Loss of one HuD allele on chromosome #1p selects for amplification of the N-myc proto-oncogene in human neuroblastoma cells.

In human neuroblastoma tumors, amplification of the N-myc proto-oncogene and loss of all or part of the short arm of chromosome #1 are both associated with a poor prognosis. Accruing evidence indicates that it is the absence of one allele of the HuD (ELAVL4) gene, encoding the neuronal-specific RNA-binding protein HuD and localized to 1p34, that is linked to amplification. In 12 human neuroblastoma cell lines, N-myc amplification correlates with loss of one HuD allele and decreased HuD expression. Transfection experiments demonstrate that modulating HuD expression affects N-myc gene copy number as well as expression. Introduction of a sense HuD construct into two N-myc amplified cell lines considerably increases N-myc expression whereas gene copy number decreases. Conversely, expression of antisense HuD in N-myc nonamplified SH-SY5Y cells reduces HuD and N-myc mRNA levels even as cells show amplification of the N-myc gene. Thus, N-myc gene copy number is modulated by alteration of HuD expression. We propose that haploinsufficiency of HuD due to chromosome #1p deletion in neuroblastoma selects for cells that amplify N-myc genes. Application of these findings could lead to more effective therapies in the treatment of those patients with the worst prognosis.

The human non-muscle alpha-actinin protein encoded by the ACTN4 gene suppresses tumorigenicity of human neuroblastoma cells.

alpha-Actinins are actin-binding proteins important in organization of the cytoskeleton and in cell adhesion. We have cloned and characterized a cDNA from human neuroblastoma cell variants which encodes the second non-muscle alpha-actinin isoform designated ACTN4 (actinin-4). mRNA encoded by the ACTN4 gene, mapped to chromosome 4, is abundant in non-tumorigenic, substrate-adherent human neuroblastoma cell variants but absent or only weakly expressed in malignant, poorly substrate-adherent neuroblasts. It is also present in many adherent tumor cell lines of diverse tissue origins. Cell lines typically co-express ACTN4 and ACTN1, a second non-muscle alpha-actinin gene. Expression is correlated with substrate adhesivity. Analysis of deduced amino acid sequences suggests that the two isoforms may differ in function and in regulation by calcium. Moreover, ACTN4 exhibits tumor suppressor activity. Stable clones containing increased levels of alpha-actinin, isolated from highly malignant neuroblastoma stem cells [BE(2)-C] after transfection with a full-length ACTN4 cDNA, show decreased anchorage-independent growth ability, loss of tumorigenicity in nude mice, and decreased expression of the N-myc proto-oncogene.

HuD, a neuronal-specific RNA-binding protein, is a putative regulator of N-myc pre-mRNA processing/stability in malignant human neuroblasts.

N-myc gene copy numbers and transcription rates are similar in N (neuroblastic, tumorigenic) and S (non-neuronal, non-tumorigenic) neuroblastoma cells with chromosomally integrated amplified N-myc genes. However, N cells show significantly higher N-myc mRNA levels than S cells. Therefore, post-transcriptional control of N-myc gene expression must differ between these cell types. Since no differences in N-myc mRNA half-life were found between N and S cells from two cell lines, steady-state levels of N-myc pre-mRNA processing intermediates were analysed. Results suggest that the differences in N-myc expression arise primarily at the nuclear post-transcriptional level. The neuronal-specific RNA-binding Hu proteins are present in cytoplasmic and nuclear fractions of N cells and one of them, HuD, binds specifically to both exonic and intronic N-myc RNA sequences. In sense and antisense HuD-transfected N cells, there are coordinate changes in HuD and N-myc expression levels. Thus, we propose that HuD plays a role in the nuclear processing/stability of N-myc pre-mRNA in N-type neuroblastoma cells.

Chromosomal localization of the murine RFC-1 gene encoding a folate transporter and its amplification in an antifolate resistant variant overproducing the transporter.

