t(10;11)-acute leukemias with MLL-AF10 and MLL-ABI1 chimeric transcripts: specific expression patterns of ABI1 gene in leukemia and solid tumor cell lines

DeepCDR: a hybrid graph convolutional network for predicting cancer drug response

MOTIVATION

Accurate prediction of cancer drug response (CDR) is challenging due to the uncertainty of drug efficacy and heterogeneity of cancer patients. Strong evidences have implicated the high dependence of CDR on tumor genomic and transcriptomic profiles of individual patients. Precise identification of CDR is crucial in both guiding anti-cancer drug design and understanding cancer biology.

RESULTS

In this study, we present DeepCDR which integrates multi-omics profiles of cancer cells and explores intrinsic chemical structures of drugs for predicting CDR. Specifically, DeepCDR is a hybrid graph convolutional network consisting of a uniform graph convolutional network and multiple subnetworks. Unlike prior studies modeling hand-crafted features of drugs, DeepCDR automatically learns the latent representation of topological structures among atoms and bonds of drugs. Extensive experiments showed that DeepCDR outperformed state-of-the-art methods in both classification and regression settings under various data settings. We also evaluated the contribution of different types of omics profiles for assessing drug response. Furthermore, we provided an exploratory strategy for identifying potential cancer-associated genes concerning specific cancer types. Our results highlighted the predictive power of DeepCDR and its potential translational value in guiding disease-specific drug design.

AVAILABILITY AND IMPLEMENTATION

DeepCDR is freely available at https://github.com/kimmo1019/DeepCDR.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Acute myeloid leukemia with t(10;11)(p11-12;q23.3): Results of Russian Pediatric AML registration study

INTRODUCTION

Translocations involving the KMT2A gene (also known as MLL) are frequently diagnosed in pediatric acute leukemia cases with either lymphoblastic or myeloid origin. KMT2A is translocated to multiple partner genes, including MLLT10/AF10 localizing at chromosomal band 10p12. KMT2A-MLLT10 is one of the common chimeric genes diagnosed in acute leukemia with KMT2A rearrangement (8%), especially in acute myeloid leukemia (AML; 18%). MLLT10 is localized in very close proximity to two other KMT2A partner genes at 10p11-12-NEBL and ABI1, so they could not be distinguished by conventional cytogenetics.

METHODS

In this work, we present a cohort of 28 patients enrolled into Russian Pediatric AML registration study carrying rearrangements between chromosomal regions 11q23.3 and 10p11-12. G-banding, FISH, reverse transcription PCR, and long-distance inverse PCR were used to characterize the KMT2A gene rearrangements in these patients.

RESULTS

We demonstrate that 25 patients harbor the KMT2A-MLLT10 rearrangement, while three patients show the rare KMT2A rearrangements (2× KMT2A-NEBL; 1× KMT2A-ABI1).

CONCLUSIONS

Therefore, the combination of cytogenetic and molecular genetic methods is of high importance in diagnosing cases with t(10;11)(p11-12;q23.3).

The tumor suppressor SHIP1 colocalizes in nucleolar cavities with p53 and components of PML nuclear bodies

The inositol 5-phosphatase SHIP1 is a negative regulator of signaling processes in haematopoietic cells. By converting PI(3,4,5)P3 to PtdIns(3,4)P2 at the plasma membrane, SHIP1 modifies PI3-kinase mediated signaling. We have recently demonstrated that SHIP1 is a nucleo-cytoplasmic shuttling protein and SHIP1 nuclear puncta partially colocalize with FLASH, a component of nuclear bodies. In this study, we demonstrate that endogenous SHIP1 localizes to intranucleolar regions of both normal and leukemic haematopoietic cells. In addition, we report that ectopically expressed SHIP1 accumulates in nucleolar cavities and colocalizes with the tumor suppressor protein p53 and components of PML nuclear bodies (e.g. SP100, SUMO-1 and CK2). Moreover, SHIP1 also colocalizes in nucleolar cavities with components of the ubiquitin-proteasome pathway. By using confocal microscopy data, we generated 3D-models revealing the enormous extent of the SHIP1 aggresomes in the nucleolus. Furthermore, treatment of cells with the proteasome inhibitor MG132 causes an enlargement of nucleolar SHIP1 containing structures. Unexpectedly, this accumulation can be partially prevented by treatment with the inhibitor of nuclear protein export Leptomycin B. In recent years, several proteins aggregating in nucleolar cavities were shown to be key factors of neurodegenerative diseases and cancerogenesis. Our findings support current relevance of nuclear localized SHIP1.

