Research ArticleCell Biology

Ablation of the Kinase NDR1 Predisposes Mice to the Development of T Cell Lymphoma

See allHide authors and affiliations

Science Signaling  15 Jun 2010:
Vol. 3, Issue 126, pp. ra47
DOI: 10.1126/scisignal.2000681

Abstract

Defective apoptosis contributes to the development of various human malignancies. The kinases nuclear Dbf2–related 1 (NDR1) and NDR2 mediate apoptosis downstream of the tumor suppressor proteins RASSF1A (Ras association domain family member 1A) and MST1 (mammalian Ste20-like kinase 1). To further analyze the role of NDR1 in apoptosis, we generated NDR1-deficient mice. Although NDR1 is activated by both intrinsic and extrinsic proapoptotic stimuli, which indicates a role for NDR1 in regulating apoptosis, NDR1-deficient T cells underwent apoptosis in a manner similar to that of wild-type cells in response to different proapoptotic stimuli. Analysis of the abundances of NDR1 and NDR2 proteins revealed that loss of NDR1 was functionally compensated for by an increase in the abundance of NDR2 protein. Despite this compensation, NDR1−/− and NDR1+/− mice were more prone to the development of T cell lymphomas than were wild-type mice. Tumor development in mice and humans was accompanied by a decrease in the overall amounts of NDR proteins in T cell lymphoma samples. Thus, reduction in the abundance of NDR1 triggered a decrease in the total amount of both isoforms. Together, our data suggest that a reduction in the abundances of the NDR proteins results in defective responses to proapoptotic stimuli, thereby facilitating the development of tumors.

Introduction

Apoptosis is a highly regulated form of programmed cell death that is important for developmental processes and tissue homeostasis. Defects in apoptosis result in the development of diseases such as autoimmunity and cancer (1). The apoptotic program is regulated by two distinct pathways: the extrinsic and the intrinsic (1). Whereas the extrinsic pathway is initiated upon stimulation of the members of the tumor necrosis factor (TNF) receptor family by extracellular ligands such as TNF-α or Fas ligand (FasL), the intrinsic pathway senses intracellular cues such as stress, deprivation of growth signals, or DNA damage, which result in the release of cytochrome c from the mitochondria. Both pathways activate caspases, a family of cysteine proteases, which cleave a wide range of cellular proteins and induce the distinct morphological changes that are associated with apoptosis, such as cell rounding, membrane blebbing, cytoskeletal disassembly, chromatin condensation, and DNA fragmentation (1).

The activities of caspases as central effectors of the apoptotic program are also either increased or decreased by kinase-mediated phosphorylation (2). Phosphorylation of caspase 2 by DNA-dependent protein kinase and of caspase 9 by c-Abl enhances their activities, whereas phosphorylation of caspases by kinases such as p38 mitogen-activated protein kinase, extracellular signal–regulated kinase, or protein kinase B (PKB; also known as Akt) suppresses their activities (2). In addition to being regulated by kinases, caspases themselves can cleave other kinases, which results in their activation in the case of Rho-activated kinase 1, hematopoietic progenitor kinase 1, and Abl, or inactivation, in the case of PKB, receptor-interacting kinase 1, and epidermal growth factor receptor (2). Activation of the Ste20-like kinase MST1 by caspase-mediated cleavage occurs in response to various proapoptotic stimuli (3). Among the substrates phosphorylated by activated MST1 upon the induction of apoptosis are members of the nuclear Dbf2–related (NDR) family of protein kinases (4, 5).

Four members of the NDR family exist in mammals: LATS1, LATS2, NDR1, and NDR2 (6). The NDR family members LATS1 and LATS2 function as tumor suppressors as part of the HIPPO pathway in Drosophila and mice (7, 8). Experiments performed in cell culture systems implicated the other mammalian NDR family members, NDR1 and NDR2, in the regulation of centrosome duplication, mitotic chromosome alignment, and induction of apoptosis (4, 9, 10). NDR1 and NDR2 display a high level of sequence identity, and so far no differences in their biochemical regulation or function have been described, although both kinases have distinct distribution patterns in mammals (11). Whereas NDR1 is highly abundant in organs of the immune system, especially in the thymus, NDR2 is mostly found in tissues of the gastrointestinal tract, which is suggestive of tissue-specific functions for each isoform. Given the predominance of NDR1 in lymphocytes and its suggested role in apoptosis, we generated mice lacking NDR1 to analyze its role in regulating apoptosis in these cells.

Here, we present evidence that NDR1 functions as a haploinsufficient tumor suppressor protein. Reduction in the expression of NDR1 was functionally compensated for by the increased abundance of NDR2. Despite this apparent compensation, complete loss or heterozygosity of NDR1 predisposed mice to the development of T cell lymphoma, which was accompanied by a reduction in the abundances of NDR1 and NDR2 proteins. Consistent with this finding, the abundance of NDR1 in samples of human T cell lymphomas was frequently reduced compared to that in normal human T cells, which suggests a conserved role for NDR1 as a tumor suppressor in preventing T cell lymphoma.

Results

Generation of NDR1-deficient mice

To obtain mice deficient in NDR1 (NDR1−/− mice), we targeted exon 4 of NDR1 (fig. S1A). Successful recombination was demonstrated by Southern blotting and polymerase chain reaction (PCR) assay (fig. S1, B and C). The absence of NDR1 protein was confirmed in the NDR1−/− mice by Western blotting analysis with an antibody against murine NDR1 (mNDR1) (fig. S1D). There was also a decrease in the amount of NDR1 protein in heterozygous mice (NDR1+/− mice) compared to that in wild-type (NDR1+/+) mice, which was suggestive of a gene dose effect. Because ablation of the gene encoding the respective NDR ortholog tricornered in Drosophila results in embryonic lethality (12), we analyzed variation in the genotype ratio among the progeny of matings between NDR1+/− mice (fig. S1E). Because the progeny showed the expected Mendelian ratio, we concluded that deletion of NDR1 did not lead to obvious developmental defects. In addition, the architectures and weights of various organs were similar in young NDR1+/+, NDR1+/−, and NDR1−/− mice. Thus, ablation of NDR1 in mice resulted in viable, fertile animals without obvious defects.

NDR1 is activated in thymocytes in response to different apoptotic stimuli

An earlier report described the activation of NDRs in transformed cells in response to FasL or TNF-α. Furthermore, short hairpin RNA (shRNA)–mediated knockdown of both isoforms results in increased resistance to apoptosis (4). To confirm these findings in untransformed cells, we treated freshly isolated thymocytes with stimuli to activate extrinsic and intrinsic apoptotic pathways (Fig. 1A). Two hours after induction of apoptosis with an antibody against Fas, there was an increase in the extent of phosphorylation of the hydrophobic motif in NDR (Thr444), which coincided with the activation of MST1, as assessed by detection of the cleavage and phosphorylation of MST1 (Fig. 1A). A similar activation of NDR was observed after γ-irradiation and treatment with dexamethasone (Fig. 1A). These results showed that NDR1 and NDR2 were activated in response to extrinsic as well as intrinsic apoptotic stimuli.