A variant of the L1210 cell (L1210/R83) selected in the presence of the lipophilic antifolate, metoprine, and a concentration of the natural diastereoisomer of 5-formyltetrahydrofolate, lL5CHO-folateH4, suboptimum for growth exhibited a 35-fold increase compared to parental L1210 cells in one-carbon, reduced folate transport. This was evidenced by the increase in Vmax for [3H]MTX (methotrexate) influx and a commensurate increase in the amount of the 46 kilodalton (kDa) transport protein and reduced folate carrier (RFC-1) mRNA. The variant is resistant to lipophilic antifolates, but shows collateral sensitivity to classical folate analogues. Karyotype analysis of L1210/R83 cells revealed the presence of several new chromosome abnormalities. One of these was a large, submetacentric marker chromosome comprising a normal #10 and a longer, abnormally banded arm of uncertain origin which exhibited an interstitial, palely staining, HSR-like segment. The results of Southern and Northern blotting showed that the RFC-1 gene copy number and RNA transcript level were markedly increased (30-35 fold) in L1210/R83 cells. Fluorescence in situ hybridization (FISH) analysis revealed that the HSR-like segment in these cells was the site of amplified RFC-1 genes. Independent revertant subclones, obtained following growth in the absence of selection pressure, showed four- to 12-fold decreases in [3H]MTX influx Vmax and in amount of NHS (N-hydroxysuccinimide)-[3H]MTX affinity labeled one-carbon, reduced folate transporter compared to L1210/R83 cells. RFC-1 gene copy number also decreased, and the mean length of the HSR in these revertants declined 1.6- to 5-fold. Based upon genomic nucleotide sequencing, the RFC-1 gene in the normal mouse genome was localized to chromosome 10 in close association with the alpha 1 (Col18a1) collagen gene at 10B3(locus 41cM). The close association of these genes was confirmed by other data showing that the alpha 1 collagen gene was co-amplified in L1210/R83 cells. These results document the amplification at the site of a putative HSR in an L1210 cell variant of the RFC-1 gene regulating expression of the one-carbon, reduced folate transporter.

Gene rearrangement: a novel mechanism for MDR-1 gene activation.

Drug resistance, a major obstacle to cancer chemotherapy, can be mediated by MDR-1/P-glycoprotein. Deletion of the first 68 residues of MDR-1 in an adriamycin-selected cell line after a 4;7 translocation, t(4q;7q), resulted in a hybrid mRNA containing sequences from both MDR-1 and a novel chromosome 4 gene. Further selection resulted in amplification of a hybrid gene. Expression of the hybrid mRNA was controlled by the chromosome 4 gene, providing a model for overexpression of MDR-1. Additional hybrid mRNAs in other drug-selected cell lines and in patients with refractory leukemia, with MDR-1 juxtaposed 3' to an active gene, establishes random chromosomal rearrangements with overexpression of hybrid MDR-1 mRNAs as a mechanism of acquired drug resistance.

Cell lineage and differentiation state are primary determinants of MYCN gene expression and malignant potential in human neuroblastoma cells.

Neuroblastoma cell lines typically exhibit multiple cell phenotypes, counterparts of those comprising the embryonic neural crest. Expression of the MYCN gene, usually amplified in cell lines, differs markedly among the various differentiation phenotypes. Whereas neuroblastic (N-type) and stem cell (I-type) sublines have abundant MYCN RNA and protein, S-type cells (nonneuronal neural crest precursors) have a 4- to 9-fold lower level of cytoplasmic mRNA and a 2- to 36-fold lower protein content. N and S sublines with chromosomally integrated MYCN genes have similar gene copy numbers. Thus, in these S cells, MYCN expression is downregulated. Nuclear run-on and mRNA stability assays have revealed similar transcription rates and mRNA half-lives in N and S cells from two cell lines, indicating that downregulation occurs posttranscriptionally prior to mRNA degradation in the cytoplasm. S-type cells derived from double minute chromosome-containing lines show 4- to 10-fold lower gene copy numbers than N counterparts. Experimental induction of differentiation to neuronal/neuroendocrine or to S-type cells results in a marked reduction of MYCNexpression and, in double minute chromosome-containing N-type sublines, in gene loss as well. Malignant potential as indicated by soft agar growth capacity and tumor formation in nude mice is markedly diminished in S cells and, generally, is directly proportional to MYCN mRNA levels. The most plausible relationship suggested by our data is that MYCN expression, regulated by cell lineage and/or differentiation state, directly modulates the malignant potential of human neuroblastoma cells.

HuD, a neuronal-specific RNA-binding protein, is a potential regulator of MYCN expression in human neuroblastoma cells.

