Dephosphorylation of Vav is associated with the induction of mouse erythroleukemia cell differentiation: effects of orthovanadate and levamisole.

Mouse erythroleukemia (MEL) cell erythroid differentiation induced by dimethyl sulfoxide or hexamethylene bisacetamide (HMBA) is accompanied by the production of hemoglobin, terminal cell division and decreases in lactate production and fructose 2,6-bisphosphate levels. A number of studies have suggested that decreases in the cellular level of protein phosphotyrosine content may play a role in MEL cell differentiation. In particular, it was shown that the expression of several protein tyrosine phosphatase genes accompany this process and that the transfection of one of these genes into MEL cells followed by its subsequent expression induced eythroid differentiation. However, none of the physiological substrates for these protein tyrosine phosphatases have been identified. It is shown here that MEL cell differentiation is accompanied by decreases in tyrosine phosphorylation of Vav and possibly of the erythropoietin receptor (EpoR). Immunoprecipitation of the EpoR and analysis of co-precipitated proteins, indicates that the EpoR associates with Vav, STAT5 and an unidentified 60 Kd protein, . HMBA-induced erythroid differentiation abrogates these associations. The phosphatase inhibitors, Na3VO4 and levamisole, inhibit HMBA-induced differentiation as well as the association of the EpoR with Vav, STAT5 and the 60 Kd protein. This is of interest since Na3VO4, at the concentrations used here, has been shown to be a potent inhibitor of the activity of protein tyrosine phosphatases. These results suggest that levamisole, at least indirectly, acts by a molecular mechanism similar to that of Na3VO4 and that the loss of the association of the EpoR with Vav, STAT5, and and/or the reduction in the level of tyrosine phosphorylation of these proteins may play a role in MEL cell differentiation.

The phosphatase inhibitors, orthovanadate and levamisole, inhibit induction of erythroid differentiation and abrogate the associated inhibition of glycolysis.

Alterations in protein tyrosine phosphate (PTP), lactate, and fructose 2,6-bisphosphate (F-2,6-P2) levels have been associated with induced MEL cell differentiation and commitment to terminal cell division (TCD). The possible relationships of perturbations in PTP metabolism and reduction in lactate formation during differentiation were investigated utilizing sodium orthovanadate, Na3VO4, primarily an inhibitor of PTP phosphatases, and levamisole, considered an alkaline phosphatase inhibitor. Both of these compounds were found to effectively inhibit the TCD-associated differentiation induced by DMSO, HMBA, and Na butyrate and to abrogate the differentiation-associated reduction in lactate accumulation due to these agents. However, they were found not to inhibit hemin-induced hemoglobin synthesis which is independent of TCD and does not alter lactate metabolism. Two brominated levamisole analogs, L-p-bromotetramisole and D-p-bromotetramisole, were also found to be inhibitors of TCD-associated differentiation and to be effective at even lower concentrations than levamisole. The changes in TCD-associated differentiation and lactate production exhibited the same concentration-dependence with respect to the inhibitors. These findings strengthened the theory that TCD-associated differentiation, decreased lactate production, and sensitivity to phosphatase inhibitors are all associated. Since the induction of MEL cell differentiation has been shown to be associated with, and is thought to be due to, the induction of PTP phosphatase activity and Na3VO4 is thought to inhibit the differentiation by inhibiting PTP phosphatase activity, the effect of levamisole on PTP levels was determined. Levamisole, like Na3VO4, was found to increase the tyrosine phosphate levels of proteins of similar molecular weights in intact cells in both the presence and absence of a differentiation inducer. Several phosphotyrosine-containing, similarly sized proteins were particularly affected by differentiation induction and by Na3VO4 and levamisole treatment. Changes in the levels of tyrosine phosphate-containing proteins of approximately 92-96, 60, and 38 kd were particularly noticeable. The induction of differentiation reduced PTP levels and inhibition of differentiation due to treatment with either Na3VO4 or levamisole increased their levels. These data suggest relationships between signal transduction pathways involved in differentiation and TCD, the regulation of lactate and F-2,6-P2 metabolism, and PTP levels.

Inhibition of mouse GATA-1 function by the glucocorticoid receptor: possible mechanism of steroid inhibition of erythroleukemia cell differentiation.

