Loss of H3K27 methylation identifies poor outcomes in adult-onset acute leukemia.

BACKGROUND: Acute leukemia is an epigenetically heterogeneous disease. The intensity of treatment is currently guided by cytogenetic and molecular genetic risk classifications; however these incompletely predict outcomes, requiring additional information for more accurate outcome predictions. We aimed to identify potential prognostic implications of epigenetic modification of histone proteins, with a focus on H3K4 and H3K27 methylation marks in relation to mutations in chromatin, splicing and transcriptional regulators in adult-onset acute lymphoblastic and myeloid leukemia. RESULTS: Histone 3 lysine 4 di- and trimethylation (H3K4me2, H3K4me3) and lysine 27 trimethylation (H3K27me3) mark expression was evaluated in 241 acute myeloid leukemia (AML), 114 B-cell acute lymphoblastic leukemia (B-ALL) and 14T-cell ALL (T-ALL) patient samples at time of diagnosis using reverse phase protein array. Expression levels of the marks were significantly lower in AML than in B and T-ALL in both bone marrow and peripheral blood, as well as compared to normal CD34+ cells. In AML, greater loss of H3K27me3 was associated with increased proliferative potential and shorter overall survival in the whole patient population, as well as in subsets with DNA methylation mutations. To study the prognostic impact of H3K27me3 in the context of cytogenetic aberrations and mutations, multivariate analysis was performed and identified lower H3K27me3 level as an independent unfavorable prognostic factor in all, as well as in TP53 mutated patients. AML with decreased H3K27me3 demonstrated an upregulated anti-apoptotic phenotype. In ALL, the relative quantity of histone methylation expression correlated with response to tyrosine kinase inhibitor in patients who carried the Philadelphia cytogenetic aberration and prior smoking behavior. CONCLUSION: This study shows that proteomic profiling of epigenetic modifications has clinical implications in acute leukemia and supports the idea that epigenetic patterns contribute to a more accurate picture of the leukemic state that complements cytogenetic and molecular genetic subgrouping. A combination of these variables may offer more accurate outcome prediction and we suggest that histone methylation mark measurement at time of diagnosis might be a suitable method to improve patient outcome prediction and subsequent treatment intensity stratification in selected subgroups.

JAK/STAT pathway inhibition overcomes IL7-induced glucocorticoid resistance in a subset of human T-cell acute lymphoblastic leukemias.

While outcomes for children with T-cell acute lymphoblastic leukemia (T-ALL) have improved dramatically, survival rates for patients with relapsed/refractory disease remain dismal. Prior studies indicate that glucocorticoid (GC) resistance is more common than resistance to other chemotherapies at relapse. In addition, failure to clear peripheral blasts during a prednisone prophase correlates with an elevated risk of relapse in newly diagnosed patients. Here we show that intrinsic GC resistance is present at diagnosis in early thymic precursor (ETP) T-ALLs as well as in a subset of non-ETP T-ALLs. GC-resistant non-ETP T-ALLs are characterized by strong induction of JAK/STAT signaling in response to interleukin-7 (IL7) stimulation. Removing IL7 or inhibiting JAK/STAT signaling sensitizes these T-ALLs, and a subset of ETP T-ALLs, to GCs. The combination of the GC dexamethasone and the JAK1/2 inhibitor ruxolitinib altered the balance between pro- and anti-apoptotic factors in samples with IL7-dependent GC resistance, but not in samples with IL7-independent GC resistance. Together, these data suggest that the addition of ruxolitinib or other inhibitors of IL7 receptor/JAK/STAT signaling may enhance the efficacy of GCs in a biologically defined subset of T-ALL.

IDENTIFYING CANCER SPECIFIC METABOLIC SIGNATURES USING CONSTRAINT-BASED MODELS.

Cancer metabolism differs remarkably from the metabolism of healthy surrounding tissues, and it is extremely heterogeneous across cancer types. While these metabolic differences provide promising avenues for cancer treatments, much work remains to be done in understanding how metabolism is rewired in malignant tissues. To that end, constraint-based models provide a powerful computational tool for the study of metabolism at the genome scale. To generate meaningful predictions, however, these generalized human models must first be tailored for specific cell or tissue sub-types. Here we first present two improved algorithms for (1) the generation of these context-specific metabolic models based on omics data, and (2) Monte-Carlo sampling of the metabolic model ux space. By applying these methods to generate and analyze context-specific metabolic models of diverse solid cancer cell line data, and primary leukemia pediatric patient biopsies, we demonstrate how the methodology presented in this study can generate insights into the rewiring differences across solid tumors and blood cancers.