Crk and ABI1: binary molecular switches that regulate abl tyrosine kinase and signaling to the cytoskeleton

The nonreceptor tyrosine kinases Abl and Arg are among the most well-characterized tyrosine kinases in the human genome. The activation of Abl by N-terminal fusions with Bcr (Bcr-Abl) or Gag (v-Abl) is responsible for chronic myeloid leukemia or Ph+ acute lymphoblastic leukemia and mouse leukemia virus, respectively. In addition, aberrant Abl and Arg activation downstream of several oncogenic growth factor receptors contributes to the development and progression of a variety of human cancers, often associated with poor clinical outcome, drug resistance, and tumor invasion and metastasis. Abl activation can occur by a variety of mechanisms that include domain interactions involving structural remodeling of autoinhibited conformations as well as direct phosphorylation by upstream kinases and phosphatases. Constitutive activation of Abl plays a significant role in regulating the actin cytoskeleton by modulating cell adhesion, motility, and invadopodia. This review addresses the role of Abl and Arg in tumor progression with particular emphasis on the roles of Crk and Abi1 adapter proteins as distinct molecular switches for Abl transactivation. These insights, combined with new insights into the structure of these kinases, provide the rationale to envision that Crk and Abi1 fine-tune Abl regulation to control signaling to the cytoskeleton.

Chromosome 8 pentasomy with partial tandem duplication of 11q23 in a case of de novo acute myeloid leukemia

Polysomy 8 is a rare abnormality, one that has been reported as associated with secondary evolution, monocytic differentiation, or poor prognosis in myeloid neoplasm. In contrast to tetrasomy 8, which is most commonly observed, pentasomy 8 is a minority component of polysomy 8. To date, only three cases of pentasomy 8 accompanied with 11q23 rearrangement have been reported. Reported here is a novel case of pentasomy 8 with partial tandem duplication of 11q23 in de novo acute myeloid leukemia. The findings contribute to understanding of the relation between the two abnormalities, which have their own individual leukemogenic potencies.

Abelson interactor protein-1 positively regulates breast cancer cell proliferation, migration, and invasion

Abelson interactor protein-1 (ABI-1) is an adaptor protein involved in actin reorganization and lamellipodia formation. It forms a macromolecular complex containing Hspc300/WASP family verprolin-homologous proteins 2/ABI-1/nucleosome assembly protein 1/PIR121 or Abl/ABI-1/WASP family verprolin-homologous proteins 2 in response to Rho family-dependent stimuli. Due to its role in cell mobility, we hypothesized that ABI-1 has a role in invasion and metastasis. In the present study, we found that weakly invasive breast cancer cell lines (MCF-7, T47D, MDA-MB-468, SKBR3, and CAMA1) express lower levels of ABI-1 compared with highly invasive breast cancer cell lines (MDA-MB-231, MDA-MB-157, BT549, and Hs578T), which exhibit high ABI-1 levels. Using RNA interference, ABI-1 was stably down-regulated in MDA-MB-231, which resulted in decreased cell proliferation and anchorage-dependent colony formation and abrogation of lamellipodia formation on fibronectin. Down-regulation of ABI-1 decreased invasiveness and migration ability and decreased adhesion on collagen IV and actin polymerization in MDA-MB-231 cells. Additionally, compared with control parental cells, ABI-1 small interfering RNA-transfected cells showed decreased levels of phospho-PDK1, phospho-Raf, phospho-AKT, total AKT, and AKT1. These data suggest that ABI-1 plays an important role in the spread of breast cancer and that this role may be mediated via the phosphatidylinositol 3-kinase pathway.