HuD is one of a family of neural antigens recognised by the sera of patients with antibody-associated paraneoplastic encephalomyelitis. Localised exclusively to neurons, these proteins are among the earliest markers of the developing nervous system. Sequence analysis suggests that HuD is an RNA-binding protein. Hu protein levels were determined for the three cell types characterising human neuroblastoma cell lines: sympathoadrenal neuroblasts (N), substrate-adherent Schwann/glial/melanoblastic precursors (S) and stem cells (I) which can give rise to both N and S cells. Western blot analysis showed similar levels of protein in three N-type cell lines; S cells have no detectable Hu protein. Northern blot analysis indicated that N cells express all three Hu genes, HuD, HuC and Hel-N1. N cells, mostly from MYCN-amplified cell lines, have consistently higher steady-state levels of MYCN mRNA than S cell counterparts. Nuclear run-on and mRNA half-life experiments revealed no differences in transcription rate or mRNA stability between N and S cells from the LA-N-1 cell line, implicating differences in post-transcriptional regulation. HuD is postulated to be instrumental in splicing/processing and/or stabilisation of mRNAs involved in cell growth and neuronal differentiation. As determined by gel-mobility shift assays, HuD fusion protein binds to the 3'UTR of human MYCN mRNA. Analysis of HuD deletion mutants has demonstrated that the first and second RNA-recognition motifs (RRMs) are required for binding. Whether HuD regulates MYCN expression and thereby influences tumour aggressiveness is of major interest.

Brain-derived neurotrophic factor protects neuroblastoma cells from vinblastine toxicity.

Brain-derived neurotrophic factor (BDNF) and its receptors are necessary for the survival and development of many neuronal cells. Because BDNF and TrkB are expressed in many poor-prognosis neuroblastoma (NB) tumors, we evaluated the role of BDNF in affecting sensitivity to chemotherapeutic agents. We investigated the effects of activation of the BDNF-TrkB signal transduction pathway in two NB cell lines, 15N and SY5Y. 15N cells lack the high-affinity receptor p145TrkB and express BDNF; 15N cells were used along with 15N-TrkB cells, a subline transfected with a TrkB expression vector. In cytotoxicity assays, 15N-TrkB cells were consistently 1.4-2 fold more resistant to vinblastine than 15N cells. Drug accumulation assays showed a 50% reduction in[3H]vinblastine accumulation in 15N-TrkB cells compared with control 15N cells. Addition of 30 ng/ml BDNF resulted in a reduction to 46% of control in 15N cells and a reduction to 28% of control in 15N-TrkB cells. SY5Y cells were chosen as a second model because they lack both endogenous BDNF and TrkB expression. p145TrkB expression is induced by 1 nM retinoic acid. Vinblastine accumulation was not significantly affected by 1 nM retinoic acid in SY5Y cells. Addition of 30 ng/ml BDNF decreased [3H]vinblastine accumulation to 58% of control in SY5Y cells and decreased [3H]vinblastine accumulation to 62% of control in TrkB-expressing SY5Y cells. Although an increase in BDNF expression in seen in multidrug-resistant sublines of SY5Y and BE(2)-C NB cells, the protective effect of BDNF in vinblastine toxicity may be unrelated to mdr-1, because the activity of other agents transported by P-glycoprotein was not affected. There was no increase in mdr-1 expression in 1 nM RA SY5Y cells and 15N-TrkB cells, as assessed by Northern blot analysis. In addition to the effects of BDNF on vinblastine cytotoxicity and accumulation, there was an inhibition in the ability of vinblastine to depolymerize tubulin in BDNF-treated cells. Thus, BDNF and TrkB may partially rescue NB cells from vinblastine toxicity and thereby may contribute to a more chemoresistant phenotype.

Measurement of P-glycoprotein expression in multidrug-resistant human neuroblastoma cell lines using self-competitive binding assay.

