Graph models of oncogenesis with an application to melanoma.

We describe several analytical techniques for use in developing genetic models of oncogenesis including: methods for the selection of important genetic events, construction of graph models (including distance-based trees, branching trees, contingency trees and directed acyclic graph models) from these events and methods for interpretation of the resulting models. The models can be used to make predictions about: which genetic events tend to occur early, which events tend to occur together and the likely order of events. Unlike simple path models of oncogenesis, our models allow dependencies to exist between specific genetic changes and allow for multiple, divergent paths in tumor progression. A variety of genetic events can be used with the graph models including chromosome breaks, losses or gains of large DNA regions, small mutations and changes in methylation. As an application of the techniques, we use a recently published cytogenetic analysis of 206 melanoma cases [Nelson et al. (2000), Cancer Genet. Cytogenet.122, 101-109] to derive graph models for chromosome breaks in melanoma. Among our predictions are: (1) breaks in 6q1 and 1q1 are early events, with 6q1 preferentially occurring first and increasing the probability of a break in 1q1 and (2) breaks in the two sets [1p1, 1p2, 9q1] and [1q1, 7p2, 9p2] tend to occur together. This study illustrates that the application of graph models to genetic data from tumor sets provide new information on the interrelationships among genetic changes during tumor progression.

Chromosome abnormalities in malignant melanoma: clinical significance of nonrandom chromosome abnormalities in 206 cases.

We report the cytogenetic abnormalities from a series of 206 primary malignant melanoma specimens referred to a single institution. A total of 169 out of 206 unique cases had chromosome breakpoints. A previously described statistical method was used to detect nonrandom distribution of chromosome breakpoints at the level of chromosome regions. Nonrandom occurrence of chromosome breakpoints (indicating that the observed number of breaks significantly exceeded the expected number of breaks) was detected in 28 regions, suggesting a hierarchy of genetic abnormalities in melanoma. Clinical variables and tumor characteristics were analyzed for associations with the presence of any nonrandom chromosome breakpoints; with individual, nonrandomly involved chromosome regions; and with paired, nonrandomly involved chromosome regions. No nonrandomly involved chromosome regions or pairs of regions appeared to significantly affect survival. These results identify recurring, nonrandom chromosome abnormalities in malignant melanoma. These results suggest that recurring, nonrandom chromosome alterations play a key role in the etiology and/or progression of malignant melanoma and identify targets within the genome for molecular genetic studies.

Chromosome abnormalities in ovarian adenocarcinoma: III. Using breakpoint data to infer and test mathematical models for oncogenesis.

Cancer geneticists seek to identify genetic changes in tumor cells and to relate the genetic changes to tumor development. Because single changes can disrupt the cell cycle and promote other genetic changes, it is extremely hard to distinguish cause from effect. In this article we illustrate how 7 techniques from statistics, theoretical computer science, and phylogenetics can be used to infer and test possible models of tumor progression from single genome-wide descriptions of aberrations in a large sample of tumors. Specifically, we propose 4 tree models for tumor progression inferred from the large ovarian cancer data set described in the first 2 articles in this series. The models are derived from 2 different methods to select the non-random genetic aberrations and 2 different methods to infer the trees, given a set of events. Various aspects of the tree models are tested and extended by 5 methods: overall tests of independence, likelihood ratio tests, principal components analysis, directed acyclic graph modeling, and Bayesian survival analysis. All our methods lead to strikingly consistent conclusions about chromosomal breakpoints in ovarian adenocarcinoma, including (1) the non-random breakpoints in ovarian adenocarcinoma do not occur independently; (2) breakpoints in regions 1p3 and 11p1 are important early events and distinguish a class of tumors associated with poor prognosis; and (3) breakpoints in 1p1, 3p1, and 1q2 distinguish a class of ovarian tumors, and the breaks at 1p1 and 3p1 are associated with poor prognosis.

Characterization of MGI 114 (HMAF) histiospecific toxicity in human tumor cell lines.