Treatment of mouse erythroleukemia (MEL) cells with hexamethylene bisacetamide induces a program of erythrodifferentiation, as judged by an increase in the synthesis of globins and other erythroid-specific products. This induction can be inhibited by glucocorticoids, e.g. dexamethasone. All globin and other erythroid-specific genes tested contain GATA response elements (GATA-RE) and can be transactivated by GATA-1, a transcription factor. GATA-1 is highly expressed in erythroid cells, including MEL cells. We noted a glucocorticoid receptor (GR) response element motif near a GATA-RE motif in the promoter region of the mouse beta-major and beta-minor globin genes and about 130 bases away from a GATA-RE in the alpha 1-globin gene promoter and, therefore, investigated the possibility that the dexamethasone-induced inhibition of induced MEL cell differentiation may involve effects of the GR on GATA-1 activity. Evidence obtained from transfection assays and DNA electrophoretic mobility shift assays indicates that the GR binds GATA-1 and interferes with its function before any interaction with DNA, but that the presence of a glucocorticoid response element near a GATA-RE augments the GR effect. The N-terminal 106-amino acid domain of the GR was found to be essential for the effect, possibly by binding to GATA-1. Since GATA-1 is autoregulatory, i.e. it has been shown by others to bind to its own promoter and up-regulate its own transcription, the finding that activated GR can interfere with GATA-1 function may provide an explanation for the inhibition by glucocorticoids of the entire program of erythroid differentiation in MEL cells. That is, by interfering with GATA-1 function, the GR inhibits not only the expression of erythroid structural genes, but may also inhibit the expression of a primary erythroid regulatory gene, GATA-1. It was also shown that the GATA-RE in each of the beta-globin promoters responds to mouse GATA-1 in a functional transfection assay.

A possible role for nitric oxide in glutamate (MSG)-induced Chinese restaurant syndrome, glutamate-induced asthma, 'hot-dog headache', pugilistic Alzheimer's disease, and other disorders.

Endogenous glutamate is thought to be a major neurotransmitter. After binding to a cell membrane receptor there can be a stimulation of what can be called the nitric oxide (NO)-mediated neurotransmission pathway (NO-MNP). The activity of the enzyme that produces NO from arginine, NO synthase, and the level of NO become elevated. NO has little activity within the cell in which it is produced, but it rapidly leaks out of that cell and produces effects in neighboring cells. The NO-MNP can be activated to release NO in endothelial cells which in turn acts on neighboring vascular smooth muscle cells to induce vasodilation. Therefore, we suggest that exogenous, ingested glutamate, like endogenous glutamate, can lead to the same stimulation of the NO-MNP in sensitive individuals which would then cause the symptoms of the Chinese restaurant syndrome and/or glutamate-induced asthma. Further, since ingested nitrite and related compounds can be metabolized to NO, NO may more directly cause the symptoms of 'hot dog headache'. In addition, it has been suggested that NO production can also be controlled in endothelial cells by fluid forces that stimulate pressure receptors. Therefore, elevations of NO and stimulation of the NO-MNP may occur due to sudden, local, alterations of blood pressure during pugilistic activities and play a role in the symptoms of pugilistic Alzheimer's disease. If these ideas are correct, then inhibitors of the NO-MNP and/or temporary reduction of the plasma level of arginine may be useful in preventing at least some of the symptoms of these disorders.

Enhancement of differentiation and cytotoxicity of leukemia cells by combinations of fluorinated pyrimidines and differentiation inducers: development of DNA double-strand breaks.

