The Pattern of GH Action in the Mouse Brain.

GH acts in numerous organs expressing the GH receptor (GHR), including the brain. However, the mechanisms behind the brain's permeability to GH and how this hormone accesses different brain regions remain unclear. It is well-known that an acute GH administration induces phosphorylation of the signal transducer and activator of transcription 5 (pSTAT5) in the mouse brain. Thus, the pattern of pSTAT5 immunoreactive cells was analyzed at different time points after IP or intracerebroventricular GH injections. After a systemic GH injection, the first cells expressing pSTAT5 were those near circumventricular organs, such as arcuate nucleus neurons adjacent to the median eminence. Both systemic and central GH injections induced a medial-to-lateral pattern of pSTAT5 immunoreactivity over time because GH-responsive cells were initially observed in periventricular areas and were progressively detected in lateral brain structures. Very few choroid plexus cells exhibited GH-induced pSTAT5. Additionally, Ghr mRNA was poorly expressed in the mouse choroid plexus. In contrast, some tanycytes lining the floor of the third ventricle expressed Ghr mRNA and exhibited GH-induced pSTAT5. The transport of radiolabeled GH into the hypothalamus did not differ between wild-type and dwarf Ghr knockout mice, indicating that GH transport into the mouse brain is GHR independent. Also, single-photon emission computed tomography confirmed that radiolabeled GH rapidly reaches the ventral part of the tuberal hypothalamus. In conclusion, our study provides novel and valuable information about the pattern and mechanisms behind GH transport into the mouse brain.

Biophysical properties of electrospun chitosan-grafted poly(lactic acid) nanofibrous scaffolds loaded with chondroitin sulfate and silver nanoparticles.

The aim of this work was to study the biophysical properties of the chitosan-grafted poly(lactic acid) (CH-g-PLA) nanofibers loaded with silver nanoparticles (AgNPs) and chondroitin-4-sulfate (C4S). The electrospun CH-g-PLA:AgNP:C4S nanofibers were manufactured using the electrospinning technique. The microstructure of the CH-g-PLA:AgNP:C4S nanofibers was investigated by proton nuclear magnetic resonance (1H-NMR), scanning electron microscopy (SEM), UV-Visible spectroscopy (UV-Vis), X-ray diffraction (XRD), and Fourier transform infrared (ATR-FTIR) spectroscopy. ATR-FTIR and 1H-NMR confirm the CH grafting successfully by PLA with a substitution degree of 33.4%. The SEM measurement results indicated apparently smooth nanofibers having a diameter range of 340 ± 18 nm with porosity of 89 ± 3.08% and an average pore area of 0.27 µm2. UV-Vis and XRD suggest that silver nanoparticles with the size distribution of 30 nm were successfully incorporated into the electrospun nanofibers. The water contact angle of 12.8 ± 2.7° reveals the hydrophilic nature of the CH-g-PLA:AgNP:C4S nanofibers has been improved by C4S. The electrospun CH-g-PLA:AgNP:C4S nanofibers are found to release ions Ag+ at a concentration level capable of rendering an antimicrobial efficacy. Gram-positive bacteria (S.aureus) were more sensitive to CH-g-PLA:AgNP:C4S than Gram-negative bacteria (E. coli). The electrospun CH-g-PLA:AgNP:C4S nanofibers exhibited no cytotoxicity to the L-929 fibroblast cells, suggesting cytocompatibility. Fluorescence microscopy demonstrated that C4S promotes the adhesion and proliferation of fibroblast cells onto electrospun CH-g-PLA:AgNP:C4S nanofibers.

Periplasmic synthesis and purification of the human prolactin antagonist Δ1-11-G129R-hPRL.