PURPOSE: The acylfulvenes are a class of antitumor agents derived from the fungal toxin illudin S. One acylfulvene derivative, MGI 114 (HMAF), demonstrates marked efficacy in xenograft carcinoma models when compared to the parent acylfulvene or related illudin compounds. The maximum tolerated dose (MTD) of the two analogs in animals, however, is similar. To help elucidate the basis of the increased therapeutic efficacy of MGI 114, we determined the in vitro cytotoxicity, cellular accumulation and DNA incorporation of this drug and compared the results with those from the parent acylfulvene analog. METHODS: The cytotoxicity of acylfulvene analogs was tested in vitro against a variety of tumor cell lines. Radiolabeled MGI 114 was used for cellular accumulation and DNA incorporation studies. RESULTS: MGI 114 retained relative histiospecific toxicity towards myeloid leukemia and various carcinoma cell lines previously noted with the parent acylfulvene compound. Markedly fewer intracellular molecules of MGI 114 were required to kill human tumor cells in vitro as compared to the parent acylfulvene, indicating that MGI 114 was markedly more toxic on a cellular level. At equitoxic concentrations, however, the incorporation of MGI 114 into genomic tumor cell DNA was equivalent to that of acylfulvene. Analysis of cellular accumulation of MGI 114 into tumor cells revealed a lower Vmax for tumor cells, and a markedly lower Vd for diffusion accumulation as compared to acylfulvene. CONCLUSIONS: The addition of a single methylhydroxyl group to acylfulvene to produce MGI 114 results in a marked increase in cytotoxicity in vitro towards tumor cells as demonstrated by the reduction in IC50 values. There was a corresponding decrease in the number of intracellular molecules of MGI 114 required to kill tumor cells, but no quantitative alteration in covalent binding of the drugs to DNA at equitoxic concentrations. This indicates that cellular metabolism plays a role in the in vitro cytotoxicity of MGI 114. The equivalent incorporation into genomic DNA at equitoxic doses suggests that DNA damage produced by acylfulvene and MGI 114 is equivalent in regard to cellular toxicity and ability to repair DNA. This increased cellular toxicity, together with the decrease in diffusion rate, may explain the increased therapeutic efficacy of MGI 114 as compared to the parent acylfulvene analog.

Chromosome abnormalities in ovarian adenocarcinoma: I. Nonrandom chromosome abnormalities from 244 cases.

Cytogenetics provides important insights into the molecular pathogenesis of human cancers. Although extensive data exist on recurring cytogenetic abnormalities in hematologic cancers, data on individual solid tumor types remain limited. Previous studies of ovarian carcinoma indicated the presence of multiple, complex clonal chromosome abnormalities. Cytogenetics remains one of a few techniques capable of detecting these multiple, simultaneously occurring genetic abnormalities. We describe cytogenetic abnormalities from a series of 244 primary ovarian cancer specimens referred to a single institution. A total of 201/244 cases had fully characterized clonal chromosome abnormalities, of which 134 showed clonal chromosome breakpoints. We used a novel statistical technique to detect nonrandom chromosome breakpoints at the level of chromosome regions. Nonrandom occurrence of chromosome breakpoints was detected at regions 1p1*, 1q1*, 1p2*, 1q2*, 1p3*, 1q3, 3p1*, 1q4*, 6q1*, 6p2, 6q2, 7p1*, 7q1, 7p2*, 11p1*, 11q1, 11q2*, 12p1, 12q2*, 13p1, and 19q1. Simultaneous occurrence of multiple abnormalities was common. However, 120/134 cases had breakpoints at one or more of 13 commonly involved regions (*), suggesting a hierarchy of genetic abnormalities. Among clinical and tumor variables that predict patient survival, tumor grade was significantly associated with the presence of chromosome breakpoints. In additional studies, we show that nonrandom chromosome abnormalities are associated with impaired survival in ovarian cancer and that specific, nonrandomly involved chromosome regions retain significant effects on survival when analyses are controlled for important clinical variables. Additional specific chromosome abnormalities in this series are described, including chromosome gains and losses in near-diploid cases and homogeneously staining regions. These results suggest that recurring, nonrandom chromosome abnormalities are important in the pathogenesis and/or progression of ovarian cancers, and target areas of the genome for molecular genetic studies.

