Structural Conversion from Non-native to Native form of Recombinant Human Epidermal Growth Factor by Brevibacillus choshinensis

Brevibacillus choshinensis (Bacillus brevis) HPD31 is a very efficient producer of recombinant human epidermal growth factor (EGF). The produced EGF is secreted into the medium with high efficiency. However part of the EGF that accumulates in the medium, exists as multimeric forms which are biologically inactive. We found the bacterium has the activity to structurally convert multimeric forms to the monomeric, native ones. Optimal temperature and pH for the conversion were 40°C and pH 9, respectively. The reaction was promoted in the presence of reduced glutathione or cysteine. But the cells which had been sonicated or exposed to moderate heat treatment completely lost the activity. Thus, it was presumed that the activity might be due to the enzyme(s) that catalyze the protein disulfide exchanging reaction, and that they resides on the surface of viable cells.     

The effect of the dimeric and multimeric forms of fibronectin on the adhesion and growth of primary glomerular cells

Primary glomerular cells placed in a chemically defined medium containing Waymouth's medium MB 752/1 supplemented with insulin, transferrin, fibroblast growth factor, nonessential amino acids, sodium pyruvate, and antibiotics showed rapid outgrowth of cells which morphologically resembled well differentiated visceral epithelial cells followed by outgrowth of poorly differentiated cells; morphologic evidence suggests these latter cells are precursor cells of the epithelial cell lineage. Whereas the well differentiated glomerular epithelial cells were never observed to divide by sequential phase microscopic observations, a chemically defined medium was developed for optimal growth of the poorly differentiated cell type. This serum-free medium contained Waymouth's medium MB 752/1 supplemented with insulin, transferrin, selenium, and fibronectin (plus non-essential amino acids, sodium pyruvate, and antibiotics). Using this chemically defined medium, we have compared the effects of dimeric and multimeric fibronectin (high molecular weight disulfide-bonded fibronectin produced by incubation of dimeric fibronectin with 3 M guanidine followed by dialysis against 0.05 M cyclohexylaminopropane sulfonic acid (CAPS) buffer, pH 11) on the adhesion and growth of the poorly differentiated primary glomerular cell type. Dimeric fibronectin (FN) was twice as effective as multimeric FN in promoting glomerular cell adhesion, although both forms of FN promoted cell adhesion better than an uncoated substratum. In contrast, cell growth studies demonstrated that multimeric FN was a more potent growth stimulant than dimeric FN. The differential effects of dimeric and multimeric forms of FN in vitro suggests that these molecules may have different functions in vivo.     

Anchorage-independent growth of normal human mesothelial cells: A sensitive bioassay for EGF which discloses the absence of this factor in fetal calf serum

Summary This laboratory recently reported that normal human mesothelial cells require epidermal growth factor (EGF) and hydrocortisone (HC), in addition to fetal calf serum and a complex defined medium component, in order to grow optimally in surface culture (9). We report here that this normal cell type also forms large colonies at high efficiency in semi-solid medium, but exhihits more stringent serum and EGF requirements for anchorage-independent than for surface growth. Mesothelial cells are unable to divide at all in semi-solid medium with added EGF or with less than 2% serum, whereas they grow slowly but progressively in surface culture under such conditions. In semi-solid medium containing 20% serum and HC, mesothelial cells are stimulated to divide by the addition of as little as 30 pg/ml purified EGF. Human urine or male mouse plasma could substitute for purified EGF, yielding growth commensurate with the levels of EGF in these biological fluids previously measured by others using radioreceptor and radioimmune assays. Thus growth of mesothelial cells in semi-solid medium can serve as a highly sensitive assay of EGF biological activity which is unaffected by the presence of serum proteins. In addition, our results demonstrate that fetal calf serum does not provide mitogenic levels of EGF to cultured cells, raising the question of the identity of plasma and serum mitogens. This work was supported by NIH grants RO1 AG02048 and RO1 CA26656 to James G. Rheinwald and by NIH postdoctoral fellowship F32 AG05303 to Paul J. La Rocca.     

