Identification of two related markers for common acute lymphoblastic leukemia as heat shock proteins.

By direct analysis of the polypeptide constituents of leukemic cells, we have previously detected several polypeptides that are restricted in their expression to acute lymphoblastic leukemia (ALL). In this study, we provide evidence that two polypeptides designated L2 and L4 are structurally related and represent novel markers for common ALL. Partial amino acid sequence analysis did not uncover differences between L2 and L4. The sequences obtained correspond to a previously cloned human gene designated hsp 27 that is expressed, following heat shock treatment, in a variety of cells. 32Pi incorporation studies indicate that L4 is an unphosphorylated form and L2 is a phosphorylated form of hsp27. The two forms were inducible by heat shock in leukemic and nonleukemic lymphoid cells. Thus, in acute leukemia, the common ALL subtype is uniquely characterized by the constitutive expression of a polypeptide that represents a major cellular phosphoprotein.

Hsp27 expression in neuroblastoma: correlation with disease stage.

BACKGROUND: The 27-kd heat shock protein (Hsp27) is differentially expressed in some malignancies, including breast carcinoma, leukemia, and malignant fibrous histiocytoma. In breast carcinoma, a high-level expression of Hsp27 has been associated with shorter disease-free survival in patients with localized disease. PURPOSE: We have observed variable levels of Hsp27 among neuroblastoma tumors. Our aim in this study was to investigate the relationship between Hsp27 expression and stage of the disease and N-myc gene copy number. METHODS: We determined Hsp27 protein levels in 53 neuroblastoma tumors representing different stages of the disease and in 17 neuroblastoma cell lines by quantitative two-dimensional polyacrylamide gel electrophoresis (PAGE). We also performed statistical analysis of Hsp27 levels in relation to stage of the disease and to N-myc gene copy number. RESULTS: Increased Hsp27 expression in neuroblastomas was associated with limited stage disease and inversely correlated with N-myc gene amplification, a feature known to predict poor clinical outcome. An inverse correlation was also observed between N-myc gene amplification and Hsp27 protein levels among the neuroblastoma cell lines analyzed. Immunohistochemical staining of sections of neuroblastomas showed that Hsp27 was most prominently expressed in the cytoplasm of large ganglionic tumor cells present in neuronally differentiated areas of the tumors. Induction of neuronal differentiation in SMS-KCNR neuroblastoma cells using retinoic acid resulted in an increase in Hsp27. CONCLUSION: High level expression of Hsp27 in neuroblastoma is a feature of limited stage, differentiated tumors. IMPLICATION: Hsp27 may play a part in the biology of neuroblastomas with a favorable outcome.

Identification and characterization of a 19q12 amplicon in esophageal adenocarcinomas reveals cyclin E as the best candidate gene for this amplicon.

Genomic DNA amplification in tumors is frequently associated with an increased gene copy number of oncogenes or other cancer-related genes. We have used a two-dimensional whole-genome scanning technique to identify gene amplification events in esophageal adenocarcinomas. A multicopy genomic fragment from a tumor two-dimensional gel was cloned, and genomic amplification encompassing this fragment was confirmed by Southern blot analysis. The corresponding DNA sequence was matched by BLAST to a BAC contig, which allowed the use of electronic-PCR to localize this amplicon to 19q12. Sequence tagged site-amplification mapping, an approach recently implemented in our laboratory (Lin, L. et al., Cancer Res., 60: 1341-1347,2000), was used to characterize the amplicon. Genomic DNA from 65 esophageal and 11 gastric cardia adenocarcinomas were investigated for 19q12 amplification using quantitative PCR at 11 sequence tagged site markers neighboring the cloned fragment. The amplicon was narrowed from >8 cM to a minimal critical region spanning <0.8 cM, between D19S919 and D19S882. This region includes the cyclin E gene. Fourteen expressed sequence tags (ESTs) covering the minimal region were then assayed for potential gene overexpression using quantitative reverse transcription-PCR. Seven of the selected ESTs were found to be both amplified and overexpressed. Among these seven ESTs, cyclin E showed the highest frequency of gene amplification and overexpression in the tumors examined, which allowed us to finalize the core-amplified region to <300 kb. These results indicate that cyclin E is the likely target gene selected by the amplification event at 19q12. The fact that cyclin E overexpression was found only in the amplified tumors examined indicates that gene amplification underlies the cyclin E gene overexpression. Our study represents the first extensive analysis of the 19q12 amplicon, and is the first to physically map the core-amplified domain to a region of <300 kb that includes cyclin E. Amplification of 19q12 was found neither in the 28 esophageal squamous cancers nor in the 39 lung adenocarcinomas examined but was observed in 13.8% of esophageal and 9.1% of gastric cardia adenocarcinomas.

