Combined effect of gene dosage and process optimization strategies on high-level production of recombinant human interleukin-3 (hIL-3) in Pichia pastoris fed-batch culture.

In this work, the combined effects of gene dosage and process optimization strategies were studied to achieve higher hIL-3 expression in Pichia system. The in-vitro multimerization method was used to generate various Pichia X-33 transformants having multi-copy expression cassettes. The quantitative polymerase chain reaction (qPCR) strategy was used to further confirm the genome integration of hIL-3 expression cassette. From shake flask expression studies, the recombinant hIL-3 concentration in culture supernatant increased upto 8 copies to a level of 310mg/L, thereafter a considerably lower expression was observed. The small scale optimization experiments at shake flask level resulted in an improved product concentration of 350mg/L. The batch and fed-batch fermentation runs in complex medium showed a product concentration of 1.81 and 1.49g/L, respectively. To further enhance the production level, the fermentation runs were conducted in modified minimal media where a maximum hIL-3 protein level of 2.23g/L was obtained in batch fermentation. The specific product yield (YP/X) was at a level of 25.65mg/g DCW, whereas the overall volumetric productivity of the process was 27.31mg/L/h. The biological activity of the partially purified hIL-3 protein was confirmed via the proliferation of human erythroleukemia TF-1 cells using MTT assay.

High-level expression and efficient refolding of therapeutically important recombinant human Interleukin-3 (hIL-3) in E. coli.

Human interleukin-3 (hIL-3) is a pleiotropic cytokine that stimulates the differentiation and proliferation of multipotent hematopoietic cells thus making it a therapeutically important molecule. In this study, its poor expression yield was improved by addressing various upstream bottlenecks in E. coli heterologous system. The codon-optimized hIL-3 gene was cloned under various signal sequences and solubility enhancer fusion tags for its hyper-expression under a strong T7 promoter. The optimization of shake flask expression studies resulted in a hIL-3 protein concentration of 225 mg/L in the form of inclusion bodies (IBs). Lowering of inducer concentration and cultivation temperature did not improve its solubility. The hIL-3 protein was refolded from IBs and resulted a protein recovery yield of 53% after optimization of refolding conditions. The refolded protein was subsequently purified using Ni-NTA affinity chromatography and gave ∼95% pure protein. The conformational properties of the refolded hIL-3 protein were studied by CD and fluorescence spectrometry where protein showed 40% α-helix and 12% ß-sheets with a fluorescence emission maxima at 344 nm. The molecular identity was further confirmed by MALDI-TOF/TOF and western blot analysis. The biological activity of refolded protein was confirmed via cell proliferation assay on human erythroleukemia TF-1 cells where commercial hIL-3 was taken as a standard control.

Plant-produced human recombinant erythropoietic growth factors support erythroid differentiation in vitro.

Clinically available red blood cells (RBCs) for transfusions are at high demand, but in vitro generation of RBCs from hematopoietic stem cells requires significant quantities of growth factors. Here, we describe the production of four human growth factors: erythropoietin (EPO), stem cell factor (SCF), interleukin 3 (IL-3), and insulin-like growth factor-1 (IGF-1), either as non-fused proteins or as fusions with a carrier molecule (lichenase), in plants, using a Tobacco mosaic virus vector-based transient expression system. All growth factors were purified and their identity was confirmed by western blotting and peptide mapping. The potency of these plant-produced cytokines was assessed using TF1 cell (responsive to EPO, IL-3 and SCF) or MCF-7 cell (responsive to IGF-1) proliferation assays. The biological activity estimated here for the cytokines produced in plants was slightly lower or within the range cited in commercial sources and published literature. By comparing EC50 values of plant-produced cytokines with standards, we have demonstrated that all four plant-produced growth factors stimulated the expansion of umbilical cord blood-derived CD34+ cells and their differentiation toward erythropoietic precursors with the same potency as commercially available growth factors. To the best of our knowledge, this is the first report on the generation of all key bioactive cytokines required for the erythroid development in a cost-effective manner using a plant-based expression system.