Activity and properties of alpha-L-fucosidase are dependent on the state of enterocytic differentiation of HT-29 colon cancer cells

Previously we have demonstrated an impairment in the activity of alpha-L-fucosidase in colon tumours. In order to establish an in vitro model to study this enzyme in colon cancer, we have determined the activity and properties of the enzyme during the differentiation of HT-29 colon cancer cells. Cultures were committed to differentiate into enterocyte-like cells by placing them in a culture medium without glucose for 18-21 days. The state of differentiation was evaluated by assaying the activity of enterocytic marker enzymes, and the acquisition of enterocyte morphology was assessed by electron microscopy. The alpha-L-fucosidase activity was determined using a fluorometric method. Intracellular levels of alpha-L-fucosidase activity are lower in non-differentiated cells (3.0 +/- 1.01 U/mg) than in differentiated ones (9.2 +/- 4.09 U/mg) (P < 0.001). This variation is not due to a greater secretion of the enzyme to the culture medium, and properties such as pH optimum or the affinity towards substrate are not dependent on differentiation. The enzyme however, is more stable at acidic pH and at high temperatures, and V(max) is higher in differentiated cells. Moreover, in undifferentiated cells the enzyme is mainly in a monomeric form whereas multimeric forms of the enzyme appear only upon differentiation. Most of these changes are very similar to those previously observed between normal colon tissue and colon tumours. Thus, we suggest that differentiation of HT-29 colon cancer cells could be used as a model to study the alterations of the enzyme alpha-L-fucosidase during the progression of the tumoural process.     

Recombinant human epidermal growth factor precursor is a glycosylated membrane protein with biological activity.

NIH 3T3 cells were transfected with cDNA corresponding to human kidney prepro-epidermal growth factor (preproEGF) under control of the inducible mouse metallothionein promoter. The synthesis of recombinant human EGF precursor by these cells has provided us with a model system for analysis of the structure and activity of this precursor. In transfected cells, the precursor was present as an intrinsic 170-kilodalton membrane protein as well as a soluble protein in the extracellular medium; both forms were N glycosylated. Glycosylation of the EGF precursor was determined by (i) the direct incorporation of [3H]mannose and [3H]glucosamine, (ii) metabolic labeling in the presence or absence of glycosylation inhibitors, (iii) enzymatic cleavage of the precursor by N-glycanase or endoglycosidase II, and (iv) lectin chromatography. Recombinant human preproEGF was purified by affinity chromatography, using wheat germ lectin and antibodies to human EGF. The intact precursor was biologically active. Purified preparations of preproEGF (i) competed with 125I-labeled EGF for binding to the EGF receptor in intact fibroblast cells, (ii) activated the intrinsic tyrosine kinase activity of the EGF receptor in membrane preparations, and (iii) sustained the growth of a mouse keratinocyte cell line that is dependent on EGF for growth. These results suggest that proteolytic processing of the precursor may not be essential for its biological function.     

Dimerization of internalized epidermal growth factor receptors.

Binding of epidermal growth factor (EGF) to cell surface EGF receptors initiates the formation of the receptor homodimers that can be detected by covalent cross-linking in intact cells or in detergent-solubilized cell extracts. Low pH dissociation of EGF from surface receptors results in immediate monomerization of receptor dimers. Using chemical cross-linking during mild permeabilization or cell solubilization, we have detected dimers of internalized EGF receptors in human carcinoma A-431 cells and transfected NIH 3T3 cells that express human EGF receptors. The percentage of internalized cross-linked receptor dimers was similar to that observed for surface EGF receptors. Furthermore, at the time of maximal accumulation of EGF-receptor complexes within the endosomal compartment (10-15 min of incubation at 37 degrees C), both the dimeric and monomeric forms of the EGF receptor are tyrosine-phosphorylated to the same extent as surface dimer and monomer species. In transfected NIH 3T3 cells, the level of dimerized and internalized kinase-negative EGF receptors was not different from that observed for wild-type receptors. These data suggest that for some time after internalization EGF does not dissociate from its receptor and indicate that a receptor conformation is preserved intracellularly that allows maintenance of receptor-receptor interactions and tyrosine kinase activity.     

