Abstract
Despite important advances in human therapeutics, no specific treatment for both non-functioning gonadotroph and resistant somatotroph adenomas is available. Gene transfer by viral vectors can be considered as a promising way to achieve a specific and efficient treatment. Here we show the possibility of efficient gene transfer in human pituitary adenoma cells in vitro using a human immunodeficiency virus (HIV)-type 1-derived vector. Using enhanced green fluorescent protein (eGFP) gene as a marker placed under the phosphoglycerate kinase (PGK) promoter, gonadotroph and somatotroph adenomas were transduced even with moderate viral loads. The expression started at day 2, reached a peak at day 5, and it was still present at day 90. For targeting somatotroph and gonadotroph adenomas, human growth hormone (GH) promoter (GH −481, +54 bp) and two fragments of the human glycoprotein hormone α-subunit promoter (α-subunit 1 −520, +33 bp, and α-subunit 2 −907, +33 bp) were tested. In gonadotroph adenomas, the percentage of identified fluorescent cells and the fluorescence intensity analyzed by fluorescence-activated cell sorting indicated that the strength of the α-subunit 1 and α-subunit 2 promoters were comparable to that of the PGK promoter. Primary cultures of rat pituitary cells showed that α-subunit 1 is more selective to thyreotroph and gonadotroph phenotypes than α-subunit 2. GH promoter activity appeared weak in somatotroph adenomas. The human GH enhancer did not increase the GH promoter activity at all but the human prolactin promoter (−250 bp) allowed 4-fold more fluorescent cells to be obtained than the GH promoter. Several cell lines appeared too permissive to test cell-specificity of pituitary promoters. However, on human non-pituitary cell cultures, the tested pituitary promoters seemed clearly selective to target endocrine pituitary phenotypes. This study gives a starting point for a gene-therapy program using lentiviral vectors to transfer therapeutic genes in human pituitary adenomas.
Introduction
Pituitary adenomas account for 10–15% of all intracranial tumors. Ninety percent are represented by somatotroph adenomas (growth hormone (GH) secretors), clinically non-functioning adenomas (mainly gonadotroph) and prolactinomas (prolactin (PRL) secretors). Although often considered as benign, they can induce serious neurological and metabolic complications. Despite important advances in transphenoidal surgery, radiotherapy and receptor-mediated pharmacology, there is no specific therapy for non-functioning tumors, nor for some somatotroph adenomas resistant to somatostatin agonists, the main therapeutic agents nowadays available (Freda 2002).
Gene transfer by viral vectors can be considered as a promising way to achieve a specific, efficient treatment. The human pituitary tumors have a slow growth rate in vivo and are practically quiescent in vitro (Atkin et al. 1997). Adenoviral vectors could be therefore considered as an appropriate choice for genetically modifying pituitary cells, given their high infecting capacity in non-dividing cells (Castro et al. 1997, Lee et al. 1999, Southgate et al. 2000, 2001). However, serious concerns have been raised about the safety of using the first generation of adenoviral vectors because of their capacity to elicit severe inflammatory responses. In fact, pituitary cellular infiltration has been reported after regional in vivo stereotaxic pituitary injection of adenovirus in sheep (Davis et al. 2002), and, contrary to what has been observed after the infection of brain parenchyma (Kajiwara et al. 2000), circulating anti-adenovirus-neutralizing antibodies have been detected consistently in rat receiving such injections. Furthermore, a significant decrease of transgene expression was seen after 3 months (Southgate et al. 2001). Newer ‘high-capacity adenoviral vectors’ could open new perspectives founded on both long-term transgene expression and absence of inflammatory reactions (Thomas et al. 2000).
Lentiviral vectors have recently been used to transduce a variety of low-proliferating or quiescent cells, such as differentiated neurons (Naldini et al. 1996a, b). They have enough cloning capacity to allow more complex expression than vectors derived from the murine leukemia virus (Kafri et al. 2000), and transduction in vivo of a variety of tissues, such as brain, liver, skeletal muscle or pancreas, has shown a sustained expression lasting for 6 months (Blomer et al. 1997). These qualities render lentiviral vectors potentially useful for the treatment of human pituitary adenomas.
Beside direct therapeutic effects, a major aim of gene therapy is to prevent damage of non-target tissues. Taking advantage of the fact that hormonal specificity is preserved in adenomas, these tumors represent an interesting model in the search of targeted expression of genes using cell-specific promoters. Several promoters of human pituitary hormones driving marker or toxic genes in adenovirus have been tested on animals (Lee et al. 1999, 2002, Southgate et al. 2000, 2001, Davis et al. 2001). These experiments have shown a restricted expression of reporter gene under the control of pituitary promoters, though they have not been applied to human pituitary adenomatous cells; only one study used human prolactinoma cells in vitro to transfer a gene by an adenoviral vector (Freese et al. 1996).
Human pituitary adenomas are difficult to obtain. Fortunately, the cells can be put into in vitro conditions mimicking some aspects of the natural disease, provided various stringent requirements be fulfilled to preserve for several weeks the molecular and pharmacological properties of tumoral cells (Jaquet et al. 1985). Using this in vitro model, we present for the first time a lentiviral vector derived from human immunodeficiency virus type-1 (HIV-1) as an effective vehicle to genetically manipulate human pituitary adenoma cells. The strength and the specificity of human promoters of both α-subunit of glycoprotein hormones (α-subunit) and GH were evaluated on gonadotroph and somatotroph tumors as well as in other cell populations. Considering that the utilization of strong promoters can reduce both the viral doses employed and side effects, an increment of GH-promoter strength was attempted. Considering that the majority of somatotroph adenomas are in fact somatolactotroph tumors, a human GH-promoter enhancer (Jin et al. 1999) was used and compared with the PRL promoter.
Materials and Methods
Construction of lentiviral vector plasmids expressing enhanced green fluorescent protein (eGFP) under the control of human specific pituitary promoters
The lentiviral vector pRRLT-PGK-eGFPsin18 (Naldini et al. 1996a, b) encodes eGFP placed under control of phosphoglycerate kinase (PGK), an ubiquitous gene promoter. In the present study, pRRLT-PGK-eGFPsin18 containing a central DNA flap sequence (Zennou et al. 2000) is called lenti-PGK-eGFP. From this plasmid, six different lentiviral vectors containing human pituitary promoters at the place of PGK promoter sequence were generated (Fig. 1). The PGK promoter was removed using BamHI/BstXI digestion. Human promoters of the glyco-protein hormone α-subunit (α-subunit 1 −520, +33 bp, and α-subunit 2 −907, +33 bp) and of GH (−481, +54 bp) were obtained after PCR amplification from human genomic DNA, using the following primers. For α-subunit 1: 5′-CAACTTTGCGTTCTTTGG-3′ and 5′-CTTCGTCTTATGAGTTCTCAGTAAC-3′; for α-subunit 2: 5′-CCTTCTTAGAATTGCCTCATACCT-3′ and 5′-CTTCGTCTTATGAGTTCTCAGTAAC-3′; for GH: 5′-AATCATGCCCAGAACCCCCGCAATC-3′ and 5′-TAGTGAGCTGTCGACAGGACC-3′. These fragments were subcloned into the PCRscript Amp SK(+) plasmid at an SrfI site (Stratagene, Amsterdam, The Netherlands). The fragments of α-subunit 1 and α-subunit 2 were excised from the PCRscript Amp SK(+) plasmid using NotI and EcoRI enzymes and ligated to RRLT-X-eGFPsin18 to generate the new vectors, RRLT-α-subunit 1-eGFPsin18 and RRLT-α-subunit 2-eGFPsin18, named lenti-α-subunit 1-eGFP and lenti-α-subunit 2-eGFP, respectively. A NotI-blunted XhoI fragment was generated to recover the GH promoter from PCRscript Amp SK(+) and inserted instead of PGK into the lenti-PGK-eGFP vector plasmid at a BamHI-blunted XhoI site. This new RRLT-hGH-eGFPsin18 vector was named lenti-GH-eGFP. The plasmid containing a GH enhancer was a gift from Dr P Cattini (University of Manitoba, Canada). The 1.6 G fragment was excised using EcoRV and ClaI enzymes and inserted upstream of the GH promoter. The vector was named lenti-1.6 GH-eGFP. A region of 327 bp at the 3′ end of the 1.6 G fragment was obtained by PCR amplification from the 1.6 G plasmid using the primers 5′-ACACTAGCCCCAAAGTTAATGAAG-3′ and 5′-CTTAACCCTCCTGCCCCTGACTT-3′. The vector containing this portion upstream of the GH promoter was named lenti-enhGH-eGFP. The human PRL promoter (−250 bp) was a gift from Dr J Martial (Université de Liège, Liège, Belgium; Quentien et al. 2002). The fragment was excised using HindIII and BglII enzymes and used instead of the PGK promoter to generate the RRLT-PRL-eGFPsin18 vector, named lenti-PRL-eGFP. Each lentiviral plasmid was verified by DNA sequencing (Ceq 8000; Beckman Coulter, Roissy, France).