Fig. 1

NDR1 is activated in response to different proapoptotic stimuli but is largely dispensable for the induction of apoptosis. (A) Thymocytes were isolated from young wild-type (WT) mice (4 to 6 weeks old), and apoptosis was induced by γ-irradiation (γ-IRR), dexamethasone (Dex), or antibody against Fas (anti-Fas) in the presence of cycloheximide (CHX). Cells were lysed after the indicated incubation times and analyzed for the activation of NDR1, NDR2, and MST1 (double asterisks indicate the uncleaved form of MST1; a single asterisk indicates the cleaved form). A representative analysis of three independent experiments is shown. (B) Thymocytes were isolated and treated with γ-irradiation, dexamethasone, antibody against Fas in the presence of cycloheximide and etoposide, or antibodies against CD3 and CD28, and apoptosis was measured 10 hours later (or 24 hours later in the case of treatment with antibodies against CD3 and CD28) by staining with propidium iodide (PI) and Annexin V. The percentage of surviving cells was calculated as the ratio of PI and Annexin V double-negative cells with and without treatment. *P < 0.005; n = 3 independent experiments. (C) Isolated thymocytes (1 × 106 cells) were seeded into 24-well dishes containing IMDM and 10% FCS and analyzed after the indicated times with a Vi-CELL automated cell counter. *P < 0.005; n = 3 independent experiments. (D) Activation of NDRs in NDR1+/+ and NDR1−/− thymocytes. Freshly isolated thymocytes were treated with γ-irradiation and lysed after the indicated times. Activation of NDR was assessed by Western blotting analysis of lysates with an antibody against phosphorylated NDR (T444-P), and the total amount of kinase was analyzed with an antibody against NDR1 for WT thymocytes and an antibody against NDR2 for NDR1-deficient samples. A representative analysis of three independent experiments is shown.

Next, we tested whether loss of NDR1 could result in the increased resistance of thymocytes to apoptosis. Freshly isolated thymocytes from NDR1+/+, NDR1+/−, and NDR1−/− mice were treated with γ-irradiation, etoposide, dexamethasone, or antibody against Fas, and apoptosis was measured 10 hours later (Fig. 1B). Furthermore, thymocyte cell death upon treatment with antibodies against CD3 and CD28 or upon cytokine withdrawal was assessed (Fig. 1, B and C). No statistically significant differences in the numbers of cells that survived dexamethasone, antibody against Fas, or antibodies against CD3 and CD28 were detected (Fig. 1B). Although not statistically significant, NDR1+/− and NDR1−/− thymocytes displayed an increased tendency for survival after treatment with antibodies against CD3 and CD28 relative to that of NDR1+/+ thymocytes. However, we observed a statistically significant decrease in the number of NDR1+/+ thymocytes that survived exposure to γ-irradiation and etoposide relative to that of NDR1−/− cells (Fig. 1B; P < 0.005). Furthermore, NDR1−/− thymocytes showed increased survival after cytokine withdrawal relative to that of NDR1+/+ thymocytes (Fig. 1C). Treatments that resulted in the increased survival of NDR1−/− thymocytes compared to that of wild-type cells seemed to also result in the increased survival of NDR1+/− cells. In addition, mature T cells were tested for a defect in response to activation-induced cell death (AICD) and cytokine withdrawal (fig. S2). As before, loss of NDR1 (and in the case of cytokine withdrawal, also heterozygosity in NDR1) resulted in a slight but statistically significant resistance of cells to the induction of apoptosis relative to that of wild-type cells (P < 0.05). These results were confirmed by analyzing the activation of NDR in NDR1−/− thymocytes after the induction of apoptosis. We found that a delay in the activation of NDR in these cells was only observed after γ-irradiation (Fig. 1D and fig. S3). Together, these experiments showed that although NDR was activated in thymocytes upon treatment with several apoptotic stimuli, loss of NDR1 alone did not result in stronger resistance to apoptosis.

NDR2 compensates for the deficiency in NDR1

NDR1 and NDR2 exhibit >80% amino acid identity (11). Thus, the absence of developmental phenotypes and the overall normal apoptotic responses in NDR1+/− and NDR1−/− mice might be due to compensation by NDR2. The effects of NDR1 heterozygosity and deficiency on the abundance of NDR2 were analyzed in tissues in which NDR1 is highly expressed, such as the thymus, spleen, and lymph nodes, and in organs in which its expression is low, such as the colon (Fig. 2A and fig. S4). The abundance of NDR2 protein was increased upon knockout of NDR1 in a tissue-specific manner (Fig. 2A). Compared with a minor ~1.5-fold increase in the abundance of NDR2 in the colon, we found a ~2.5- to 3-fold increase in its abundance in tissues in which NDR1 was highly expressed (Fig. 2B). NDR2 was slightly increased in abundance in NDR1+/− mice relative to that in NDR1+/+ mice, whereas the amount of NDR1 protein was lower, reflecting the gene dosage effect observed earlier (fig. S1D). These changes, however, did not result from the increased expression of NDR2 or from an increase in the stability of NDR2 protein (fig. S5). Thus, NDR1 deficiency in mice appeared to be compensated for by an increase in the abundance of NDR2 protein through a posttranscriptional mechanism.

Fig. 2

Reduction in NDR1 gene dose is compensated for by the increased abundance of NDR2 protein. (A) Analysis of the abundances of NDR1 and NDR2 in representative tissues from NDR1+/+, NDR1+/−, and NDR1−/− mice with specific antibodies against NDR1 and NDR2. A representative analysis of three independent experiments is shown. (B) Quantification of changes in the abundances of NDR1 and NDR2 upon heterozygosity and deficiency of NDR1 (n = 3 independent experiments). Changes are expressed as fold changes in the ratio of NDR to Hsc70 relative to that in the WT. *P < 0.03; **P < 0.005. (C) Primary MEFs transfected with constructs expressing shLUC or shNDR2 were seeded for the induction of apoptosis 72 hours after transfection. Apoptosis was induced 24 hours later with etoposide (200 μM) or antibody against Fas (1 μg/ml) in the presence of cycloheximide (1 μg/ml). Apoptosis in GFP+ cells was measured by flow cytometry after 22 hours for treatment with etoposide and 18 hours for treatment with antibody against Fas. Results obtained were normalized to those of NDR1+/+-shLUC for better comparison of the two treatments *P < 0.03; **P < 0.005; n = 3. (D) Primary MEFs were transfected as described in (C), sorted for the presence of GFP, and further expanded for 48 hours. Apoptosis was induced with etoposide (200 μM) for 22 hours. A representative analysis of three independent experiments is shown. (E) MEFs were pretreated as described in (D), but apoptosis was induced with antibody against Fas (1 μg/ml) in the presence of cycloheximide (1 μg/ml) for 18 hours. A representative analysis of three independent experiments is shown. (F) Quantification of cleaved caspase 3 for the analysis shown in (D) and (E). Quantification of cleaved caspase 3 and actin after apoptosis induction was performed with LI-COR Odyssey technology. The ratio of the abundance of cleaved caspase 3 to that of actin was normalized to that of treated NDR1+/+-shLUC samples.