Expression and purification of non-N-glycosylated porcine interleukin 3 in yeast Pichia pastoris.

Yeast Pichia pastoris has been widely utilized to express heterologous recombinant proteins. P. pastoris expressed recombinant porcine interleukin 3 (IL3) has been used for porcine stem cell mobilization in allo-hematopoietic cell transplantation models and pig-to-primate xeno-hematopoietic cell transplantation models in our lab for many years. Since the yeast glycosylation mechanism is not exactly the same as those of other mammalian cells, P. pastoris expressed high-mannose glycoprotein porcine IL3 has been shown to result in a decreased serum half-life. Previously this was avoided by separation of the non-glycosylated porcine IL3 from the mixture of expressed glycosylated and non-glycosylated porcine IL3. However, this process was very inefficient and lead to a poor yield following purification. To overcome this problem, we engineered a non-N-glycosylated version of porcine IL3 by replacing the four potential N-glycosylation sites with four alanines. The codon-optimized non-N-glycosylated porcine IL3 gene was synthesized and expressed in P. pastoris. The expressed non-N-glycosylated porcine IL3 was captured using Ni-Sepharose 6 fast flow resin and further purified using strong anion exchange resin Poros 50 HQ. In vivo mobilization studies performed in our research facility demonstrated that the non-N-glycosylated porcine IL3 still keeps the original stem cell mobilization function.

High level expression of active recombinant human interleukin-3 in Pichia pastoris.

Interleukin-3 (IL-3) is a hematopoietic growth factor involved in the survival, proliferation and differentiation of multipotent hematopoietic cells. A DNA fragment containing the mature human IL-3 sequence was cloned into pPICZαA, generating a fusion protein with the alpha factor signal sequence in the N-terminus and 6×His as well as c-Myc tags in the C-terminus. The resulting plasmid was integrated into the genome of Pichia pastoris strain X-33. Recombinant yeast transformants with high-level rhIL-3 production were identified, secreting as much as 26mg/L rhIL-3 after 4days of induction by methanol in flask. The rhIL-3 was purified by Ni(+)-NTA affinity chromatography, followed by DEAE anion exchange, yielding over 95% highly purified rhIL-3 preparation at about 21mg/L. Mass spectrometry and MALDI-TOF-TOF analysis of the purified rIL-3 showed molecular weights of 18995.694Da and 22317.469Da, due to different degrees of N-linked glycosylation. The biological activity of the rhIL-3 proteins was confirmed by its ability to support ba/f3 cells proliferation and activate the ERK signaling pathways. The results demonstrate that the experimental procedure we have developed can produce a large amount of active recombinant human IL-3 from P. pastoris.

A strategic crossflow filtration methodology for the initial purification of promegapoietin from inclusion bodies.

A novel crossflow filtration methodology is demonstrated for the initial purification of the therapeutic protein, promegapoietin-1a (PMP), produced as inclusion bodies (IBs) in a recombinant Escherichia coli bioprocess. Two strategic separation steps were performed by utilizing a filtration unit with a 1000 kDa polyethersulphone membrane. The first step, aiming for separation of soluble contaminants, resulted in a 50% reduction of the host cell proteins, quantified by total amino acid analysis and a 70% reduction of all DNA, quantified by fluorometry, when washing the particulate material with a 10mM EDTA in 50mM phosphate buffer, pH 8. The second step, aiming for separation of particulate contaminants from solubilized IBs, resulted in a 97-99.5% reduction of endotoxin, used as a marker for cell debris, and was quantified by the kinetic turbidimetric LAL endotoxin assay. The overall PMP yield was 58% and 33% respectively for the two solubilizations investigated, guanidine hydrochloride and arginine, as measured by RP-HPLC. The scope was also to investigate the physical characteristics of the intermediate product/s with regard to the choice of IB solvent. Preliminary results from circular dichroism spectroscopy measurements indicate that the protein secondary structure was restored when arginine was used in the second step.

