Pearce D., Interactions between mTORC2 core subunits Rictor and mSin1 dictate selective and context-dependent phosphorylation of substrate kinases SGK1 and Akt. J. Biol. Chem. 298, 102288 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Yoon C., Lu J., Yi B. C., Chang K. K., Simon M. C., Ryeom S., Yoon S. S., PI3K/Akt pathway and Nanog maintain cancer stem cells in sarcomas. Oncogenesis 10, 12 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar] [Retracted]
- 48.Malak P. N., Dannenmann B., Hirth A., Rothfuss O. C., Schulze-Osthoff K., Novel AKT phosphorylation sites identified in the pluripotency factors OCT4, SOX2 and KLF4. Cell Cycle 14, 3748–3754 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Schaefer T., Wang H., Mir P., Konantz M., Pereboom T. C., Paczulla A. M., Merz B., Fehm T., Perner S., Rothfuss O. C., Kanz L., Schulze-Osthoff K., Lengerke C., Molecular and functional interactions between AKT and SOX2 in breast carcinoma. Oncotarget 6, 43540–43556 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Watanabe S., Umehara H., Murayama K., Okabe M., Kimura T., Nakano T., Activation of Akt signaling is sufficient to maintain pluripotency in mouse and primate embryonic stem cells. Oncogene 25, 2697–2707 (2006). [DOI] [PubMed] [Google Scholar]
- 51.Lin Y., Yang Y., Li W., Chen Q., Li J., Pan X., Zhou L., Liu C., Chen C., He J., Cao H., Yao H., Zheng L., Xu X., Xia Z., Ren J., Xiao L., Li L., Shen B., Zhou H., Wang Y. J., Reciprocal regulation of Akt and Oct4 promotes the self-renewal and survival of embryonal carcinoma cells. Mol. Cell 48, 627–640 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Kharas M. G., Okabe R., Ganis J. J., Gozo M., Khandan T., Paktinat M., Gilliland D. G., Gritsman K., Constitutively active AKT depletes hematopoietic stem cells and induces leukemia in mice. Blood 115, 1406–1415 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53.Bertacchini J., Guida M., Accordi B., Mediani L., Martelli A. M., Barozzi P., Petricoin E. III, Liotta L., Milani G., Giordan M., Luppi M., Forghieri F., de Pol A., Cocco L., Basso G., Marmiroli S., Feedbacks and adaptive capabilities of the PI3K/Akt/mTOR axis in acute myeloid leukemia revealed by pathway selective inhibition and phosphoproteome analysis. Leukemia 28, 2197–2205 (2014). [DOI] [PubMed] [Google Scholar]
- 54.Park S., Chapuis N., Tamburini J., Bardet V., Cornillet-Lefebvre P., Willems L., Green A., Mayeux P., Lacombe C., Bouscary D., Role of the PI3K/AKT and mTOR signaling pathways in acute myeloid leukemia. Haematologica 95, 819–828 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55.Ren Z., Shrestha M., Sakamoto T., Melkman T., Meng L., Cairns R. A., Zacksenhaus E., Mak T. W., Stambolic V., Minden M. D., Wang J. A., Opposing effects of NPM1wt and NPM1c mutants on AKT signaling in AML. Leukemia 34, 1172–1176 (2020). [DOI] [PubMed] [Google Scholar]
- 56.Gallo A., Lo Sterzo C., Mori M., di Matteo A., Bertini I., Banci L., Brunori M., Federici L., Structure of nucleophosmin DNA-binding domain and analysis of its complex with a G-quadruplex sequence from the c-MYC promoter. J. Biol. Chem. 287, 26539–26548 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 57.Oka M., Mura S., Otani M., Miyamoto Y., Nogami J., Maehara K., Harada A., Tachibana T., Yoneda Y., Ohkawa Y., Chromatin-bound CRM1 recruits SET-Nup214 and NPM1c onto HOX clusters causing aberrant HOX expression in leukemia cells. eLife 8, e46667 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 58.Wang L., Qiao C., Cao L., Cai S., Ma X., Song X., Jiang Q., Huang C., Wang J., Significance of HOXD transcription factors family in progression, migration and angiogenesis of cancer. Crit. Rev. Oncol. Hematol. 179, 103809 (2022). [DOI] [PubMed] [Google Scholar]
- 59.Suthon S., Perkins R. S., Bryja V., Miranda-Carboni G. A., Krum S. A., WNT5B in physiology and disease. Front. Cell Dev. Biol. 9, 667581 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 60.R. Nusse, X. Lim, Signaling components as Wnt targets. The Wnt homepage (1997–2004); https://web.stanford.edu/group/nusselab/cgi-bin/wnt/target_genes_components.