Inactivation of KLF4 promotes T-cell acute lymphoblastic leukemia and activates the MAP2K7 pathway.

T-cell acute lymphoblastic leukemia (T-ALL) is an aggressive hematological malignancy with a high incidence of relapse in pediatric ALL. Although most T-ALL patients exhibit activating mutations in NOTCH1, the cooperating genetic events required to accelerate the onset of leukemia and worsen disease progression are largely unknown. Here, we show that the gene encoding the transcription factor KLF4 is inactivated by DNA methylation in children with T-ALL. In mice, loss of KLF4 accelerated the development of NOTCH1-induced T-ALL by enhancing the G1-to-S transition in leukemic cells and promoting the expansion of leukemia-initiating cells. Mechanistically, KLF4 represses the gene encoding the kinase MAP2K7. Our results showed that in murine and pediatric T-ALL, loss of KLF4 leads to aberrant activation of MAP2K7 and of the downstream effectors JNK and ATF2. As a proof-of-concept for the development of a targeted therapy, administration of JNK inhibitors reduced the expansion of leukemia cells in cell-based and patient-derived xenograft models. Collectively, these data uncover a novel function for KLF4 in regulating the MAP2K7 pathway in T-ALL cells, which can be targeted to eradicate leukemia-initiating cells in T-ALL patients.

Endogenous knockdown of survivin improves chemotherapeutic response in ALL models.

Although the cure rate of newly diagnosed acute lymphoblastic leukemia (ALL) has improved over the past four decades, the outcome for patients who relapse remains poor. New therapies are needed for these patients. Our previous global gene expression analysis in a series of paired diagnosis-relapse pediatric patient samples revealed that the antiapoptotic gene survivin was consistently upregulated upon disease relapse. In this study, we demonstrate a link between survivin expression and drug resistance and test the efficacy of a novel antisense agent in promoting apoptosis when combined with chemotherapy. Gene-silencing experiments targeting survivin mRNA using either short-hairpin RNA (shRNA) or a locked antisense oligonucleotide (LNA-ON) specifically reduced gene expression and induced apoptosis in leukemia cell lines. When used in combination with chemotherapy, the survivin shRNA and LNA-ON potentiated the chemotherapeutic antileukemia effect. Moreover, in a mouse primary xenograft model of relapse ALL, the survivin LNA-ON decreased survivin expression in a subset of animals, and produced a statistically significant decrease in tumor progression. Taken together, these findings suggest that targeting endogenous levels of survivin mRNA by LNA-ON methods may augment the response to standard chemotherapy by sensitizing otherwise resistant tumor cells to chemotherapy.

Aseptic meningitis in a child after systemic treatment with high dose cytarabine.

Cytarabine was temporally associated with aseptic meningitis syndrome in an 8-year-old Hispanic girl being treated for acute lymphoblastic leukemia.

Novel mitochondrial DNA deletion found in a renal cell carcinoma.

Polymerase chain reaction (PCR) was used to analyze a rarely deleted region of mitochondrial DNA (mtDNA) from 39 human renal cell carcinomas (RCC) and matched normal kidney tissue removed during radical nephrectomy. One tumor specimen (E.R.) had a unique PCR product approximately 250 base pairs (bp) smaller than the PCR product found in the normal E.R. kidney. Sequence analysis of the tumor-specific PCR fragment revealed a 264 bp deletion in the first subunit (NDI) of NADH:ubiquinone oxidoreductase (complex I) of the electron transport chain. Southern analysis of the RCCs demonstrated that approximately 50% of the mtDNA molecules in the primary RCC contained a unique 3.2 kb EcoRV restriction fragment found only in E.R. tumor mtDNA. Northern analysis demonstrated preferential transcription of the truncated NDI mRNA. None of the five metastases or any normal tissue from E.R. contained levels of the NDI deletion detectable by PCR. This is the first reported case of an intragenic NDI mtDNA deletion.

Marked increase in mitochondrial DNA deletion levels in the cerebral cortex of Huntington's disease patients.

To determine if somatic mtDNA mutations might contribute to the neurodegeneration observed in Huntington's disease (HD), we quantitated the amount of the common mitochondrial 4977 nucleotide pair deletion (mtDNA4977) in cortex and putamen of HD patients and age-matched controls by the serial dilution-polymerase chain reaction method. Cortical deletion levels were analyzed in the temporal, frontal, and occipital lobes. HD temporal lobes had an 11-fold greater mean mtDNA4977 deletion level than age-matched controls, and HD frontal lobes had fivefold greater levels. HD occipital lobe and putamen deletion levels were comparable with control levels. These results support the hypothesis that HD is associated with elevated cortical mtDNA damage.