Distinctive molecular profiles of high-grade and low-grade gliomas based on oligonucleotide microarray analysis.

Astrocytomas are heterogeneous intracranial glial neoplasms ranging from the highly aggressive malignant glioblastoma multiforme (GBM) to the indolent, low-grade pilocytic astrocytoma. We have investigated whether DNA microarrays can identify gene expression differences between high-grade and low-grade glial tumors. We compared the transcriptional profile of 45 astrocytic tumors including 21 GBMs and 19 pilocytic astrocytomas using oligonucleotide-based microarrays. Of the approximately 6800 genes that were analyzed, a set of 360 genes provided a molecular signature that distinguished between GBMs and pilocytic astrocytomas. Many transcripts that were increased in GBM were not previously associated with gliomas and were found to encode proteins with properties that suggest their involvement in cell proliferation or cell migration. Microarray-based data for a subset of genes was validated using real-time quantitative reverse transcription-PCR. Immunohistochemical analysis also localized the protein products of specific genes of interest to the neoplastic cells of high-grade astrocytomas. Our study has identified a large number of novel genes with distinct expression patterns in high-grade and low-grade gliomas.

Co-amplification of a novel gene, NAG, with the N-myc gene in neuroblastoma.

Substantial evidence implicates amplification of the N-myc gene with aggressive tumor growth and poor outcome in neuroblastoma. However some evidence suggests that this gene alone is not the sole determinant of outcome in N-myc amplified tumors. We have searched for genes that co-amplify with N-myc in neuroblastoma by means of two-dimensional analysis of genomic restriction digests. Using this approach, we have identified and cloned a novel genomic fragment which is co-amplified with N-myc in neuroblastomas. This fragment was mapped in close vicinity to N-myc on chromosome arm 2p24. It was amplified in 5/8 N-myc amplified neuroblastoma cell lines and in 9/13 N-myc amplified tumors. Using a PCR-based approach we isolated a 4.5 kb c-DNA sequence that is partly contained in the genomic fragment. The open reading frame of the cDNA encodes a predicted protein of 1353 amino acids (aa). The homology of the predicted protein, which we designated NAG (neuroblastoma amplified gene), to a C. elegans protein of as yet unknown function, and its ubiquitous expression suggest that NAG may serve an essential function. By Northern blot analysis we showed that amplification of the cloned gene correlates with over-expression in neuroblastoma cell lines. Amplification and consequent over-expression of NAG may, therefore, contribute to the phenotype of a subset of neuroblastomas.

Estimation of mutation rates based on the analysis of polypeptide constituents of cultured human lymphoblastoid cells.

A subclone of a human diploid lymphoblastoid cell line, TK-6, with consistently high cloning efficiency has been used to estimate the rates of somatic mutations on the basis of protein variation detected by two-dimensional polyacrylamide gel electrophoresis. A panel of 267 polypeptide spots per gel was screened, representing the products of approximately 263 unselected loci. The rate of human somatic mutation in vitro was estimated by measuring the proportion of protein variants among cell clones isolated at various times during continuous exponential growth of a TK-6 cell population. Three mutants of spontaneous origin were observed, giving an estimated spontaneous rate of 6 x 10(-8) electrophoretic mutations per allele per cell generation (i.e., 1.2 x 10(-7) per locus per cell generation). Following treatment of cells with N-ethyl-N-nitrosourea, a total of 74 confirmed variants at 54 loci were identified among 1143 clones analyzed (approximately 601,000 allele tests). The induced variants include 65 electromorphs which exhibit altered isoelectric charge and/or apparent molecular weight and nine nullimorphs for each of which a gene product was not detected at its usual location on the gel. The induced frequency for these 65 structural gene mutants is 1.1 x 10(-4) per allele. An excess of structural gene mutations at ten known polymorphic loci and repeat mutations at these and other loci suggest nonrandomness of mutation in human somatic cells. Nullimorphs occurring at three heterozygous loci in TK-6 cells may be caused by genetic processes other than structural gene mutation.