We have previously shown that pretreatment of mouse erythroleukemia (MEL) cells with the fluorinated pyrimidines 5-fluorouracil (FUra) or 5-fluorodeoxyuridine (FUdR) followed by the differentiation inducer hexamethylene bisacetamide (HMBA) greatly enhanced the magnitude of their differentiation and caused extensive cell death. We have now extended these studies to address the mechanism that may be responsible for this enhancement and have also examined a human leukemic cell line (HL-60) for its sensitivity to this combination cytotoxic-differentiation therapy. We found that in HL-60 cells, pretreatment with FUdR, but not FUra, followed by 1.2% dimethylsulfoxide (DMSO) led to an 8 to 10-fold enhancement of cell death as compared to FUdR alone. When all-trans-retinoic acid (ATRA) was used instead of DMSO, the enhancement of differentiation and cytotoxicity was 5-fold. The percent of cells induced to differentiate was dependent on the concentration of both FUdR and ATRA. In HL-60 cells resistant to ATRA-induced differentiation, the combination of FUdR and ATRA did not result in enhanced cytotoxicity. Leucovorin (LV), a compound known to enhance the inhibitory effect of FUra or FUdR on DNA synthesis, increased the effectiveness of the cytotoxic-differentiation therapy, whereas thymidine inhibited its effectiveness. This suggests that inhibition of DNA metabolism may be an integral part of the differentiation-enhancing cytotoxic mechanism. To further explore inhibition of DNA synthesis, DNA was extracted under alkaline or neutral conditions from 3H-thymidine-labelled cells that were treated with FUra/LV and HMBA individually or in combination. The emergence of single and double-strand DNA breaks was monitored by agarose gel electrophoresis. In parallel to the enhancement of cytotoxicity, the combination treatment (FUra/LV followed by HMBA) also produced a 2.5-3-fold increase in the DNA breaks when compared to the same effect obtained by the agents applied individually. Thus, we propose that DNA degradation may be the mechanism responsible for the enhanced loss of cell viability. In summary, we report here an approach which is targeted to increasing the death rate of leukemic cells through the combined use of low doses of cytotoxic drugs and differentiation inducers.

Reduction in lactate accumulation correlates with differentiation-induced terminal cell division of leukemia cells.

Lactate accumulation in the medium and glucose utilization decreased during the induction of in vitro differentiation of mouse erythroleukemia (MEL) and human myeloid leukemia (HL-60) cells. The decrease in lactate accumulation occurred as early as 24 h after inducer treatment was initiated and occurred prior to the decrease in glucose utilization. The decrease in lactate accumulation was greater than that predicted by the decrease in glucose utilization, i.e., the ratio of glucose used glycolytically, as measured by lactate accumulation, to glucose used in other pathways ('glycolytic ratio') markedly decreased during differentiation in these cell lines. Differentiation correlated with the abrogation of the high levels of lactate accumulation first described by Warburg as characteristic of some transformed and neoplastic cells. Studies on both parental and differentiation-resistant variant MEL cell lines indicated that the changes in lactate accumulation were not dependent on the changes in glucose utilization and could be dissociated from them. Moreover, the changes in lactate accumulation only occurred in cells able to undergo differentiation-induced terminal cell division. This regulatable expression of lactate accumulation in MEL and HL-60 cells in vitro may make them useful model systems for the elucidation of the molecular mechanisms controlling lactate formation in malignant cells.

Combination cytotoxic-differentiation therapy of mouse erythroleukemia cells with 5-fluorouracil and hexamethylene bisacetamide.

The effects of 5-fluorouracil (5-FUra), in combination with various differentiation inducers on the growth and differentiation of mouse erythroleukemia (MEL) cells were investigated. The cells were first treated with 5-FUra, washed, and then treated with various concentrations of differentiation inducers: hexamethylene bisacetamide (HMBA), dimethyl sulfoxide (DMSO), and N-methylformamide. Pretreatment with 5-FUra, shown here to be a weak inducer of MEL cell differentiation, enhanced the subsequent HMBA induction of differentiation. The three inducers of differentiation markedly inhibited cell growth and increased cell death in a dose- and time-dependent manner if given immediately after cells were exposed to 5-FUra. In contrast, 5-FUra at similar concentrations inhibited cell growth, but only slightly increased cell death, while inducers without 5-FUra had little effect on cell growth or viability. When placed in fresh drug-free medium for 6 days following drug treatments, the cells completely recovered from the growth inhibition of 5-FUra as a single agent, whereas in cells previously treated with only HMBA there was a inhibition of cell growth without loss of viability. In contrast, a profound and prolonged growth inhibition with 98% cell death occurred in cells previously treated with 5-FUra followed by HMBA. The enhancement of 5-FUra cytotoxicity appeared to be directly related to the degree of differentiation and to biochemical events that occur during the commitment to terminal cell division induced by N-methylformamide, DMSO, or HMBA. An increase in Okazaki fragments was found in MEL cells treated with HMBA or DMSO when committed to terminal cell division. DNA breaks also follow 5-FUra treatment (A. Yoshioka et al., J. Biol. Chem., 262: 8235-8241, 1987) and may be the events that lead to cell death. The marked increase in cell death resulting from 5-FUra/HMBA treatment may be, at least partly, a consequence of increased DNA breaks due to 5-FUra followed by inhibition of DNA repair which is known to occur following the HMBA or DMSO induction of differentiation and commitment to terminal cell division. This combined sequential cytotoxic-differentiation therapy resulting in synergistic cytotoxicity and differentiation may be the basis of a new approach to cancer therapy and may aid in reducing the amounts of chemotherapeutic agents required for effective treatment, while maintaining or even increasing their therapeutic effects.