Production of lentiviral vectors
Lentiviral particles were generated by using HEK 293T cells submitted to a triple polyethylenimine-mediated cotransfection with each lentiviral vector plasmid, conjointly with the vesicular stomatitis virus-G-encoding plasmid pMDG and the packaging plasmid pCMVΔR8.91 (Naldini et al. 1996a, 1996b). The supernatants were collected twice at 1-day intervals, treated with DNase I (37 °C for 15 min), concentrated by ultracentrifugation, aliquoted and stored at −80 °C. All the supernatants were free of replication-competent virus as determinated by the replication-competent retrovirus (RCR) assay. HeLaP4 cells which stably contain the lacZ gene under the control of the tat-dependent HIV promoter were transduced and cultivated for 2 weeks, trypsinization being done every 3–4 days to favour the free expansion of cells. After infection with replicating HIV, HeLaP4 cells constantly show a clear-cut expression of β-galactosidase (β-gal) and they form syncitia. In our material, the consistent absence of β-gal expression and the absence of syncitia indicated the absence of HIV-like replication-competent viruses.
Titration of lentiviral vector supernatants
Titration of lenti-PGK-eGFP, lenti-α-subunit 1-eGFP and lenti-α-subunit 2-eGFP vector supernatants was performed on HeLaT cells. In this cell line, both α-subunit promoters as well as the PGK promoter were able to induce eGFP transcription. No eGFP expression under GH and PRL promoters was found in HeLaT cells. Consequently, titration of lenti-GH-eGFP, lenti-enhGH-eGFP, lenti-1.6 GH-eGFP and lenti-PRL-eGFP vector supernatants was performed on GH4C1 cells (a rat mammosomatotroph cell line) following the same procedure employed on HeLaT cells. Briefly, 1 × 105 cells were plated in 12-well dishes; 24 h later, cells were transduced overnight in 500 μl appropriate medium (Gerolami et al. 2000, Le Pechon-Vallée et al. 2000) containing 8 μg/ml polybrene and 0.1–1 μl concentrated viral supernatant: next day, the transduction medium was removed and cells were grown in 10% appropriate medium for 3 days. The percentage of transduced cells was determined by fluorescence-activated cell sorting (FACS), and the viral titer was calculated as previously described (Limon et al. 1997). Titers around 1 × 108 transducing units per ml were obtained. The titers of lenti-PGK-eGFP supernatants were similar when evaluated on HeLaT or GH4C1 cells and so the titrations obtained from these two cell lines could be safely compared.
Cell cultures
Human pituitary adenomas Fragments of human pituitary adenomas were obtained from patients submitted to trans-sphenoidal surgery. Gonadotroph and somatotroph tumors were selected according to the clinical hormonal status and immunocytochemical data (Table 1; Barlier et al. 1998). The present study was approved by the Ethics Commitee of the University of Aix-Marseille (France) and was undertaken after informed consent of patients and participants.
No later than 2 h after surgery, the fragments were mechanically and enzymatically dissociated with collagenase and trypsin (Jaquet et al. 1985); 2 × 105 cells were plated on 12-well dishes coated with extracellular matrix of bovine corneal epithelial cells (Jaquet et al. 1985). The culture medium consisted of Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal calf serum (FCS), penicillin (100 U/ml), streptomycin (100 U/ml) and fungizone (0.25 mg/ml). Culture chambers were maintained at 37 °C in an atmosphere containing 7% CO2.
Cell lines Various cell lines were tested to check the strength and specificity of pituitary promoters. MCF7 (a human breast cancer cell line), U118 (a human glioblastoma cell line), αT3 (a mouse gonadotroph cell line; Windle et al. 1990) and LβT2 (a mouse gonadotroph cell line; Turgeon et al. 1996) were cultured in DMEM supplemented with 10% FCS and penicillin/streptomycin at 37 °C in humidified air atmosphere containing 5% CO2. GH4C1 and GC (a rat somatotroph cell line) cells were cultured in Ham’s F-10 supplemented with 2.5% FCS and 7.5% horse serum with penicillin/streptomycin at 37 °C in humidified atmosphere containing 7% CO2 (Le Pechon-Vallée et al. 2000).
Human non-pituitary primary cultures Human thyroid tissue were obtained from surgical fragments of patients bearing Grave’s disease. The fragments were collected in Coon’s modification of Ham’s F-12 medium (Eurobio, Les Ulis, France) and were dissociated mechanically and enzymatically (Rasmussen et al. 1996). Cells were washed and plated in Coon’s modification of Ham’s F-12 medium supplemented with 5% FCS enriched with bovine thyroid-stimulating hormone (TSH; 1 IU/l), human insulin (10 mg/l), somatostatin (10 μg/l), human transferrin (6 mg/l), hydrocortisone (10−8 M) and glycylhistidyl-acetate (10 μg/l), at 37 °C, in a humidified atmosphere containing 5% CO2 (Rasmussen et al. 1996). Human lymphocytes were cultured in RPMI medium supplemented with 10% FCS and penicillin/streptomycin at 37 °C in humidified atmosphere containing 7% CO2. Cells from primary culture of human glioma were cultured in DMEM/Ham’s F-12 supplemented with 10% FCS and penicillin/streptomycin, at 37 °C in a humidified atmosphere containing 7% CO2. Finally, primary cultures of stroma cells from human adenomas were tested. These cells were obtained from primary cultures of human pituitary adenomas cultured for 3 months, which allowed stroma cells to proliferate as adenomatous cells progressively died. After this period, only stroma cells, mainly fibroblasts, were present, according to both immunocytochemical identification and the absence of detectable hormone in the medium. These cells were collected in new plates and cultured in DMEM supplemented with 10% FCS and penicillin/streptomycin at 37 °C in a humidified atmosphere containing 7% CO2.
Rat anterior pituitary primary culture As previously described (Moreau et al. 1997), 50 female Wistar rats (201–225 g; Charles River, Iffa Credo, L’Arbresle, France) were decapitated and the pituitaries rapidly dissected out, cut into small pieces and enzymatically dispersed in trypsin-DMEM at 37 °C. DNase (1 mg/ml) was added for 1 min. Trypsin activity was stopped and a Ca2+/Mg2+-free medium containing EDTA was used to facilitate cell dissociation. Cells were suspended and centrifuged for 10 min at 300 g. After counting, they were seeded over glass coverslips in 12-well dishes at 2 × 105 cells/well.