Evaluation of refolding conditions for a human recombinant fusion cytokine protein, promegapoietin-1a.

Conditions to obtain correctly folded PMP-1a (promegapoietin-1a), an engineered fusion IL-3 (interleukin-3) and thrombopoietin receptor agonist from recombinant Escherichia coli IBs (inclusion bodies), were defined to generate sufficient amounts of protein for evaluation as a potential therapeutic compound. Several ionic and non-ionic detergents, as well as the chaotrope urea, in combination with selected additives, were screened for their ability to dissolve IB protein and promote formation of monomeric, oxidized protein. Upon dissolution, soluble aggregates constituted 50-60% of total protein in detergent-solubilized IBs depending on the level of detergent used, whereas use of urea increased aggregation to approx. 70%. Subsequent addition of 5 mM cysteine or DTT (dithiothreitol) reduced the levels of aggregation, but never lower than approx. 20%. Refolds from detergent-solubilized IBs with or without organic modifiers characteristically produced multiple persistent misfolded species. However, the addition of a 12:1 molar excess of cystine (cystine/DTT) to urea-dissolved IBs containing DTT, followed by dilution, promoted the formation of correctly oxidized, disulfide-paired PMP-1a monomer with minimal misfolds present. Thus treatment of urea-dissolved proteins with thiol-group-containing additives and control of dilution, pH, protein concentration and order of addition were able to produce a maximum refold efficiency of 40-50% of correctly paired protein monomer.

Expression and purification of the recombinant diphtheria fusion toxin DT388IL3 for phase I clinical trials.

A genetically engineered fusion toxin targeted to acute myeloid leukemia (AML) blasts was designed with the first 388 amino acid residues of diphtheria toxin with an H-M linker fused to human interleukin-3. The cDNA was subcloned in the pRK bacterial expression plasmid and used to transform BLR (DE3) Escherichia coli. A single transformed colony was grown in Superbroth with ampicillin; bacteria were centrifuged at an OD(650) of 1.3; master cell bank aliquots of bacteria in 30% glycerol/Superbroth were frozen and stored at -80 degrees C. Master cell bank bacteria were diluted 1500-fold into Superbroth and recombinant protein was induced with 1 mM IPTG at an OD(650) of 0.6. After two additional hours of fermentation, inclusion bodies were isolated, washed, and denatured in guanidine hydrochloride and dithioerythritol. Recombinant protein was refolded by diluted 100-fold in cold buffer with arginine and oxidized glutathione. After dialysis, purified protein was obtained after anion-exchange, size exclusion on FPLC, and polymyxin B affinity chromatography. The final material was filter sterilized, aseptically vialed, and stored at -80 degrees C. Seventy-five 3-L bacterial culture preparations were made and pooled for the AT-1 batch (568 mL) and twenty-four 3-L bacterial culture preparations were made and pooled for the AT-2 batch (169 mL). The final product was characterized by Coomassie Plus protein assay, Coomassie-stained SDS-PAGE, limulus amebocyte lysate endotoxin assay, human AML TF/H-ras cell cytotoxicity assay, sterility, tandem mass spectroscopy, IL3 receptor binding affinity, ADP ribosylation activity, inhibition of normal human CFU-GM, disulfide bond analysis, immunoblots, peptide mapping, stability, HPLC TSK3000, N-terminal sequencing, E. coli DNA contamination, C57BL/6 mouse toxicity, cynomolgus monkey toxicity, and immunohistochemistry. Yields were 25.7+/-5.6 mg/L bacterial culture of denatured fusion toxin. After refolding and chromatography, final yields were 20+/-11% or 5 mg/L. Vialed product was sterile. Batches were in 0.25 M sodium chloride/5 mM Tris, pH 8, and had protein concentrations of 1.8-1.9 mg/mL. Purity by SDS-PAGE was 99+/-1%. Aggregates by HPLC were <1 %. Potency revealed a 48 h IC(50) of 6-8 pM on TF/H-ras cells. Endotoxin levels were 1 eu/mg. The remaining chemical and biologic assays confirmed the purity, composition, and functional activities of the molecule. The LD(10) in mice was 250 microg/kg/day every other day for six doses. The MTD in monkeys was 60 microg/kg/day every other day for six doses. Drug did not react with tested frozen human tissue sections by immunohistochemistry. There was no evidence of loss of solubility, proteolysis aggregation, or loss of potency over 6 months at -80 and -20 degrees C. Further, the drug was stable at 4 and 25 degrees C in the plastic syringe and administration tubing for 24 h and at 37 degrees C in human serum for 24 h. The synthesis of this protein drug should be useful for production for clinical phase I/II clinical trials and may be suitable for other diphtheria fusion toxins indicated for clinical development.