We next analyzed whether interfering with the observed compensation mechanism would result in decreased apoptosis. Early-passage NDR1+/+ and NDR1−/− mouse embryonic fibroblasts (MEFs) that expressed shRNA specific for NDR2 (shNDR2) or a firefly luciferase control (shLUC) were treated with either etoposide or antibody against Fas. Knockdown of NDR2 in NDR1−/− MEFs resulted in a significant (P < 0.005) decrease in the number of apoptotic cells relative to that of wild-type MEFs expressing shLUC (Fig. 2C and fig. S6). Although not consistently observed in thymocytes (Fig. 1, B and C), deficiency in NDR1 in MEFs resulted in increased resistance to both stimuli. This might indicate that the efficiency of the compensation mechanism was cell type–dependent or was disturbed upon ex vivo culture. However, knockdown of NDR2 in the NDR1-deficient background further reduced the occurrence of apoptosis in this setting (Fig. 2C). In addition, analysis of apoptotic extracts revealed that knockdown of NDR2 in NDR1-deficient MEFs resulted in substantially reduced cleavage of caspase 3 in response to etoposide and antibody against Fas (Fig. 2, D to F). These data showed that NDR2 functionally compensated for the reduced expression of NDR1 and for NDR1 deficiency in untransformed cells and that defects in this compensation mechanism resulted in increased resistance to apoptotic stimuli.

Aged NDR1+/− and NDR1−/− mice develop high-grade peripheral T cell lymphoma

Although our initial analysis revealed that loss of NDR1 was functionally compensated for by NDR2 in a gene dose–dependent manner, with only minor apoptotic defects resulting from NDR1 deficiency and heterozygosity, it seemed possible that targeting NDR1 could contribute to the development of age-related diseases. To this end, we analyzed aged, 17- to 27-month-old NDR1+/+, NDR1+/−, and NDR1−/− mice (Fig. 3A). About 70% of all of the NDR1+/− and NDR1−/− mice exhibited lymphomas in the various tissues analyzed (Fig. 3B and table S1). Female mice were more prone to lymphomas (80%) than were male NDR1+/− mice (25%) and NDR1−/− mice (45%) (Fig. 3B). Immunophenotyping defined the lesions as high-grade peripheral T cell lymphomas, with all tumors containing cells that had CD3, a marker for T cells (Fig. 3C). Flow cytometric analysis of T cell lymphomas from aged NDR1+/− and NDR1−/− mice revealed that the lesions were characterized either by the infiltration of CD4+CD8+ double-positive (DP) cells into peripheral immunological organs or by the expansion of CD4+ single-positive (SP) cells (Fig. 3D). Thus, deficiency or heterozygosity in NDR1 appeared to predispose mice to the development of T cell lymphoma later in life, which suggested that NDR1 acted as a haploinsufficient tumor suppressor for T cell lymphoma.

Fig. 3

Aged NDR1+/− and NDR1−/− mice develop high-grade peripheral T cell lymphoma. (A) Tumor spectrum in aged NDR1+/+, NDR1+/−, and NDR1−/− mice (18 to 27 months of age). Mice were dissected, and H&E-stained tissue sections were analyzed for the development of tumors. (B) Bar chart representing the rates of lymphoma detected in aged NDR1+/+, NDR1+/−, and NDR1−/− mice. In addition, the gender-specific rates of lymphoma development are shown. Note that female mice seemed to be more prone to develop lymphoma than were male mice. The numbers of total mice used for the analysis are given in the lower panel. (C) Example of immunohistochemical characterization of identified lymphatic lesions. Tumor cells (indicated by arrows) that infiltrated kidney tissue were stained with antibodies against Pax5 and CD3 to discriminate between B cells and T cells. (D) Further characterization of T cell lymphoma. Single-cell suspensions from the lymph nodes of aged animals were incubated with antibodies against CD4 and CD8 and analyzed by flow cytometry. Examples are given of a normal flow cytometry profile from an aged NDR1−/− mouse (left panel), of the infiltration of CD4+CD8+ DP cells into lymph nodes (middle panel), and of the expansion of the CD4+ SP population (right panel).

NDR1+/− and NDR1−/− mice are highly susceptible to carcinogen-induced lymphomagenesis

To examine whether young mice show increased susceptibility to carcinogen-induced development of lymphoma, we chose to use N-ethyl-N-nitrosourea (ENU) as a carcinogen in our experiments because it is thought to induce mainly T cell lymphomas in the C57BL/6 background (13). Mice that were 4 to 5 weeks old were injected intraperitoneally with a single dose of ENU (100 mg/kg) and monitored for a period of 9 months for the development of cancer. Nine months after injection, 88% of NDR1−/− mice and 79% of NDR1+/− mice developed tumors relative to only 50% of wild-type mice (Fig. 4A and table S2). Analogous to the case in aged mice, heterozygosity in NDR1 resulted in a rate of tumor development similar to that caused by deficiency in NDR1. We observed mainly hematopoietic tumors, with T cell lymphoma being the most frequent. In addition, myeloproliferative diseases (MPDs) were observed at later time points (table S2). In terms of T cell lymphomas, 53% of the ENU-treated NDR1−/− mice and 47% of the ENU-treated NDR1+/− mice developed tumors. In contrast, only 30% of wild-type mice developed T cell lymphomas after treatment with ENU (Fig. 4B), as reported previously (13, 14). NDR1+/− and NDR1−/− animals showed not only an increased penetrance of disease, relative to that of wild-type mice, but also an earlier onset in the development of T cell lymphoma after treatment with carcinogen; tumor development in NDR1−/− mice began ~2 months earlier than that in wild-type mice (Fig. 4B). The T cell lymphomas were characterized by a massive increase in the size and weight of the thymus (Fig. 4C). Infiltration of other organs, such as lymph nodes, spleen, and kidneys, which led to disruption of normal tissue architecture, was frequently observed (Fig. 4D). Flow cytometric analysis of the tumors showed that they were characterized by CD4+CD8+ DP cells, thereby confirming that the tumors originated from cells of the T cell lineage (Fig. 4E). Thus, although the development of precursor, but not peripheral, T cell lymphoma occurred after treatment with ENU, NDR1+/− and NDR1−/− mice had a higher susceptibility to carcinogen-induced lymphomagenesis than did wild-type mice.

