Gene combinations inducing neoplasms in Drosophila.

Riede, Isolde.  Tumorine AG, Im Amann 7, D-88662 Überlingen, Germany;  Tel. 49-7551-916850;  Fax. 916849. 

Abstract

Drosophila mutant Malignant Brain Tumor carries mutations both in proliferative genes and developmental genes. Mutant proliferative genes break the restriction of the cell cycle, induce somatic pairing gaps of polytene chromosomes, allow replication and genome instability. Neoplasms are coupled to a proliferative gene defect. Tumor suppressor activity maps to 85cM on the third chromosome and was identified as tolloid allele. Transgenic Malignant Brain Tumor with addition of wild type tolloid partially rescue tumor formation: neuroblasts regain their ability to differentiate and lose malignancy. Still hyperplasia occurs, indicating that the tumor suppressor mutation adds to tumor formation, but that the initial tumor forming mutation, loss of cell cycle control, is not due to tumor suppression. Toll and pelle alleles induce tumor formation in trans over Malignant Brain tumor as oncogenes. Oncogenes or tumor suppressor are unable to induce overgrowth of cells, they give rise to tissue specificity of tumor formation. Neoplasms can be induced by a minimum of one dominant proliferative gene defect, two recessive proliferative gene defects, one proliferative gene defect over one oncogene defect, one proliferative gene defect over one tumor suppressor gene defect, or one proliferative gene defect over one deficiency.

Introduction

Both in humans and in Drosophila, malignant tumor formation is due to a multigenic process. Proto-oncogenes which can change their activity to become dominant cell growth promoting genes, are instrumental in developmental processes such as cell communication, signal transduction and regulation of gene expression. A number of human cancers are associated with mutation or loss of both copies of a tumor suppressor gene. Their protein products keep in check an otherwise uncontrollable ability of cells to proliferate (Bishop, 1982; Levine et al., 1991; Weinberg, 1995). Cancer causing alleles of tumor suppressors are usually recessive, whereas oncogenic alleles can act in a dominant fashion to promote cell overgrowth. Tumors in humans are rarely associated with single gene defects (Foulds, 1958; Nowell, 1976; Peto, 1977). Two major features are altered in tumor cells. First they lose control over the cell cycle and proliferate in an unregulated fashion. Second they lose or alter differentiation. It is not yet completely understood which gene defects are involved in tumor induction, promotion and progression. Although general tumor suppressor genes are known (Levine et al., 1991; Kamb et al., 1994; Nobori et al., 1994), no model system has been established so far, which can distinguish between single effects on differentiation, proliferation, and malignancy.

            The Drosophila line Malignant Brain Tumor (MBT) has been subject to genetic analysis. This has shown the existence of proliferative genes involved in tumor formation (Riede, 1996). Mutant proliferative genes break the restriction of the cell cycle and induce somatic pairing gaps of polytene chromosomes (Riede, 1997). They induce cell overgrowth in different tissues and polytenization of brain cells, allow replication. Proliferative genes do not induce a brain tumor; accordingly a different class of genes induces differentiation defects, causing undetermined brain cells which are competent to induce the cell cycle.

            The early events during Drosophila embryogenesis, from the view of a single cell, involve a modulation of two features. First, the expression pattern of the genes active in the G1 or G2 phase, as well as the functions of the cell within the organism has to be defined. This type of modulation determines the cell fate and differentiates the tissue. Second, a programmed number of cell cycles has to be driven. Differentiation and proliferation of cells require an equilibrium. Excessive proliferation will lead to more cells than programmed, leading to hyperplasic growth. Premature differentiation of several cells, before the programmed number of cell cycles is driven, will lead to missing cells: organs or structures are incomplete. A lack of differentiation in addition with activity of the cellular program to proliferate will lead to an overgrowth of undifferentiated tissue and, accordingly, to a tumor.