Mutation and expression analyses of the MET and CDKN2A genes in rhabdomyosarcoma with emphasis on MET overexpression

Rhabdomyosarcoma (RMS) is the most common soft-tissue sarcoma of childhood. The simultaneous loss of Ink4a/Arf function and disruption of Met signaling in Ink4a/Arf-/- mice transgenic for hepatocyte growth factor/scatter factor (HGF/SF) induces RMS with extremely high penetrance and short latency. To address the roles of MET and CDKN2A (p16INK4A/p14ARF) in human RMS, we performed mutational analyses in 39 samples of RMS by PCR-SSCP. No mutations were detected in exons 14-21 of MET whereas a nonsense mutation at codon 80 of p16(INK4A) was identified in an alveolar RMS cell line. We also quantified the relative expression levels and DNA copy numbers of these genes in seven cell lines and 17 fresh tumors by real-time quantitative PCR. Expression of MET was detected in all samples; however, more than 10-fold difference was found in the samples with higher or lower expression level, despite a normal DNA copy number. The protein expression level was consistent with that of mRNA, and in cell lines with a higher expression level, MET was constitutively activated. Notably, the expression level of MET was significantly higher in patients who died (P = 0.02), in patients with stage IV (P = 0.04), as well as in patients with PAX3-FKHR chimeric transcript (P = 0.04). On the other hand, reduced or absent expression of p16INK4A and/or p14(ARF) showed no significant correlation with the clinicopathological parameters, except for the age at diagnosis. Our data suggest that MET plays a role in the progression of RMS.

The mixed-lineage leukemia fusion partner AF4 stimulates RNA polymerase II transcriptional elongation and mediates coordinated chromatin remodeling

AF4 gene, frequently translocated with mixed-lineage leukemia (MLL) in childhood acute leukemia, encodes a putative transcriptional activator of the AF4/LAF4/FMR2 (ALF) protein family previously implicated in lymphopoiesis and Purkinje cell function in the cerebellum. Here, we provide the first evidence for a direct role of AF4 in the regulation of transcriptional elongation by RNA polymerase II (Pol II). We demonstrate that mouse Af4 functions as a positive regulator of Pol II transcription elongation factor b (P-TEFb) kinase and, in complex with MLL fusion partners Af9, Enl and Af10, as a mediator of histone H3-K79 methylation by recruiting Dot1 to elongating Pol II. These pathways are interconnected and tightly regulated by the P-TEFb-dependent phosphorylation of Af4, Af9 and Enl which controls their transactivation activity and/or protein stability. Consistently, increased levels of phosphorylated Pol II and methylated H3-K79 are observed in the ataxic mouse mutant robotic, an over-expression model of Af4. Finally, we confirm the functional relevance of Af4, Enl and Af9 to the regulation of gene transcription as their over-expression strongly stimulates P-TEFb-dependent transcription of a luciferase reporter gene. Our findings uncover a central role for these proteins in the regulation of transcriptional elongation and coordinated histone methylation, providing valuable insight into their contribution to leukemogenesis and neurodegeneration. Since these activities likely extend to the entire ALF protein family, this study also significantly inputs our understanding of the molecular basis of FRAXE mental retardation syndrome in which FMR2 expression is silenced.

Mutations of the PTPN11 and RAS genes in rhabdomyosarcoma and pediatric hematological malignancies

PTPN11 has been identified as a causative gene in Noonan syndrome (NS), responsible for about 50% of cases of NS. Given the association between NS and an increased risk of some malignancies, notably leukemia and probably some solid tumors including neuroblastoma (NB) and rhabdomyosarcoma (RMS), recent studies have reported that gain-of-function somatic mutations in PTPN11 occur in some hematological malignancies, especially de novo juvenile myelomonocytic leukemia (JMML) and in some solid tumors such as NB, although at a low frequency. In a screen for mutations of PTPN11 in 7 cell lines and 30 fresh tumors of RMS and in 25 cell lines and 40 fresh tumors of NB, we identified a missense mutation (A72T) in an embryonal RMS patient. In the RMS samples, we also detected mutations of NRAS in 1 cell line and 1 patient; both mutations were in embryonal RMSs and had no PTPN11 mutations. No mutations of PTPN11 were detected in NB. In 95 leukemia cell lines and 261 fresh leukemia samples including 22 JMMLs, 9 kinds of missense mutations were detected in 17 leukemia samples, which included 11 (50.0%) mutations in JMML samples and lower frequencies in other hematological malignancies. Furthermore, we identified 4 (18.2%) NRAS mutations and 1 (4.5%) KRAS mutation in 5 JMML samples, 1 of which had a concomitant PTPN11 mutation. Our data suggest that mutations of PTPN11 as well as RAS play a role in the pathogenesis of not only myeloid hematological malignancies but also a subset of RMS malignancies.