The human prolactin antagonist Δ1-11-G129R-hPRL is a 21.9 kDa recombinant protein with 188 amino acids that downregulates the proliferation of a variety of cells expressing prolactin receptors. Periplasmic expression of recombinant proteins in E. coli has been considered an option for obtaining a soluble and correctly folded protein, as an alternative to cytoplasmic production. The aim of this work was, therefore, to synthesize for the first time, the Δ1-11-G129R-hPRL antagonist, testing different activation temperatures and purifying it by classical chromatographic techniques. E. coli BL21(DE3) strain was transformed with a plasmid based on the pET25b( +) vector, DsbA signal sequence and the antagonist cDNA sequence. Different doses of IPTG were added, activating under different temperatures, and extracting the periplasmic fluid via osmotic shock. The best conditions were achieved by activating at 35 °C for 5 h using 0.4 mM IPTG, which gave a specific expression of 0.157 ± 0.015 µg/mL/A600 at a final optical density of 3.43 ± 0.13 A600. Purification was carried out by nickel-affinity chromatography followed by size-exclusion chromatography, quantification being performed via high-performance size-exclusion chromatography (HPSEC). The prolactin antagonist was characterized by SDS-PAGE, Western blotting, reversed-phase high-performance liquid chromatography (RP-HPLC) and MALDI-TOF-MS. The final product presented > 95% purity and its antagonistic effects were evaluated in vitro in view of potential clinical applications, including inhibition of the proliferation of cancer cells overexpressing the prolactin receptor and specific antidiabetic properties, taking also advantage of the fact that this antagonist was obtained in a soluble and correctly folded form and without an initial methionine.

Suppression of Prolactin Secretion Partially Explains the Antidiabetic Effect of Bromocriptine in ob/ob Mice.

Previous studies have shown that bromocriptine mesylate (Bromo) lowers blood glucose levels in adults with type 2 diabetes mellitus; however, the mechanism of action of the antidiabetic effects of Bromo is unclear. As a dopamine receptor agonist, Bromo can alter brain dopamine activity affecting glucose control, but it also suppresses prolactin (Prl) secretion, and Prl levels modulate glucose homeostasis. Thus, the objective of the current study was to investigate whether Bromo improves insulin sensitivity via inhibition of Prl secretion. Male and female ob/ob animals (a mouse model of obesity and insulin resistance) were treated with Bromo and/or Prl. Bromo-treated ob/ob mice exhibited lower serum Prl concentration, improved glucose and insulin tolerance, and increased insulin sensitivity in the liver and skeletal muscle compared with vehicle-treated mice. Prl replacement in Bromo-treated mice normalized serum Prl concentration without inducing hyperprolactinemia. Importantly, Prl replacement partially reversed the improvements in glucose homeostasis caused by Bromo treatment. The effects of the Prl receptor antagonist G129R-hPrl on glucose homeostasis were also investigated. We found that central G129R-hPrl infusion increased insulin tolerance of male ob/ob mice. In summary, our findings indicate that part of Bromo effects on glucose homeostasis are associated with decrease in serum Prl levels. Because G129R-hPrl treatment also improved the insulin sensitivity of ob/ob mice, pharmacological compounds that inhibit Prl signaling may represent a promising therapeutic approach to control blood glucose levels in individuals with insulin resistance.

Brain STAT5 signaling modulates learning and memory formation.

The signal transducer and activator of transcription 5 (STAT5) is a transcription factor recruited by numerous cytokines. STAT5 is important for several physiological functions, including body and tissue growth, mammary gland development, immune system and lipid metabolism. However, the role of STAT5 signaling for brain functions is still poorly investigated, especially regarding cognitive aspects. Therefore, the objective of the present study was to investigate whether brain STAT5 signaling modulates learning and memory formation. For this purpose, brain-specific STAT5 knockout (STAT5 KO) mice were studied in well-established memory tests. Initially, we confirmed a robust reduction in STAT5a and STAT5b mRNA levels in different brain structures of STAT5 KO mice. STAT5 KO mice showed no significant alterations in metabolism, growth, somatotropic axis and spontaneous locomotor activity. In contrast, brain-specific STAT5 ablation impaired learning and memory formation in the novel object recognition, Barnes maze and contextual fear conditioning tests. To unravel possible mechanisms that might underlie the memory deficits of STAT5 KO mice, we assessed neurogenesis in the hippocampus, but no significant differences were observed between groups. On the other hand, reduced insulin-like growth factor-1 (IGF-1) mRNA expression was found in the hippocampus and hypothalamus of STAT5 KO mice. These findings collectively indicate that brain STAT5 signaling is required to attain normal learning and memory. Therefore, STAT5 is an important downstream cellular mechanism shared by several cytokines to regulate cognitive functions.