Studies of the inheritance of human ribosomal DNA variants detected in two-dimensional separations of genomic restriction fragments.

We have investigated the variation in human ribosomal DNA repeat units as revealed in two-dimensional electrophoretic separates of genomic restriction fragments that were end-labeled at NotI cleavage sites. The transcribed portion of the ribosomal DNA results in approximately 20 labeled fragments visible on each gel as multicopy spots. We have mapped these spots to the sequences responsible for their appearance on the gels, based on their migration positions and direct sequencing of spots, and describe several previously unreported sources of variation. By studying mother/father/child families we gained information on how much of the between-repeats variation is due to differences between and within repeat arrays on homologous chromosomes. Two instances in which a child exhibited more copies of a particular fragment than were present in the parents are described and hypothesized to be due to events such as multiple unequal sister-chromatid exchanges or gene conversions.

Arcuate nucleus transcriptome profiling identifies ankyrin repeat and suppressor of cytokine signalling box-containing protein 4 as a gene regulated by fasting in central nervous system feeding circuits.

The arcuate nucleus of the hypothalamus is a primary site for sensing blood borne nutrients and hormonal messengers that reflect caloric status. To identify novel energy homeostatic genes, we examined RNA extracts from the microdissected arcuate nucleus of fed and 48-h fasted rats using oligonucleotide microarrays. The relative abundance of 118 mRNA transcripts was increased and 203 mRNA transcripts was decreased during fasting. One of the down-regulated mRNAs was ankyrin-repeat and suppressor of cytokine signalling box-containing protein 4 (Asb-4). The predicted structure of Asb-4 protein suggested that it might encode an intracellular regulatory protein, and therefore its mRNA expression was investigated further. Reverse transcription quantitative polymerase chain reaction was used to validate down-regulation of Asb-4 mRNA in the arcuate nucleus of the fasted Sprague-Dawley rat (relative expression of Asb-4 mRNA: fed = 4.66 +/- 0.26; fasted = 3.96 +/- 0.23; n = 4, P < 0.01). Down-regulation was also demonstrated in the obese fa/fa Zucker rat, another model of energy disequilibrium (relative expression of Asb-4 mRNA: lean Zucker = 3.91 +/- 0.32; fa/fa = 2.93 +/- 0.26; n = 5, P < 0.001). In situ hybridisation shows that Asb-4 mRNA is expressed in brain areas linked to energy homeostasis, including the arcuate nucleus, paraventricular nucleus, dorsomedial nucleus, lateral hypothalamus and posterodorsal medial amygdaloid area. Double in situ hybridisation revealed that Asb-4 mRNA colocalises with key energy homeostatic neurones. In the fed state, Asb-4 mRNA is expressed by 95.6% of pro-opiomelanocortin (POMC) neurones and 46.4% of neuropeptide Y (NPY) neurones. By contrast, in the fasted state, the percentage of POMC neurones expressing Asb-4 mRNA drops to 73.2% (P < 0.001). Moreover, the density of Asb-4 mRNA per fasted POMC neurone is markedly decreased. Conversely, expression of Asb-4 mRNA by NPY neurones in the fasted state is modestly increased to 52.7% (P < 0.05). Based on its differential expression, neuroanatomical distribution and colocalisation, we hypothesise that Asb-4 is a gene involved in energy homeostasis.

Quantitative analysis of Op18 phosphorylation in childhood acute leukemia.