Frequent alterations of evolutionarily conserved regions of chromosome 1 in human malignant melanoma.

Recurring alterations of chromosome 1 represent the most frequent site of structural chromosome abnormalities across all human solid tumors, including human cutaneous malignant melanoma. In melanoma, breakpoints involving chromosome 1 appear to accumulate most frequently at the paracentromeric regions, and secondly, to cluster within 1p36. Of interest, these three band regions (1p11-12, 1q21, and 1p36) were simultaneously recognized by a single YAC clone which was isolated from sequences mapping to 1q21. This observation indicates the common and highly conserved nature of sequences residing within these three bands. Because of this finding, we have examined the possible association of these recurring sites of rearrangements of chromosome 1 in malignant melanoma. To elucidate genomic alterations in these regions, we have analyzed melanoma samples simultaneously by fluorescence in situ hybridization (FISH) using both the YAC clone encoding 1p11, 1q21, and 1p36 homologous sequences, and an alpha-satellite probe for the chromosome 1 centromere. Twelve of 20 (60%) randomly selected melanoma cell lines showed detectable rearrangements in one or more of the chromosome 1 band regions. These results provide support for the notion that the homology between these regions is associated with chromosomal instability, and possibly, is of biologic relevance in malignant melanoma.

Chromosome abnormalities in ovarian adenocarcinoma: II. Prognostic impact of nonrandom chromosome abnormalities in 244 cases.

In a large series of ovarian carcinomas from 244 patients, 134 cases had chromosome rearrangements. We showed before that the pattern of chromosome breakpoints involved 21 separate chromosome regions nonrandomly and, in 90% of cases with breaks, the breakpoints occurred within 13 commonly involved regions. Log-rank and proportional hazards regression analyses showed that the aggregate presence of a chromosome breakpoint in any of 21 nonrandomly involved regions and breaks in 9 distinct regions (1p1, 1q2, 1p3, 3p1, 6p2, 11p1, 11q1, 12q2, and 13p1) were associated with reduced patient survival. Breakpoints in other areas of the genome, including other nonrandomly involved regions, were not associated with decreased survival. Because many cases had breakpoints in more than one nonrandomly involved region, proportional hazards regression was also used to analyze for effects of each nonrandomly involved region, controlling for effects of other regions. With this approach, only breakpoints within 1p1 and 3p1 retained independent, deleterious effects on survival. Similarly, when nonrandomly involved regions were entered into a proportional hazards model containing clinical variables associated with altered patient survival (tumor grade, tumor stage, and residual disease > 1 cm after resection), only 1p1 (P = 0.007) and 3p1 (P = 0.04) were associated with independent, negative effects on survival. These studies demonstrate that chromosome breakpoints within specific, nonrandomly involved chromosome regions are associated with impaired survival in ovarian cancers. Regions 1p1 and 3p1 are identified as areas of particular significance and are appropriate targets for analytical techniques such as SAGE and microarray analysis.

Limited erythropoietic response to combined treatment with recombinant human interleukin 3 and erythropoietin in myelodysplastic syndrome.

Twenty-two patients with myelodysplastic syndrome were treated with combined recombinant human erythropoietin and recombinant human interleukin 3 (rHuIL-3). All 22 patients were evaluable for toxicity and 21/22 for response. Thirteen patients (62%) required rHuIL-3 dose reduction because of toxicity. Nineteen experienced a 50% or greater rise in neutrophil count. Of seven patients with initial platelet counts of > 100,000, three experienced increases of > 15,000/ml while an equal number had a comparable decline. Five patients (21%) experienced a significant rise in reticulocyte count, and two transfusion-dependent patients experienced a significant decrease in transfusions. Erythroid burst-forming units were increased by > or = 50% in nine of 11 patients after combined therapy. Similar changes were seen in multipotential colony-forming units. Clinical responses were comparable to results obtained with epo alone while toxicities due to IL-3 were significant.