Determination of recombinant Interferon-α2 in E. coli periplasmic extracts by reversed-phase high-performance liquid chromatography.

Reversed-phase high-performance liquid chromatography (RP-HPLC) has been used to analyze Interferon α-2 (IFN-α2) as a pure protein or as a pharmaceutical preparation: a method for analyzing periplasmic IFN-α2 directly in osmotic shock extract has, however, never been reported. This work describes an RP-HPLC methodology for the qualitative and quantitative analysis of human IFN-α2a and IFN-α2b directly in bacterial periplasmic extracts or in purified preparations. The analytical method has been set up and validated for accuracy, precision, linearity, sensitivity and specificity. A recovery test indicated an average bias of ∼1%, intra-day and inter-day quantitative determinations presented relative standard deviations always≤5%, while the working sensitivity was of ∼0.3µg of IFN-α2 (RSD=5%). The method proved to be suitable for detecting and quantifying also glycosylated and oxidized forms and N-methionylated IFN-α2 molecules, it was, however, not able to distinguish between IFN-α2a and IFN-α2b. This rapid methodology allows the application of RP-HPLC as a powerful tool to monitor the production yield and quality of IFN-α2 in osmotic shock fluids, right after, or even during the fermentation process.

Expression, purification and characterization of the authentic form of human growth hormone receptor antagonist G120R-hGH obtained in Escherichia coli periplasmic space.

The human growth hormone receptor antagonist G120R-hGH precludes dimerization of GH and prolactin receptors and consequently JAK/STAT signaling. Some modifications in this antagonist resulted in a drug specific for the GH receptor, called Pegvisomant (Somavert®). However, the original G120R-hGH is usually synthesized in bacterial cytoplasm as inclusion bodies, not being a commercial product. The present work describes the synthesis and characterization of G120R-hGH secreted into bacterial periplasm and obtained with a vector based on a constitutive lambda-PL promoter. This antagonist can be useful for studies aiming at investigating the effects of a simultaneous inhibition of GH and prolactin signaling, as a potential anti-tumoral or anti-diabetic compound. G120R-hGH, synthesized using the W3110 E. coli strain, showed a yield of 1.34 ± 0.24 µg/ml/A600 (∼0.79 mg G120R-hGH/g of wet weight cells) after cultivation at 30 °C up to 3 A600 units and induction at 37 °C, for 6 h, with final 4.3 ± 0.3 A600. A laboratory scale purification was carried out using three chromatographic steps with a total yield of 32%, reaching 98% purity. The obtained protein was characterized by SDS-PAGE, Western Blotting, Mass spectrometry, RP-HPLC, HPSEC and in vitro proliferation bioassay. The proliferation assay, based on Ba/F3-LLP cells, shows that G120R-hGH (100 ng/ml) significantly inhibited (64%) the proliferative action of hGH (1 ng/ml). This is the first time that G120R-hGH is synthesized in bacterial periplasmic space and therefore correctly folded, without the initial methionine. The reasons for a divergent efficacy for antagonizing hGH versus hPRL is currently unknown and deserves further investigation.

Distribution of growth hormone-responsive cells in the mouse brain.