- 61.Fukumoto S., Hsieh C. M., Maemura K., Layne M. D., Yet S. F., Lee K. H., Matsui T., Rosenzweig A., Taylor W. G., Rubin J. S., Perrella M. A., Lee M. E., Akt participation in the Wnt signaling pathway through Dishevelled. J. Biol. Chem. 276, 17479–17483 (2001). [DOI] [PubMed] [Google Scholar]
- 62.Falini B., Nicoletti I., Martelli M. F., Mecucci C., Acute myeloid leukemia carrying cytoplasmic/mutated nucleophosmin (NPMc+ AML): Biologic and clinical features. Blood 109, 874–885 (2007). [DOI] [PubMed] [Google Scholar]
- 63.Dovey O. M., Cooper J. L., Mupo A., Grove C. S., Lynn C., Conte N., Andrews R. M., Pacharne S., Tzelepis K., Vijayabaskar M. S., Green P., Rad R., Arends M., Wright P., Yusa K., Bradley A., Varela I., Vassiliou G. S., Molecular synergy underlies the co-occurrence patterns and phenotype of NPM1-mutant acute myeloid leukemia. Blood 130, 1911–1922 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 64.Issa G. C., Bidikian A., Venugopal S., Konopleva M., DiNardo C. D., Kadia T. M., Borthakur G., Jabbour E., Pemmaraju N., Yilmaz M., Short N. J., Maiti A., Sasaki K., Masarova L., Pierce S., Takahashi K., Tang G., Loghavi S., Patel K., Andreeff M., Bhalla K., Garcia-Manero G., Ravandi F., Kantarjian H., Daver N., Clinical outcomes associated with NPM1 mutations in patients with relapsed or refractory AML. Blood Adv. 7, 933–942 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 65.Konopleva M. Y., Walter R. B., Faderl S. H., Jabbour E. J., Zeng Z., Borthakur G., Huang X., Kadia T. M., Ruvolo P. P., Feliu J. B., Lu H., Debose L. K., Burger J. A., Andreeff M., Liu W., Baggerly K. A., Kornblau S. M., Doyle L. A., Estey E. H., Kantarjian H. M., Preclinical and early clinical evaluation of the oral AKT inhibitor, MK-2206, for the treatment of acute myelogenous leukemia. Clin. Cancer Res. 20, 2226–2235 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 66.Nam H. S., Benezra R., High levels of Id1 expression define B1 type adult neural stem cells. Cell Stem Cell 5, 515–526 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 67.Gundry M. C., Brunetti L., Lin A., Mayle A. E., Kitano A., Wagner D., Hsu J. I., Hoegenauer K. A., Rooney C. M., Goodell M. A., Nakada D., Highly efficient genome editing of murine and human hematopoietic progenitor cells by CRISPR/Cas9. Cell Rep. 17, 1453–1461 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 68.Moreno-Mateos M. A., Vejnar C. E., Beaudoin J. D., Fernandez J. P., Mis E. K., Khokha M. K., Giraldez A. J., CRISPRscan: Designing highly efficient sgRNAs for CRISPR-Cas9 targeting in vivo. Nat. Methods 12, 982–988 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 69.Mitrea D. M., Cika J. A., Guy C. S., Ban D., Banerjee P. R., Stanley C. B., Nourse A., Deniz A. A., Kriwacki R. W., Nucleophosmin integrates within the nucleolus via multi-modal interactions with proteins displaying R-rich linear motifs and rRNA. eLife 5, e13571 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 70.Wang X., Shi Z., Lu H. Y., Kim J. J., Bu W., Villalobos J. A., Perera D. N., Jung S. Y., Wang T., Grimm S. L., Taylor B. C., Rajapakshe K., Park H., Wulfkuhle J., Young N. L., Li Y., Coarfa C., Edwards D. P., Huang S., High-throughput profiling of histone post-translational modifications and chromatin modifying proteins by reverse phase protein array. J. Proteomics 262, 104596 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 71.Coarfa C., Grimm S. L., Rajapakshe K., Perera D., Lu H.-Y., Wang X., Christensen K. R., Mo Q., Edwards D. P., Huang S., Reverse-phase protein array: Technology, application, data processing, and integration. J. Biomol. Tech. (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Legends for movies S1 and S2
Legend for data file S1
Figs. S1 to S6
Movies S1 and S2
Data file S1