Transduction of cells
After 5 days of culture, pituitary adenoma cells were transduced using various multiplicities of infection (MOIs) of each lentiviral vector in the presence of 6 μg/ml polybrene and appropriate medium to get a final volume of 500 μl. Cells were incubated overnight in this medium. The following day, considered as day 1 post-transduction, the medium was removed and replaced with 1 ml DMEM supplemented with 10% FCS. The expression of eGFP was followed up by fluorescence microscopy and/or FACS. For the long-term studies, the culture medium was renewed every week and cells were trypsinized and plated onto new coated wells to prevent fibroblastic invasion.
Cell lines, human non-pituitary cells and human pituitary-derived fibroblasts were plated at 105 cells/well in 12-well dishes. Then, 24 h later, cells were infected with each lentiviral construct in the presence of polybrene at 8 μg/ml in 500 μl appropriate medium. At day 1 post-transduction, the medium was changed. The fluorescence was analyzed by FACS 3 days later.
Primary cultures of rat pituitary were transduced following the same procedure. Briefly, after 48 h of culture, dishes were infected overnight with a 5-MOI load of viral particles conveying either the PGK, the α-subunit 1 or the α-subunit 2 promoter. At day 5 the cultures were fixed in 10% buffered nascent formaldehyde, cryo-protected in 2 M buffered sucrose and stored at −80 °C until analysis.
Cytochemical identification of cells
Cytospinning of human pituitary adenoma cells In order to get human adenoma cells without fragments of culture matrix and fibroblasts, a gentle enzymatic dissociation was performed for 1 min with 100 μl trypsin-EDTA (Invitrogen). Detached cells were resuspended in 100 μl of appropriate medium. Each aliquot was dropped down into a disposable cytofunnel (Shandon, Pittsburgh, PA, USA) and centrifuged against a gelatin-coated cytoslide (Shandon, Eragny, France) at 1000 r.p.m. for 5 min in a Cytospin 2 apparatus (Shandon, Cheshire, UK). After centrifugation the slides were immersed for 1 h in 10% buffered nascent formaldehyde, pH 7.4. Fixed cells were either conserved at −80 °C in 2 M buffered sucrose or mounted immediately in Mowiol for microscopic observation and cell counting.
Immunocytochemistry Cultured tumoral cells and fibroblasts were fixed for 30 min in buffered nascent formaldehyde; residual free aldehyde groups were quenched by buffered 50 mM ammonium chloride; unspecific epitopes were neutralized by BSA. After cell permeabilization with 0.2% Triton-X100 (Sigma-Aldrich) the cells were incubated overnight (at 4 ×C) with 1:100 dilutions of either a monoclonal anti-human α-subunit of pituitary glycoprotein hormones antibody (Dako, Trappes, France), a polyclonal anti-human chromogranin A antibody (Immunotech Beckman Coulter, Roissy, France) or a monoclonal anti-human fibroblast antibody (Cymbus Biotechnology, Hants, UK). Specificity was controlled by substituting the main antibody with buffered BSA. A second incubation followed for 2 h at 4 °C with either FITC-conjugated F(ab′)2 donkey-anti-mouse IgG fragments (Jackson ImmunoResearch Laboratories, West Grove, PA, USA) for detecting the α-subunit, with Texas Red-conjugated similar donkey anti-rabbit fragments for detecting chromogranin A, or with Texas Red-conjugated donkey anti-mouse fragments for the identification of fibroblasts.
Immunofluorescence labelling was also performed to identify the hormone phenotype of infected rat adenohypophyseal cells (Moreau et al. 1997). Thawed cells were refixed in 10% buffered nascent formaldehyde, quenched, permeabilized and neutralized. Primary antibody incubation was done overnight at 4 °C with polyclonal antibodies against rat PRL, luteinizing hormone, GH and TSH elaborated by Dr A.F. Parlow (Harbor-UCLA Medical Center, Torrance, CA, USA) and with a polyclonal anti-rat adrenocorticotrophin (ACTH) 1–24 antibody elaborated by Dr M. Grino (Institut Jean Roche, Faculty of Medicine, Marseille, France). TRITC (tetramethylrhodamine β-isothiocyanate)-conjugated F(ab′)2 anti-rabbit IgG fragments were used as secondary layers to detect PRL, β-TSH and ACTH; GH and β-luteinizing hormone were detected with Protein A coupled to Texas Red (Jackson ImmunoResearch Laboratories).
Confocal microscopy Intracellular localization of eGFP was visualized 3 and 5 days after infection. Fixed and permeabilized cells were counterstained on slides with 2 μg/ml propidium iodide (Molecular Probes P-3566) in Mowiol mounting medium; this procedure gives a relatively low background fluorescence from RNA and counterstains cell nuclei in optimal contrast conditions with respect to eGFP. Specimens were observed on a Leica TCS laser-scanning confocal device coupled to a DMR inverted microscope equipped with planachromatic 20/0.4 and 63/1.4 objectives (Leica, Heidelberg, Germany). The argon/krypton laser source generated spectral lines at 488 and 568 nm; signals were recovered through bandpass filters centered at 520 nm (green emission) and 600 nm (red emission). The Leica Scanware algorithm was used to store optical sections. Infected adenoma cells labelled for human chromogranin A were also observed under confocal microscopy.
Estimation of eGFP-expressing cells
FACS To follow up eGFP expression, FACS analysis was done. Cells were trypsinized and resuspended in 3 μl of iodure propidium (50 μg/ml; BD Biosciences, Pharmingen, San Diego, CA, USA) in 500 of appropriate medium and run on a FACSsort (Becton Dickinson, San Jose, CA, USA). Data were analyzed with the CellQuest program (Becton Dickinson). 10 000 events were acquired for each analysis.
Quantitative microscopy For nine non-functioning gonadotroph adenomas and for primary cultures of rat pituitary, quantitative microscopy of fluorescent cells was performed using a Nikon TE-300 inverted microscope supplied with a Plan Fluor ELWD 20×/0.45 PH1 DM objective and a Hamamatsu C4742–95 digital camera monitored with the Twain Two driver. For each slide, 10–102 microscopic fields (430 μm long×345 μm wide) were digitized according to a systematic random sampling procedure (Gundersen & Jensen 1987). For human adenomas, each field was recorded twice: once under phase contrast to count the total number of cells per field, and once under epifluorescence (barrier filter centered at 480 nm) for identifying and counting eGFP-emitting cells. The capture time was constantly fixed at 111 ms, the time at which any fluorescent cell was neither visualized nor detected by the camera in non-infected cells; the weakest eGFP emission was 37 times more brilliant than the background noise. The total number of cells counted per experimental condition ranged from 521 to 4136. For cultured rat pituitary cells, each field was captured three times: once under phase contrast, once under epifluorescence with barrier filter centered at 480 nm (eGFP cells), and once with a barrier filter centered at 560 nm to detect hormone-labelled cells. The total number of cells counted per hormone phenotype ranged from 2460 to 7341. All cell counts were performed directly by one single investigator, without any image-processing algorithm. Doubtful decisions were cleared by overlaying fluorescence and phase-contrast images with Photoshop Software (Adobe System, Mountain View, CA, USA).
Statistical analysis
For quantitative microscopy, the statistical analysis was done taking into consideration the first 500 cells systematically random-sampled and computed for each experimental condition. Statistical significance was determined by one-factor ANOVA followed by Dunnett’s test. For FACS quantitation, significance was determined by non-parametric Wilcoxon test or Mann–Whitney test. Data were expressed as the mean±s.e.m.