Contact-dependent growth inhibition and apoptosis of epidermal growth factor (EGF) receptor-expressing cells by the membrane-anchored form of heparin-binding EGF-like growth factor.

Heparin-binding epidermal growth factor-like growth factor (HB-EGF) transduces mitogenic signals through the EGF receptor (EGFR). There are two forms of HB-EGF, the membrane-anchored form (pro-HB-EGF) and the soluble form (sHB-EGF). We studied the biological activity of pro-HB-EGF by using a model in which pro-HB-EGF-expressing effector cells was co-cultured with EGFR-expressing target cells. The DER cell, an EGFR-expressing derivative of the interleukin-3-dependent hematopoietic 32D cell line, grows well in the presence of EGF or sHB-EGF without IL-3. When DER cells were co-cultured on a monolayer of Vero-H cells overexpressing pro-HB-EGF, growth inhibition and subsequent apoptosis were induced in the DER cells even in the presence of excess amounts of EGF or sHB-EGF. Such growth inhibition of DER cells was abrogated when specific antagonists for pro-HB-EGF were added in the culture medium or when direct contact of DER cells with Vero-H cells was prevented, indicating that pro-HB-EGF is involved in this inhibitory effect. Pro-HB-EGF-induced apoptosis of DER cells was also observed even in the presence of IL-3. This rules out the possibility of simple competition between soluble EGFR ligands and pro-HB-EGF. Moreover, 32D cells expressing EGFR mutant composed of the extracellular and the transmembrane domain of EGFR and the cytoplasmic domain of erythropoietin receptor did not undergo apoptosis by co-culture with Vero-H cells, indicating that the inhibitory signal induced by pro-HB-EGF-expressing Vero-H cells is mediated to DER cells via EGFR and that the cytoplasmic domain of EGFR is essential for pro-HB-EGF-induced apoptosis. From these results, we concluded that pro-HB-EGF has unique biological activity through cell-cell contact that is distinct from the activity of sHB-EGF.     

Second messenger modulation of epidermal growth factor receptor function does not occur at the level of receptor dimerization.

Epidermal growth factor (EGF)-induced receptor dimerization may provide a mechanism for activation of the receptor protein tyrosine kinase and for initiation of post-receptor signalling pathways. We have examined whether second messengers and agents that modulate EGF receptor function act at the level of receptor dimerization. Both the Ca2+ ionophore ionomycin and the tumour promotor tetradecanoylphorbol acetate (TPA), added shortly before EGF, inhibit EGF receptor protein tyrosine kinase activity in intact cells. In permeabilized cells, elevation of Ca2+ similarly inhibits EGF receptor function. The inhibitory effect of Ca2+, unlike that of TPA, appears not to be dependent on protein kinase C activity. Neither ionomycin nor phorbol ester affects EGF-induced receptor dimerization, as shown by cross-linking and immunoblotting techniques, although the phosphotyrosine content of both monomeric and dimeric receptors is strongly decreased. Furthermore, we show that EGF receptor dimerization is not affected by increases in cyclic AMP or intracellular pH, nor by changes in transmembrane potential, medium osmolarity or the glycosylation state of the receptor. These result suggest that modulation of EGF receptor function occurs at a step other than receptor dimerization.     