            The establishment of the basic features of the dorsal-ventral embryonic pattern requires specifically the action of maternal effect genes (Anderson and Nüsslein-Volhard, 1984). Signal transduction is triggered on the ventral side of the Drosophila embryo through the binding of an extracellular ligand to the Tl (Toll) receptor (Stein and Nüsslein-Volhard, 1992). Tl encodes a transmembrane protein, homologous to the human Interleukin-1 Receptor (Hashimoto et al., 1988). spz (spätzle) acts immediately upstream of Tl in the genetic pathway and represents a component of the extracellular signaling pathway (Morisato and Anderson, 1994; Chasan and Anderson, 1993). Downstream of Tl, the action of pll (pelle) encoding a protein kinase, is required (Shelton and Wassermann, 1993). The product of tube (tub) is a protein that disrupts the interaction of Tl with pll (Letsou et al., 1991, 1993). pll leads to dissociation of the dorsal (dl)-cactus (cact) complex and to dl nuclear import (Roth et al., 1991). dl is a member of the rel family of transcription factors (Steward, 1987). Accordingly, proto-oncogenes are members of the dorsal-ventral patterning pathway. The gene activities of spz, Tl, pll and tub result in a ventral to dorsal gradient of dl protein in the nuclei of the syncytial blastoderm of the embryo. This gradient gives rise to region specific expression of zygotic genes. One of them, tld (tolloid), is transcribed by nuclei in the dorsal-most 40 % of the blastoderm embryo.

tld is homologous to Human Bone Morphogenetic Protein 1 (Shimell et al., 1991). The protein consists of an amino terminal metalloprotease domain followed by an interaction domain including complement protein repeat motifs and epidermal growth factor type three domains. Two domains of the protein contribute to its genetic interaction with decapentaplegic (dpp), a member of the TGFß family (Childs and O'Connor, 1994). This family of extracellular factors can stimulate or inhibit cell growth or differentiation, depending on the cell type involved (Massagué, 1990).

Here I show that tld appears as the tumor suppressor gene in MBT. In MBT tumor formation, the role of tld⁺ lies in differentiation of neuroblasts. Another mutant gene, spz^MBT, belongs to the pathway inducing dorso-ventral polarity of the embryo. Out of the pathway, Tl and pll alleles are able to induce brain tumor formation in trans over MBT. The oncogenes Tl and pll and the tumor suppressor gene tld encode proteins that determine cell differentiation: developmental genes. Mutations in developmental genes alone do not break the restriction of the cell cycle. Distinct combinations of oncogenes, tumor suppressor, proliferative genes and deficiencies induce neoplasm formation in Drosophila.

Results

Mapping of tumor suppressor activity to vicinity of tld

To avoid interference with mutant genes causing lethality and not being involved in tumor formation, the third chromosome of MBT was recombined twice with wild type. From fifty recombinants one strain was selected, MBT*, which was able to induce brain tumor as heterozygote over MBT. To map tumor suppressor activity, MBT* was recombined with ru st e ca. Recombinants were tested for their ability to induce brain tumor formation as heterozygotes over MBT. Out of 280 recombinants, 21 ru st e and 22 ca marked recombinants induced brain tumor formation over MBT. This localizes tumor suppressor activity to 85cM on the third chromosome. Independently, the same crosses were performed with markers st e tx. Tumor formation in 7% of MBT/recombinant MBT tx, based on 200 recombinants, indicates, that the tumor suppressor locus is 7 cM proximal from tx, i.e., at 84cM.

 

Interactions of tld alleles and MBT

Deficiencies Df(3R)X18E(tld^-) and Df(3R)XTA1(tld^-), that cover region 96B (85cM) are lethal over MBT. Df(3R)tld^68-62(tld-) deletes about 4 kb within tld (Shimell et al., 1991), and does not complement MBT. Tumors are induced in 20% of Df(3R)X18E/MBT and in 30% of Df(3R)XTA1/MBT. The primary tld alleles have been ordered into an allelic series based on complementation behavior and phenotype (Ferguson and Anderson 1992) (Table 1). Several alleles have been sequenced, and the interaction domains with dpp have been identified. Two domains of tld contribute to its genetic interaction with dpp, one was identified within the metalloprotease domain, and one in the first CUB repeat (Childs and O'Connor 1994). tld alleles were crossed over MBT and the hybrids screened for lethality and tumor induction. tld^5H65, tld^7M89, tld^7H41 and tld^9D36 are able to induce brain tumor formation over MBT (Table 1). None of the tld alleles analyzed does complement MBT, indicating that tld^MBT represents a strong allele. The strong alleles 10E95, 10F102 and 6P41 show a strong phenotype over MBT. The strong alleles 7H41 and 6P117 exhibit weak phenotypes over MBT. The weak allele 7M89 shows strong interaction in MBT. Thus, any allele being mutant in region around amino acid 250 shows strong interrelation in MBT. This region confers one of the tld-dpp interaction domains, within the metalloprotease region of tld. tld-1 and mali-2 are recombinants of MBT containing tld^MBT in different genetic environment (Riede 1996). They were crossed over tld^5H56. In hybrids