Benzyl alcohol alters the hemoglobin phenotype of mouse erythroleukemia cells.

The hemoglobin minor/hemoglobin major ratio expressed in mouse erythroleukemia (MEL) cells grown in vitro varies according to the differentiation inducer utilized. For example, butyrate and hemin induce higher hemoglobin minor/hemoglobin major ratios than do dimethyl sulfoxide (DMSO) or hexamethylene bisacetamide (HMBA). Benzyl alcohol in non-toxic concentrations was found to markedly reduce the hemoglobin minor/hemoglobin major ratio and to moderately reduce the total hemoglobin induced by DMSO or HMBA in MEL cells, while only slightly decreasing the ratio induced by hemin or butyrate. The addition of dexamethasone (another and more potent inhibitor of the induction of hemoglobin synthesis than benzyl alcohol) to the media during HMBA induction of differentiation increased the hemoglobin minor/hemoglobin major ratio. This is similar to other "inhibitory" treatments (i.e., treatments that result in sub-optimal hemoglobin production) that have been previously reported. Therefore, although benzyl alcohol and dexamethasone both partly inhibit the induction of total hemoglobin production, they have opposite effects on the induced hemoglobin phenotype: benzyl alcohol decreases the hemoglobin minor/hemoglobin major ratio while dexamethasone increases it. The mechanism(s) of the alteration in the hemoglobin phenotype is unknown as is the mechanism of induction by any of the various inducing agents or of the inhibition of induction by any treatment. However, it appears that if the signal for the induction of hemoglobin minor is sufficiently potent (as it is during butyrate or hemin induction), it cannot be overcome by benzyl alcohol at a "non-toxic" concentration.

A possible effect of heme on the fate of DNA ligase activity extracted from differentiating mouse erythroleukemia cells.

When mouse erythroleukemia (MEL) cells were induced to differentiate by growth in the presence of dimethyl sulfoxide, hexamethylene bisacetamide (HMBA), or hemin, the apparent activity of DNA ligase extractable from inducer-treated cells decreased 70 to 80% when compared to untreated cells. Earlier work had indicated that these changes did not occur in a differentiation-resistant MEL cell variant and suggested that the decrease in the level of DNA ligase activity might be related to the differentiation process. Since the MEL cells accumulate high levels of both hemoglobin-bound and non-hemoglobin-bound heme, the effect of both hemoglobin and hemin on DNA ligase activity of MEL cell extracts was tested. When cell-free extracts containing DNA ligase activity were preincubated with hemin at concentrations up to 150 microM, an 80% or greater inhibition of the DNA ligase activity resulted. The ATP-dependent DNA ligase from bacteriophage T4 was also inhibited by hemin, but the NAD-dependent DNA ligase from Escherichia coli was not sensitive to this treatment. Preincubation of these same extracts with hemoglobin at levels comparable to those present in differentiating cells did not result in inhibition of any of the ATP-dependent DNA ligases tested. Culturing the cells with dimethyl sulfoxide in the presence of imidazole resulted in a marked decrease in globin chain accumulation but did not reverse the dimethyl sulfoxide-related decrease in DNA ligase activity. These data suggest the possibility that heme or its metabolites, but not globin or hemoglobin, could serve to modify the process of DNA replication and/or repair in differentiating MEL cells via inhibition of DNA ligase activity. These data are consistent with the findings of Lo et al. (S.C. Lo, R. Aft, and G.C. Mueller, Cancer Res., 41: 864-870, 1981) which correlated the onset of differentiation-related terminal cell division in MEL cells with the levels of nonhemoglobin heme present in these cells.

Induction of differentiation in mouse erythroleukaemia cells by the action of papain at the cell surface.

The addition of one of several proteases to cultures of mouse erythroleukaemia (MEL) or human K-562 leukaemia cells can induce a substantial portion of the cells to undergo erythroid differentiation. This effect is due, at least in part, to the proteolytic action of these enzymes. The critical substrate(s) for this proteolytic action is not a component of the medium or a long-lived substance(s) released from the cells. In order to determine if the substrate(s) is located on the cell surface or intracellularly, a comparison of the ability of non-immobilized papain and immobilized papain (i.e. covalently linked to Sepharose beads which were larger than the cells) to induce MEL cell differentiation was undertaken. Both papain preparations induced the same level of differentiation. The proteolytic activity of the bead-linked papain remained associated with the beads. Therefore, proteases induce erythroid differentiation in these cells by acting proteolytically on a substrate(s) that is exterior to the cell.