Common cancer biomarkers

There is an increasing interest in complementing conventional histopathologic evaluation with molecular tools that could increase the sensitivity and specificity of cancer staging for diagnostic and prognostic purposes. This study strove to identify cancer-specific markers for the molecular detection of a broad range of cancer types. We used 373 archival samples inclusive of normal tissues of various lineages and benign or malignant tumors (predominantly colon, melanoma, ovarian, and esophageal cancers). All samples were processed identically and cohybridized with an identical reference RNA source to a custom-made cDNA array platform. The database was split into training (n = 201) and comparable prediction (n = 172) sets. Leave-one-out cross-validation and gene pairing analysis identified putative cancer biomarkers overexpressed by malignant lesions independent of tissue of derivation. In particular, seven gene pairs were identified with high predictive power (87%) in segregating malignant from benign lesions. Receiver operator characteristic curves based on the same genes could segregate malignant from benign tissues with 94% accuracy. The relevance of this study rests on the identification of a restricted number of biomarkers ubiquitously expressed by cancers of distinct histology. This has not been done before. These biomarkers could be used broadly to increase the sensitivity and accuracy of cancer staging and early detection of locoregional or systemic recurrence. Their selective expression by cancerous compared with paired normal tissues suggests an association with the oncogenic process resulting in stable expression during disease progression when the presently used differentiation markers are unreliable.

The MYO1F, unconventional myosin type 1F, gene is fused to MLL in infant acute monocytic leukemia with a complex translocation involving chromosomes 7, 11, 19 and 22

We analysed a complex translocation involving chromosomes 7, 11, 19 and 22 in infant acute monocytic leukemia, and identified that the MLL gene on 11q23 was fused to the unconventional myosin type 1F, MYO1F, gene on 19p13.2-13.3. MYO1F consists of at least 28 exons and was predicted to encode a 1098-amino-acid with an N-terminal head domain containing both ATP-binding and actin-binding sequences, a neck domain with a single IQ motif, and a tail with TH1, TH2 and SH3 domains. Northern blot analysis of RNAs prepared from multiple tissues showed that the expression of approximately 4-kb transcripts appeared constant in most tissues examined. However, MYO1F was expressed in only three of 22 leukemic cell lines. The MLL-MYO1F fusion protein contains almost the entire MYO1F, however, C-terminal MYO1F has neither the transactivation domain nor the dimerization domain found in various MLL fusion partners. Further analysis of this novel type of MLL fusion protein would provide new insights into leukemogenesis. MYO1F is the fourth partner gene of MLL on 19p13. At the cytogenetic level, it may be difficult to distinguish MLL-ENL, MLL-ELL, MLL-EEN and MLL-MYO1F fusions created by t(11;19)(q23;p13), and it is likely that cases of t(11;19) lacking a known fusion gene may result in this gene fusion.

Polysomy 8 defines a clinico-cytogenetic entity representing a subset of myeloid hematologic malignancies associated with a poor prognosis: report on a cohort of 12 patients and review of 105 published cases

Tetrasomy, pentasomy, and hexasomy 8 (polysomy 8) are relatively rare compared to trisomy 8. Here we report on a series of 12 patients with acute myeloid leukemia (AML), myelodysplastic syndrome (MDS), or myeloproliferative disorder (MPD) associated with polysomy 8 as detected by conventional cytogenetics and fluorescence in situ hybridization (FISH). In an attempt to better characterize the clinical and hematological profile of this cytogenetic entity, our data were combined with those of 105 published patients. Tetrasomy 8 was the most common presentation of polysomy 8. In 60.7% of patients, polysomy 8 occurred as part of complex changes (16.2% with 11q23 rearrangements). No cryptic MLL rearrangements were found in cases in which polysomy 8 was the only karyotypic change. Our study demonstrates the existence of a polysomy 8 syndrome, which represents a subtype of AML, MDS, and MPD characterized by a high incidence of secondary diseases, myelomonocytic or monocytic involvement in AML and poor overall survival (6 months). Age significantly reduced median survival, but associated cytogenetic abnormalities did not modify it. Cytogenetic results further demonstrate an in vitro preferential growth of the cells with a high level of aneuploidy suggesting a selective advantage for polysomy 8 cells.