tld phenotype should occur, additional recessive defects should be complemented by the heterozygote wild type function of tld5H56. Hybrids exhibit rough eyes at the permissive temperature. At the restrictive temperature 30% die as pupae. The optic lobes are not hyperplasic, brain cells are differentiated but do not adhere. Wild type brain cells adhere. Non-adhesion of the upper and the lower surface of the wings occurs, resulting in haltere-like wings of adults. Thus, cell adhesion is affected by tld in MBT. Adhesion of cells helps them to communicate and is a precondition for contact inhibition.

Rescue with tld+ transgene

            pMBO1366 contains DNA en-coding for the 3.5 kb transcript of tld⁺. Chromosome 1 or 2 with pMBO1366 were combined accordingly with the third chromosome of MBT (MBT-III) (Table 2). tld⁺ induces growth disadvantage at the permissive tempera-ture. MBT-III is viable and sterile. The addition of tld⁺ leads to 80% pupal lethality, to a three days longer generation time and to defects in eclosed flies, these move very slowly and are unable to fly. Accordingly tld⁺ is crucial in the speed of cell division and differentiation of MBT. At the restrictive temperature, MBT exhibits unsegmented optic lobes with undifferentiated malignant neuroblasts. The enlargement of the brain is reduced by tld⁺, the ventral ganglion is elongated. Ingrowth of tumor tissue into the ventral ganglion was not observed. Thus, cells have lost their ability to invade. Hyperplasic growth of the tissue was not completely rescued. 80% of the transgenes' brains are two to three times larger than wild type brains. tld⁺ transgenic MBT regain wild type pattern of ARD expression in the brain, indicating that tld⁺ drives differentiation of neuroblasts. tld is expressed in larval wild type brain. Expression occurs in the segment where cells undergo cell divisions before they differentiate (Finelle et al., 1995; Nguyen et al., 1994). tld-1 induces brain hyperplasia formation, which could be rescued by recombinant tld⁺. tld-2 exhibits brain tumor formation, which was partially rescued.

 

Oogenesis and Embryogenesis are temperature sensitive periods for brain tumor induction in MBT.

     Three mutant genes coopera-tively induce brain tumor for-mation in MBT: tld, yeti and spz. In 100% of the L3 larvae brain tumor formation is induced, when parents were shifted to the restrictive temperature 24 hours prior to egg deposition. Temperature sensitive period for spz is the oogenesis and the first hour of embryogenesis. tld  temperature sensitivity encompasses 4 to 6 hours of embryogenesis (Lindsley and Zimm 1992). Oogenesis at the permissive temperature reduces the rate of tumor formation to 30% of the larvae, thus spz contributes to tumor formation. If the period for the tld effect occurs at the permissive temperature, 70% of the larvae induce brain tumor. Thus, both oogenesis and embryogenesis are temperature sensitive periods for MBT, both, spz and tld, add to tumor formation in MBT which includes temperature sensitivity in both cases.

 

Tl and pll promote tumor formation.

            Tl and pll are members of the signalling pathway that is initiated by spz. Deficiencies comprising Tl and pll are complemented by MBT and spz-1, accordingly both genes are not expected to be mutant in MBT. Several alleles were found interrelating with MBT (Table 3). Seven alleles are able to induce tumor formation over MBT, indicating oncogenic potential of this pathway. Tlr632 and pll⁰⁷⁸, both recessive alleles, are able to induce malignant brain tumor over MBT. This defines Tl and pll as proto-oncogenes.