Fig. 4

NDR1+/− and NDR1−/− show a high degree of susceptibility to ENU-induced lymphomagenesis (A) Kaplan-Meier tumor-free survival curves of ENU-treated mice starting after injection with ENU (100 μg/g) at 4 weeks of age and followed for up to 10 months. (B) Kaplan-Meier lymphoma-free survival curves of ENU-treated mice. Fatal thymic T cell lymphoma occurred at the highest frequency and at the earliest time points. Other hematopoietic malignancies occurred at later time points (fig. S1 and table S2). (C) Examples of T cell lymphoma from an ENU-treated NDR1−/− mouse (middle panel) and an NDR1+/− mouse (lower panel). For comparison, a normal untreated thymus from a WT mouse is shown (upper panel). (D) Example of macroscopic (upper left panel) and microscopic infiltration of kidney tissue. Sections from affected organs were stained with H&E and analyzed for infiltrating tumor cells. Magnification: ×20 (right panels), ×40 (lower left panel). For comparison, an H&E-stained section from an unaffected WT kidney is shown (lower right panel). (E) CD4 and CD8 profiles from thymus (upper panel), spleen (middle panel), and lymph nodes (lower panel) of ENU-treated NDR1−/− mice without T cell lymphoma (left) or with fatal T cell lymphoma (right). Note that CD4+CD8+ DP cells infiltrated the spleen and lymph nodes.

MPDs were observed in addition to T cell lymphomas (14). As before, NDR1+/− and NDR1−/− mice were more prone to these diseases (32% and 35%, respectively) than were wild-type mice (20%) (fig. S7A). Some mice affected by MPDs showed massive splenomegaly (fig. S7B). Analysis of these lesions by flow cytometry showed an increase in the number of cells that belonged to neither the T cell nor the B cell lineage. This population was instead composed of erythroid (Ter119+/CD71+) and myeloid (Gr-1+/Mac1+) cells (fig. S7C). In summary, loss of or heterozygosity in NDR1 predisposed younger mice to the development of T cell lymphoma and MPDs after treatment with carcinogen, which further confirmed that NDR1 functioned as a haploinsufficient tumor suppressor.

The development of lymphoma is associated with a decrease in the abundances of NDR1 and NDR2

The apparent gene dose effect of NDR1 on the abundance of NDR2 protein and the predisposition of NDR1-deficient mice to the development of T cell lymphoma prompted us to examine the amounts of NDR1 and NDR2 proteins in tumors from ENU-treated mice. Because changes in the abundance of NDR2 protein upon the loss of NDR1 were not reflected at the level of NDR2 messenger RNA (mRNA) (fig. S5), our analysis focused on the abundances of NDR proteins through the use of antibodies against NDR1 and NDR2. Thymocyte extracts from healthy and untreated NDR1+/+, NDR1+/−, and NDR1−/− mice were used as controls (Fig. 5A). We used the loss of detectable p27 protein as an indicator of tumor formation (15, 16). Comparison of the amounts of NDR1 and NDR2 in tumors and in normal thymocytes of the respective genotypes showed in most cases that both kinases were decreased in abundance relative to that in untransformed thymocytes (Fig. 5A). For example, in NDR1-deficient tumors, there was a marked decrease in the abundance of NDR2 protein; only one of seven tumors retained NDR2 protein in amounts similar to that found in untransformed NDR1−/− thymocytes (Fig. 5A). The same was true for most NDR1+/+ and NDR1+/− tumors (Fig. 5A). We also determined the amounts of NDR1 and NDR2 in tumor material from MPDs obtained during the ENU experiment (Fig. 5B). As before, there was a reduction in the total amount of NDR proteins in the tumor samples compared to that in normal splenocytes; however, in tumors from NDR1+/+ and NDR1+/− mice, it was the abundance of NDR1 that was mainly affected. In samples from mice lacking NDR1, there was a striking reduction in the amount of NDR2 protein (Fig. 5B). In summary, most murine tumors showed a reduction in the abundance of at least one of the NDR isoforms, which suggested that the development of tumors was accompanied by a reduction in the total amount of NDR protein.

Fig. 5

NDR1 and NDR2 proteins in murine and human tumors. (A) Analysis of the abundance of NDR1 and NDR2 in tumor material from ENU-treated NDR1+/+ (upper panel), NDR1+/− (middle panel), and NDR1−/− (lower panel) mice. Single-cell suspensions from normal thymi of the indicated genotype or tumors were lysed and analyzed by Western blotting for the presence of NDR1 and NDR2 with isoform-specific antibodies, with actin as a loading control. Note that the abundances of both isoforms were decreased in most of the tumors compared with those in normal thymocytes of the respective genotypes. The samples were also analyzed by Western blotting for the presence of p27, whose loss is indicative of a tumor. (B) NDR1 and NDR2 proteins in MPD samples obtained from NDR1+/+ (upper panel), NDR1+/− (lower panel), and NDR1−/− mice. Single-cell suspensions from spleens were lysed and analyzed by Western blotting for the presence of NDR1 and NDR2, with actin as loading control. (C) NDR1 and NDR2 proteins in human T cell lymphoma samples. Tumor tissue (malignant T cells >70% as judged by histopathology) was extracted and analyzed by Western blotting for the presence of NDR1 and NDR2 with specific antibodies. Purified CD3+ human T cells purified from whole blood were used as a control.

To determine the relative amounts of NDR1 and NDR2 in samples of human T cell lymphoma with different characteristics (table S3), we analyzed tumor biopsies with >70% malignant T cells or purified CD3+ human T cells by Western blotting with specific antibodies for the presence of NDR1 and NDR2 (Fig. 5C). Most human tumors showed a substantial decrease in the abundance of NDR1 relative to that in normal human T cells, similar to the situation observed in murine MPD samples. Thus, the total amount of NDR protein was reduced in samples of murine and human T cell lymphomas, which suggested a conserved role for NDRs in tumor biology.

The reduced amounts of NDR proteins in tumors correlate with changes in the expression of genes implicated in the regulation of cell death

We found that tumor development was associated with the decreased abundance of the NDR isoforms in most samples (Fig. 5A). Therefore, we categorized the tumors according to their abundance of NDR protein into three groups—NDR-high, NDR-middle, and NDR-low—and performed microarray analysis (table S4). We found 46 genes that showed at least a 1.5-fold change (P < 0.05) in expression in NDR-low tumors relative to that in normal thymocytes, NDR-high tumors, and NDR-middle tumors (Fig. 6A and table S5). The correlations between the abundance of NDR isoforms in tumors and changes in gene expression were validated by analyzing the abundance of Pou2af1, because Pou2af1 mRNA was highly abundant in NDR-low tumors relative to that in other samples (fig. S8).

Fig. 6

The low abundances of NDR1 and NDR2 in tumors correlates with changes in the expression of genes associated with cell death. (A) Expression profiles of genes that were changed at least 1.5-fold (P < 0.05) in NDR-low tumors (table S4; red arrows indicate selected genes discussed in the text). (B) Functional analysis of genes associated with the low abundance of NDR in tumors. The list of genes obtained in (B) was imported into the Ingenuity Pathways Analysis program, and functional analysis was performed. Functions with a −log(P) of >3 are shown. (C) Analysis of Ki67 and TUNEL staining in NDR-high and NDR-low tumors. Colocalization (red) of DAPI (4′,6-diamidino-2-phenylindole; blue) and TUNEL (green) was analyzed with Imaris imaging software. (D) Quantification of TUNEL-positive cells in NDR-high tumors (n = 2 tumors) and NDR-low tumors (n = 3). For each sample, 27 fields of view were analyzed. **P < 0.001. (E) HeLa cells transfected with constructs encoding inducible shRNA against NDR1 and NDR2 were treated with tetracycline (TET) for 72 hours and transfected with siRNA against MEF2c (siMEF2c) or control siRNA (siCon). Twenty-four hours later, cells were treated with etoposide (100 μM) for 48 hours and analyzed by staining for Annexin V and PI. *P < 0.002; n = 3.