Roles of epidermal growth factor and Na+/H+ exchanger-1 in esophageal epithelial defense against acid-induced injury

Epidermal growth factor (EGF) is predominantly secreted by salivary glands and activates Na(+)/H(+) exchanger-1 (NHE-1), which regulates intracellular pH (pH(i)). We investigated the roles of EGF and NHE-1 in esophageal epithelial defense against acid using human esophageal epithelial cell lines and a rat chronic esophagitis model. Esophageal epithelial cells were incubated with acidified medium in the absence or presence of EGF. Cell viability and changes in pH(i) were measured. Chronic acid reflux esophagitis was induced in rats with and without sialoadenectomy. Esophageal lesion index, epithelial proliferation, and expression of EGF receptors and NHE-1 were examined. EGF protected esophageal epithelial cells against acid in a dose-dependent manner, and the cytoprotective effect of EGF was completely blocked by treatment with NHE-1 inhibitors. Tyrosine kinase, calmodulin, and PKC inhibitors significantly inhibited cytoprotection by EGF, whereas MEK, phosphatidylinositol 3-kinase, and PKA inhibitors had no effect. EGF significantly increased pH(i) recovery after NH(4)Cl pulse acidification, and this increase in pH(i) recovery was significantly blocked by inhibitors of calmodulin and PKC. Sialoadenectomy led to an increase in the severity of chronic esophagitis but affected neither epithelial proliferation nor expression of EGF receptors. Expression of NHE-1 mRNA was increased in esophagitis and upregulated in rats with sialoadenectomy. The increasing severity of esophagitis in rats with sialoadenectomy was prevented by exogenous administration of EGF. In conclusion, EGF protects esophageal epithelial cells against acid through NHE activation via Ca(2+)/calmodulin and the PKC pathway. Deficiency in endogenous EGF is associated with increased severity of esophagitis. EGF and NHE-1 play crucial roles in esophageal epithelial defense against acid.     

Epidermal growth factor (EGF) and somatomedin C regulate G1 progression in competent BALB/c-3T3 cells

The activity in platelet-poor plasma that allowed density-arrested BALB/c-3T3 cells rendered competent by a transient exposure to platelet-derived growth factor (PDGF) to traverse G1 and enter the S phase has been termed progression activity. Epidermal growth factor (EGF) and somatomedin C-supplemented medium was shown to be capable of replacing the progression activity of 5% platelet-poor plasma (PPP) for competent density-inhibited BALB/c-3T3 cells. Exposure of competent cells to medium supplemented with EGF and somatomedin C reduced the 12 h minimum G1 lag time found in plasma-supplemented medium by 2 h. It is suggested that the reduction in the minimum time required for progression through G1 is due to the availability of free, unbound somatomedin C. Complete G1 traverse required both EGF and somatomedin C; however, the traverse of the last 6 h of G1 and entry into the S phase required only somatomedin C. Though EGF and somatomedin C could replace the G1 phase progression activity of plasma, medium supplemented with EGF and somatomedin C did not support complete cell cycle traverse or growth of sparse cultures of BALB/c-3T3 cells.     

Rat Sertoli cells secrete a growth factor that blocks epidermal growth factor (EGF) binding to its receptor

The conditioned medium from Sertoli cells contains a potent mitogen(s) that can markedly stimulate the proliferation of 4 different cell lines of endoderm or mesoderm origin in the presence or absence of serum. With A431 cells, conditioned medium produced in a dose-dependent manner up to a 5.2-fold increase in cell number after 5 days in culture. Addition of follicle-stimulating hormone (FSH), testosterone, retinol, and insulin to the Sertoli cells increased the secretion of the mitogenic activity. The ability of Sertoli cell conditioned medium (SCCM) to displace 125I-labeled epidermal growth factor (125I-EGF) from formalin-fixed A431 cells was also examined. The SCCM from Sertoli cells incubated with insulin contained 1.42 ng eq of EGF/ml; testosterone, retinol, and FSH (in the presence of insulin) further increased the secretion of this EGF competing activity to 2.09, 2.56, and 3.22 ng eq/ml, respectively. The amount of EGF competing activity was positively correlated with mitogenic activity. Separation of SCCM by gel filtration on Bio-Gel P-10 produced three major peaks of EGF-competing activity at apparent Mr = 1800-2100, 3800-4200, and 8000-9500. Chromatographing SCCM (in the presence of protease inhibitors) on size exclusion high performance liquid chromatography revealed two peaks of EGF competing activity at Mr about 8000 and 2000 coincident with and proportional to peaks of mitogenic activity. This activity was heat-sensitive and resistant to reducing agents, and addition of an equivalent amount of EGF as that present in SCCM produced an inhibition in growth of the A431 cells compared to a 3-fold stimulation with SCCM. Thus, the Sertoli cells secrete a potent mitogen that is distinct from EGF and alpha TGF. This factor that we have termed Sertoli cell-secreted growth factor is hormonally regulated by FSH, testosterone, and retinol and may play an important role in controlling spermatogenesis.     