            In pll⁶²⁸/MBT a second eye on one or both sides occurs in 50% of the flies. The same phenotype exhibits in pll⁶²⁸/spz-1, but not in pll⁶²⁸/+ or MBT/+. Other pll⁶²⁸/spz-1 reveal asymmetries at the front head, leading to tissue outgrowth. This indicates, that pll⁶²⁸ might further destabilize the head and brain development, already disorganized by spz-1.

Proliferative alleles interact with developmental genes over srn

            Mutations in proliferative genes break the restriction of the cell cycle, induce melanotic tumors in 100% of stage three larvae and somatic pairing gaps in salivary gland chromosomes. Lethality of homozygous animals occurs at different developmental stages. Adult viability of heterozygotes is reduced. Over MBT, Aus⁹, mer¹⁴ and srn⁸⁸ induce lethality and brain tumor formation. efe alleles over MBT induce brain hyperplasia and partial pupal lethality (Riede, 1997). Developmental genes spz and tld are mutant in MBT, pll and Tl act as oncogenes over MBT. The genetic interlink between the proliferative genes and the signal transduction cascades can be identified, by crossing proliferative alleles over alleles in differentiation genes and screening the hybrids for lethality and growth aberrations. MBT does not complement a number of developmental genes (Table 4). Alleles interfering with MBT were crossed over proliferative alleles. srn⁸⁸ reveals as a major interactor. Full lethality occurs over alleles of spz, Tl, cact, Ser, and gro. Partial lethality was observed over pll, tub, and dl alleles. mbt^P interacts with several alleles of differentiation genes. Complete lethality of heterozygotes is rare, and is seen only with pll and N alleles. srn⁸⁸ does not complement efe alleles. As srn⁸⁸ interferes with differentiation genes and with proliferative genes, it represents the genetic link between both.

Neoplasm formation (Tables 5 -7)

One strong proliferative gene defect induces neoplasm for-mation

            Proliferative alleles show tumor formation as melanomes. The strong proliferative alleles Aus⁹, srn⁸⁸ and mer¹⁴ induce lethality and melanome for-mation in homozygous and heterozygous animals. Variable expression shows that individual factors influence the phenotypic expression of the genotype. 50% of the heterozygote Aus⁹ larvae show melanomes. Accordingly, one dominant proliferative mutation gives rise to melanomes and is sufficient for tumorformation. 

Two recessive proli-ferative gene defects induce neoplasm for-mation

            All recessive alleles of proliferative genes, when homo-zygous, induce melan-omes. Two recessive mutations in cis induce neoplasm formation. efe⁸⁹ is a recessive proliferative allele, mutant in a gene at 92cM on the third chromosome far from srn⁸⁸ (56cM). efe⁸⁹ does not complement srn⁸⁸, all hybrids die as third instar larvae with melanomes. Accordingly, two reces-sive alleles in trans induce melanome for-mation.

One proliferative gene defect over one oncogene defect in-duces neoplasm formation

pll⁶²⁸/spz-2, pll⁰⁷⁸/spz-2, mer¹⁴/ pll⁶²⁸ and rig^P/pll⁶²⁸ induce neoplasms. Accordingly, the com-bination of an onco-gene over a proliferative allele in trans induces cell overgrowth. The tissue most frequently involved is oncogene specific, while proliferative alleles do not exhibit tissue specificity.

One proliferative gene defect over one tumor suppressor gene defect induces neoplasm formation

            tld^T/spz-2 and srn⁸⁸/tld^7M89 induce neoplasms. Accordingly, trans heterozygotes of proliferative alleles over tumor suppressor genes induce cell overgrowth. Heterozygotes for alleles of the tumor suppressor gene tld and a wild type allele induce neoplasm formation if a proliferative mutation is present.

One proliferative gene defect over one deficiency induces neoplasm fo-rmation

            8D06/spz-2, 8D06/tld-1, R1/tld-1, btx^P/8D06 induce neo-plasms. Accordingly, het-erozygotes of a proli-ferative allele over a deficiency in cis induce cell overgrowth. roeST1/ tld-1, roeST1/mali-2, mbt^P/tld-1 and mer¹⁴/8D06, proliferative alleles over deficiencies in trans, induce neoplasms.