To date, all reported measurements of multidrug resistance have been semiquantitative. The purpose of the present study is to establish and validate the self-competitive binding assay technique utilizing monoclonal antibody for quantitative estimation of multidrug resistance in tumor cells. This technique is used for P-glycoprotein concentration measurement in BE(2)-C human neuroblastoma cell line and its sublines with primary resistance to colchicine and actinomycin D. Monoclonal antibody MRK-16 was used in this study. It was labeled with iodine-125 (125I) to trace the concentration of antibody-antigen complexes. The binding data were obtained by varying the concentration of the unlabeled antibody. The results were fitted to a model equation to estimate the number of binding sites and antibody-antigen dissociation constant. The P-glycoprotein concentration was significantly higher in the resistant sublines than in the sensitive line. The highest levels were achieved in actinomycin D-resistant cells: 2.1 x 10(6) binding sites/cell versus 5.4 x 10(4) binding sites/cell in the sensitive cells. The consistency of the results was verified by repeating the study three times for each cell line. The binding assay results were confirmed by Western blot experiments performed on the same cell lines.

In vivo imaging and specific targeting of P-glycoprotein expression in multidrug resistant nude mice xenografts with [125I]MRK-16 monoclonal antibody.

Multidrug resistance (MDR) in tumors is associated with P-glycoprotein (Pgp) expression. In vivo quantitation of Pgp may allow MDR to be evaluated noninvasively prior to treatment planning. The purpose of this study was to radiolabel MRK-16, a monoclonal antibody that targets an external epitope of P-glycoprotein, and perform in vivo quantitation of P-glycoprotein in a MDR xenograft nude mouse model. MRK-16 was labeled with 125I by the iodogen method, with subsequent purification by size exclusion chromatography. Groups of 10 Balb c mice were each xenografted with colchicine-resistant or sensitive neuroblastoma cell lines, respectively. Whole body clearance and tumor uptake over time was quantitated by gamma camera imaging, and biodistribution studies were performed with [125I]MRK-16 and an isotype matched control antibody, A33. Quantitative autoradiography and immunohistochemistry analysis of tumors was also evaluated to confirm specific targetting of [125I]MRK-16. Peak tumor uptake was at 2-3 days post-injection, and was significantly greater in resistance compared to sensitive tumors (mean % injected dose/g +/- SD) (18.76 +/- 2.94 vs 10.93 +/- 0.96; p < 0.05). Quantitative autoradiography verified these findings (19.13 +/- 0.622 vs 12.08 +/- 0.38, p < 0.05). Specific binding of [125I]MRK-16 was confirmed by comparison to [131I]A33 in biodistribution studies, and localized to cellular components of tissue stroma by comparison of histologic and autoradiographic sections of sensitive and resistant tumors. Immunoblot analysis demonstrated a 4.5-fold difference in P-glycoprotein expression between sensitive and resistant cell lines without colchicine selective pressure. We conclude that in vivo quantitation of P-glycoprotein in MDR tumors can be performed with [125I]MRK-16.(ABSTRACT TRUNCATED AT 250 WORDS)

Human neuroblastoma I-type cells are malignant neural crest stem cells.

Human neuroblastoma I-type cells isolated from cell lines in vitro are morphologically intermediate between neuroblastic (N) cells, with properties of embryonic sympathoblasts, and substrate-adherent (S) cells having properties of embryonic Schwann/glial/melanocytic cells of the neural crest. I cells have biochemical features of both N and S cells. We propose that the I-type cell represents a malignant neural crest stem cell. The strongest evidence in support of this hypothesis is that: (a) I cells can generate progeny that have neuronal properties, i.e., are committed neuroblasts, or properties of nonneuronal, embryonic neural crest-derived cells; and (b) I-type cells can generate multipotent I-type progeny, indicating their capacity for self-renewal, a feature of stem cells. We report here that I-type cells, derived from four different human neuroblastoma cell lines and experimentally induced to differentiate, give rise to cells with distinct N or S cell phenotypes, indicative of I cell multipotentiality. Experiments with a large panel of I-type subclones, isolated from clonal I-type BE(2)-C cells and exposed to retinoic acid to induce neuronal differentiation or 5-bromo-2'-deoxyuridine to obtain S-type cells, demonstrated that differentiation occurs via induction and selection and not by selection of spontaneously arising variants. The differentiation phenotype was stable. We conclude that human neuroblastoma I-type cells are multipotent embryonic precursor cells of the peripheral nervous system, capable of either neuronal or nonneuronal neural crest cell differentiation.

Expression of stem cell factor and c-kit in human neuroblastoma. The Children's Cancer Group.