Results
Characterization of cells from human pituitary adenomas in culture
Human pituitary adenomas contain non-dividing tumoral cells proper and stroma cells, mainly fibroblasts, actively proliferating in vitro. Considering that the primary culture of adenoma cells could be invaded by fibroblasts, we first characterized the cells actually present after several days of culture using cytospinning and immunocytochemistry. Cells collected from the culture after gentle trypsinization and cytospun were arranged mainly into clusters composed of a few cells to several tens of cells, displaying two different morphologies (Fig. 2A). The great majority consisted of small, circular or elliptical cells, about 12–15 μm in diameter, positively labelled with antibodies directed against chromogranin A and/or α-subunit of pituitary glycoprotein hormones; these cells were considered as adenomatous cells proper (Fig. 2B). Conjointly, some clusters included a few much larger and polymorphic cells (25–40 μm) positively labelled with the anti-human fibroblast antibody (Fig. 2A) or with the anti-α-subunit antibody; the first group was considered to be composed of fibroblasts and the second of large adenoma cells. Cell counts showed that fibroblasts did not exceed 0.3% of the total number of cells at day 20; large adenoma cells did not exceed 0.8%.
eGFP expression driven by PGK promoter in transduced adenomatous cells
To evaluate transduction effciency only the small eGFP-positive cells were quantified (Fig. 2B). After transduction, variable eGFP fluorescence intensity from one adenoma to another was observed. Gonadotroph cells expressed eGFP as a green cytoplasmic and nuclear fluorescence of variable intensity (Fig. 3A and B). Both cytoplasmic and nuclear fluorescences consistently increased with the concentration of lentiviral particules. At day 5, the mean percentage of eGFP-fluorescent cells was 6.1% at 1 MOI. The percentage of eGFP-positive cells rose abruptly to 43.4% at 2.5 MOI, attaining 67.8% after a viral load of 100 MOI (Fig. 2C–H and Fig. 4). The augmentation of the nuclear fluorescence followed a more regular course: 0.6, 21, 30, 59 and 74% of cells with an eGFP-emitting nucleus were found at 1, 2.5, 10, 50 and 100 MOI, respectively (Fig. 4).
When the lentiviral load was kept at 10 MOI and the post-infection culture time was prolonged, eGFP expression was first detected at day 2 (Fig. 2I and J). The outburst of expression happened between days 2 and 5 (Fig. 2K and L, and Fig. 5) and was still present at day 20 (Fig. 2M and N). Fibroblast-free adenoma cells derived from two clinically non-functioning tumors (NF8, NF10) showed a strong eGFP fluorescence (similar to Fig. 2N) 90 days after a 10 MOI infection.
For comparison and to extend results from gonadotroph adenomas to other pituitary tumors, eGFP-expressing cells were evaluated by the FACS method, after gentle trypsinization as previously used for quantitative microscopy, in three somatotroph adenomas (A1, A2, A3) and three gonadotroph adenomas (NF8, NF9, NF10) transduced with 10 MOI. Some 77, 90 and 97% of cells of the three somatotroph tumors, and 46, 86, 67% of cells of three gonadotroph adenomas expressed eGFP, results similar to those obtained by quantitative microscopy. These results allow us to validate FACS as a procedure to safely quantify the transduction of human adenoma cells.
eGFP expression driven by α-subunit 1, α-subunit 2 and GH promoters in human gonadotroph and somatotroph adenomas
The α-subunit 1 promoter was found to be a strong promoter in gonadotroph adenomas. In fact, the mean percentages of fluorescent cells were not significantly different from values obtained after infection with lenti-PGK-eGFP. Transduction effciency significantly decreased when the α-subunit 2 promoter was used (Fig. 6). Driven by the GH promoter, eGFP was detected in only 5.7% of gonadotroph cells. In somatotroph adenomas, the mean percentage of fluorescent cells was not significantly different under PGK, α-subunit 1 or α-subunit 2 promoters and reached only 31% under the GH promoter.
As for the PGK promoter, with α-subunit 1 and α-subunit 2 promoters a great variability in fluorescence intensity was observed from one tumor to another (Table 2). In all cases, fluorescence intensity under the GH promoter was significantly lower than that observed under the PGK promoter. In human somatotroph cells, it represented only 4% of the mean fluorescence intensity under the PGK promoter. Under the α-subunit 2 promoter, both types of tumor showed a lower, but not significant, fluorescence than that observed under the α-subunit 1 promoter.
In one gonadotroph (NF10) and one somatotroph (A2) adenoma, 5 days after transduction, the percentage of eGFP-positive cells remained similar under doses of lenti-α-subunit 1-eGFP and lenti-α-subunit 2-eGFP at 5 and 10 MOI (Table 3). In the same tumors, increasing doses of lenti-GH-eGFP did not improve the percentage of fluorescent cells. The time-course variation of eGFP expression driven by the α-subunit 1 and α-subunit 2 promoters in two gonadotroph adenomas (NF11, NF12) showed comparable fluorescent cell percentages (about 50%) and brilliances at days 8 and 20 (about 250 arbitary units).
Cell type-specific eGFP expression driven by α-subunit 1, α-subunit 2 and GH promoters in pituitary and non-pituitary cell lines
To test the cell specificity of α-subunit 1, α-subunit 2 and GH promoters, eGFP expression was assessed in gonadotroph (αT3, LβT2), somatotroph (GH4C1, GC) and non-pituitary (MCF7, U118) cell lines, and compared with the eGFP expression driven by the ubiquitous PGK promoter. Considering the mean percentages of fluorescent cells (Fig. 7A), α-subunit 1 and α-subunit 2 promoters showed a similar activity to the PGK promoter in gonadotroph cell lines and, surprisingly, in somatotroph cell lines. These percentages were significantly lower in non-pituitary cell lines (MCF7, U118) but it is noteworthy that they remained high: more than 50% of cells expressed eGFP under the α-subunit 1 and α-subunit 2 promoters. The mean percentages of fluorescent cells under the GH promoter were similar to that observed under the PGK promoter in somatotroph cell lines and was signficantly lower in gonadotroph cell lines and in non-pituitary cell lines.
The results obtained from the mean fluorescence intensities were similar (Fig. 7B). The activity of gonadotroph promoters was the same as that of the PGK promoter in gonadotroph cell lines and was only a little lower than what was observed with the PGK promoter in other cell lines, except in U118 in which both lenti-α-subunit 1-eGFP and lenti-α-subunit 2-eGFP exhibited only 7% of the activity detected for lenti-PGK-eGFP. The activity of the GH promoter was found constantly to be lower than that of the PGK promoter even in somatotroph cell lines, in which the intensities under the GH promoter represented only 14 and 24% of the intensities observed for PGK in GH4C1 and GC cells.
Cell-type-specific eGFP expression driven by α-subunit 1, α-subunit 2 and GH promoters in human non-pituitary cells and human pituitary-derived fibroblasts
To test the cellular specificity of pituitary promoters, four human primary cultures were used: fibroblasts, lymphocytes, glioma cells and thyrocytes. The experiments were run under the same conditions as for cell lines. In all cultures, the mean percentages of fluorescent cells as well as the mean fluorescence intensities dramatically decreased under pituitary promoters compared with the PGK promoter (Fig. 8). The mean percentages of fluorescent cells under pituitary promoters did not rise beyond 6%, except in thyrocytes where it attained 21% with both the α-subunit 1 and α-subunit 2 promoters; however, the corresponding fluorescent intensities in this type of cells were low, representing only 10% of the intensity measured after the PGK promoter. In fibro-blasts, under increasing loads from 5 to 50 MOI of vectors containing eGFP driven by the pituitary promoters, the number of fluorescent cells as well as the fluorescence intensities remained unchanged (Fig. 9). Similar results were obtained in the rest of human non-pituitary cultures.