Abnormalities in the p34cdc2-related PITSLRE protein kinase gene complex (CDC2L) on chromosome band 1p36 in melanoma.

The two genes encoding the PITSLRE protein kinase isoforms, CDC2L1 and CDC2L2, are localized to human chromosome band 1p36. The PITSLRE protein kinases are a part of the p34cdc2 supergene family. Several protein products of the CDC2L locus may be effector(s) in apoptotic signaling. The larger PITSLRE p110 isoforms appear to regulate some aspect of RNA splicing/transcription during the cell cycle. One or more of these genes may function as tumor suppressor genes in melanoma. Using fluorescence in situ hybridization, one allele of the CDC2L gene complex on chromosome 1 was either deleted or translocated in 8 of 14 different melanoma cell lines. We also observed mutations in the 5' promoter region of the CDC2L1 gene in four different cell lines relative to normal melanocytes using PCR-SSCP analysis and direct DNA sequencing. Western blot analysis revealed decreased level of PITSLRE protein expression in several cell lines, as well as in four surgical malignant melanoma specimens relative to normal melanocytes. Thus, the decreased PITSLRE protein expression appears to result from deletion of the CDC2L alleles and possibly by mutations within the 5' promoter region. We propose that aberrations in the CDC2L genes may contribute to the pathogenesis or progression of melanoma.

Detection of chromosome 6 abnormalities in melanoma cell lines by chromosome arm painting probes.

Chromosome 6 abnormalities, particularly the deletion of 6q, are among the most frequent chromosomal changes in human malignant melanoma. In this study, chromosome 6 rearrangements in 21 melanoma cell lines were identified by fluorescence in situ hybridization (FISH) with the use of chromosome 6 arm specific painting probes (CAPs). Structural abnormalities of chromosome 6 were detectable in 18 of 21 (86%) cases, including the identification of structural abnormalities that were not detectable by routine G-banding analysis. The results of this study correlate the frequency of 6q alterations and demonstrate that FISH with CAPs can be of considerable use for rapid screening of specific chromosome abnormalities in melanoma.

Efficacy of MGI 114 (6-hydroxymethylacylfulvene, HMAF) against the mdr1/gp170 metastatic MV522 lung carcinoma xenograft.

Illudins are a novel class of agents with a chemical structure entirely different from current chemotherapeutic agents. A new semisynthetic derivative, MGI 114 (NSC 683,863, 6-hydroxymethyl-acylfulvene, HMAF), is markedly effective in a variety of lung, breast and colon carcinoma xenograft models. This analogue, MGI 114, is currently in phase I human clinical trials, and is scheduled for two different phase II trials. To determine if MGI 114 could be effective in vivo against mdr tumour cells, we generated an mdr1/gp170-positive clone of the metastatic MV522 human lung carcinoma line by transfecting a eukaryotic expression vector containing the cDNA encoding for the human gp170 protein. This MV522/mdr1 daughter line retained the metastatic ability of parental cells. The parental MV522 xenograft is mildly responsive in vivo to mitomycin C and paclitaxel, as evidenced by partial tumour growth inhibition and a small increase in life span, whereas MV522/mdr1 xenografts were resistant to these agents. In contrast to mitomycin C and paclitaxel, MGI 114 produced xenograft tumour regressions in 32 of 32 animals and completely eliminated tumours in more than 30% of MV522/mdr1 tumour-bearing mice. Thus, MGI 114 should be effective in vivo against mdr1/gp170-positive tumours.

Treatment with low-dose oral etoposide in patients with myelodysplastic syndromes.