Oncoprotein 18 (Op18) is a major cystolic phosphoprotein constituent of leukemia cells. There is cumulative evidence that suggests a role for Op18 in integrating signals from diverse pathways involved in cell growth. Op18 phosphorylation is induced with proliferation in a variety of cell types, and is essential for cell cycle progression. In this study we analyzed the level of unphosphorylated Op18 and of its major phosphorylated forms, Op18a and Op18b, in a series of 177 childhood acute leukemias by means of quantitative two-dimensional polyacrylamide gel electrophoresis (2-D PAGE). Op18 phosphorylation was significantly correlated with the white blood count at the time of diganosis, and with a high percentage of cells in the S phase. Our findings suggest that strategies to inhibit Op18 expression or phosphorylation may be effective in inhibiting leukemia cell proliferation.

PCNA levels in neuroblastoma are increased in tumors with an amplified N-myc gene and in metastatic stage tumors.

N-myc oncogene amplification in neuroblastoma has been found to be significantly associated with advanced stage disease and tumor progression. However, there is a lack of data on tumors, regarding the relationship between N-myc gene amplification and proliferation activity. Proliferating cell nuclear antigen (PCNA) is a proliferation-induced 36 kD nuclear protein that is the auxiliary component of DNA polymerase delta. PCNA levels in tissues have been found to correlate with proliferative activity. We have examined PCNA levels in neuroblastomas in relation to N-myc gene amplification and tumor stage. Statistically, significantly higher levels of PCNA were observed in tumors with an amplified N-myc gene relative to tumors with a single gene copy. The highest levels of PCNA were observed in advanced stage tumors with an amplified N-myc gene. Treatment of neuroblastoma cells in culture with retinoic acid, which induces differentiation, resulted in a substantial decrease in PCNA. Our results suggest that PCNA levels may reflect differences in proliferative activity between neuroblastomas, related to stage of the disease and to N-myc gene copy number.

Genetic analysis of children of atomic bomb survivors.

Studies are under way for the detection of potential genetic effects of atomic bomb radiation at the DNA level in the children of survivors. In a pilot study, we have examined six minisatellites and five microsatellites in DNA derived from 100 families including 124 children. We detected a total of 28 mutations in three minisatellite loci. The mean mutation rates per locus per gamete in the six minisatellite loci were 1.5% for 65 exposed gametes for which mean parental gonadal dose was 1.9 Sv and 2.0% for 183 unexposed gametes. We detected four mutations in two tetranucleotide repeat sequences but no mutations in three trinucleotide repeat sequences. The mean mutation rate per locus per gamete was o% for the exposed gametes and 0.5% for the unexposed gametes in the five microsatellite loci. No significant differences in the mutation rates between the exposed and the unexposed gametes were detected in these repetitive sequences. Additional loci are being analyzed to increase the power of our study to observe a significant difference in the mutation rates at the 0.05 level of significance.

Quantitative analysis of a new marker for common acute lymphoblastic leukemia detected by two-dimensional electrophoresis.

A major subtype of childhood lymphoblastic leukemia has been previously defined on the basis of expression of a cell surface marker referred to as the common acute lymphoblastic leukemia antigen. In this study we provide evidence that leukemic cells from most of the patients with common acute lymphoblastic leukemia contain a cytosolic polypeptide designated L4, the occurrence of which is strongly correlated with the expression of the common acute lymphoblastic leukemia antigen. The detection and quantitation of L4 was accomplished by analysis of cellular polypeptides resolved by two-dimensional polyacrylamide gel electrophoresis. This approach provides a powerful tool for the delineation of distinct polypeptides corresponding to different types of leukemia and complements immunological analysis of the cell surface marker phenotype.

TGF-ß-induced IRAK-M expression in tumor-associated macrophages regulates lung tumor growth.