Identified oncogenes or tumor suppressor genes are unable to induce neoplasms

            Alleles tld^7M89, pll⁰⁷⁸, pll⁶²⁸ and Tl^r632 were screened for neoplasms as homozygotes, heterozygotes inter se, heterozygotes over wild type or heterozygotes over deficiencies in cis. None of these combinations induces a neoplasic growth of cells.

Neoplasms without proliferative gene defect

            pll⁶²⁸/spz-1 and gro^E75/spz-1 induce neoplasm formation. Other gene combinations with oncogenic potential like spz⁶⁷/pll⁰⁷⁸, pll⁶²⁸/spz⁶⁷, spz67/pll⁶²⁸, spz⁶⁷/pll⁰¹⁹ do not induce neoplasms. Thus, neoplasm formation induced by gene combinations without identified proliferative gene defect is coupled to spz-1. This strain has a complete somatic pairing of the giant chromosomes (Riede, 1997).

Discussion

A tumor cell circumvents the restrictions of the cell cycle. Mutations in proliferative genes change the chromatin structure, allow replication of DNA and lead to hyperplasic growth of tissue. Thus, a proliferative gene mutation is thought to be the primary initiative event in tumor formation. A lack of differentiation provides a cell the competence to divide and migrate. Differentiation genes lead to determination of the tissue. Mutations of tld and spz in MBT and the oncogenic potential of pll and Tl alleles over MBT show that the determination process is destabilized in tumor formation. Mutation of tld in MBT is a secondary event: the addition of wild type tld in transgenes reduce vitality of MBT at the permissive temperature. Thus, tumor suppression is a secondary event, an adaptive mutation which suppresses lethality induced from a primary defect in a proliferative gene.

A human autosomal disorder, nevoid basal cell carcinoma syndrome, that predisposes to both cancer and developmental defects, is associated with mutants of the human homolog of Drosophila patched (Hahn et al., 1996). patched plays a role in segment polarity, and interferes with TGFß gene family members (Hooper and Scott, 1989). tld, the identified tumor suppressor gene of MBT interferes as well with TGFß family members. Thus, tumor predisposition and tumor suppression can act on similar biochemical levels.

            The differentiation genes and proliferative genes interact genetically. To identify genetic interactors, proliferative alleles were screened for lethality over differentiation genes. One allele, srn⁸⁸, exhibits most interactive potential. This allele does not complement a number of proliferative alleles (Riede, 1997). Accordingly, it represents the genetic link between proliferative alleles and differentiation genes. All alleles of the proliferative genes show an unusual feature: the somatic pairing of the chromosomes is incomplete. Shorter or longer stretches of the chromosomes are involved, depending on the allele. srn⁸⁸ frequently induces somatic pairing defects of long distances, up to half a chromosome.

            Mutations in human BRCA1 are responsible for about 10% of breast cancers and ovarian cancers. Its protein associates with Rad51, a member of a protein family mediating homologous pairing (Scully et al., 1997). BRCA1 breast tumors are characterized by a high degree of genome plasticity (Marcus et al., 1996). Proliferative genes in Drosphila induce somatic pairing gaps and replication initiation errors (Riede 1997, 1998). Accordingly, the phenotype of Drosophila reflects the molecular interaction of BRCA1: defect chromosome pairing and genome plasticity. This reflects, that BRCA belongs to the class of proliferative genes. Proliferative genes are the only genes that are causally related to cell proliferation in cancer formation. Developmental genes only add to the event by changing the differentiation pattern of the cells.