MLL-MLLT10 fusion in acute monoblastic leukemia: variant complex rearrangements and 11q proximal breakpoint heterogeneity

Cytogenetic studies of acute monoblastic leukemia cases presenting MLL-MLLT10 (alias MLL-AF10) fusion show a broad heterogeneity of chromosomal breakpoints. We present two new pediatric cases (French-American-British type M5) with MLL-MLLT10 fusion, which we studied with fluorescence in situ hybridization. In both we detected a paracentric inversion of the 11q region that translocated onto chromosome 10p12; one case displayed a variant complex pattern. We review the cytogenetic molecular data concerning the proximal inversion breakpoint of 11q and confirm its heterogeneity.

Detection of cryptic MLL insertions using a commercial dual-color fluorescence in situ hybridization probe

Involvement of the MLL gene located at chromosome region 11q23 is a frequent occurrence in both acute myelocytic leukemia and acute lymphoblastic leukemia. More than 30 loci have now been associated with MLL, usually by reciprocal translocation. Deletions, insertions, and more complex rearrangements of MLL are rarely seen. We present three cases of AML M5 showing no cytogenetic evidence of 11q23 rearrangement, in which a commercial MLL dual-color fluorescence in situ hybridization probe revealed a nonstandard abnormal signal pattern, suggesting cryptic insertion of the MLL gene into its partner gene site.

Cytogenetics, fluorescence in situ hybridization, and reverse transcriptase polymerase chain reaction are necessary to clarify the various mechanisms leading to an MLL-AF10 fusion in acute myelocytic leukemia with 10;11 rearrangement

In acute myelocytic leukemia (AML), predominantly in AML M5a, a recurrent chromosome aberration involves 11q23/MLL and the short arm of chromosome 10. Molecular studies have shown that the AF10 gene at 10p12 is consistently a partner gene in cases with 10;11 rearrangement. A simple reciprocal translocation cannot lead to the known MLL-AF10 fusion transcript because the 3' part of the MLL gene is orientated to the telomere and the 3' part of the AF10 gene to the centromere. In a series of 1897 AML samples, 14 cases (0.74%) showed 10;11 rearrangements leading to a MLL-AF10 fusion transcript. These cases were analyzed in detail with G banding analyses, fluorescence in situ hybridization, and molecular investigation in a single center. Five different mechanisms of (10;11) rearrangements leading to a MLL-AF10 fusion transcript can be observed (i.e., reciprocal translocations, insertions of either 10p into 11q or 11q into 10p, as well as complex and cryptic rearrangements). Compared to translocations involving MLL and other partner genes, complex rearrangements are unique for MLL-AF10 fusions. This may result from the opposite orientation of MLL and AF10.

Fusion of an AF4-related gene, LAF4, to MLL in childhood acute lymphoblastic leukemia with t(2;11)(q11;q23)

We showed that the LAF4 gene on 2q11.2-12 was fused to the MLL gene on 11q23 in a pediatric patient with CD10 positive acute lymphoblastic leukemia (ALL) having t(2;11)(q11;q23). The LAF4 gene, which encodes a lymphoid nuclear protein of 1227 amino acids with transactivation potential, is thought to have a role in early lymphoid development. The LAF4 protein was homologous to AF4 and AF5q31 proteins that are fused to MLL in infant early pre-B ALL and the breakpoint of LAF4 was located within the region homologous to the transactivation domain of AF4 and AF5q31. Expression of the 8.5-kb LAF4 transcript was detected in the adult heart, brain, and placenta and in the fetal brain. LAF4 expression was found to be higher in ALL cell lines than in AML and Epstein-Barr virus-transformed B-lymphocyte cell lines. These findings suggest that LAF4, AF4 and AF5q31 might define a new family particularly involved in the pathogenesis of 11q23-associated ALL.

Is t(10;11)(p11.2;q23) involving MLL and ABI-1 genes associated with congenital acute monocytic leukemia?