Tumor-associated macrophages (TAMs) constitute a major component of the immune cell infiltrate observed in the tumor microenvironment (TME). Factors present in the TME, including tumor growth factor-ß (TGF-ß), allow tumors to circumvent host-mediated immune responses to promote tumor progression. However, the molecular mechanism(s) involved are not clear. Toll-like receptors (TLRs) are important mediators of innate immune responses by immune cells, whose activation triggers the production of molecules required for anti-tumoral responses. Interleukin (IL) receptor-associated kinase (IRAK)-M is an inactive serine/threonine kinase, predominantly expressed in macrophages and is a potent negative regulator of TLR signaling. In this study, we show that TAMs express significantly higher levels of IRAK-M compared with peritoneal macrophages in a syngeneic mouse model of lung cancer. Subcutaneous implantation of Lewis lung carcinoma cells in IRAK-M(-/-) mice resulted in a five-fold reduction in tumor growth as compared with tumors in wild-type (WT) animals. Furthermore, compared with WT TAMs, TAMs isolated from IRAK-M(-/-) mice displayed features of a classically activated (M1) rather than alternatively activated (M2) phenotype, as manifest by greater expression of IL-12, interferon-γ (IFN-γ) and inducible nitric oxide synthase. Human lung cancer cells induced IRAK-M expression in human peripheral blood mononuclear cells (PBMCs) when co-cultured together. Tumor cell-induced expression of IRAK-M was dependent on the activation of TGF-ß pathway. Similarly, treatment of human PBMCs or mouse macrophage cell line, RAW 264.4, with TGF-ß, induced IRAK-M expression. Interestingly, IRAK-M gene expression in 439 human lung adenocarcinoma tumors correlated with poor survival in patients with lung cancer. Together, our data demonstrates that TGF-ß-dependent induction of IRAK-M expression is an important, clinically relevant mechanism by which tumors may circumvent anti-tumor responses of macrophages.

Diminished phosphorylation of a heat shock protein (HSP 27) in infant acute lymphoblastic leukemia.

We have previously reported lack of expression of a polypeptide designated L3 in infant acute lymphoblastic leukemia (ALL). Expression of L3 occurred predominantly in older children with pre-B ALL. We have recently reported the expression during B cell ontogeny of two other polypeptides, designated L2 and L4 with a similar Mr as L3, which were identified as phosphorylated and non-phosphorylated forms respectively of the low Mr heat shock protein. hsp27. In this study we have characterized L3 and identified it as another phosphorylated form of hsp27. The two phosphorylated forms appear to be differentially expressed in acute leukemia. L3 levels in infants who expressed hsp27 isoforms L2 and L4 were significantly diminished compared to levels in older children with an equivalent amount of hsp27. We conclude that leukemic cells in infant ALL exhibit a unique pattern of phosphorylation of hsp27 expressed at a pre-B cell stage of differentiation.

Somatic mutation studies in human lymphoid cells: the detectability of quantitative genetic variants in two-dimensional gels.

The feasibility of detecting quantitative genetic variants based on a decrease in the integrated intensity of polypeptide spots in two-dimensional polyacrylamide gels of human lymphoblastoid cell clones was investigated. A battery of 65 spots on 115 gels was studied to determine the distribution of quantitative measures for spots where no mutation had occurred. The corresponding distribution for spots which have decreased integrated intensity as a result of a mutation at one of two alleles coding for the spot was investigated by quantitating spots for which mutations were known to have occurred. These two distributions allowed the estimation of the rates of false positive and false negative errors for any particular strategy aimed at detecting null mutations, and thus provides a basis for the design of efficient strategies. Our silver stained gels have sufficient reproducibility of spot integrated intensities so that, for situations in which the mutation rate is relatively high, it is practical to monitor a subset of spots for null variants using the same digitized images as are used to detect structural variants.

Two-dimensional separations of the genome and proteome of neuroblastoma cells.

Two-dimensional (2-D) electrophoretic methods have been available that allow separation of the protein constituents of a cell population. It has also become feasible to electrophoretically separate in two dimensions and to display DNA fragments derived from genomic digests. Through the appropriate choice of restriction enzymes, the functional component of the genome that encompasses CpG islands can be preferentially visualized in 2-D gels. The same computerized approach for the analysis of 2-D patterns can be applied to investigations at either the protein or DNA levels. Our group has utilized 2-D electrophoresis to investigate both protein and DNA changes in cancer. The emphasis to date has been on the identification of proteins, the abundance of which is related to specific biological features of the tumors analyzed and of DNA fragments encompassed in genomic amplifications, as the latter commonly contain growth-related genes. Findings derived from our analysis of neuroblastoma tumors and cell lines using 2-D approaches are reviewed. Data for four proteins observed in 2-D gels are presented because of our demonstrated association of these proteins with differentiation and proliferation properties of neuroblastoma. At the genomic level, the detection of amplifications using 2-D gels has necessitated an understanding of the variability displayed by multi-copy genomic fragments, which we have accomplished to a large part and which we present. An important benefit of 2-D approaches is the efficiency of scale and the ease with which abundant proteins or multicopy genomic fragments can be detected, identified and quantitatively analyzed.