Role of extracellular electrolytes in the activation of ribosomal protein S6 kinase by epidermal growth factor, insulin-like growth factor 1, and insulin in ZR-75-1 cells

Activation of ribosomal protein S6 kinase by epidermal growth factor (EGF), insulin, and insulin-like growth factor 1 (IGF1) was studied in the human mammary tumor cell line ZR-75-1 in isotonic buffers. In contrast to growth factor-dependent S6 phosphorylation which is strongly dependent on extracellular pH (Chambard, J. C., and J. Pouyssegur. 1986. Exp. Cell Res. 164:282-294.) preincubation of cells in buffers with different pH values ranging from 7.5 to 6.5 had no effect on basal or EGF-stimulated S6 kinase activity. Replacement of extracellular Na+ with choline or replacement of extracellular Ca++ with EGTA also did not inhibit stimulation of S6 kinase by EGF. When intracellular Ca++ was buffered with the permeable Ca++ chelator quin2, EGF stimulation was reduced 50%. A similar inhibition of the EGF response was observed when cells were incubated in buffers with high K+ concentrations or in the presence of the K+ ionophore valinomycin. Insulin and IGF1 stimulation of S6 kinase were also inhibited by high K+ concentrations and by buffering intracellular Ca++. In contrast to the responses to EGF, insulin- and IGF1-activation of S6 kinase was enhanced when glucose was present and depended on the presence of bicarbonate in the medium. The results indicate that ionic signals generated by growth factors and insulin, such as increases in intracellular pH or Na+, do not seem to be involved in the activation of S6 kinase. However, effects of growth factors or insulin on membrane potential and/or K+ fluxes and redistribution of intracellular Ca++ may play a role in the activation process. Furthermore, the mechanism of insulin activation of S6 kinase is distinct from the growth factors by its dependency on extracellular bicarbonate.     

Synthesis and secretion of an epidermal growth factor (EGF) by human fibroblast cells in culture

Human fibroblast (WS-1) cells in culture synthesized and secreted an epidermal growth factor which cross-reacted with human epidermal growth factor (hEGF) purified from human urine. hEGF secreted by the cells (designated as WS-1 EGF or fibroblast EGF) and hEGF isolated from urine (designated as urine EGF) were immunologically indistinguishable. The molecular weight of fibroblast EGF estimated by gel filtration was identical with that of hEGF from urine. On chromatofocusing chromatography, fibroblast EGF was eluted mainly at pH 4.26 as a sharp symmetric peak with a minor peak at pH 4.62, like urine EGF. These results suggested that EGF synthesized and secreted by human fibroblast cells is an identical molecule to that of hEGF in human urine.     