Characterization of mutants within the gene for the adenovirus E3 14.7-kilodalton protein which prevents cytolysis by tumor necrosis factor.

The 14,700-Da protein (14.7K protein) encoded by the E3 region of adenovirus has previously been shown to protect mouse cells from cytolysis by tumor necrosis factor (TNF). Delineating the sequences in the 14.7K protein that are required for this activity may provide insight into the mechanism of protection from TNF by 14.7K as well as the mechanism of TNF cytolysis. In the present study, we examined the ability of 14.7K mutants to protect cells from lysis by TNF. In-frame deletions as well as Cys-to-Ser mutations in the 14.7K gene were generated by site-directed mutagenesis and then built into the genome of a modified adenovirus type 5 (dl7001) that lacks all E3 genes. dl7001, which replicates to the same titers as does adenovirus type 5 in cultured cells, has the largest E3 deletion analyzed to date. 51Cr release was used to assay TNF cytolysis. Our results indicate that most mutations in the 14.7K gene result in a loss of function, suggesting that nearly the entire protein rather than a specific domain functions to prevent TNF cytolysis.

The 10,400- and 14,500-dalton proteins encoded by region E3 of adenovirus function together to protect many but not all mouse cell lines against lysis by tumor necrosis factor.

We have reported that the E3 14,700-dalton protein (E3 14.7K protein) protects adenovirus-infected mouse C3HA fibroblasts against lysis by tumor necrosis factor (TNF) (L. R. Gooding, L. W. Elmore, A. E. Tollefson, H. A. Brady, and W. S. M. Wold, Cell 53:341-346, 1988). We have also observed that the E1B 19K protein protects adenovirus-infected human but not mouse cells against TNF lysis (L. R. Gooding, L. Aquino, P. J. Duerksen-Hughes, D. Day, T. M. Horton, S. Yei, and W. S. M. Wold, J. Virol. 65:3083-3094, 1991). We now report that, in the absence of E3 14.7K, the E3 10.4K and E3 14.5K proteins are both required to protect C127 as well as several other mouse cell lines against TNF lysis. The 14.7K protein can also protect these cells from TNF in the absence of the 10.4K and 14.5K proteins. This protection by the 10.4K and 14.5K proteins was not observed in the C3HA cell line. These conclusions are based on 51Cr release assays of cells infected with virus E3 mutants that express the 14.7K protein alone, that express both the 10.4K and 14.5K proteins, and that delete the 14.7K in combination with either the 10.4K or 14.5K protein. The 10.4K protein was efficiently coimmunoprecipitated together with the 14.5K protein by using an antiserum to the 14.5K protein, suggesting that the 10.4K and 14.5K proteins exist as a complex in the infected mouse cells and consistent with the notion that they function in concert. Considering that three sets of proteins (E3 14.7K, E1B 19K, and E3 10.4K/14.5K proteins) exist in adenovirus to prevent TNF cytolysis of different cell types, it would appear that TNF is a major antiadenovirus defense of the host.

The E1B 19,000-molecular-weight protein of group C adenoviruses prevents tumor necrosis factor cytolysis of human cells but not of mouse cells.