Growth hormone (GH) exerts important biological effects primarily related to growth and metabolism. However, the role of GH signaling in the brain is still elusive. To better understand GH functions in the brain, we mapped the distribution of GH-responsive cells and identified the receptors involved in GH central effects. For this purpose, mice received an acute intraperitoneal challenge with specific ligands of the GH receptor (mouse GH), prolactin receptor (prolactin) or both receptors (human GH), and their brains were subsequently processed immunohistochemically to detect the phosphorylated form of STAT5 (pSTAT5). GH induced pSTAT5 immunoreactivity in neurons, but not in astroglial cells of numerous brain regions, including the cerebral cortex, nucleus accumbens, hippocampus, septum and amygdala. The most prominent populations of GH-responsive neurons were located in hypothalamic areas, including several preoptic divisions, and the supraoptic, paraventricular, suprachiasmatic, periventricular, arcuate, ventromedial, dorsomedial, tuberal, posterior and ventral premammillary nuclei. Interestingly, many brainstem structures also exhibited GH-responsive cells. Experiments combining immunohistochemistry for pSTAT5 and in situ hybridization for GH and prolactin receptors revealed that human GH induced pSTAT5 in most, but not all, brain regions through both prolactin and GH receptors. Additionally, males and females exhibited a similar number of GH-responsive cells in forebrain structures known to be sexually dimorphic. In summary, we found GH-responsive cells primarily distributed in brain regions implicated in neurovegetative, emotional/motivational and cognitive functions. Our findings deepen the understanding of GH signaling in the brain and suggest that central GH signaling is likely more ample and complex than formerly recognized.

N-glycoprofiling analysis in a simple glycoprotein model: a comparison between recombinant and pituitary glycosylated human prolactin.

Human prolactin (hPRL) is a polypeptide hormone occurring in the non-glycosylated (NG-hPRL) and glycosylated (G-hPRL) forms, with MM of approximately 23 and 25kDa, respectively. It has a single, partially occupied N-glycosylation site located at Asn-31, which makes it a particularly simple and interesting model for glycosylation studies. The bioactivity of G-hPRL is lower than that of NG-hPRL (by ca. 4-fold) and its physiological function is not clear. However, carbohydrate moieties generally play important roles in the biosynthesis, secretion, biological activity, and plasma survival of glycohormones and can vary depending on the host cell. The main objective of this study was to determine the N-glycan structures present in native, pituitary G-hPRL and compare them with those present in the recombinant hormone. To obtain recombinant G-hPRL, genetically modified Chinese hamster ovary cells (CHO), adapted to growth in suspension, were treated with cycloheximide, thus increasing the glycosylation site occupancy from 5.5% to 38.3%, thereby facilitating G-hPRL purification. CHO cell-derived G-hPRL (CHO-G-hPRL) was compared to pituitary G-hPRL (pit-G-hPRL) especially with regard to N-glycoprofiling. Among the main differences found in the pituitary sample were an extremely low presence of sialylated (1.7%) and a high percentage of sulfated (74.0%) and of fucosylated (90.5%) glycans. A ∼6-fold lower in vitro bioactivity and a higher clearance rate in mice were also found for pit-G-hPRL versus CHO-G-hPRL. N-Glycan profiling proved to be a useful and accurate methodology also for MM and carbohydrate content determination for the two G-hPRL preparations, in good agreement with the values obtained directly via MALDI-TOF-MS.

Expression, purification, and characterization of authentic mouse prolactin obtained in Escherichia coli periplasmic space.