Basic principles for utilizing combination differentiation agents.

The induction of differentiation in several tumor lines serves as a basis for a new approach to cancer treatment. In vitro studies in the mouse erythroleukemia (MEL) cell system have identified about 300 agents capable of inducing differentiation by mechanisms that remain to be elucidated. The design of differentiation therapy will depend on the specific tumor cell type, an effective time course, and the synergistic interaction among combinations of two or more inducers. The induction of differentiation may be followed by terminal cell division (TCD) or programmed cell death in several tumor cell systems. This mechanism for the destruction of tumor cells is one goal of differentiation therapy and differs from nonspecific cytotoxic therapy. To evaluate the effect of differentiation therapy, a clear distinction must be made between nonspecific cytotoxicity and the programmed TCD of induced cytodifferentiation. One possible parameter for assessing the commitment to TCD in the MEL cell system is a selective decrease in DNA ligase activity, which does not appear to occur following treatment with nonspecific cytotoxic agents. These biological and biochemical parameters should be helpful in designing agents capable of inducing TCD in vivo.

Lack of evidence for activation of a serum factor in protease-induced differentiation of mouse erythroleukemia cells.

The addition of certain proteases to cultures of Friend virus-infected mouse erythroleukemia cells can induce up to 90% of the cells in culture to become hemoglobin-containing, as assessed by positive staining for benzidine (B+). Because the mechanism of this protease action is unknown, media components were studied as possible targets for protease activity. Aliquots of medium plus serum were incubated for various times with levels of protease sufficient to induce approximately 50% of the cells to the B+ state. Cells were added to protease-pretreated serum either before or after inactivation of the protease. In all cases, enzymatically active protease had to be present with the cells to induce B+ cells to form. Serum and other components of the medium pretreated with protease were inactive. Mouse erythroleukemia cells grown in the absence of serum were also induced by proteases to form B+ cells. These data imply that the inducing action of proteases cannot be passively transferred by protease-pretreated serum or medium nor is serum required for protease-mediated induction of B+ cells. Taken together, these conclusions suggest that the protease action is on the cells or on cellular products intimately associated with cells.

The proportions of hemoglobin types induced in mouse erythroleukemia cells vary with the inducer or combination of inducers, the inducer concentration and the time of induction.

The relative amounts of hemoglobin (Hb) major and Hb minor accumulated during induction of erythrodifferentiation in mouse erythroleukemia (MEL) cells were studied. The ratio of major to minor was found to depend not only upon the inducer tested (as reported previously by others), but also upon the concentration of the inducer and the time of exposure to the inducer as well as the specific cell line of MEL cells studied. At concentrations required for optimal induction of differentiation, certain agents led to the accumulation of predominantly Hb major, but suboptimal concentrations of the same inducers led to predominantly Hb minor accumulation. After a relatively short induction time (2 da) utilizing a given inducer either the level of Hb minor was higher than that of Hb major or the levels of the two Hb's were approximately equal, but after longer induction periods (3-7 da) Hb major was more abundant than Hb minor. In addition, it was found that the three proteases tested induced predominantly Hb minor. The addition of suboptimal concentrations of low molecular weight inducers acted synergistically with a given protease to produce a high yield of Hb-containing cells. When these agents were added singly they induced relatively low Hb major/Hb minor ratios, but when a low molecular weight inducer was added together with a protease in a "synergistic" combination, elevated ratios were induced. The proportions of hemoglobin types induced in MEL cells may be related in part to the intensity of the induction response. In view of these data, classifications of inducers based solely on the ratios of Hb types produced must be guarded.

G-actin and F-actin levels at different stages of mouse erythroid differentiation.

Monomeric (G-) actin and filamentous (F-) actin levels were determined in Triton X-100 extracts prepared from mouse erythroid cells at various stages of differentiation. G-actin and F-actin were found in the Triton-soluble fraction and in the Triton-insoluble fraction, respectively. G-actin levels in untreated and dimethyl sulfoxide-treated (differentiated) erythroleukemia cells, reticulocytes, and erythrocytes were 48, 33, 2.8, and 0.37 microgram/mg protein, respectively, and F-actin levels were 17, 35, 45, and 59 micrograms/mg protein, respectively. G-actin/F-actin ratios were successively lower in cells representing the more mature stages of development.