Application of the Ingenuity Pathways Analysis program to functionally annotate the genes that correlated with a low abundance of NDR protein revealed an enrichment of genes implicated in the regulation of cell death and the cell cycle (Fig. 6B). Indeed, we observed the increased expression of genes whose products exert antiapoptotic functions, including CD5L, MEF2c, and Pou2af1 (1719). Furthermore, genes such as CDKN2C, TOP2A, and CCND3, whose products exert proapoptotic functions (2023), were decreased in expression. Because we identified genes whose products are implicated in regulation of the cell cycle as correlating with the low abundance of NDR in tumors, we tested whether loss of NDR1 had any effect on the proliferation of T cells (fig. S9). Loss of NDR1 resulted in reduced proliferation of T cells after stimulation with antibody against CD3 alone or in combination with antibody against CD28; however, stimulation with a higher concentration of antibody against CD3 abolished this effect. Next, we attempted to confirm our findings in NDR-high and NDR-low tumor samples (Fig. 6C and fig. S10). We observed no major differences in Ki67 staining between NDR-high and NDR-low tumors, suggesting that proliferation was similar in cells from both samples; however, the numbers of apoptotic cells were significantly decreased in NDR-low tumors relative to that in NDR-high tumors (Fig. 6D; P < 0.001), confirming that cells with a decreased abundance of NDR had increased resistance to apoptosis, as observed earlier.

An increase in the abundance of MEF2c correlates with increased resistance to apoptosis (19). Therefore, we tested whether knockdown of MEF2c by small interfering RNA (siRNA) could rescue the apoptosis defects in HeLa cells in which both NDR1 and NDR2 have been knocked down (Fig. 6E). Decreasing the abundance of MEF2c in cells depleted of NDR1 and NDR2 resulted in a slight, but statistically significant, rescuing effect in response to etoposide but not after Fas-induced apoptosis (fig. S11), indicating that additional factors mediated the effects downstream of NDR. Together, these data suggest that the low abundance of NDR proteins in tumors correlated with changes in the expression of genes whose products are implicated in cell death and proliferation. Further analysis showed that the apoptotic response of cells, and not their proliferation, seemed to be mainly affected by the low abundance of NDR in tumors, thereby fully confirming the results obtained earlier.

A decrease in the abundance of E47 mediates resistance to apoptosis in the context of decreased amounts of NDR protein

To further understand the effect of NDR proteins on the expression of genes that correlate with the low abundance of NDR in tumors, we performed an analysis to identify common upstream factors that regulated gene expression in our settings (Fig. 7A). Targets of E2A (E12/E47) were enriched in our data set. E2A encodes two basic helix-loop-helix transcription factors, which are important for the regulation of the development, proliferation, and survival of lymphocytes (24). Loss or functional inactivation of E2A results in the development of T cell lymphoma and leukemia in mice and humans (2527). We therefore tested whether the abundance of E2A protein was reduced in tumors in which the abundance of NDR proteins was reduced. Only tumors with a normal abundance of NDR retained a high abundance of the E2A isoform E47, whereas its abundance was substantially reduced when the abundance of NDR was reduced, and it was even absent in ~40% of the NDR-low tumors (Fig. 7B). To validate the effect of diminished amounts of NDR protein on the abundance of E47, we analyzed the amount of E47 in HeLa cells stably transfected with validated shRNA against NDR1 and NDR2 (Fig. 7C). Indeed, the abundance of E47 was reduced upon knockdown of NDR1 and NDR2 in these cells (Fig. 7C). Finally, we attempted to rescue the apoptosis defects in NDR-depleted cells by increasing the abundance of E47 (Fig. 7, D and E). Indeed, the increased abundance of E47 by transfection of cells rescued the apoptotic defects to ~50% in cells in which NDR1 and NDR2 were knocked down after treatment with both etoposide and antibody against Fas (Fig. 7, D and E). Together, these results suggest that the reduced abundances of NDR proteins in tumors and cells affected the abundance of the E2A product E47. Restoring the abundance of E47 reduced the apoptosis defects after treatment with etoposide and antibody against Fas, thereby giving mechanistic insight into the tumor-suppressive function of NDR proteins.

Fig. 7

Reduction of the abundances of NDRs results in decreased amounts of E47 in tumors and in cells, which leads to resistance to apoptosis. (A) Analysis of common upstream regulators of correlating genes. Through use of the Ingenuity Pathways Analysis program Knowledge Base, upstream regulators of the expression of each correlating gene and the total known regulated genes were extracted. P values were calculated by hypergeometric distribution. Regulators with a −log(P) of >3 are shown. (B) Analysis of the abundance of E47 in ENU-induced tumors. (C) Reduced abundance of E47 upon knockdown of NDR1 and NDR2 in HeLa cells. Knockdown of NDR1 and NDR2 in HeLa cells stably expressing specific shRNAs was induced for 4 days by tetracycline. (D) Rescue of the apoptotic defects in NDR1- and NDR2-depleted HeLa cells by the increased abundance of E47. Knockdown of NDR1 and NDR2 in HeLa cells was induced for 72 hours, and cells were transfected with plasmid encoding E47 coupled to an IRES-GFP to monitor transfection. Apoptosis was induced 24 hours later with etoposide (100 μM) for 48 hours. Apoptosis was assessed by staining cells with Annexin V and PI and gating on GFP-positive cells. *P < 0.005; n = 3. (E) HeLa cells were pretreated and transfected as described in (D). Twenty-four hours later, cells were treated with antibody against Fas (0.5 μg/ml) in the presence of cycloheximide (10 μg/ml) for 6 hours and apoptosis was assessed as described in (D). **P < 0.025; n = 3.

Discussion

The NDR family members NDR1 and NDR2 are implicated in the regulation of apoptosis downstream of Fas (4). We generated mice deficient in NDR1 to further analyze the role of NDR1 in regulating signaling processes that induce apoptosis, mainly in lymphocytes, in which NDR1 is highly abundant. We showed that NDR1 in thymocytes was activated not only upon stimulation of cells with antibody against Fas but also after the induction of apoptosis by intrinsic stimuli, such as DNA damage and dexamethasone. This finding is suggestive of a broader role for NDR1 in apoptosis and shows that the activation of NDR1 is not limited to the stimulation of death receptors. However, although induction of apoptosis results in the activation of NDR, loss of NDR1 did not result in major defects in apoptosis in thymocytes and T cells. In conditions in which loss of NDR1 resulted in apoptosis defects, heterozygosity in NDR1 also increased resistance to apoptosis (Fig. 1 and fig. S2). By analyzing the abundance of the second NDR kinase isoform in mammals, NDR2, we found that NDR2 was increased in abundance in tissues from NDR1−/− and NDR1+/− mice. Interfering with this increase in NDR2 abundance resulted in increased resistance to apoptosis. Thus, NDR2 functionally compensated for the loss of NDR1, which indicated a tight, gene dosage–dependent regulation of the total abundance of NDR proteins in vivo.