Materials and Methods

Genetics of MBT strains

            The temperature sensitive (ts) Drosophila mutant line Malignant Brain Tumor (MBT) forms malignant neuroblasts in the brain of larvae. It carries interrelative mutant genes: höcker^MBT hedera^MBT (second chromosome) mali^MBT tld^MBT yeti^MBT spz^MBT (third chromosome) (Riede, 1996). Of the list, mali and yeti are proliferative genes, i.e. induce cell overgrowth and somatic pairing gaps of polytene chromosomes. Not proliferative is spz, this gene defect alone does not induce cell proliferation or somatic pairing gaps. The polygenic defect in MBT was analyzed by recombination analysis. In principle, first phenotypes had been identified and reference strains exhibiting this phenotype were obtained. Second, the phenotypes were mapped. Third, the phenotype had to appear with the deficiencies of the region in question over MBT and the reference strain carrying the mutation. Fourth alleles of suspected genes had to react with MBT and the reference strain carrying the gene defect. With defined deficiencies and alleles all other strains were tested, to evaluate the genotype of all strains. mali^MBT has been localized to 87B (Riede, 1997), yeti^MBT to 96F (Riede, 1996). One of the loci being involved in lethality has been localized to 97F by the P-element insertional deletion Df(3R)mbt^P (Wismar et al., 1995). MBT/Df(3R)mbt^P hybrids are 100% temperature sensitive pupal lethal. In MBT/Df(3R)mbt^P larvae grown at 29ºC, no brain tumor is observed. MBT was recombined with Df(3R)mbt^P. If the deficiency would cover a mutation causing 100% temperature sensitive lethality, wild type recombinants should not appear. One per cent of the recombinants eclose. Therefore, this deficiency does not harbor the tumor suppressor gene causing 100% temperature sensitive lethality.

To obtain tld^MBT stocks, MBT and MBT* was recombined with ru st e ca. Recombinants were screened for temperature sensitivity over tld^5H56, tld^7M89, Df(3R)X18E and Df(3R)XTA1. Strains were selected that exhibit different phenotypes: tld-1 (two identified mutant genes inducing ts pupal lethality and hyperplasia formation of the brain ru st e tld^MBT yeti^MBT), tld-2 (ru st e tld^MBT yeti^MBT spz^MBT, brain tumor formation) and mali-2 (two identified mutant genes inducing partial ts lethality, mali^MBT tld^MBT ca). All induce brain tumor formation over MBT. Partial pupal lethals (spz-1, one identified mutant gene, ru st e spz^MBT) and pupal lethals inducing hyperplasia formation of the brain (spz-2, two identified genes, ru st e yeti^MBT spz^MBT) have been isolated. spz^MBT has been identified at 92cM, i.e. region 97 (Riede, 1996). spz-1, spz-2 and MBT do not complement spz⁶⁷.

Genes and fly stocks

            Three strong proliferative recessive lethal alleles were selected for this study (56cM): Aus⁹, srn⁸⁸ and merlin (mer)¹⁴ (Riede, 1997). They have been induced with EMS, give rise to brain tumor over MBT and reveal long unpaired chromosome regions. Lethality in trans over many proliferative alleles define them as interactors. srn⁸⁸ over proliferative alleles of amanda, drache, efendi (efe) or Aus are not viable. efe (92cM) is a  proliferative gene that does not express a mutant phenotype in MBT; efe alleles are recessive. They exhibit weak interactive potential, as they complement, in part, each other. A number of P-element insertions are lethal over MBT and disrupt the somatic pairing process. Two P-element insertions of this kind were chosen for this study. They have non expressed bellatrix (btx)^P and rigel (rig)^P (Cooley et al., 1988).

Within the embryo, the neurogenic ectoderm is fixed by ventral-laterally located cells. Within the segmented germ band, the neurogenic ectoderm becomes subdivided, the neuroblasts segregate and proliferate in a defined manner (Campos-Ortega 1993). Products of proneural genes and neurogenic genes, such as Delta (Dl), are involved in the determination process that includes the proliferation and differentiation of cells. Notch (N), neuralized (neu), E(Spl)-C and gro prevent neural hyperplasic growth. N is an embryonic tumor suppressor gene that acts through a lateral inhibition of neuroblasts (Gateff 1994). Serrate (Ser) is involved in the control of cell proliferation (Speicher et al., 1994). Deficiencies roeST1 (84A6B1;84D4-9, tub^-), T32 (86E2-4;87B9-10, mali-), XTA1 (96A22-23B1;96D2-3, tld^-), X18E (96A17-20;96C1, tld^-), 8D06 (96E10-12;97A3-4, yeti^- gro^-), R1 (96F2;96F12-14, yeti^- gro^-), ro82b (96F11-14;97F3-11 yeti^- gro^- Tl^- pll^- spz^- efe^-) and mbt^P (97F, efe^-) induce lethality over MBT. Stocks were obtained from E. Gundelfinger, Magdeburg (Df(3R)XTA1, Df(3R)X18E, Df(3R)Ser^+82f24) and M. B. O'Connor, Irvine (tld+ transgene stocks 1366-67, 1366-68, Df(3R)tld^68-62 (Shimell et al., 1991)). spz⁶⁷, cact, dl, tub, pll, tld, neu, N, Dl, and Tl mutants were obtained from C. Nüsslein-Volhard and I. Koch, Tübingen, SerRX106 (Thomas et al., 1991) from U. Thomas, Magdeburg, gro^E75 (Preiss et al., 1988) and gro- deficiencies from A. Preiss, Hohenheim. All other stocks were obtained from the Bloomington Stock Center. The balancer chromosomes, deficiencies and markers are described (Lindsley and Zimm, 1992).