Prolactin (PRL) is a pleiotropic hormone produced by lactotroph cells of the anterior pituitary gland and is mainly related to lactation control and reproduction. Recombinant mouse prolactin (r-mPRL), never obtained in its authentic form, can be very useful for research and tests in animal models, in which human prolactin (hPRL) is usually employed in a heterologous mode. Synthesis of r-mPRL was carried out here via secretion in Escherichia coli periplasmic space using a plasmid containing mPRL cDNA joined to the DsbA signal peptide sequence under the control of a constitutive major leftward promoter of the bacteriophage λ (λPL). Fermentation in a pilot bioreactor was carried out at 30°C, with 6 H of induction at 37°C, reaching an optical density of 23 A600 units, a specific yield of 0.06-0.1 µg mPRL/(mL A600), and a concentration of up to 2.2 µg/mL. Even with such a low yield and a poor mass fraction, r-mPRL was purified via a three-step laboratory process based on hydrophobic chromatography, reversed-phase high-performance liquid chromatography, and high-performance size-exclusion chromatography (HPSEC). The purified hormone was then characterized using SDS-PAGE, Western blotting, and HPSEC and showed, by Nb2 rat lymphoma cell proliferation assay, a bioactivity of 39.5 IU/mg, determined against the International Standard of recombinant hPRL [World Health Organization (WHO)-97/714].

Improved bioprocess with CHO-hTSH cells on higher microcarrier concentration provides higher overall biomass and productivity for rhTSH.

Since the recombinant thyroid-stimulating hormone (rhTSH) is secreted by stably transfected Chinese hamster ovary (CHO-hTSH) cells, a bioprocess consisting of immobilizing the cells on a substrate allowing their multiplication is very suitable for rhTSH recovering from supernatants at relative high degree of purity. In addition, such a system has also the advantage of easily allowing delicate manipulations of culture medium replacement. In the present study, we show the development of a laboratory scale bioprocess protocol of CHO-hTSH cell cultures on cytodex microcarriers (MCs) in a 1 L bioreactor, for the preparation of rhTSH batches in view of structure/function studies. CHO-hTSH cells were cultivated on a fetal bovine serum supplemented medium during cell growth phase. For rhTSH synthesis phase, 75% of supernatant was replaced by animal protein-free medium every 24 h. Cell cultures were monitored for agitation (rpm), temperature (°C), dissolved oxygen (% DO), pH, cell concentration, MCs coverage, glucose consumption, lactate production, and rhTSH expression. The results indicate that the amount of MCs in the culture and the cell concentration at the beginning of rhTSH synthesis phase were crucial parameters for improving the final rhTSH production. By cultivating the CHO-hTSH cells with an initial cell seeding of four cells/MC on 4 g/L of MCs with a repeated fed batch mode of operation at 40 rpm, 37 °C, 20% DO, and pH 7.2 and starting the rhTSH synthesis phase with 3 × 10(6) cells/mL, we were able to supply the cultures with enough glucose, to maintain low levels of lactate, and to provide high percent (∼80%) of fully covered MCs for a long period (5 days) and attain a high cell concentration (∼9 × 10(5) cells/mL). The novelty of the present study is represented by the establishment of cell culture conditions allowing us to produce ∼1.6 mg/L of rhTSH in an already suitable degree of purity. Batches of produced rhTSH were purified and showed biological activity.

A molecular mimic of phosphorylated prolactin (S179D PRL) secreted by eukaryotic cells has a conformation with an increased positive surface charge compared to that of unmodified prolactin.

S179D prolactin (S179D PRL) is a pseudophosphorylated form of human PRL which has potent antitumor and anti-angiogenic activities in vivo. This molecule binds to the same forms of the PRL receptor (PRLR) as unmodified PRL, yet this binding results in different intracellular signaling and biological end points. Since it is now clear that PRLRs are predimerized and therefore that ligand binding must initiate signaling by inducing a conformational change in the receptor dimer, we hypothesized that S179D PRL had an altered conformation compared to unmodified PRL. The conformation of the ligand-receptor ternary complex would therefore also have an altered conformation, and thus, different signaling molecules would be activated. Here we present evidence in support of this hypothesis by demonstrating, in contrast to unmodified PRL, that S179D PRL has reduced nickel and zinc binding capacity and a higher affinity for heparin and DEAE. Conformational changes have occurred since these features are counterintuitive on the basis of the simple substitution of a serine with a negatively charged aspartate residue. To demonstrate that these particular properties of S179D PRL were not due to misfolding of the molecule during production, S179D PRL was expressed in two different mammalian cell lines. Also investigated was the potential for production of S179D PRL as a soluble cytoplasmic, or secreted periplasmic, protein in Escherichia coli.