During development, mice with mutations of stem cell factor (SCF) or its receptor c-kit exhibit defects in melanogenesis, as well as hematopoiesis and gonadogenesis. Because melanocytes derive from neural crest cells, the role of SCF and c-kit was investigated in the neural crest-derived childhood tumor neuroblastoma. Using reverse transcription-polymerase chain reaction analysis, simultaneous expression of steady-state mRNA for the SCF ligand and its receptor c-kit was found in 14 of 14 (100%) human neuroblastoma cell lines and clones and in 8 of 18 (45%) human neuroblastoma tumor samples. Functional blockade of c-kit receptors in the cell lines SK-N-BE(2) and SH-SY5Y using the mouse monoclonal anti-c-kit antibody SR-1 resulted in a significant decrease in cellular growth rate when measured by either 3H-thymidine incorporation or clonogenicity. In addition, higher levels of c-kit mRNA expression were associated with parental neuroblastoma cell lines and subclones with a neuronal (N) differentiation phenotype, whereas lower levels of c-kit mRNA were associated with neuroblastoma cell line subclones having a schwannian/glial/melanocytic pattern of differentiation. However, the differentiation phenotype of neuroblastoma cell lines was not directly altered when c-kit expression was blocked using the SR-1 antibody. In summary, these data indicate that c-kit receptor expression may play a significant role in the growth regulation of the two neuroblastoma cell lines examined and suggest that c-kit may also play a similar role in neuroblastoma growth regulation in vivo. Simultaneous expression of SCF and c-kit mRNA in both neuroblastoma cell lines and tumors implies that c-kit may act as part of an autocrine growth loop in conjunction with endogenous production of SCF in this disease.

Biochemical and pharmacological characterization of mu, delta and kappa 3 opioid receptors expressed in BE(2)-C neuroblastoma cells.

Total opioid binding in the human neuroblastoma cell line BE(2)-C has a density similar to that found in brain, with a Bmax value of 383 +/- 60 fmol/mg protein and a KD of 0.4 +/- 0.07 nM for the nonselective opioid antagonist 3H-diprenorphine. Selective assays reveal a binding distribution of mu (38%), delta (16%) and kappa 3 (43%) opioid receptors. There is no observable kappa 1 or kappa 2 binding. The sum of the Bmax values in the selective binding assays (370 +/- 39 fmol/mg protein) approximates closely that observed with 3H-diprenorphine, suggesting that mu, delta and kappa 3 sites account for most of the binding. The binding selectivities of various opiates and opioid peptides in the BE(2)-C cells are similar to those in rat brain. Delta and mu binding are defined easily by traditional selective ligands. The binding profiles also distinguish clearly mu from kappa 3 binding. The selective mu ligand DAMGO competes with mu binding over 35-fold more potently than kappa 3 binding, whereas morphine shows a 10-fold selectivity. Functionally, selective mu, delta and kappa 3 agonists inhibit forskolin-stimulated cAMP accumulation through distinct receptor mechanisms that are pertussis toxin-sensitive. In addition to demonstrating that BE(2)-C cells provide a useful model system for studying mu, kappa 3 and delta receptors, these studies confirm that kappa 3 receptors represent a pharmacologically distinct receptor class in this cell line.

In vivo uptake of carbon-14-colchicine for identification of tumor multidrug resistance.

UNLABELLED: A major limitation in the treatment of cancer with natural product chemotherapeutic agents is the development of multidrug resistance (MDR). Multidrug resistance is attributed to enhanced expression of the multidrug resistance gene MDR1. Colchicine (CHC) is known to be one of the MDR drugs. We have previously demonstrated that it is possible to distinguish multidrug-resistant tumors from multidrug-sensitive tumors in vivo on the basis of tritium (3H) uptake following injection of 3H-CHC. METHODS: The present studies were carried out in xenografted animals using 14C-CHC which may be more indicative of 11C-labeled CHC distribution with regard to circulating metabolites, since metabolic processes following injection of (ring C, methoxy-11C)-CHC may produce significant amounts of circulating 1-carbon fragments (i.e., methanol and/or formaldehyde). Experiments were carried out at a dose of 2 mg/kg. RESULTS: Activity concentration per injected dose was approximately twice as great in sensitive as in resistant tumors (p < 0.05) at 60 min following intravenous injection of 14C-CHS. About 75% of total activity was CHC in the sensitive tumors. The findings are further confirmed by the quantitative autoradiographic evaluation of resistant and sensitive tumors. CONCLUSIONS: These studies confirm our previous observations that it is possible to noninvasively distinguish multidrug-resistant tumors from sensitive tumors in vivo based on uptake of an injected MDR drug using a 14C-labeled CHC at the same position and of comparable specific activity to a 11C-CHC tracer used for PET imaging.