Forty-three patients with myelodysplastic syndromes (MDS) received treatment with oral etoposide 50 mg/day for 21 consecutive days every 4 weeks. Eighteen patients (42%) experienced hematological responses, including 12 of 17 (70%) patients with chronic myelomonocytic leukemia (CMML). Three of five CMML patients who failed treatment with hydroxyurea experienced major hematological responses with oral etoposide. Median response duration exceeded 9 months (range: 4-49 + months), and one patient remains in an unmaintained complete remission for 4 years. Toxicity included nausea/vomiting in five patients, fever (four patients), infection (three patients), mucositis (two patients), and anorexia (two patients). Two patients had grade 4 neutropenia with sepsis necessitating treatment withdrawal. We conclude that low-dose oral etoposide has remitting activity in MDS and is an effective treatment alternative for patients with CMML.

Characterization of acylfulvene histiospecific toxicity in human tumor cell lines.

PURPOSE: Acylfulvene derivatives demonstrate marked efficacy in xenograft carcinoma models as compared with the parent illudin compounds. To elucidate the increased therapeutic efficacy of acylfulvene analogs, we compared them with the illudin compounds in terms of their in vitro cytotoxicity, cellular accumulation and DNA incorporation. METHODS: The cytotoxicity of various acylfulvene analogs was tested in vitro against a variety of tumor cell lines. Radiolabelled acylfulvene analog was prepared and used for cellular accumulation and DNA incorporation studies. RESULTS: The prototype acylfulvene analog retained selective histiospecific toxicity towards myeloid leukemia and various carcinoma cell lines. In vitro killing of tumor cells by acylfulvene required up to a 30-fold increase in molecules per cell, as compared with illudin S, indicating that acylfulvene was less toxic on a cellular level. At equitoxic concentrations, acylfulvene incorporation into genomic tumor cell DNA was equivalent to illudin S suggesting that cellular metabolism has a role in acylfulvene cytotoxicity. Analysis of cellular accumulation of acylfulvene into tumor cells revealed a markedly higher Vmax for tumor cells, and a lower Vd for diffusion accumulation into other cells. CONCLUSIONS: The combination of higher Vmax and lower Vd may explain the increased in vivo efficacy of acylfulvene.

Stimulation of hematopoiesis by amifostine in patients with myelodysplastic syndrome.

The aminothiol, amifostine (Ethyol; U.S. Bioscience, West Conshohocken, PA), is a cytoprotective agent that ameliorates the toxicities of anticancer therapy. In vitro, amifostine promotes the formation and survival of primitive hematopoietic progenitors derived from myelodysplastic bone marrow (BM) specimens. To evaluate the hematological effects of amifostine, 18 patients with myelodysplastic syndrome (MDS) and one or more refractory cytopenias received treatment with amifostine in a Phase I/II study. Four cohorts received intravenous treatment with 100, 200, or 400 mg/m2 amifostine three times a week, or 740 mg/m2 weekly for three consecutive weeks followed by 2 weeks observation. Nonresponding patients received a second course of therapy at the next higher dose level depending upon drug tolerance. Bone marrow (BM) progenitor growth was assessed before treatment and after day 21. Diagnoses included refractory anemia (7), refractory anemia with ringed sideroblasts (5), refractory anemia with excess blasts (RAEB) (4), and RAEB-in transformation (RAEB-t) (2). Single- or multi-lineage hematologic responses occurred in 15 patients (83%) treated with the three-times-a-week dose schedule. Fourteen patients had a 50% or greater increase in absolute neutrophil count with amifostine treatment (range, 426 to 11,348/microL). Platelet count increased in 6 (43%) of 14 patients with thrombocytopenia (absolute increase, 16, 000 to 110,000/microL), and 5 of 15 red blood cell transfusion-dependent patients had a 50% of greater reduction in transfusion needs. Assayable hematopoietic progenitors increased in 13 of 15 evaluable patients; including CFU-GEMM (12), BFU-E (8), and CFU-GM (6). Amifostine doses less than or equal to 200 mg/m2 were well tolerated, whereas grade II nausea, vomiting, and fatigue was limiting at higher doses. Three patients with excess blasts before enrollment experienced an increase in BM blast percentage and two patients had evolution to acute leukemia that persisted after treatment withdrawal. We conclude that amifostine administered at doses

Increased chromosome 20 copy number detected by fluorescence in situ hybridization (FISH) in malignant melanoma.