Tumor necrosis factor (TNF) is a multifunctional immunoregulatory protein that is secreted by activated macrophages and is believed to have antiviral activities. We reported earlier that when mouse C3HA fibroblasts are infected with human adenoviruses, the 289R and 243R proteins encoded by region E1A render the cells susceptible to lysis by TNF, and a 14,700-molecular-weight protein (14.7K protein) encoded by region E3 protects the cells against lysis by TNF. We now report that the 19,000-molecular-weight (19K) (176R) protein encoded by the E1B transcription unit can protect human HEL-299 fibroblasts and human ME-180 cervical carcinoma cells against lysis by TNF. This was determined by infecting cells with adenovirus double mutants that lack region E3 and do or do not express the E1B-19K protein and by measuring cytolysis by using a short-term (18-h) 51Cr-release assay. Under these assay conditions, the 51Cr release was specific to TNF and was not a consequence of the cyt phenotype associated with E1B-19K protein-negative mutants. Also, by using virus double mutants that lack E3 in combination with other early regions, we found that E1A, the E1B-55K protein-encoding gene, E3, and E4 are not required to protect HEL-299 cells against TNF cytolysis. Three additional human cancer cell lines (HeLa, HCT8, and RC29) and a simian virus 40-transformed WI38 cell line (VA-13) also required E1B for protection against TNF cytolysis, indicating that the E1B-19K protein is required to protect many if not all human cell types against lysis by TNF when infected by adenovirus. The E1B-19K protein was not able to protect six different adenovirus-infected mouse cell lines against TNF lysis, even though the protein was shown to be efficiently expressed in one of the cell lines. HEL-299 or ME-180 cells infected by a mutant that lacks the E1B-19K protein but retains region E3 were not lysed by TNF, indicating that one or more of the E3 proteins can protect these cells against TNF lysis in the absence of the E1B-19K protein. Thus, the E3-14.7K but not the E1B-19K protein can protect adenovirus-infected mouse cells against TNF cytolysis, whereas the E1B-19K protein as well as one or more of the E3 proteins can protect adenovirus-infected human cells against TNF cytolysis.

Adenovirus E3 14.7K protein functions in the absence of other adenovirus proteins to protect transfected cells from tumor necrosis factor cytolysis.

A 14,700-kDa protein (14.7K) encoded by the E3 region of adenovirus has been shown to protect adenovirus-infected mouse C3HA cells from lysis by tumor necrosis factor (TNF) (L. R. Gooding, L. W. Elmore, A. E. Tollefson, H. A. Brady, and W. S. M. Wold, Cell 53:341-346, 1988). These infected cells are sensitized to TNF by expression of the adenovirus E1A proteins (P. Duerksen-Hughes, W. S. M. Wold, and L. R. Gooding, J. Immunol. 143:4193-4200, 1989). In this study we show that 14.7K suppresses TNF cytolysis independently of adenovirus infection. Mouse C3HA and C127 cells were transfected with the 14.7K gene controlled by the mouse metallothionein promoter, and permanent 14.7K-expressing cell lines were tested for sensitivity to TNF cytolysis. Transfected cells which were sensitized to TNF either by inhibitors of protein synthesis, microfilament-destabilizing agents, or adenovirus infection were found to be resistant to TNF cytolysis. Two monoclonal antibodies were isolated and used to quantitate 14.7K in transfected and infected cells. Enzyme-linked immunosorbent assay (ELISA) analysis with these monoclonal antibodies and 14.7K immunoblots showed that 14.7K expression can be induced with cadmium in C3HA and C127 transfectants. The 14.7K induction correlated with a dose-dependent decrease in sensitivity to TNF cytotoxicity. The 14.7K protein does not substantially alter cell surface TNF receptor numbers or affinity on C3HA mouse fibroblasts, as determined by Scatchard analysis of 125I-TNF binding. The 14.7K protein also does not alter TNF signal transduction in general, because TNF induction of cell surface class I major histocompatibility complex molecules on 14.7K transfectants was unmodified. Our findings indicate that the adenovirus 14.7K protein functions as a specific inhibitor of TNF cytolysis in the absence of other adenovirus proteins and thus is a unique tool to study the mechanism of TNF cytotoxicity.

A protein serologically and functionally related to the group C E3 14,700-kilodalton protein is found in multiple adenovirus serotypes.

A 14.7-kilodalton protein (14.7K protein) encoded by the E3 region of group C adenoviruses has been shown to protect virus-infected fibroblasts from lysis by tumor necrosis factor (TNF) (L.R. Gooding, L.W. Elmore, A.E. Tollefson, H.A. Brady, and W.S.M. Wold, Cell 53:341-346, 1988). In this study we show that adenoviruses of other groups are also protected from TNF-induced cytolysis. Representative serotypes of groups A, B, D, and E produce a protein analogous to the 14.7K protein found in human group C adenoviruses. Deletion of this protein in group C viruses permits virus infection to induce cellular susceptibility to TNF killing. As with group C adenoviruses, cells infected with wild-type adenoviruses of other serotypes are not killed by TNF and are protected from lysis induced by TNF plus cycloheximide. However, cells are susceptible to TNF-induced lysis when infected with adenovirus type 4 mutants from which the 14.7K gene has been deleted. Although all known adenovirus serotypes infect epithelial cells, adenoviruses cause several diseases with various degrees of pathogenesis. Our findings suggest that the 14.7K protein provides a function required for the in vivo cytotoxicity of many adenoviruses independent of the site of infection or degree of pathogenesis.