Proteases act synergistically with low molecular weight inducers to stimulate mouse erythroleukemia cell differentiation.

Proteases stimulate mouse erythroleukemia (MEL) cell differentiation and multiplication. The stimulation of differentiation is synergistically increased by low concentrations of dimethyl sulfoxide. Synergism of other low molecular weight inducers with representative proteases, alpha-chymotrypsin and protease V8, was tested. Hemin. hypoxanthine, actinomycin D, aminonucleoside of puromycin, hexamethylene bisacetamide, and 5-azacytidine were also found to act synergistically with this protease to augment MEL cell hemoglobin production, but not cell multiplication. Fatty acids (acetate, propionate, butyrate, isobutyrate, and valerate), prostaglandins A1 and E1, amino acids, and amino acid analogs and metabolites did not act synergistically with chymotrypsin. Some physiologic amino acids were found to be weak inducers. Several of the low molecular weight inducers also acted synergistically with protease V8 in inducing differentiation, and, as with chymotrypsin, did not act synergistically in stimulating cell multiplication. Like chymotrypsin, protease V8 did not act synergistically with butyrate. The earlier finding that proteases, but not low molecular weight inducers, stimulate cell multiplication during the induction of differentiation was confirmed. Carboxypeptidase A also was found to be an inducer.

DNA ligase and DNase activities in mouse erythroleukemia cells during dimethyl sulfoxide-induced differentiation.

DNA ligase and DNase levels were measured in cell-free extracts from untreated mouse erythroleukemia (MEL) cells and from cells treated with dimethyl sulfoxide (Me2SO) to induce erythroid differentiation. The DNase activity present in the extracts was sensitive to inhibition by G-actin and was, therefore, presumed to be DNase I. When the MEL cells were induced to differentiate by culturing in the presence of 1.8% Me2SO for 3 or 4 days, the apparent activity of the DNA ligase decreased to approximately 12% of the value in untreated MEL cells. In contrast, the apparent DNase I activity of the extracts from Me2SO-treated cells increased over that in extracts from untreated cells by a factor of 2. The activity of acid phosphatase, a lysosomal enzyme, remained unchanged. When strain DR-10, a mutant of the MEL cells which does not undergo Me2SO-induced differentiation, was treated with Me2SO, the DNA ligase and DNase activities of extracts from these cells remained unchanged as compared to extracts from untreated DR-10 cells. Therefore, the marked increase in the level of DNA ligase activity appeared to be related to the process of differentiation in the Me2SO-treated MEL cells.

Correlated genetic and EcoRI cleavage map of Bacillus subtilis bacteriophage phi105 DNA.

The seven previously identified EcoRI cleavage fragments of phi 105 DNA were ordered with respect to their sites of origin on the phage genome by marker rescue. One fragment, H, did not carry any determinants essential for replication. This fragment was totally missing in a deletion mutant which exhibited a lysogenization-defective phenotype. There is a nonessential region on the phi 105 genome which begins in fragment B, spans fragment H, and ends in fragment F. The size of the nonessential region, as estimated by alterations observed in the fragmentation patterns of deletion mutant DNAs, is approximately 2.7 X 10(6) daltons. Two new EcoRI cleavage fragments with molecular weights of approximately 0.2 X 10(6) were detected by autoradiography of 32P-labeled DNA. These small fragments were not located on the cleavage map.

Polymorphous presentations in vitelliform macular dystrophy: subretinal neovascularisation and central choroidal atrophy.

Two dominantly inherited macular dystrophies demonstrate the difficulty in establishing a diagnosis based on the fundus appearance. In 1 family the propositus presented with unilateral retinal haemorrhage associated with subretinal choroidal neovascularisation which remained unilateral over an 8-year period. In the other family the propositus presented with bilateral central choroidal atrophy. All affected family members had an abnormal electro-oculogram and a normal electroretinogram, suggesting the diagnosis of vitelliform macular dystrophy. Since vitelliform macular dystrophy has a wide range of expressivity, with polymorphous appearances of the fundus, the diagnosis is best made by the presence of a dominant mode of inheritance and an abnormal electro-oculogram.