Expression, purification, refolding, and characterization of recombinant human interleukin-13: utilization of intracellular processing.

Interleukin-13 (IL-13) is a pleiotropic cytokine that elicits both proinflammatory and anti-inflammatory immune responses. Recent studies underscore its role in several diseases, including asthma and cancer. Solution studies of IL-13 and its soluble receptors may facilitate the design of antagonists/agonists which would require milligram quantities of specifically labeled protein. A synthetic gene encoding human IL-13 (hIL-13) was inserted into the pMAL-c2 vector with a cleavage site for the tobacco etch virus (TEV) protease. Coexpression of the fusion protein and TEV protease led to in vivo cleavage, resulting in high levels of hIL-13 production. hIL-13, localized to inclusion bodies, was purified and refolded to yield approximately 2 mg per liter of bacteria grown in minimal media. Subsequent biochemical and biophysical analysis of both the unlabeled and (15)N-labeled protein revealed a bioactive helical monomer. In addition, the two disulfide bonds were unambiguously demonstrated to be Cys29-Cys57 and Cys45-Cys71 by a combined proteolytic digestion and mass spectrometric analysis.

Diphtheria toxin fused to human interleukin-3 is toxic to blasts from patients with myeloid leukemias.

Leukemic blasts from patients with acute phase chronic myeloid leukemic and refractory acute myeloid leukemia are highly resistant to a number of cytotoxic drugs. To overcome multi-drug resistance, we engineered a diphtheria fusion protein by fusing human interleukin-3 (IL3) to a truncated form of diphtheria toxin (DT) with a (G4S)2 linker (L), expressed and purified the recombinant protein, and tested the cytotoxicity of the DTLIL3 molecule on human leukemias and normal progenitors. The DTLIL3 construct was more cytotoxic to interleukin-3 receptor (IL3R) bearing human myeloid leukemia cell lines than receptor-negative cell lines based on assays of cytotoxicity using thymidine incorporation, growth in semi-solid medium and induction of apoptosis. Exposure of mononuclear cells to 680 pM DTLIL3 for 48 h in culture reduced the number of cells capable of forming colonies in semi-solid medium (colony-forming units leukemia) > or =10-fold in 4/11 (36%) patients with myeloid acute phase chronic myeloid leukemia (CML) and 3/9 (33%) patients with acute myeloid leukemia (AML). Normal myeloid progenitors (colony-forming unit granulocyte-macrophage) from five different donors treated and assayed under identical conditions showed intermediate sensitivity with three- to five-fold reductions in colonies. The sensitivity to DTLIL3 of leukemic progenitors from a number of acute phase CML patients suggests that this agent could have therapeutic potential for some patients with this disease.

Characterization of the isoforms of PIXY321, a granulocyte-macrophage-colony stimulating factor-interleukin-3 fusion protein, separated by preparative isoelectric focusing on immobilized pH gradients.