Cell-type-specific eGFP expression driven by α-subunit 1, α-subunit 2 and GH promoters in rat pituitary primary culture
To test the cellular specificity of pituitary promoters with respect to pituitary endocrine cells, the eGFP expression was assessed on rat pituitary primary culture transduced by lenti-α-subunit 1-eGFP, lenti-α-subunit 2-eGFP and lenti-GH-eGFP and compared with the transduction obtained by lenti-PGK-eGFP. Under the PGK promoter, all endocrine cells of the anterior pituitary were able to be transduced. Under the α-subunit 1 promoter, the maximal percentage of fluorescent cells was observed in thyreotroph and gonadotroph cells. Thyreotroph cells attained 86% and gonadotroph cells 43% of PGK-fluorescent cells. In other cell types, the percentages of α-subunit 1-fluorescent cells represented only 8% of PGK-fluorescent PRL cells, 5% of PGK-fluorescent GH cells and 5% of PGK-fluorescent ACTH cells. The α-subunit 2 promoter appared less efficient than α-subunit 1 in both thyreotroph and gonadotroph phenotypes. Indeed, the percentages of fluorescent cells under α-subunit 2 were higher than those under α-subunit 1 in PRL, GH and ACTH cells: they represented 29, 33 and 52% of PGK-fluorescent cells, respectively (Fig. 10). No fluorescent cells were seen under the GH promoter.
Improving somatotroph cell targeting
To increase the percentage of fluorescent cells after transduction of human somatotroph cells by lentiviral vector containing pituitary promoter, the 1.6 G fragment from the locus-control region in the human GH gene, known to be a distal enhancer, and a shorter fragment, enhGH containing Pit-1 sequences (Jin et al. 1999), were put upstream of the GH promoter. Any significant increase of the percentage of fluorescent cells was observed in GC and GH4C1 cell lines as well as in somatotroph adenoma cells. With the majority of human somatotroph adenomas being somatolactotroph adenoma, the PRL promoter, inserted instead of the GH promoter, was evaluated in the same conditions. In these cells, the mean percentages of fluorescent cells increased by a factor of 4 (Fig. 11); a 2-fold mean fluorescence intensity was observed under the PRL promoter in comparison to the GH promoter (data not shown). These results were obtained on two adenomas (A5, A6) considered as ‘pure’ somatotroph adenomas since they contained only 10% of PRL cells, according to data from the Surgical Pathology Department of Timone Hospital, Marseille, France (D Figarella-Branger, unpublished data). The PRL promoter appeared highly specific: the mean percentages of fluorescent cells (Fig. 11) as well as the mean fluorescence intensities (data not shown) dramatically decreased under the PRL promoter compared with the PGK promoter in human non-pituitary cells, human pituitary-derived fibroblasts and non-pituitary cell lines (U118 and MCF7) as well as in one human gonadotroph adenoma cell line (NF13). The mean fluorescence intensities represented only 4% of the mean intensities observed for PGK in human gonadotroph adenoma cells.
Discussion
For the first time, we present the genotypical modification of human pituitary adenoma cells in vitro using an HIV-derived lentiviral vector as a gene-transfer tool. Considering that the primary culture of human pituitary adenomas shows consistent contamination by cells derived from the stroma, the first step was to determine the phenotype of cell populations actually infected by the lentiviral vector. The proliferative potential of adenoma cells is very weak: less than 5% of cells at 4 days of primary culture (Atkin et al. 1997). On the contrary, the fibroblasts show a much higher proliferative rate. The immunocytochemical identification of cells showed that the percentage of fibroblasts at the beginning of the culture was lower than 1% of the total cell population, rendering these cells not relevant for the estimation of eGFP-labelled cells. In fact, when cells are collected from culture boxes, the great majority of fibroblasts remains attached to the extracellular matrix coating the wells. Thus, microscopy allowed us to determine that cells subsequently detected by FACS were clearly representive of the adenoma cell population.
After using eGFP expression as a marker gene under the control of the PGK promoter to assess gene-transfer efficiency, both gonadotroph and somatotroph adenomas were found efficiently transduced even with moderate viral loads: up to 97% of cells were able to express eGFP 5 days after infection. The percentage of eGFP-positive cells increased drastically between days 2 and 5 and was still present 90 days after infection. This sudden increment cannot be attributed to proliferation of adenoma cells (Atkin et al. 1997) and/or infected fibroblasts. Nevertheless, it is well known that two mechanisms can artificially increase eGFP: first, pseudotransduction, i.e. the direct transfer of marker gene protein by its presence in vector supernatants or its incidental incorporation into vector particules (Liu et al. 1996), and secondly, the transient expression of the marker gene inside a target cell prior to the vector’s integration into the specific target cell’s genome (Haas et al. 2000). In the present study, two facts contribute to excluding biased eGFP expression: 1, the fluorescence at day 1 post-infection was absent in both transduced and non-transduced cells; this observation eliminates pseudotransduction; 2, the time course of events showed that the increment of adenomatous cells expressing eGFP was quite in advance of any increment produced by cell proliferation; this elimates the second bias.
If the safety of neighbouring tissues (optic nerve or hypothalamus) is considered, targeting expression of the transgene is the major challenge in gene therapy of pituitary adenomas. Tissue specificity could be attempted by in situ intratumoral injection of the viral vector bearing tissue-specific promoters during transphenoidal microsurgery as suggested by some experiments on rats (Lee et al. 2000). For the first time, in order to target gonadotroph adenomas (for which no specific therapy exists) and somatotroph adenomas resistant to somatostatin agonists, the α-subunit of glycoprotein hormone promoter and the GH promoter were tested. The fragments of the promoters were chosen according to the functional sequences involved in both the expression of the gene and the cell specificity.
After transduction by adenovirus, a similar fragment to α-subunit 2 ( −846 to 45 bp) has been shown to restrict to the αT3 gonadotroph cell line the expression of the β-gal marker gene or herpes simplex virus thymidine kinase toxic gene (Lee et al. 1999, 2000). However, a shorter fragment (−500 bp) seemed to be sufficient for full expression of the human α-subunit gene in pituitary cell lines: the basal expression was not affected by deletion from 1800 to 442 bp upstream of the transcription start site of the gene (Horn et al. 1992), but expression was reduced with loss of promoter sequences between −442 and −391 bp (Schoderbek et al. 1992). This part of the promoter contains a pituitary glycoprotein hormone basal element (PGBE) which is able to function as a thyrotroph/gonadotroph-specific enhancer (Aylwin & Burin 1995). Consequently, two fragments of human α-subunit promoter were tested, α-subunit 2 and the more restricted fragment α-subunit 1 (–520, +33 bp). We showed that both fragments induced eGFP expression in mouse αt3 and LβT2 gonadotroph cell lines as well as in human gonadotroph adenomatous in vitro even if they did not secrete α-subunit in vivo (cf. NF6). In fact, in nonfunctioning tumors, no correlation was seen between the fluorescence intensity and the in vivo hormonal status. In contrast in somatotroph tumors, one low-secreting tumor (A2) was associated with the weakest activities of the promoters. But our series was too short to be conclusive.
Using −336, +58 bp from the human GH 5′ flanking region inserted in adenovirus, Lee et al.(1999) showed restricted expression of a marker gene (β-gal) and of a toxic gene (thymidine kinase) in a somatotroph GH3 cell line. Some 95–100% of GH3 cells were β-gal-positive, 4–5 days after infection by adenoviruses containing this fragment of GH promoter. Based on this study and on other experiments in animals (Cattini et al. 1986), we selected a similar portion of the GH promoter (−481, +54 bp) to obtain a specific expression. After transduction of two somatotroph cell lines (GH4C1 and GC) by the lentiviral vector driven by this part of the GH promoter, up to 98% of cells expressed eGFP. However, contrary to what is observed on rat somatotroph cell lines, only 31% of cells expressed eGFP in human somatotroph adenoma cells in culture. In vitro, these tumors still remain GH-secreting and keep their responsiveness to pharmacological treatments (Jaquet et al. 1985). Whatever the model used, the strength of GH promoter was weak in our study, which is in agreement with the results of authors measuring β-gal activity (Lee et al. 1999). In that study, the β-gal activity in GH3 cells infected by adenovirus containing GH promoter represented only 16.2% of the β-gal activity under cytomegalovirus ubiquitous promoter.