DNA amplification is an important mechanism of tumor progression that allows cancer cells to up-regulate the expression of critical genes such as oncogenes and genes conferring drug resistance. Recent studies using comparative genomic hybridization (CGH) revealed increased DNA copies of 20q sequences in 7 melanoma cell lines and B archival metastatic melanoma lesions. To evaluate chromosome 20 abnormalities in more detail and to resolve discrepancies between karyotype and CGH findings, we performed FISH analysis of metaphase cells in 13 melanoma cell lines (including the 7 lines used for CGH) and 9 primary melanoma specimens by using a whole chromosome paint specific for chromosome 20. All 13 cell lines (100%) and 8/9 primary tumors (89%) showed extra copies of chromosome 20 relative to tumor ploidy. Additionally, 6/14 cell lines (43%) and 2/8 primary tumors (25%) showed translocated chromosome 20 material previously undetected by standard cytogenetics. Cytologic evidence for gene amplification was also found in one cell line, which contained an add(20)(p13), with additional DNA being derived from 20q sequences. These data suggest that overrepresentation of a gene or genes important for melanoma pathogenesis resides on the long arm of chromosome 20.

Band 1p36 abnormalities and t(1;17) in ovarian carcinoma.

In a series of 128 karyotyped ovarian carcinomas, 42% of cases with chromosome 1 clonal structural abnormalities had breaks at band 1p36 (usually involving translocations of unknown material). Fluorescent in situ hybridization (FISH) studies using combinations of 1 centromere and 1p36.3-specific probes (16 cases) or 1 centromeric and 17 whole-chromosome paint probes (11 cases with 1p+) revealed a trend toward deletion of 1pter relative to 1 centromere (63%); intratumor heterogeneity; and the origin of 1p+ in 3/11 cases (27%) from chromosome 17 [t(1;17)(p36;?)]. The frequency of this specific breakpoint and its involvement in recurrent translocations suggest that these regions are loci for genes important in the pathogenesis of a subset of sporadic ovarian carcinomas.

Expression of the Zn finger gene, EVI-1, in acute promyelocytic leukemia.

The EVI-1 gene encodes a Zn finger, DNA binding protein previously detected in some acute myelogenous leukemias (AML) and myelodysplasias (MDS), but not in normal marrow or cord blood cells. Experimental studies suggest EVI-1 blocks cellular differentiation by binding to GATA-1 or other specific DNA sequences controlling gene expression, and may be involved in the pathogenesis of some AMLs. To further define potential roles for EVI-1 in leukemia pathogenesis, we studied its regulation in acute promyelocytic leukemias (APL). Seven of 11 APL cases expressed EVI-1 RNA detected by RNA PCR at diagnosis, and expression was detected in two additional cases after treatment with all-trans retinoic acid (ATRA). Two of four cases studied at relapse also expressed EVI-1 RNA. To investigate regulation of EVI-1 expression in APL, we examined its expression in the NB4 APL cell line. NB4 cells did not express EVI-1 under basal conditions, but expressed EVI-1 after ATRA-induced differentiation. When NB4 cells were exposed to ATRA and transferred to cultures with N,N'-hexamethylene-bis-acetamide (HMBA), differentiation occurred but EVI-1 RNA was not detected, indicating that EVI-1 expression was not required for terminal, NB4 differentiation. ATRA-resistant NB4 cells were obtained by continuous culture in gradually increasing concentrations of ATRA. These cells did not express markers of differentiation but continued to express EVI-1 for several weeks even after ATRA withdrawal. To assess whether expression of the APL PML-RAR alpha fusion gene alone was sufficient for ATRA induction of EVI-1, the PML-RAR alpha gene cDNA was expressed in U937 histiocytic lymphoma cells. ATRA treatment of PML-RAR alpha-transfected or control U937 cells did not induce EVI-1 expression. In conclusion, this study demonstrates the EVI-1 gene is consistently expressed in APL cells either constitutively or after ATRA treatment. ATRA represents the first biologically active agent shown to specifically regulate EVI-1 expression in blood cells. In contrast to previous studies in AML and MDS, the pattern of EVI-1 expression suggests it may facilitate rather than inhibit myeloid differentiation during ATRA treatment. However, effects of EVI-1 expression are likely to be complex, and expression in ATRA-resistant APL cells may indicate multiple roles for this gene.