We present here the purification and the characterization of the isoforms of PIXY321, a genetically engineered fusion of granulocyte-macrophage-colony stimulating factor and interleukin-3 expressed in yeast. The isoforms of PIXY321 were isolated using preparative isoelectric focusing (IEF) on immobilized pH gradients. Analysis of the collected fractions on analytical IEF gels showed that PIXY321 was resolved into four discrete isoforms of isoelectric point (pI) 5.0, 5.1, 5.2 and 5.3 with excellent yields. Subsequent analysis of purified isoforms of PIXY321 by peptide mapping and mass spectrometry linked the microheterogeneity of the original molecule to three parameters, the presence of deamidated residues, charged glycans and the pattern of O-linked glycosylation along the peptide sequence. This last parameter emphasizes the role of conformational aspects as key factors influencing the apparent isoelectric point of protein isoforms.

Characterization of the microheterogeneities of PIXY321, a genetically engineered granulocyte-macrophage colony-stimulating factor/interleukin-3 fusion protein expressed in yeast.

PIXY321, a human cytokine analog genetically engineered by the fusion of granulocyte-macrophage colony-stimulating factor (GM-CSF) and interleukin-3 (IL-3), was expressed in yeast under the control of the alcohol dehydrogenase 2 (ADH2) promoter and the alpha-mating factor expression system. To provide the material necessary for the evaluation of PIXY321 in clinical trials, the production was scaled up to the 1200-1 scale and the PIXY321 molecule isolated by four successive steps of ion-exchange chromatography. Multiple heterogeneities, due to the presence of different patterns of glycosylation as well as multiple amino acid sequences at both N and C termini, were characterized on the purified molecule using complementary analytical techniques including electrophoresis, liquid chromatography and electrospray mass spectrometry. Four different N-terminal sequences were identified but simplified to a reproducible ratio of two sequences, the mature form and a form starting at Ala3, by adjustment of the process conditions. Molecules lacking 1-6 residues at the C-terminus were identified and their relative frequencies quantified. Amino acid modifications, such as three oxidized Met residues at positions 79, 141 and 187 and one deamidated Asn residue at position 176, were detected at low level. Microheterogeneities in glycosylation were characterized on four different sites, one located in the GM-CSF portion and three in the IL-3 portion of the molecule. The sites were shown to be differentially occupied and to carry 0-10 mannose residues according to their location in the sequence. Precise measurement of the heterogeneities at the molecular level were used to tune the process conditions and ensure reproducibility of the clinical product between lots.

Effect of formulation and freeze-drying on the long-term stability of rDNA-derived cytokines.

Lyophilized recombinant DNA-derived cytokine preparations intended for use as International Reference Material have been processed in such a way as to give batches of preparations which possess homogeneity and high stability, in order that each cytokine standard has a reasonable use of some 10 to 20 years. This has been achieved by exercising careful control at every stage of the lyophilization process. All cytokines were ampouled in microgram quantities with relatively large quantities (milligrams) of bulking agent such as albumin and carbohydrate which were shown to have no effect on the biological activity of the cytokine. Residual moisture present in the preparations and their uptake was measured by Karl-Fischer titration and found to vary from 0.5% to 4% moisture. On storage over phosphorus pentoxide these values were reduced to 0.1% to 0.5% with good reproducibility throughout the batch. Reconstitution of the freeze-dried plug was readily achieved in seconds. The stability of freeze-dried preparations of several cytokines was monitored by observing changes in biological activity after accelerated degradation studies. To date, several freeze-dried preparations have shown no loss in activity after storage at temperatures up to +56 degrees C for one year.

Purification and characterization of natural human interleukin-3.