To assess cell-specificity of these pituitary promoters, we used different cell lines. Clearly, the activity of the promoter was not restricted to those cell lines producing the hormone corresponding to the tested promoters. Even in non-pituitary cell lines, the pituitary promoter remained active; the fluorescence intensity under pituitary promoters was only low in the U118 glioma cell line. The overlapping of GH and α-subunit promoter activities in pituitary cell lines was not surprising if we consider that these pituitary cells derive from a common progenitor, opening the possibility that they share common transcription factors. A similar overlapping was also observed in the study by Lee et al.(1999). What is more surprising is the activity of pituitary hormone promoters in non-pituitary cell lines. It might reflect the incompleteness of differentiation of these cells. From a practical point of view and according to our results, cell lines seem too permissive to be used as cell-specificity testers of pituitary promoters and probably of other promoters as well. Consequently, the cell specificity of the three pituitary promoters might be tested on primary cultured human cells. As we have shown, the activity of the three promoters decreased dramatically in these cells, compared with the activity of PGK promoter, confirming the selectivity of the pituitary promoters. In view of gene therapy of pituitary adenomas using stereotaxic injection, it is noteworthy that the activities of α-subunit 1, α-subunit 2 and GH promoters were almost absent in fibroblasts coming from human pituitary adenomas and in gliomas, even at heavy doses of lentiviral vectors. This study clearly shows the limits of animal cellular models and pituitary cell lines. In a gene-therapy program it is crucial to have the possibility of testing lentiviral constructs on human tumoral cells in vitro.
Somatotroph adenomas showed a strong activity of α-subunit 1 and α-subunit 2 promoters. It is noteworthy that 36% of somatotroph adenomas secreted both α-subunit hormone and GH (Kontogeorgos et al. 1993). Correspondingly, the GH promoter activity appeared low in gonadotroph adenomas, probably due to the promoter specificity as well as to its relative intrinsic weakness. eGFP expression under the GH promoter was too low at 10 MOI to be observed by epifluorescence in cells of human pituitary adenomas as well as of rat pituitary. The results on rat pituitary cells and on human pituitary adenomas indicate α-subunit 1 promoter to be more powerful and more restricted to gonadotroph and thyreotroph phenotypes than the α-subunit 2 promoter. Specific elements have been clearly identified in the portion of the α-subunit promoter before −500 bp upstream, but the cis elements are not well established at present. Whatever it may be, our results suggest the presence of sequences limiting cell-type-specificity between −500 and −846 bp. Even if a residual activity of the α-subunit promoter remained in GH, PRL or ACTH cells, it should not be a contraindication to attempt gene therapy targeted by this promoter, considering that the majority of patients deserving this type of therapy have already developed large, invasive tumors and several pituitary deficiencies.
For improving human somatotroph cell targeting, several strategies were attempted. A 1.6 G fragment, from the locus-control region in the human GH gene and located about 15 kb upstream of the GH transcription initiation site, was identified as enhancer of the GH promoter (Jin et al. 1999). The enhancer activity of this fragment could be localized in a 260 bp region in the 3′ end of the 1.6 G which contains Pit-1-binding sites. In transfected GH3 cell line, the 1.6 G fragment induced a 5-fold increment of the human GH promoter activity (Jin et al. 1999). However, in our experiments and using lentiviral vectors as vehicles, the 1.6 G fragment as well as the 260 bp region were unable to increase the activity, whatever the model used.
In vivo, 80% of somatotroph adenomas are GH- and PRL-secreting. According to the immunocytochemical data, almost all somatotroph adenomas contain somatolactotroph and lactotroph cells (Melmed et al. 1995) During pituitary ontogenesis, somatotroph and lactotroph cells derive from a common progenitor (Dasen and Rosenfeld 2001). Consequently, the use of PRL promoter to target human somatotroph cells was attempted. Even if HIV-based vectors have a greater packaging limit than oncoretroviral vectors, the titer decreased in a semi-logarithmic manner as insert size and proviral length increased (Kumar et al. 1999). Accordingly, the first −250 bp of the PRL promoter, which contains Pit-1-binding sites, was chosen (Lemaigre et al. 1991) instead of the −4429 bp PRL promoter fragment previously tested by Davis et al.(2001). In human somatotroph adenomas transduced with lentiviral vectors the mean percentage of fluorescent cells under the PRL promoter quadrupled in comparison to the mean percentage under the GH promoter, even in somatotroph adenomas considered as ‘pure’ by immunocytochemistry. Moreover, this short fragment of PRL promoter appeared very specific to PRL and GH phenotypes as well as the −4429 bp PRL fragment in adenoviral vector on ovine pituitary (Davis et al. 2001).
In conclusion, HIV-derived lentiviral vectors are promising tools to achieve gene transfer in human pituitary adenoma cells. According to our results, an effective transduction was observed up to 3 months. The percentage of cell fluorescence reached a plateau around 5 MOI and at day 5, whichever the promoter used. The results showed that the GH promoter seems specific but poorly active in human adenoma somatotroph cells. The PRL promoter appeared more adequate to target human somatotroph cells. α-Subunit 1 and α-subunit 2 promoters are powerful and specific gonadotroph promoters, adequately targeting long-term gonadotroph cells from nonfunctioning adenomas. Results on rat pituitary showed α-subunit 1 promoter as the most appropriate choice. Considering that long-term transgene expression can be obtained using lentiviral vectors, that receptor-mediated pharmacology can actually be efficient enough for controling the behavior of adenomas expressing receptors, and that pituitary promoters are able to target selectively pituitary cells, these tumors represent good candidates for receptor-dependent gene therapy.
Human non-functioning (NF) and somatotroph (A) pituitary adenomas
| Age (years) | Gender | Immuno-cytochemistry | MRI | PRL (μg/l)* | GH (μg/l)* | α subunit (UI/l)* | FSH (UI/l)* | LH (UI/l)* | TSH (UI/l)* | |
|---|---|---|---|---|---|---|---|---|---|---|
| All pituitary hormones, including α-subunit, were tested. Only positive results were indicated ++≥50% positive cells; +<50% positive cells. Data from the Department of Surgical Pathology, Timone Hospital, Marseilles. | ||||||||||
| *In vivo secretion, normal ranges: PRL: females 1–24 μg/l; males 1–17 μg/l; GH: 0.2–2.4 μg/l; α-subunit: menopausal women <1.6 UI/l; women, follicular phase <0.9 UI/l; men <0.9 UI/l; FSH: menopausal women 19–130 UI/l; males 1–9 UI/l; women, follicular phase 3–8 UI/l; LH: menopausal women 12–58 UI/l; males 1–5 UI/l; women follicular phase 1–7 UI/l; TSH: 0.1–4 UI/l. | ||||||||||