Although the loss of NDR1 was compensated for in healthy young mice, we still observed that loss of NDR1 predisposed mice to the development of T cell lymphoma, both in older mice and in younger mice that were treated with carcinogen (Figs. 3 and 4). Given that the abundance of NDR1 protein was sensitive to gene dose, it was not surprising that NDR1+/− mice displayed a similar predisposition to tumor development, indicating that NDR1 functions as a haploinsufficient tumor suppressor to prevent T cell lymphoma (28). Indeed, under several conditions, the loss of one allele of NDR1 resulted in apoptosis defects comparable to those observed in NDR1−/− cells (Fig. 1 and fig. S2). Although the decrease in the abundance of NDR1 was compensated for by an increase in the amount of NDR2, our experiments showed that, at least in terms of apoptosis, the compensation was not always complete. Under certain conditions, loss or heterozygosity of NDR1 resulted in increased resistance to the induction of apoptosis. Could the increased tumor development observed in NDR1+/− and NDR1−/− mice be explained by a predisposition to decrease the abundance of NDR2 and thereby increase resistance to proapoptotic stimuli?

To address this hypothesis, we analyzed tumor material from NDR1+/+, NDR1+/−, and NDR1−/− mice for the abundances of NDR1 and NDR2. In T cell lymphoma and MPD samples from mice deficient in NDR1, the abundance of NDR2 was strongly decreased relative to the normal amount of NDR2 in NDR1-deficient cells (Fig. 5). This indicated that tumor development was accompanied by defects in the ability of cells to compensate for the loss of NDR1. T cell lymphoma samples from NDR1+/− mice also displayed a decrease in the abundance of NDR2 in most of the cases analyzed. The development of T cell lymphoma in these mice was also accompanied by a decrease in the abundance of NDR1 (Fig. 5). Even in tumors obtained from wild-type mice, a decrease in the amounts of NDR1 and NDR2 proteins was observed in most cases. Thus, we can conclude that the development of T cell lymphoma in mice is accompanied by a decrease in the abundances of both NDR isoforms.

Given that, in healthy tissue, the reduction in the abundance of NDR1 through the loss of one allele of NDR1 was already compensated for by an increase in the amount of NDR2, one might argue that the development of T cell lymphoma was accompanied by a decrease in total amount of NDR proteins. From this point of view, a reduction in the amount of one isoform without an increase in the amount of the other would yield a reduction in the total amount of NDR proteins and would thereby facilitate defects in the response to proapoptotic stimuli. In line with this interpretation, the abundance of NDR1 was substantially reduced without an accompanying increase in the amount of NDR2 in 75% of the human T cell lymphoma samples that we analyzed. Thus, the development of human T cell lymphoma is also associated with a decrease in the abundance of NDR proteins. In addition, the abundance of NDR proteins was reduced in MPD samples, arguing for a tumor-suppressive function of NDRs in different cell types and organs. Together, these data suggest that the development of T cell lymphoma in mice and humans is accompanied by a reduction in total amounts of NDRs. Loss of one allele of NDR1 apparently predisposes mice to the development of tumors by further decreasing the abundances of NDR1 and NDR2. Other studies have shown that predisposition to tumorigenesis can be strongly affected by even small changes in the expression and protein abundance of tumor suppressors (2931), which, in the case of death-associated protein kinase 1 in chronic lymphocytic leukemia, affects the responses of tumor cells to proapoptotic stimuli (29). A decrease in the abundance of NDR could therefore similarly predispose to the development of tumors by affecting the responses of cells to apoptotic stimuli.

Evading apoptosis is considered to be one of the major hallmarks of cancer (32). We showed here that NDR proteins were activated not only after extrinsic death signals but also after intrinsic signals such as DNA damage. A reduction in the amount of total NDR proteins in untransformed cells resulted in the increased resistance of these cells to extrinsic and intrinsic apoptotic stimuli. Consistent with this observation, tumors in which the abundances of both NDR isoforms were reduced exhibited a specific gene expression profile associated with cell death. Furthermore, this gene expression profile was enriched in genes whose products are implicated in the regulation of the cell cycle. However, NDR1-deficient T cells did not exhibit increased proliferation after stimulation. Defects in proliferation upon knockout of genes that encode several well-known tumor suppressor proteins, such as BRCA1 (breast cancer 1, early onset), SMAD4 (mothers against decapentaplegic homolog 4), and VHL (von Hippel-Lindau), have been described (3335). In addition, whereas tumors with a low abundance of NDR showed a substantial decrease in the number of apoptotic cells, proliferation in these tumors seemed to be minimally affected. During tumor development, these cells probably acquire changes in other signaling pathways to overcome the defects in proliferation that occur upon loss of NDR1. Still, it will be interesting for future studies to address in more detail the effect of NDR1 deficiency on proliferation.

The genes that correlated with the low abundances of NDR1 and NDR2 in T cell lymphoma were enriched for targets of the E2A products E12 and E47. Deficiency in E2A results in the development of T cell lymphoma in mice (25, 27). In addition, functional inactivation of E47 has been described in human T cell acute lymphoblastic leukemia/lymphoma (26). Reconstitution of E2A-deficient lymphoma cells with E12 or E47 results in cell death (36). Strikingly, knockdown of NDR1 and NDR2 in HeLa cells resulted in the decreased abundance of E47, which was also observed in the tumors that we analyzed. An increase in the abundance of E47 in HeLa cells in which NDR1 and NDR2 were knocked down substantially rescued the apoptosis defects after treatment with both etoposide and antibody against Fas, confirming the functional link between NDRs and the abundance of E47 in apoptosis. A reduction in the amounts of NDR1 and NDR2 thus may result in the resistance of tumors to apoptosis at least partially by affecting the abundance of E47 protein. Future studies are needed to further define the effects of NDR1 and NDR2 on the abundance of E47.

The data presented here suggest that NDRs have a tumor-suppressive function. Indeed, other studies have reported the loss of heterozygosity of the genomic region containing NDR1 in diffuse large B cell non-Hodgkin’s lymphoma, cervical cancer, colorectal carcinoma, lung cancer, ovarian cancer, and renal cell carcinoma (3742); however, no deletions of the NDR2 locus have been reported so far. In addition, loss or reduction in the abundance of NDR1 or NDR2 mRNAs has not been associated so far with the development of T cell lymphoma in humans (43). This might be explained by our results, because we observed changes in NDR1 and NDR2 mainly at the level of protein. First, loss of NDR1 did not result in the increased abundance of NDR2 mRNA. Second, although we observed a reduction in the abundances of NDR1 and NDR2 proteins in tumors, NDR1 and NDR2 did not appear as hits in the gene expression analysis. Therefore, it will be relevant in future studies to develop immunohistological methods to analyze the abundances of NDR1 and NDR2 proteins in tumors.