The human interleukin-3 gene was cloned in 1986 and the biochemical and biological properties of recombinant human interleukin-3 (rhuIL-3) protein were described. In this report we compare rhuIL-3 with nonrecombinant, natural huIL-3, purified from the supernatant of the human T cell leukemia line Jurkat. The main purification step, affinity chromatography, using a selected monoclonal anti-huIL-3 antibody, resulted in an approximately 40,000-fold enrichment of huIL-3. Combination of this step with ion-exchange and reverse-phase chromatography yielded natural huIL-3 of high purity (greater than 98%). A highly sensitive and specific sandwich ELISA, comprising two epitope-mapped monoclonal anti-huIL-3 antibodies, was used to quantitate huIL-3 during purification. Amino acid sequence determination revealed that the 38 N-terminal amino acids of Jurkat-derived huIL-3 are identical to the published sequence deduced from human fetal liver genomic DNA but differ in one residue from that derived from human T cell clones. The degree of glycosylation of Jurkat-derived huIL-3 was similar to Chinese hamster ovary cell-expressed rhuIL-3. Natural huIL-3 showed very similar biological activities to rhuIL-3 in proliferation and receptor binding assays utilizing huIL-3 responsive primary cells and cell lines, as well as in the human bone marrow colony assay.

Heterogeneity in human interleukin-3 receptors. A subclass that binds human granulocyte/macrophage colony stimulating factor.

125I-Labeled recombinant human interleukin-3 (IL-3) was used to study the characteristics and distribution of receptors for IL-3 on human cells. Receptors were found on primary monocytes, on some strains of KG-1 cells, and on pre-B cell lines. Binding was rapid at 37 degrees C, while requiring several hours to reach equilibrium at 4 degrees C. Equilibrium binding studies indicated that IL-3 bound to a single class of high affinity receptor (less than 500 receptors/cell) with a Ka of approximately 1 x 10(10) M-1. Inhibition studies revealed that human granulocyte/macrophage colony stimulating factor partially inhibited the binding of 125I-IL-3 to human monocytes but not JM-1 cells. Additional analysis showed that on KG-1 cells, both IL-3 and GM-CSF partially competed specific binding of heterologous radiolabeled ligand, with approximately equivalent capacities. This competition occurred at both 37 and 4 degrees C. These results suggest heterogeneity in the binding sites for IL-3 and GM-CSF in which a subset of receptors binds only IL-3, a subset only GM-CSF, and another subset can bind both, all with high affinity. Additional heterogeneity was suggested by equilibrium binding of 125I-IL-3 to KG-1 cells which revealed a biphasic Scatchard plot containing a low affinity component not observed on monocytes and JM-1 cells.

Interleukin-3 down-regulates its own receptor.

To gain insight into the mechanisms involved in regulating murine interleukin-3 (mIL-3) receptor expression, we have examined the effects of mIL-3 and murine granulocyte-macrophage colony-stimulating factor (mGM-CSF) on mIL-3 receptor internalization and re-expression and studied the relationship between mIL-3 cell surface receptor density and growth factor sensitivity. As a source of cells for our studies, we used a B6SUtA clone, B6SUtA1, which grows equally well in mIL-3 or mGM-CSF when supplemented with 20% fetal calf serum (FCS) in RPMI 1640. Intracellular processing studies carried out in the presence and absence of methylamine suggested that mIL-3 is cleaved at two specific sites before its complete digestion within lysosomes. However, unlike its ligand, cycloheximide studies indicated that internalized mIL-3 receptors are recycled to the cell surface. When B6SUtA1 cells were continuously passaged in mIL-3, cell populations allowed to exhaust the mIL-3 in the medium (high density cells) expressed more than ten times (ie, approximately 100,000/cell) the mIL-3 receptor number of those growing exponentially at low cell concentrations (low density cells). Since the high density cells were no larger than the low density cells, the marked increase in mIL-3 receptor number per cell reflects a true up-regulation of receptor expression. A kinetic analysis of this up-regulation revealed that it begins within one hour of mIL-3 exhaustion. Moreover, proliferation assays with these two cell populations, using 3H-thymidine incorporation, suggested that the high density cells were 30-fold more responsive to mIL-3. However, when B6SUtA1 cells were passaged in mGM-CSF, there was no difference in mIL-3 receptor number between high density and low density cells (ie, approximately 100,000/cell). Identical studies carried out with another mIL-3 dependent cell line, 32D C3, demonstrated that this phenomenon was not unique to B6SUtA1 cells.