Epidermal growth factor-like transforming growth factor. I. Isolation, chemical characterization, and potentiation by other transforming factors from feline sarcoma virus-transformed rat cells

An acid-stable transforming growth factor (TGF) that interacts with epidermal growth factor (EGF) receptors and is structurally related to EGF was isolated from serum-free culture fluids of Snyder-Theilen feline sarcoma virus-transformed rat embryo (FeSV-Fre) cells. Purification of this EGF-like TGF (eTGF) was achieved by molecular filtration chromatography and successive reverse-phase high pressure liquid chromatography steps on octadecyl support eluted with acetonitrile and 1-propanol gradients, respectively. Rat eTGF consists of a 7.4-kD single polypeptide chain that co-migrates with biological activity in dodecyl sulfate-polyacrylamide electrophoresis gels. Like preparations of a related TGF from human melanoma cells (Marquardt, H., and Todaro, G.J. (1982) J. Biol. Chem. 257, 5220-5225), but unlike EGF from rat, human, or mouse, rat eTGF has phenylalanine and lacks methionine. However, the sequence of the first 30 amino acid residues in rat eTGF is H2N-Val-Val-Ser-His-Phe-Asn-Lys-Cys-Pro-Asp-Ser-His-Thr-Gln-Tyr-Cys-Phe-His-Gly - Thr-(x)-Arg-Phe-Leu-Val-Gln-Glu-Glu-(Lys)-(Lys)-, which is significantly (20% and 28%) homologous to the NH2-terminal region of mouse EGF and human EGF, respectively. In addition to eTGF, molecular filtration chromatography of acid-soluble extracts from medium conditioned by FeSV-Fre cells resolved a 14-kD transforming factor(s) apparently devoid of intrinsic mitogenic activity but able to elicit a strong anchorage-independent growth response in the presence of eTGF or EGF. These results show that: 1) a 7.4-kDa TGF structurally and functionally related to EGF has been isolated from FeSV-Fre cells and 2) the full anchorage-independent growth-promoting activity of medium conditioned by FeSV-Fre cells is due to the coordinate action of at least two types of factors, the 7.4-kDa eTGF and a second 14-kDa transforming factor(s).     

Murine epidermal growth factor: heterogeneity on high resolution ion-exchange chromatography

We have shown that epidermal growth factor (EGF) purified either by the classical method of Savage and Cohen, or solely by h.p.l.c. techniques can be resolved into two species, EGF alpha and EGF beta. However, despite the apparent purity of such materials, as determined both chromatographically and by amino acid analysis, they failed to give homogeneous products on radioiodination. Analysis by isoelectric focusing on agarose gels followed by transfer to nitrocellulose and silver staining showed that EGF alpha could be further resolved into three sub-species which focused at pH 4.6, 4.3 and 4.1. EGF beta (which also focused at pH 4.6) contained very small amounts of the species with isoelectric points of 4.1 and 4.3, probably due to slight contamination of this preparation by EGF alpha. Preparative separation of the sub-species of EGF alpha was achieved by high performance anion-exchange chromatography at pH 6.5 on a Pharmacia Mono Q column. Radioiodination of these purified sub-species did not produce significant charge heterogeneity. However, two slightly different forms of [125I]EGF alpha 1 (pH 4.6 species) were separable by anion-exchange chromatography on the Mono Q column. All of the EGF species competed for binding to EGF receptors on A431 cells and were active mitogens for BALB/c 3T3 fibroblasts.     

ras and neu oncogenes reverse serum inhibition and epidermal growth factor dependence of serum-free mouse embryo cells

Serum-free mouse embryo cells, cultured in basal nutrient medium supplemented with insulin, transferrin, epidermal growth factor, fibronectin, and high-density lipoprotein, do not exhibit growth crisis, lack detectable chromosomal aberrations, are nontumorigenic in vivo, are dependent on epidermal growth factor for survival, and are growth inhibited by serum or platelet-free plasma. These cells after transfection with the human Ha-ras or rat neu oncogenes no longer required epidermal growth factor for survival, were tumorigenic in vivo, and also proliferated in serum-containing medium. Autocrine activity capable of replacing epidermal growth factor was detected in conditioned medium from ras-transformed cultures, but little such activity was detected in medium from neu-transformed cultures. In addition, the capability of ras or neu-transformed cells to grow in serum-containing medium could not be mimicked in untransformed cells by the addition of growth factors or conditioned medium from transformed cells. These results suggest that the known structural similarity of the neu gene product to the EGF receptor is also reflected in a functional similarity by which the mutationally activated neu protein can replace the ligand-activated EGF receptor. These results also suggest that the ability of ras- and neu-transformed cells to escape the effect of the inhibitory serum activity is a nonautocrine property distinct from the acquisition of EGF autonomy.     