High-level secretion of growth hormone by retrovirally transduced primary human keratinocytes: prospects for an animal model of cutaneous gene therapy.

A gene therapy clinical trial for treatment of growth hormone (GH) deficiency has not been reached yet, but several strategies using different gene transfer methodologies and animal models have been developed and showed successful results. We have set up an ex vivo gene therapy protocol using primary human keratinocytes transduced with an efficient retroviral vector (LXSN) encoding the human (hGH) or mouse GH (mGH) genes. These stably modified cells presented high in vitro expression levels of hGH (7 microg/106 cells/d) and mGH (11 microg/106 cells/d) after selection with geneticin. When the hGH-secreting keratinocytes were grafted onto immunodeficient dwarf mice (lit/scid), hGH levels in the circulation were about 0.2-0.3 ng/mL during a 12-d assay and these animals presented a significant body weight increase (p < 0.01) compared to the control. Substitution of conventional grafting methodologies with organotypic raft cultures revealed a peak value of up to 20 ng mGH/mL in the circulation of grafted lit/scid mice at 1 h postimplantation, followed by a rapid decline to baseline (approximately 2 ng/mL) within 24 h. One week after grafting, however, the cultured excised implants still presented approx 45% of their original in vitro secretion efficiency. Further studies are being carried out to identify the main factor(s) that still constitute one of the major impediments to the success of this promising model of cutaneous gene therapy.

Human macroprolactin displays low biological activity via its homologous receptor in a new sensitive bioassay.

CONTEXT: Macroprolactinemia is a frequent finding in hyperprolactinemic individuals, usually without clinical impact. Data on biological activity of macroprolactin (bbPRL) are controversial and mostly based on a heterologous rat Nb2 cell bioassay. Biological activity of bbPRL observed in vitro but not in vivo may be due to its high molecular weight, preventing its passage through capillary barrier. Alternatively, bbPRL bioactivity may differ depending on the prolactin (PRL) receptor (PRLR) species specificity. OBJECTIVE: The objective of the study was to characterize the bioactivity of bbPRL in a homologous bioassay: Ba/F-3 cells stably expressing the human PRLR. DESIGN/SETTING/PATIENTS: Chromatography-purified bbPRL from macroprolactinemic individuals (group I, n = 18) and monomeric PRL from hyperprolactinemic patients without macroprolactinemia (group II, n = 5) were tested in Nb2 and Ba/F-LLP bioassays. Both groups were followed up at the neuroendocrinology outpatients' clinic. MAIN OUTCOME MEASURE: Biological activity of bbPRL presented in the two bioassays was measured. RESULTS: In group I, no patient had hypogonadism. Mean ratio bioactivity to immunoactivity of bbPRL in the Nb2 assay was 0.69. There was no dose-response in 15 of the 18 samples tested in Ba/F-LLP assay. In group II, three patients had galactorrhea and all five had hypogonadism. Mean ratio bioactivity to immunoactivity of monomeric PRL samples was 1.35 in Nb2 and 0.91 in Ba/F-LLP assay. CONCLUSION: Whereas both bioassays achieve similar results with respect to monomeric PRL activity, our results indicate that the activity displayed by bbPRL toward the rat receptor may be inappropriate because it is not observed in the human PRLR-mediated assay, consistent with the apparent absence of bioactivity in vivo.

Periplasmic expression of human growth hormone via plasmid vectors containing the lambdaPL promoter: use of HPLC for product quantification.