Reverse transformation of multidrug-resistant cells.

Spontaneously transformed Chinese hamster lung cells with high levels of resistance (approximately 100-fold to 70,000-fold) to actinomycin D, daunorubicin, or vincristine exhibit morphology and growth patterns characteristic of normal cells in vitro and reduced tumorigenicity in vivo. These reverse transformed, multidrug-resistant cells amplify and highly overexpress one or more genes encoding P-glycoprotein. Similarly, hydrocarbon-induced mouse sarcoma cells selected with actinomycin D, vincristine, or ethidium bromide developed high levels of resistance associated with reduced drug accumulation and suppression of malignancy. To determine whether human tumor cells would undergo similar changes and whether reverse transformation reflected an altered state of differentiation, nine multidrug-resistant sublines were selected with four agents from human neuroblastoma cells with well defined pathways of differentiation. Those five with resistance levels above about 125-fold showed a reduced tumor frequency as compared to control cells. All resistant sublines showed altered differentiation. The changes in transformation phenotype appear to be intrinsic and not the result of altered immunogenicity. Two additional consequences of high level multidrug resistance have been observed: change in ganglioside composition in the Chinese hamster cells, manifested as a block in higher ganglioside biosynthesis and/or a relative increase in GM3, and increase in epidermal growth factor receptor in all three cell systems. A tentative hypothesis links ganglioside and growth factor receptor changes to the change in transformation phenotype. The basis of the reverse transformation phenomenon is not known, but the major alterations in expression of P-glycoprotein, gangliosides, and the epidermal growth factor receptor implicate, in some way, the plasma membrane.

Regulation of Bcl-2 oncoprotein levels with differentiation of human neuroblastoma cells.

When established in culture, human neuroblastoma cell lines typically are comprised of heterogeneous cellular subpopulations, including neuroblastic (N-type), substrate-adherent (S-type), and intermediate (I-type) cells that can be distinguished by their characteristic morphologies and expression of differentiation-associated antigens. Here we examined the relative levels of the Bcl-2 oncoprotein in 15 clones derived from four different neuroblastoma cell lines. Among six clones isolated from the SK-N-SH line, levels of p26-Bcl-2 correlated with morphology and differentiation markers with the hierarchy of bcl-2 expression being: N-type cells > N/I-type > I-type > S-type. Furthermore, stimulation of one of the N-type clones, SH-SY5Y, with the phorbol ester, 12-O-tetradecanoylphorbol-13-acetate, induced differentiation toward a more neuronal-like phenotype and resulted in a 5- to 10-fold elevation in the relative levels of Bcl-2 protein. High relative amounts of p26-Bcl-2 protein were also found in an N-type clone derived from the SMS-KCN line. In two N-type clones derived from the LA-N-1 line, however, levels of Bcl-2 protein were only moderately elevated, and in one N-type clone from the SK-N-BE(2) line the levels of Bcl-2 protein were low. Thus, high relative levels of Bcl-2 oncoprotein are not a universal feature of N-type cells (three of six clones tested). In contrast, all 5 of the S-type clones evaluated contained relatively low levels of Bcl-2 protein, suggesting that these cells (which may represent embryonic precursors of Schwann, glial, and melanocytic cells) do not typically express the bcl-2 gene at high levels. Consistent with this inverse correlation between Bcl-2 protein levels and S-type characteristics, stimulation of an I-type clone derived from the SK-N-BE(2) line with 5-bromodeoxyuridine was accompanied by an accumulation of S-type cells in these cultures, decreased Bcl-2 protein, diminutions in the neuronal markers neurofilament-M and neuron-specific enolase, and an increase in the relative levels of the S-type marker proteins vimentin and beta-2-microglobulin. Conversely, stimulation of this I-type clone with retinoic acid resulted in an accumulation of N-type cells (which are thought to represent embryonic precursors of sympathetic neurons), decreased vimentin and beta-2-microglobulin, increased neurofilament-M, and a marked elevation in p26-Bcl-2.(ABSTRACT TRUNCATED AT 400 WORDS)