Multiple glycosylated forms of T cell-derived interleukin 3 (IL-3). Heterogeneity of IL-3 from physiological and nonphysiological sources.

Interleukin 3 (IL-3) derived from mouse T cells was biosynthetically labeled with either [35S]methionine or [3H]mannose, affinity-purified using various anti-IL-3 antibodies, and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Autoradiography revealed the same three major bands with Mr values of 21,500-22,500, 27,000-31,000, and 32,000-36,000, irrespective of whether the anti-IL-3 antibody had been directed to the N or C termini of the IL-3 polypeptide. Bioassay of eluates from the gels confirmed that all three bands exhibited IL-3 bioactivity. IL-3 produced from two nonphysiological sources, the myelomonocytic leukemia WEHI-3B or Cos 7 cells that had been transfected with an IL-3 cDNA clone, had in each case a different pattern of microheterogeneity. Treatment with either tunicamycin or N-glycanase resulted in IL-3 running as one band with Mr 16,000, corresponding to its 140-amino acid polypeptide chain. No evidence for proteolytic processing was detected. These results show that the Mr heterogeneity of IL-3 was highly dependent on the cellular source and is due to N-linked glycosylation.

Production of recombinant human colony stimulating factors in yeast.

Efficient yeast expression and purification systems for production of recombinant human GM-CSF, IL-3 and G-CSF have been established. Though yeast-derived production of recombinant CSFs (through the use of secretion based system) allows for generation of native molecules which can then be readily separated from fermentation broth, in many instances, natural cDNAs have had to be altered to allow for efficient expression, as well as production of a less heterogeneous product. In the case of CSFs described herein, beneficial mutations (made through site-directed mutagenesis) have included elimination of potential N-linked glycosylation sites, removal of KexII protease recognition sites (notably alterations in dibasic sequences) and elimination of extraneous cysteine residues which might complicate isolation of a homogeneous product due to intermolecular disulfide bonding.

Keratinocyte-derived granulocyte/macrophage colony-stimulating factor induces DNA synthesis by peritoneal macrophages.

Keratinocytes have been demonstrated to produce a number of cytokines, including growth factors such as the CSF IL-3. Circulating blood monocytes and some elicited macrophages retain a significant proliferative potential in response to colony-stimulating activity. Because a macrophage response is prominent in a variety of cutaneous immune reactions, we have studied the ability of conditioned media (CM) from a transformed murine keratinocyte cell line (PAM 212) and from normal murine keratinocytes to induce growth of peritoneal macrophages. CM from both normal and transformed keratinocyte cultures induces [3H]thymidine incorporation by thioglycollate-elicited, but not resident, peritoneal macrophages. IEF of PAM 212 CM reveals peaks of activity at pI 4.8 and less than or equal to 4.2. Analysis of CM by reversed-phase HPLC demonstrates active fractions that elute at 46 to 48% and 53 to 55% acetonitrile. The Mr of the 46 to 48% acetonitrile factor is 25 to 30 kDa by gel filtration HPLC. Polyclonal anti-granulocyte/macrophage (GM) CSF antibody blocks the induction of macrophage [3H]thymidine incorporation by factors with pI 4.8 and eluting at 46 to 48% acetonitrile but does not reduce the activity of crude CM or the factor eluting at 53 to 55% acetonitrile. Based on both physiochemical criteria and antibody neutralization, keratinocytes produce GM-CSF. Keratinocyte-derived factors, including GM-CSF, may play an important role in regulating cutaneous macrophage responses.