High yield of restriction fragment length polymorphisms in two-dimensional separations of human genomic DNA.

We have investigated the extent to which restriction fragment length polymorphism can be detected by two-dimensional electrophoresis of end-labeled genomic restriction fragments. Genomic DNA was digested with NotI and EcoRV and labeled at the NotI recognition site before first-dimension electrophoresis in disk gels. DNA in the disk gels was further digested in situ with HinfI prior to second-dimension electrophoresis, yielding patterns in which approximately 2000 end-labeled fragments were simultaneously visualized. On the basis of studies of 6 mother/father/child trios, a group of 184 fragments was organized into 85 polymorphic systems in which all allelic fragments were detectable in the 2-D patterns. Another 206 fragments varied as to their presence among individuals, but their relatedness to other fragments was not established. Our data indicate that a large number of DNA polymorphisms can be simultaneously scored in 2-D separations of genomic DNA fragments.

Quantitative and qualitative genetic variation in two-dimensional DNA gels of human lymphocytoid cell lines.

There is a continuing need for more efficient methods to examine human (and other) populations for altered germinal and somatic cell mutation rates. To this end, we have explored the potential usefulness of two-dimensional (2-D) electrophoresis of human DNA fragments obtained from restriction-enzyme-digested genomic DNA, using samples from father/mother/child trios. On a single 2-D DNA preparation, approximately 2000 DNA fragments varying in size from 1.0 to 5.0 kbp in the first dimension and 0.3 to 2.0 kbp in the second dimension are visualized. To enter into a genetic analysis of quantitative variation, these fragments must exhibit positional and quantitative stability. With respect to the latter, if spots that are the product of two homologous DNA fragments are to be distinguished with the requisite accuracy from spots that are the product of only one fragment, the coefficient of variation of spot intensity should be approximately < or = 0.12. At present, 482 of the spots in our preparations meet these standards. In an examination of preparations based on three Japanese mother/father/child trios, 43 of these 482 spots were found to exhibit variations that segregated within families according to Mendelian principles. Additionally, of the 2000 spots, 1114 (of which the aforementioned 482 are a subset) were deemed appropriate for the study of qualitative variation. A total of 142 variable spots were identified; the heterozygosity index for these DNA fragments was 4.4%. The genetic nature of the additional variants was again established by their segregation according to Mendelian principles. We have established the feasibility of cloning fragments from such gels and determining their nucleotide sequence. This technology should be highly efficient in monitoring for mutation resulting in loss/gain/rearrangement events in DNA fragments distributed throughout the genome.

A two-dimensional electrophoresis-related laboratory information processing system: spot matching.

An approach for the computer-assisted analysis of two-dimensional gels has been developed as a part of our laboratory information processing system (LIPS). This approach relies in part on an algorithm for the pairwise matching of protein spots. The matching process initially matches spots based on a cross-correlational measure of how well neighboring spots align. While this first pass correctly determines most spot correspondences and noncorrespondences, it can make errors. Higher accuracy is obtained by monitoring the consistency of spot match decisions in a second pass, which demands that neighboring spot pairs that align spatially must also have been found to match in the first pass. Pairwise comparisons of gels are combined into n-way comparisons by matching spot lists of gels to "master" gel spot lists, which in turn are matched to higher level masters, resulting in a hierarchy of matched spots. After each pairwise match the results are reviewed and corrected with the assistance of a graphical match-editor. Results are given for 19 single-cell-derived lymphoid clones in which the presence of a mutation had previously been established, each processed in duplicate. Only one of 46 spot changes failed to be detected, which demonstrates that the strategy is sensitive and efficient for detecting qualitative spot differences.