In conclusion, we have identified NDR1 as a tumor suppressor protein that prevents T cell lymphoma. NDR1 shows features of haploinsufficiency, because loss of one allele of NDR1 predisposed mice to the development of T cell lymphoma. Loss or heterozygosity of NDR1 resulted in the increased resistance of thymocytes and T cells to apoptosis under several conditions. Furthermore, our analysis showed that the loss of NDR1 triggered a reduction in the abundance of the compensating NDR2 isoform by an as yet unidentified posttranscriptional mechanism, which resulted in a decrease in the total amounts of NDR proteins in tumors. Given that NDR1 is expressed predominantly in lymphocytes and that its expression is frequently reduced in T cell lymphomas of murine and human origin, we suggest that the abundances of NDR1 (and NDR2) are likely to be reduced in other lymphoid neoplasias as well. Future studies that address the abundance of NDR1 (and NDR2) in tumors of different, also nonlymphoid, origins will therefore broaden our understanding of the role of NDRs in counteracting tumor development.

Materials and Methods

Generation of NDR1-deficient mice and induction of lymphoma

An ~9-kb Bam HI–Not I fragment containing exons 4, 5, and 6 of NDR1 was amplified from bacterial artificial chromosome clone 25140 (244/G01) (Incyte Genomics) with Expand Long Template Taq polymerase (Roche) and was subcloned into pMCS5. A 5-kb IRES/lacZ/Neo cassette was inserted into the Xho I site of exon 4. 129/Ola embryonic stem (ES) cells were electroporated with the targeting vector after it was linearized with Sal I. An external probe was used to screen Southern blots of ES cell extracts after digestion with Kpn I. An internal probe and a lacZ-Neo probe were used to characterize ES clones that had undergone homologous recombination. Correctly targeted ES cells were used to generate chimeric mice. Male chimeras were mated with wild-type C57BL/6 females to obtain NDR1+/− mice. NDR1+/− mice were backcrossed for at least four generations with pure C57BL/6 mice. The progeny of NDR1+/− intercrosses were genotyped by multiplex PCR with three primers: (i) Ex4checkb, 5′-GTCTTCTCATCGCTGTCACAGCT-3′; (ii) Neo-2, 5′-GCTGCCTCGTCCTGCAGTTCATTC-3′; and (iii) 6540bk, 5′-GCTCCCGCTCAGTTACCTGCTCC-3′. To induce the development of lymphoma, we injected 4-week-old female NDR1+/+, NDR1+/−, and NDR1−/− mice intraperitoneally with a single dose of ENU (100 mg/kg) dissolved in phosphate-buffered saline (pH 6.0) and monitored them for up to 9 months after injection. All mouse experiments were performed according to the Swiss Federal Animal Welfare Law.

Reagents and antibodies

Dexamethasone and ENU were from Sigma. Antibodies against CD3e and CD28 were from eBioscience; antibodies against CD4, CD8, and B220 were obtained from Immunotools; antibodies against p27 and actin were from Santa Cruz Biotechnology; antibodies against E47, cleaved poly(adenosine diphosphate-ribose) polymerase, and Fas (Jo-2) were obtained from BD Biosciences; antibodies against cleaved caspase 3, MST1, and phosphorylated MST were from Cell Signaling; and antibody against Hsc70 was from Stressgen. Antibodies against pThr444/442 NDR1/2 (T444-P), human NDR1, and NDR2 have been described elsewhere (4, 44). To obtain antibodies against mNDR1, we used a peptide corresponding to the C-terminal part of mNDR1 (ILKPTVTTSSHPETDYKNKD) to immunize rabbits (45). Antibodies were subjected to immunoaffinity purification and tested for their specificity (fig. S12). Antibodies against Pou2af1 were a gift from P. Matthias and have been described elsewhere (46).

Cell culture and transfections

HeLa, NIH 3T3, and IMCD3 cells were maintained in Dulbecco’s modified Eagle’s medium containing 10% fetal calf serum (FCS) and antibiotics. The generation of HeLa cells expressing shRNA against NDR1 and NDR2 has been described previously (4). To induce knockdown of NDR1 and NDR2, we treated cells with tetracycline for 72 hours. Cells were transfected with Lipofectamine 2000 (Invitrogen), as described by the manufacturer. Validated siRNA against MEF2c was obtained from Ambion, and control siRNA was from Qiagen (fig. S13A). Complementary DNA (cDNA) encoding E47 was a gift from C. Gallegos (University of Lleida, Lleida, Spain), and constructs encoding E47 with an internal ribosomal entry site–green fluorescent protein (IRES-GFP) were obtained by subcloning E47 into a pcDNA3 vector that contained an IRES-GFP cassette cut with Bam HI and Eco RI (fig. S13B).

Protein extraction and Western blotting analysis

Proteins were extracted from freshly isolated or cultured cells as previously described (47). To extract proteins from frozen tissues, we minced flash-frozen, freshly isolated organs using a tissue homogenizer with 6 μl of lysis buffer per milligram of tissue. Extracts were incubated for 1 hour at 4°C, and cellular debris were removed by centrifugation at 16,000g for 15 min at 4°C. Western blotting analysis was performed as described previously (47), except that blots analyzed with the LI-COR Odyssey System were incubated with secondary antibodies that were coupled to fluorescent dyes. Quantifications were performed with the LI-COR Odyssey software.

Apoptosis assays

For the analysis of thymocyte apoptosis, freshly isolated thymocytes from 4- to 6-week-old mice were isolated and 1 × 106 cells were seeded into 48-well plates. Apoptosis was induced by γ-irradiation (TORREX 120D, Astrophysics Research Corp.; 5 mA/120 kV and 0.13 gray/s), etoposide, dexamethasone, antibodies against CD3 and CD28, or antibody against Fas in the presence of cycloheximide. The response to cytokine withdrawal was assessed by seeding the cells in Iscove’s modified Dulbecco’s medium (IMDM) containing 10% FCS for the indicated time and analyzed with a Vi-CELL automated cell counter. For the induction of apoptosis in MEFs, cells were seeded at consistent densities 24 hours before treatment. Apoptosis was induced by treating cells with etoposide or antibody against Fas in the presence of cycloheximide. Apoptosis was assessed either by Annexin V staining (BD Biosciences) or with a cationic lipophilic dye DilC1(5) assay kit (Invitrogen). Cells were analyzed by flow cytometry. Assays to assess AICD were performed as described previously (48). In short, T cells were purified from spleens with T cell enrichment columns (eBioscience) and activated for 72 hours on plates coated with antibody against CD3 (5 μg/ml) in the presence of soluble antibody against CD28 (1 μg/ml). Activated viable cells were obtained with Lympholyte-M (Cedarlane Labs) and replated for 20 hours on plates coated with increasing concentrations of antibody against CD3 in the presence of interleukin-2 (50 U/ml; Immunotools). Analysis of AICD was performed as described previously (48, 49).