| NS, non-secreting adenoma; ND, not determined; MRI, Magnetic resonance imaging; marco, macroadenoma; SSE, suprasellar extension. | ||||||||||
| Patient | ||||||||||
| NF1 | 61 | M | βFSH++ | macro, SSE | 22 | <0.1 | 0.6 | 3 | 0.5 | 3.5 |
| NF2 | 51 | M | βFSH++; βLH+ | macro, SSE | 2.4 | ND | <0.25 | 2.4 | 1.6 | 0.13 |
| NF3 | 78 | M | βFSH++; βLH+ | macro | 45 | 0.2 | 2.4 | 20 | 7 | 6.6 |
| NF4 | 55 | F | NS | macro | 45 | 0.2 | 0.3 | 14 | 2 | 4 |
| NF5 | 57 | M | βFSH++ | macro | 6 | ND | 0.2 | 10.3 | 0.7 | 1.5 |
| NF6 | 29 | F | βFSH | macro | 49 | 1.3 | <0.25 | 6 | 1.8 | 2 |
| NF7 | 50 | M | NS | macro | 6 | ND | 0.3 | 4.6 | 2.3 | 0.7 |
| NF8 | 37 | M | βFSH++ | macro | 30 | ND | 0.25 | 5.3 | 1.5 | 2.3 |
| NF9 | 73 | M | βFSH+; βLH+ | macro | 5 | ND | <0.25 | 4.1 | 1.7 | 1.8 |
| NF10 | 48 | M | αsu+; βFSH++ | macro, SSE | 10 | ND | 1.2 | 23 | 1.4 | 2.9 |
| NF11 | 51 | F | βFSH+; βLH+ | macro, SSE | 16 | ND | 0.5 | 13.8 | 52 | 3.5 |
| NF12 | 70 | F | αsu+; βFSH+ | macro, SSE | 55 | ND | 0.5 | 7.7 | 1.2 | 5.8 |
| NF13 | 59 | F | αsu+; βFSH+ | macro, SSE | 2.5 | ND | 0.6 | 11 | 5.8 | 3.9 |
| A1 | 35 | M | GH++; PRL+ | macro | 7 | 1600 | ND | ND | ND | ND |
| A2 | 22 | F | GH++ | macro, SSE | 39 | 22 | ND | 4.9 | 3.3 | 1.89 |
| A3 | 18 | F | GH++ | macro | 19 | 207 | ND | 3.3 | <1 | 1 |
| A4 | 30 | F | GH+; PRL+ | macro | 25 | 185 | ND | ND | ND | ND |
| A5 | 25 | F | GH+ | macro | 35 | 5 | ND | ND | ND | ND |
| A6 | 32 | F | GH+ | macro | 20 | 16 | ND | ND | ND | ND |
Fluorescent intensity assessed by FACS, 5 days after infection by lentiPGK, α-subunit 1, α-subunit 2 and GH-eGFP in 4 gonadotroph and 3 somatotroph adenomas
| Fluorescent intensities | ||||
|---|---|---|---|---|
| PGK | α-subunit 1 | α-subunit 2 | GH | |
| * statistically significant P<0.05, with respect to PGK and α-subunit 1 promoters. | ||||
| Adenomas | ||||
| non functioning | ||||
| NF6 | 656 | 1836 | 373 | 15 |
| NF8 | 882 | 822 | 244 | |
| NF9 | 532 | 1166 | 361 | |
| NF10 | 184 | 147 | 218 | 24 |
| mean±s.e.m. | 563 ± 145 | 992 ± 351 | 299 ± 39 | 68 ± 48* |
| somatotroph | ||||
| A1 | 1677 | 154 | 66 | 30 |
| A2 | 402 | 73 | 50 | 15 |
| A3 | 1483 | 422 | 256 | 129 |
| mean±s.e.m. | 1187 ± 396 | 216 ± 100 | 124 ± 66 | 58 ± 35* |
Percentage of fluorescent cells assessed by FACS 5 days after infection by lentiPGK, α-subunit 1, α-subunit 2 and GH-eGFP in one gonadotroph adenoma (NF10) and one somatotroph adenoma (A3)
| PGK | α-subunit 1 | α-subunit 2 | GH | |||||
|---|---|---|---|---|---|---|---|---|
| MOI | 5 | 10 | 5 | 10 | 5 | 10 | 5 | 10 |
| NF10 | 61 | 66 | 48 | 50 | 41 | 40 | 9 | 8 |
| A3 | 92 | 98 | 92 | 98 | 73 | 88 | 10 | 9 |
Schematic drawing of the plasmids used to construct a lentivirus vector with specific promoter (Naldini et al. 1996a, b).
Citation: Journal of Endocrinology 183, 1; 10.1677/joe.1.05759
Cytospun cells from cultured human non-functioning pituitary adenomas. (A) Overlaid images of NF1 gonadotroph adenoma after epifluorescence (antibody against human fibroblast) and phase-contrast microscopy showing a typical cluster containing predominantly small, ovoidal cells and several larger cells; one large cell is positively labeled (red arrow, fibroblast) but another group of similar cells is not (black arrow). (B) Epifluorescence image after double labeling with antibodies against human α-subunit of pituitary glycoprotein hormones (green emission) and human chromogranin A (red emission) in NF1 gonadotroph adenoma; some cells are predominantly α-subunit-positive (green arrow) but the majority are double labelled, emitting in the orange-to-yellow spectrum (yellow arrow). Calibration bar for (A) and (B), 20 μm. (C)–(N) Infected NF7 gonadotroph adenoma with lenti-PGK-eGFP. Cell countings were done on systematically random sampled microscopic fields exemplified in these series. Fields were digitized under phase-contrast microscopy (C, E, G and I, K, M) and on epifluorescence (corresponding D, F, H and J, L, N). At day 5 after infection, the increment of eGFP-fluorescent cells with the lentiviral loading is clear from 2.5 (C, D), to 10 MOI (E, F) up to 100 MOI (G, H), where the majority of cells are fluorescent. When the viral load was kept at 10 MOI, the number of eGFP-fluorescent cells increased with time: at day 1 after transduction (I, J), practically no fluorescent cells were found; there were eGFP-emission outbursts at day 5 (K, L), which persisted with comparable intensity at day 20 (M, N). Calibration bar for the series in (D), 100 μm.
Citation: Journal of Endocrinology 183, 1; 10.1677/joe.1.05759
Laser confocal microscopy sections showing the intracellular distribution of eGFP in cultured, infected NF2 human non-functioning pituitary adenoma. (A) Simultaneous detection of nuclear DNA after propidium iodide staining (red emission) in clusters of adenoma cells displaying variable degrees of eGFP expression (green emission), at day 3 after infection. Some cells have been mainly labelled in the nucleus (red arrows); however, in a large majority the green fluorescent signal of eGFP appear orange or yellow when overlapping the red nuclear counterstaining (orange arrows). (B) High magnification of a large infected adenoma cell at day 5, labeled for human chromogranin A (red emission); eGFP (green emission) occupies the whole nuclear area (N) and co-localizes with chromogranin A in an extensive reticular structure of the cytoplasm (orange arrow). Calibration bars, 10 μm.
Citation: Journal of Endocrinology 183, 1; 10.1677/joe.1.05759
Transduction efficiency of lenti-PGK-eGFP vectors in nine human non-functioning adenomas (NF1–NF9) 5 days after infection. After cytospinning, direct counting of 500 cells included in systematically random-sampled fields was done by phase-contrast microscopy and epifluorescence. White columns represent the percentage of eGFP-emitting cells in the 500-cell population. Points represent the percentage of eGFP-emitting cell nuclei in the same population (means±s.e.m.). Asterisks on the white columns denote a statistically significant difference with respect to 1 MOI (*P<0.05; **P<0.01). Asterisks on the points indicate statistically significant difference with respect to the 2.5 MOI (*P<0.05). No significant difference was detected between 50 and 100 MOI.
Citation: Journal of Endocrinology 183, 1; 10.1677/joe.1.05759
Time course of eGFP-expressing cells in three non-functioning adenomas after transduction by lenti-PGK-eGFP. Two adenomas (NF8 and NF10) were infected with 10 MOI; another (NF7) received a 50 MOI load. eGFP-expressing cells were counted as for Fig. 1, and expressed as percentage of the total counted cells.
Citation: Journal of Endocrinology 183, 1; 10.1677/joe.1.05759
Cell-type-specific expression of lenti-PGK, α-subunit 1, α-subunit 2 and GH bearing the eGFP gene in human gonadotroph and somatotroph adenomas. Four gonadotroph (NF6, NF8, NF9, NF10) and three somatotroph adenomas (A1, A2, A3) were infected with 5 MOI. Percentages of fluorescent cells assessed by FACS, 5 days after infection. Bars represent the means±s.e.m. Asterisks denote statistically significant differences with respect to PGK and α-subunit 1 promoters (*P<0.05).