Brain whole mounts, ARD stain

            The non-ligand binding structural subunit of nicotinic acetylcholine receptor, ARD, is expressed in the ventral ganglion and widely distributed in neuropiles of the optic lobes. The expression of this subunit is a marker for appropriate differentiation of the neuropiles. ARD like immunreactivity in larval brains was obtained according to Schuster et al. (1993). Parents were placed on fresh medium and shifted immediately to the restrictive temperature of 29ºC. The F1 generation was analyzed. Same size L3 larvae were dissected in Ringer`s solution. A hyperplasic brain (at least twice the volume of wild type brain) shows wild type segmentation of the optic lobes. The cells of the optic lobes are small, ingrowth of the tissue into the ventral ganglion is not observed. A brain tumor shows no signs of segmentation, ARD is not expressed in neuropiles. Brain tumor cells are small and large and do not adhere. Malignancy is defined as ingrowth of cells into the ventral ganglion.

Transformation rescue of tld^MBT

            Two stocks containing pMBO1366 with recombinant tld ⁺ (Shimell et al., 1991) on the first or the second chromosome were crossed with appropriate marker/balancer strains. The F1 was subsequently crossed with MBT or tld^MBT containing recombinants. The F2 generation containing pMBO1366 on the first or second chromosome and a TM6B,Tb balanced third chromosome of MBT, tld-1 or tld-2 was shifted to 29ºC. L3 larvae, homozygous for the third chromosome were analysed. As control, in parallel the same crosses were performed with a w marked first or a Pm second chromosome without pMBO1366.

Screening for neoplasm formation

            The F1 generation of at least two independent crosses was analyzed for complementation, i.e., the occurrence of eclosed hybrids. Adults were screened for growth abnormalities or neoplasm formation. In case of lethality, percentage and stage were determined. If brain tumor was suspected, L3 larvae were dissected in PBS and the brains were screened for brain tissue overgrowth. All crosses were maintained routinely at 20ºC. For crosses with the temperature sensitive lines MBT, MBT recombinants, tld ^9D36, Dl ^6B, spz⁶⁷, Tl ^r632 and Tl ^r26 virgins were collected and kept at 18ºC over night, prior to the addition of males and immediate shift to 29ºC.

            Acknowledgments:  I wish to thank M.B. O'Connor, O. Vef, C. Nüsslein-Volhard, I. Koch, U. Thomas, E. Gundelfinger, A. Preiss, and the Bloomington Stock Center for the strains. For donating the ARD-antibody, I thank B. Phannavong.