The influence of different factors acting on Escherichia coli periplasmic expression of recombinant human growth hormone (hGH) in shake flask cultures has been investigated. Bacterial vectors containing the phage lambdaP(L) promoter, which is temperature activated, were utilized. Four different signal peptides were compared: DsbA, npr, STII and one derived from the natural hGH signal peptide, this last used as a reference. Other factors such as medium composition, optimized induction and expression conditions, and different bacterial strains were also studied. The determination of hGH, carried out directly in osmotic shock fluids, was based on an isocratic reversed-phase high-performance liquid chromatography method, which allows direct, rapid evaluation of the quality and quantity of hGH being secreted in the bacterial periplasmic space immediately after or even during fermentation. The level of hGH production increased 2.5-fold compared with the reference vector, reaching a level of 3.9 +/- 0.63 micro g/ml/A(600) (n = 6; coefficient of variation = 16.2%). The expression level was affected by the signal peptide and by the induction conditions, being more effective when activation started in the early logarithmic phase which, however, exhibited remarkably different optical density (OD) according to medium composition. Our results thus indicate that 6 h activation at 40-42 degrees C, starting with an OD (A(600)) of approximately 3 in a very rich medium, were conditions capable of providing the maximum secretion level for a vector utilizing the DsbA signal sequence and E.coli W3110 or RB791 as host cells.

Reversed-phase high-performance liquid chromatography method for the determination of prolactin in bacterial extracts and in its purified form.

Reversed-phase high-performance liquid chromatography methodology for the determination of human prolactin (hPRL) in bacterial periplasmic space or in purified preparations has been developed. The technique, based on the high hydrophobicity of the hPRL molecule, allows its separation from the bulk of bacterial proteins. The precision for periplasmic shock fluid analysis was characterized by relative standard variations of 3-7% for intra-day and of 3-25% for inter-day determinations. Accuracy, evaluated by recovery tests, was of the order of 90%, a calibration curve being constructed with the use of a lyophilized osmotic shock fluid extract, which provided a stable, readily prepared internal reference. Sensitivity was of the order of 0.5 microg of hPRL. The methodology developed also provided a tool for comparing the hydrophobicity of glycosylated and non-glycosylated prolactin molecules obtained from several different species and of different preparations of native or biosynthetic human prolactin.

High-level expression of human thyroid-stimulating hormone in Chinese hamster ovary cells by co-transfection of dicistronic expression vectors followed by a dual-marker amplification strategy.

The utilization of dicistronic mRNA expression vectors, containing the gene of interest upstream of an amplifiable marker gene, has shown success in rapidly, efficiently and reproducibly obtaining stable cell lines that express high levels of the protein of interest. For this reason, human thyroid-stimulating hormone (hTSH), a heterodimeric glycoprotein composed of non-covalently linked alpha- and beta-subunits, was expressed in Chinese hamster ovary (CHO) cells using a system based on dicistronic expression vectors. These contained the genes of interest and the amplifiable gene markers dihydrofolate reductase (DHFR) and adenosine deaminase (ADA), separated by an internal ribosome entry site isolated from the encephalomyocarditis virus. After the cells (CHO-DHFR-) had been co-transfected with the expression vectors and submitted to gene amplification in culture medium containing stepwise increments of methotrexate, it was possible to isolate clones that presented a secretion level of up to 7.2+/-1.3 microg/10(6) cells per day, the highest ever reported for the expression of this glycoprotein hormone. A second treatment, involving the utilization of deoxycoformycin, directed to amplify the ADA marker gene, provided a clone with an additional 2-3-fold increase in hTSH secretion, reaching a secretion level of 17.8+/-7.6 microg/10(6) cells per day. Cell culture and hTSH production in a hollow-fibre bioreactor were set up in order to carry out a preliminary physico-chemical, immunological and biological characterization of this hormone in comparison with pituitary-extracted hTSH (from the National Institute of Diabetes and Digestive and Kidney Diseases) and the only recombinant hTSH now available (Thyrogen). The availability of recombinant hTSH is very important in the diagnosis and therapy of thyroid carcinoma, via stimulation of radioiodine uptake.