Attenuated processing of epidermal growth factor in the face of marked degradation of transforming growth factor-alpha

T3M4 human pancreatic carcinoma cells avidly bound and internalized 125I-labeled epidermal growth factor (EGF) but did not readily degrade the ligand. Pulse-chase experiments in which the cell-bound radioactivity was allowed to dissociate into the incubation medium in the presence of unlabeled EGF indicated that the majority of the released 125I-EGF consisted of intact EGF and a slightly processed species that readily bound to the cell. Omission of unlabeled EGF during the chase period markedly decreased the amount of radioactivity in the incubation medium, mainly as a result of the rebinding of EGF to the cells. In contrast, T3M4 cells readily degraded 125I-labeled transforming growth factor-alpha (TGF-alpha), and the released radiolabeled products did not rebind to the cells. Both ligands were released from T3M4 cells under acidic conditions, complete dissociation occurring at a pH of 4.5 for EGF, and a pH of 6.5 for TGF-alpha. A 431 human epidermoid carcinoma cells and ASPC-1 human pancreatic carcinoma cells also failed to extensively degrade 125I-EGF, whereas Rat-1 fibroblasts markedly degraded the growth factor. As in the case of T3M4 cells, ASPC-1 cells extensively degraded 125I-TGF-alpha. Degradation of either ligand was blocked by the lysosomotropic compound methylamine in all the tested cell lines. Immunoprecipitation of the EGF receptor with specific polyclonal antibodies and Western blot analysis revealed the anticipated 170-kDa protein in T3M4 cells. Both EGF and TGF-alpha enhanced EGF receptor degradation, but TGF-alpha was less effective than EGF. These findings indicate that in certain cell types EGF and TGF-alpha may be differentially processed.     

Secretion of an epidermal growth factor-like growth factor by epidermal growth factor-independent rat mammary carcinoma cells.

Rat mammary carcinoma (RMC) cells derived from serially transplantable mammary tumors are independent of epidermal growth factor (EGF) for long-term growth in serum-free medium. This phenotype is in contrast to that of normal mammary epithelial cells or cells derived from nontransplantable tumors that express an absolute requirement for EGF for growth in culture. The results of the experiments reported here indicate that EGF-independent RMC cells secrete a growth factor with potent EGF-like mitogenic activity. Conditioned media obtained from these cells can substitute for EGF for the growth of the EGF-dependent cell line MCF-10. This growth factor is neither EGF nor transforming growth factor alpha and does not compete with 125I-EGF for binding to EGF receptors. Phosphotyrosine Western blot analysis of lysates obtained from EGF-independent RMC cells revealed the presence of a 190 kilodalton (kDa) protein that was distinct from the EGF receptor. Similarly, growth of MCF-10 cells to confluence in serum-free medium supplemented with conditioned medium growth factor in place of EGF resulted in the disappearance of the EGF receptor band and appearance of the 190 kDa band in phosphotyrosine Western blots. The 190 kDa tyrosine-phosphorylated protein detected in cells stimulated by the conditioned medium factor is unlikely to be the c-erbB-2 protein, as indicated by negative results in immunoprecipitation experiments and in vitro kinase assays. In summary, EGF-independent RMC cells secrete a factor with potent EGF-like mitogenic activity. This suggests that an autocrine loop involving this growth factor mediates EGF independence in these cells.