            References:  Anderson, K.V., and C. Nüsslein-Volhard 1984,  Nature 311: 223-227;  Bishop, M., 1982,  In: Advances in Cancer Research, vol. 37 (Klein, G., and S. Weinhouse, eds.). pp. 1-28.  Academic Press, New York;  Campos-Ortega, J.A., 1993,  In: The Development of Drosophila melanogaster, (Bate, M., and A.M. Arias, eds.). pp. 1091-1130. Cold Spring Harbor Laboratory press;  Chasan, R., and K.V. Anderson 1993,  In: The Development of Drosophila melanogaster, (Bate, M., and A.M. Arias, eds.). pp. 387-424. Cold Spring Harbor Laboratory press;  Childs, S.R., and M.B. O'Connor  1994,  Develop. Biol. 162 :209-220;  Cooley, L., R. Kelley, and A. Spradling 1988,  Science 239: 1121-1128;  Ferguson, E.L., and K.V. Anderson 1992,  Development 114: 583-597;  Finelli, A.L., T. Xie, C.A. Bossie, R.K. Blackman, and R.W. Padgett 1995,  Genetics 141: 271-281;  Foulds, L., 1958,  J. Chron. Dis. 8: 2-37;  Gateff, E., 1994,  Int. J. Dev. Biol. 38: 565-590;  Gateff, E., T. Löffler, and J. Wismar 1993,  Mech. Develop. 41: 15-31;  Hahn, H., C. Wicking, P.G. Zaphiropoulos, et al., 1996,  Cell 85: 841-851;   Hashimoto, C., K.L. Hudson, and K.V. Anderson 1988,  Cell 52: 269-279;  Hooper, J.E., and M.P. Scott 1989,  Cell 59: 751-765;  Kamb, A., N.A. Gruis, J. Weaver-Feldhaus, Q. Liu, K. Harshman, S.V. Tautigian, E. Stockert, R.S. Day, III, B.E. Johnson, and M.H. Skolnick 1994,  Science 264: 436-440;  Letsou, A., S. Alexander, K. Orth, and S.A. Wassermann 1991,  Proc. Natl. Acad. Sci. 88: 810-814;  Letsou, A., S. Alexander, and S.A. Wassermann 1993,  EMBO J. 12: 3449-3458;  Levine, A.J., J. Momand, and C.A. Finlay 1991,  Nature 351: 453-456;  Lindsley, D.L., and G.G. Zimm 1992,  The Genome of Drosophila melanogaster. San Diego, CA,  Academic Press;  Marcus, J.N., P. Watson, D.L. Page, et al., 1996,  Cancer 77: 697-709;  Massagué, J., 1990,  Annu. Rev. Cell. Biol. 6: 597-641;  Morisato, D., and K.V. Anderson 1994,  Cell 76: 677-688;  Nguyen, T., J. Jamal, M.J. Shimell, K. Arora, and M.B. O'Connor 1994,  Develop. Biol. 166: 569-586;  Nobori, T., K. Miura, D.J. Wu, A. Lois, K. Takabayashi, and D.A. Carson 1994,  Nature 368: 753-756;  Nowell, P., 1976,  Science 194: 23-28;  Peto, R., 1977,  In: Origins of Human Cancer. pp 1403-1428. New York: Cold Spring Harbor;  Preiss, A., D.A. Hartley, and S. Artavanis-Tsakonas 1988,  EMBO J. 7: 3917-3927;  Riede, I., 1996,  Canc. Genet. Cytogenet. 90: 135-141;  Riede, I., 1997,  Canc. Genet. Cytogenet. 97: 143-154;  Riede, I., 1998,  Cancer Detection and Prevention 22: S61;  Roth, S., Y. Hiromi, G. Godt, and C. Nüsslein-Volhard 1991,  Development 112: 371-388;  Schuster, R.,  B. Phannavong, C. Schröder, and E.D. Gundelfinger 1993,  J. comp. Neurol. 335: 149-162;  Scully, R., J. Chen, A. Plug, Y. Xiao, D. Weaver, J. Feunteun, T. Ashley, and D.M. Livingston 1997,  Cell 88: 265-275;   Shelton, C.A., and S.A. Wassermann 1993,  Cell 72: 515-525;  Shimell, M.J., E.L. Ferguson, S.R. Childs, and M.B. O'Connor 1991,  Cell 67: 469-481;  Speicher, S.A., U. Thomas, U. Hinz, and E. Knust 1994,  Development 120: 535-544;  Stein, D., and C. Nüsslein-Volhard 1992,  Cell 68: 429-440;  Steward, R., 1987,  Science 238: 692-694;  Thomas, U., S.A. Speicher, and E. Knust 1991,  Development 111: 749-761;  Weinberg, R.A., 1995,  Cell 81: 323-330;  Wismar, J., T. Löffler, N. Habtemichael, O. Vef, M. Geißen, R. Zirwes, W. Altmeyer, H. Sass, and E. Gateff 1995,  Mech. Develop. 53: 141-154.