Retrovirus-mediated knockdown of NDR2

To knock down NDR2 in MEFs, we inserted oligonucleotides targeting murine NDR2 into pTER (50) and tested them in transient transfections of IMCD3 cells (fig. S14). shNDR2#13 and shLUC (sequence provided upon request) were cloned into the pSUPER-retro.gfp.neo vector (OligoEngine) and used for further experiments. For virus production, Phoenix-Eco cells were transfected with jetPEI transfection reagent (Polyplus Transfection), and virus was harvested 48 and 72 hours later. Virus-containing supernatant was used to spin-infect MEFs or NIH 3T3 cells in the presence of polybrene (5 μg/ml; 1000g at 30°C for 1 hour). Medium was changed after 6 hours, and the cells were left to recover for 48 hours.

Flow cytometry

Single-cell suspensions from lymphatic organs or tumor tissue were obtained by pressing the organ through a 70-μm nylon mesh. Resulting suspensions were depleted of erythrocytes by incubation in Gey’s solution and subsequently incubated with antibodies covalently conjugated to fluorescein isothiocyanate, phycoerythrin, or allophycocyanin fluorophores. These included antibodies against CD3e, Gr-1, Ter119, and CD71, which were obtained from BD Biosciences; antibodies against CD4, CD8, and B220, which were from Immunotools; and antibody against Mac1, which was from SouthernBiotech. Flow cytometric analysis was performed with a FACSCalibur cell analyzer (BD Biosciences). Cell sorting was performed with a MoFlo device (DakoCytomation).

Histopathological analysis

Tissue specimens were fixed in 4% neutral buffered formalin and embedded in paraffin. Paraffin sections (4-μm thick) were stained with hematoxylin and eosin (H&E) or with antibodies against Pax5 (BD Biosciences), CD3 (Dako), or Ki67 (NeoMarkers). Terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL) staining was performed with the ApoAlert DNA Fragmentation Assay kit (Clontech) as described by the manufacturer. Acquired images were analyzed with Imaris imaging software.

T cell proliferation assays

T cells were purified from spleens with T cell enrichment columns (eBioscience) and labeled with carboxyfluorescein diacetate succinimidyl ester as described by the manufacturer (eBioscience). Labeled cells were activated for 72 hours on plates coated with antibody against CD3 in the presence of soluble antibody against CD28 and subsequently analyzed by flow cytometry.

Isolation of RNA and quantitative reverse transcription PCR

Total RNA from flash-frozen organs or tumor cells was isolated with TRIzol reagent (Invitrogen) and further purified with the RNeasy kit (Qiagen). cDNA from samples was generated from 2 μg of total RNA with M-MuLV reverse transcriptase (New England Biolabs) and oligo(dT) primers. Quantitative reverse transcription PCR to detect mNDR2 (primer sequences upon request) was performed with SYBR Green technology in an ABI Prism 7000 detection system (Applied Biosystems).

Microarray analysis

RNA extracted as described earlier was processed and hybridized onto Affymetrix 430v2 chips as described by the manufacturer. Data were analyzed with Expressionist (GenData AG). Normalized data were analyzed by one-way analysis of variance (ANOVA) (P < 0.05). Only those changes in expression greater than 1.5-fold were analyzed. Microarray data have been stored in the Gene Expression Omnibus (GEO) database under accession number GSE21902.

Statistical analysis

Statistical analyses were performed by the Student’s t test for comparison between two samples. ANOVA was used for the analysis of microarray data.

Acknowledgments

Acknowledgments: We thank L. Quintanilla-Fend (GSF, Munich, Germany) for help with the antibodies used for immunohistochemistry; A. Wodnar-Filipowicz (University of Basel, Basel, Switzerland) for providing purified human T cells; G. Holländer (University of Basel) and A. Cornils (Friedrich Miescher Institute, Basel, Switzerland) for discussion; and P. Morin Jr. (University of Moncton, Moncton, New Brunswick, Canada) for comments on the manuscript. Funding: This work was supported by the Novartis Research Foundation, the Swiss Cancer League, and the Boehringer Ingelheim Fonds. Author contributions: H.C. designed and performed the research, analyzed the data, and wrote the manuscript; M.R.S. generated NDR1-targeted mice; A.H. designed, cloned, and tested shRNA against mNDR2 and critically reviewed the manuscript; D.H. helped with mouse experiments; D.S. designed and tested quantitative PCR assays for detecting murine NDR2 mRNA; S.D. analyzed tissue samples; and B.A.H. initiated this project and assisted with research design and manuscript preparation. Competing interests: H.C. and B.A.H. have a patent application, PCT application WO2008/065391, related to this work. Accession numbers: Microarray data have been stored in the GEO database under accession number GSE21902.

Supplementary Materials

www.sciencesignaling.org/cgi/content/full/3/126/ra47/DC1

Fig. S1. Generation of NDR1-deficient mice.

Fig. S2. Mature T cells from NDR1+/− and NDR1−/− mice show slightly increased resistance to apoptosis.

Fig. S3. Activation of NDR is unchanged in NDR1-deficient thymocytes in response to induction of apoptosis by dexamethasone or antibody against Fas together with cycloheximide.

Fig. S4. NDR1 and NDR2 show distinct patterns in mice.

Fig. S5. The expression of NDR2 and the abundance of NDR2 protein remain unchanged upon loss of NDR1.

Fig. S6. Reduction of the abundance of NDR2 in NDR1-deficient MEFs results in increased resistance to apoptosis.

Fig. S7. Increased development of MPDs in NDR1+/− and NDR1−/− mice after ENU treatment.

Fig. S8. An increase in the abundance of Pou2af1 correlates with a low abundance of NDR in tumors.

Fig. S9. Loss of NDR1 results in minor defects in the proliferation of mature T cells after stimulation of the TCR.

Fig. S10. Cells from NDR-high and NDR-low tumors show no consistent differences in Ki67 staining.

Fig. S11. siRNA against MEF2c does not rescue apoptosis defects in cells depleted of NDR1 and NDR2.

Fig. S12. Characterization of a murine antibody against NDR1.

Fig. S13. Validation of siRNA against MEF2c and overexpression of E47 in transfected cells.

Fig. S14. Testing of different shNDR2 constructs.

Table S1. Aged mice analyzed for tumor development.

Table S2. Overview of tumors identified in ENU-treated mice.

Table S3. Classification of human T cell lymphoma samples.

Table S4. Classification of tumors according to the abundances of NDR1 and NDR2.

Table S5. Genes whose expression was significantly altered in NDR-low tumors.

References and Notes

View Abstract

Navigate This Article