Congenital, or perinatal, leukemias are rarely observed, but retrospective molecular studies seem to suggest a more frequent onset in prenatal life. Myelocytic types are common, and chromosome band 11q23 rearrangements at the MLL locus are characteristic genetic markers. The fusion of the MLL gene with one of its partners, ABI-1, has recently been described in two infant leukemia patients with monocytic involvement and good clinical outcome. We report a case of congenital monocytic leukemia with the same gene involvement and good response to chemotherapy. The blast metaphases were probed by fluorescence in situ hybridization, and t(10;11)(p11.2;q23) involving MLL and ABI-1 genes was demonstrated with the same breakpoint in ABI-1. The congenital presentation of this case suggests a possible relationship of this genetic event with in utero leukemogenesis.

The chromosome translocation t(7;11)(p15;p15) in acute myeloid leukemia results in fusion of the NUP98 gene with a HOXA cluster gene, HOXA13, but not HOXA9

The nucleoporin gene NUP98 has been reported to be fused to 9 partner genes in hematologic malignancies with 11p15 translocations. The NUP98-HOXA9 fusion gene has been identified in acute myeloid leukemia (AML) and chronic myelogenous leukemia with t(7;11)(p15;p15). We report here a novel NUP98 partner gene, HOXA13, in a patient with de novo AML having t(7;11)(p15;p15). The HOXA13 gene is part of the HOXA cluster genes and contains 2 exons, encoding a protein of 338 amino acids with a homeodomain. The NUP98-HOXA13 fusion protein consists of the N-terminal phenylalanine-glycine repeat motif of NUP98 and the C-terminal homeodomain of HOXA13, similar to the NUP98-HOXA9 fusion protein. Reverse transcriptase-polymerase chain reaction (RT-PCR) analysis in various leukemic cell lines showed that the HOXA13 gene was expressed significantly more frequently in acute monocytic leukemic cell lines than in other leukemic cell lines (P = 0.039). HOXA13 and three HOXA cluster genes (A9, A10, A11) located at the 5' end of the HOXA9 gene were frequently expressed in myeloid leukemic cell lines. Our results revealed that t(7;11)(p15;p15) was not a single chromosomal abnormality at the molecular level. The protein encoded by the NUP98-HOXA13 fusion gene is similar to that encoded by NUP98-HOXA9, and the expression pattern of the HOXA13 gene in leukemic cell lines is similar to that of the HOXA9 gene, suggesting that the NUP98-HOXA13 fusion protein may play a role in leukemogenesis through a mechanism similar to that of the NUP98-HOXA9 fusion protein.

Aberrations of the hSNF5/INI1 gene are restricted to malignant rhabdoid tumors or atypical teratoid/rhabdoid tumors in pediatric solid tumors

The hSNF5/INI1 gene, which encodes a subunit of the SWI/SNF family of chromatin-remodeling complexes and is located at 22q11.2, has been reported as a tumor suppressor gene inactivated in malignant rhabdoid tumors (MRTs). We analyzed this gene in varieties of pediatric solid tumors including MRTs, using the reverse transcription-polymerase chain reaction (PCR) and PCR-single strand conformation polymorphism method. We found 5 homozygous deletions, 2 truncated mutations, one missense mutation, and one silent mutation of the hSNF5/INI1 gene in 7 MRT cell lines, and one homozygous deletion, one microdeletion, one splicing acceptor site mutation, and one absence of expression in 7 fresh tumor tissues of MRT and atypical teratoid (AT)/rhabdoid tumors (RTs). Homozygous deletions were also found in one (KYM-1) of 8 rhabdomyosarcoma (RMS) cell lines. To investigate characteristics of the KYM-1 cell line, we have established KYM-1 tumors in nude mice into which KYM-1 cells were transplanted. Notably, we found that MyoD1, known as a marker for RMS, was not expressed in the KYM-1 cell line as well as MRT cell lines and fresh tumors. Histopathologic, cytogenetic, and molecular studies of the KYM-1 cell line and KYM-1 tumors in nude mice have revealed that this RMS cell line should be MRT rather than RMS. RMS-carrying aberrations of the hSNF5/INI1 gene should be reevaluated. No aberrations of this gene were found in the other 34 cell lines or 80 fresh tumor specimens except the single nucleotide polymorphisms in the 3' noncoding region. These results suggest that alterations of the hSNF5/INI1 gene were restricted to MRTs or AT/RTs in pediatric solid tumors.