Characterization of cellular accumulation and toxicity of illudin S in sensitive and nonsensitive tumor cells.

Illudins are novel low molecular weight natural products cytotoxic to human tumor cells in vitro. Illudin-derived analogs are effective against experimental human cancers nonresponsive to conventional anticancer agents. It is not known why some illudin analogs are more efficacious in vitro and in vivo than other analogs. Therefore, the in vitro cytotoxicity of the parent compound illudin S towards tumor cells was characterized using radiolabeled drug. Two cell lines sensitive at nanomolar concentrations using only a 15-min exposure period displayed a saturable, energy-dependent accumulation of illudins with relatively low K(m) and high Vmax values. A nonsensitive cell line, requiring millimolar concentrations to achieve in vitro toxicity, showed minimal illudin uptake with higher K(m) and lower Vmax values. No release of radioactivity could be demonstrated from tumor cells, indicating that there was no efflux of illudin S (or metabolites) from these cells. The number of intracellular illudin S molecules required to kill 50% of cells of different tumor cell lines varied from 78000 to 1114000 molecules per cell and was correlated with the 2-h IC50 value determined using a colony-forming assay. Illudin S was cytotoxic to a variety of multidrug-resistant tumor cell lines regardless of whether resistance was mediated by gp170/mdrl, gp180/MRP, GSHTR-pi, topoisomerase I, topoisomerase II, increased DNA repair capacity, or alterations in intracellular thiol content. Information obtained in this study could be used to design clinical phase I trials and to develop analogs with improved therapeutic indexes.

Clinical implications of cytogenetic abnormalities in melanoma.

Unlike leukemia, in which specific reciprocal translocations are frequently observed, melanomas involve complex recurring chromosome anomalies. Analysis of the constituted genome of melanoma patients should identify cancer susceptibility genes and at-risk individuals in families with a history of melanoma. The first of these genes to be cloned is the cell cycle regulatory protein inhibitor--the p16 gene-- and a second gene locus for melanoma predisposition has been linked to the chromosome 1p36 band region. Detection of the most common somatic genetic alterations in melanoma enhances our understanding of molecular mechanisms of melanoma development and may lead to genetic markers in melanoma. Some alterations may be used to identify interesting subpopulations. Others may be of prognostic value when they are considered in tandem with clinical data.

Iron deprivation increases erythropoietin production in vitro, in normal subjects and patients with malignancy.

Although tissue hypoxia is the major stimulus for erythropoietin (EPO) production, serum EPO (sEPO) levels at any given Hb in iron-deficiency anaemia are relatively higher than in other anaemias. Iron chelators stimulate erythropoiesis in anaemia of chronic disease via unknown mechanisms. A recent study suggested that deferoxamine (DFO) regulates steady-state EPO RNA. Here we report that altered intracellular iron balance regulates EPO production both in vitro and in two unique clinical trials. In vitro, both iron chelation with DFO and blockade of Tf-mediated iron uptake with anti-Tf receptor antibody 42/6, stimulated EPO production in serum-deprived hepatoma cells. Conversely, iron repletion by haemin, inhibited EPO production in these cells. In clinical studies, sEPO levels rose in adult volunteers treated with DFO coupled to hydroxyethyl starch (HES-DFO) and in patients with advanced malignancy treated with anti-Tf receptor antibody 42/6, in a time- and dose-dependent manner. These studies indicate intracellular iron balance regulates EPO production in humans.