Citation: Journal of Endocrinology 183, 1; 10.1677/joe.1.05759
Cell-type-specific expression of lenti-PGK, α-subunit 1, α-subunit 2 and GH bearing the eGFP gene in cell lines. Gonadotroph cell lines of mouse (αT3 and LβT2), somatotroph cell lines of rat (GH4C1 and GC) and human non-pituitary cell lines (MCF7 and U118) were infected with 5 MOI of each lentiviral vector. Percentages of fluorescent cells (A) and fluorescence intensity (B) assessed by FACS, 3 days after infection. Each experiment was performed in triplicate. Bars represent means±S.e.m. Asterisks denote statistically significant differences with respect to the PGK promoter (*P<0.05).
Citation: Journal of Endocrinology 183, 1; 10.1677/joe.1.05759
Cell-type-specific expression of lenti-PGK, α-subunit 1, α-subunit 2 and GH bearing the eGFP gene in human non-endocrine primary culture. Human primary cultures of fibroblasts cultured from a human pituitary adenoma, lymphocytes, glioma cells and thyrocytes were infected with 5 MOI of each lentiviral vector. Percentages of fluorescent cells (A) and fluorescence intensity (B) assessed by FACS, 3 days after infection. Each experiment was performed in triplicate. Bars represent means±s.e.m. The mean percentages and the fluorescence intensities were significantly lower (P<0.05) under pituitary promoters than under the PGK promoter in all categories of cells.
Citation: Journal of Endocrinology 183, 1; 10.1677/joe.1.05759
Human pituitary fibroblasts infected with increasing doses of lenti-PGK, α-subunit 1, α-subunit 2 and GH bearing the eGFP gene. Human primary cultures of fibroblasts were infected with increasing doses (5, 10, 20 and 50 MOI) of each lentiviral vector. Percentages of fluorescent cells (A) and fluorescence intensity (B) assessed by FACS, 3 days after infection.
Citation: Journal of Endocrinology 183, 1; 10.1677/joe.1.05759
Cell-type-specific expression of lenti-PGK, α-subunit 1 and α-subunit 2 bearing the eGFP gene in rat anterior pituitary primary culture. Five days after infection with 5 MOI of each lentiviral vector, cells from primary culture of rat pituitary were analyzed immunocytochemically. eGFP expression was assessed by quantitative microscopy and was expressed as the percentage of green fluorescent cells in the hormone-specific cell population (red fluorescent cells). Each experiment was performed in triplicate. Bars represent means±s.e.m. Asterisks denote statistically significant differences with respect to the PGK promoter (*P<0.05).
Citation: Journal of Endocrinology 183, 1; 10.1677/joe.1.05759
Transduction by lenti-enhGH, 1.6 GH and PRL bearing the eGFP gene in comparison with lenti-PGK-eGFP and lenti-GH-eGFP. (A) The cells were infected with 5 MOI of each lentiviral vector. Percentages of fluorescent cells assessed by FACS, 3 days after infection of GH4C1 and GC cell lines and 5 days after infection of somatotroph adenoma cells (A4, A5, A6). Asterisks denote statistically significant differences with respect to the GH promoter (*P<0.05). (B) Different cell types were infected with 5 MOI of lenti-PRL-eGFP and lenti-PGK-eGFP vectors. Percentages of fluorescent cells assessed by FACS 3 days after infection of human non-pituitary cell lines (U118 and MCF7), human pituitary fibroblasts and human lymphocytes and 5 days after infection of gonadotroph adenoma cells (NF13). The mean percentages were significantly lower under the PRL promoter than the PGK promoter in all categories of cells. Each experiment was performed in triplicate. Bars represent means±s.e.m.
Citation: Journal of Endocrinology 183, 1; 10.1677/joe.1.05759
The vector pRRL-PGK-eGFP was a gift from Dr D Trono, University of Geneva, Geneva, Switzerland. We thank Dr P L Mellon (University of California at San Diego, CA, USA), for kindly providing αT3 and LαT2 cell lines, Dr A Denizot (Service de Chirurgie Endocrinienne, CHU Nord, Marseille, France) for human thyroid fragments and Dr L. Ouafik (Laboratoire de Cancérologie Expérimentale, Marseille, France) for human glioma cells. We thank also Professor Peter Cattini (Manitoba, Canada) for the human GH enhancer plasmid, Dr Jean Steinberg for help in statistical analysis and Miss Henriette Gérard for supplying cytospin facilities.
Funding
This work was supported by the Association pour la Recherche sur le Cancer 2001 and the Gene Vector Production Network sponsored by the Association Française contre les myopathies.
References
Atkin SL, Burnett HE, Landolt AM, Green VL, Hipkin LJ & White MC 1997 Identification of proliferating human anterior pituitary adenoma cells in vitro. Neurological Research 19 420–425.
Aylwin SJB & Burin JM 1995 The role of transcription factors in the pituitary expression of the glycoprotein hormone α-subunit gene. Journal of Molecular Endocrinology 15 221–231.
Barlier A, Gunz G, Zamora AJ, Morange-Ramos I, Figarella-Branger D, Dufour H, Enjalbert A & Jaquet P 1998 Pronostic and therapeutic consequences of Gsα mutations in somatotroph adenomas. Journal of Clinical Endocrinology 83 1604–1610.
Blomer U, Naldini L, Kafri T, Trono D, Verma IM & Gage FH 1997 Highly efficient and sustained gene transfer in adult neurons with a lentivirus vector. Journal of Virology 71 6641–6649.
Castro MG, Goya RG, Sosa YE, Rowe J, Larregina A, Morelli A & Lowenstein PR 1997 Expression of transgenes in normal and neoplastic anterior pituitary cells using recombinant adenoviruses: long term expression, cell cycle dependency, and effects on hormone secretion. Endocrinology 138 2184–2194.
Cattini PA, Peritz LN, Anderson TR, Baxter JD & Eberhardt NL 1986 The 5′ flanking sequences of the human growth hormone gene contain a cell specific control element. DNA 5 503–509.
Dasen JS & Rosenfeld MG 2001 Signaling and transcriptional mechanisms in pituitary development. Annual Review of Neurosciences 24 327–355.
Davis JRE, McVerry J, Lincoln GA, Windeatt S, Lowenstein PR, Castro MG & McNeilly AS 2001 Cell type-specific adenoviral transgene expression in the intact ovine pituitary gland after stereotaxic delivery: an in vivo system for long-term multiple parameter evaluation of human pituitary gene therapy. Endocrinology 142 795–801.
Davis JRE, McMahon RFT, Lowenstein PR, Castro MG, Lincoln GA & McNeilly AS 2002 Adenovirus-mediated gene transfer in the ovine pituitary gland is associated with hypophysitis. Journal of Endocrinology 173 265–271.
Freda PU 2002 Somatostatin analogs in acromegaly. Journal of Clinical Endocrinology and Metabolism 87 3013–3018.
Freese A, During MJ, Davidson BL, Gennarelli TA, Kaplitt MG, Flamm ES & Snyder PJ 1996 Transfection of human lactotroph adenoma cells with an adenovirus vector expressing tyrosine hydroxylase decreases prolactin release. Journal of Clinical Endocrinology and Metabolism 81 2401–2404.
Gerolami R, Uch R, Jordier F, Chapel S, Bagnis C, Bréchot C & Mannoni P 2000 Gene transfer to hepatocellular carcinoma: transduction efficacy and transgene expression kinetics by using retroviral and lentiviral vectors. Cancer Gene Therapy 7 1286–1292.
Gundersen HJ & Jensen EB 1987 The efficiency of systematic sampling in stereology and its prediction. Journal of Microscopy 147 229–263.
Haas DL, Case SS, Crooks GM & Kohn DB 2000 Critical factors influencing stable transduction of human CD34+ cells with HIV-1-derived lentiviral vectors. Molecular Therapy 2 71–80.
Horn F, Windle JJ, Barnhart KM & Mellon PL 1992 Tissue-specific gene expression in the pituitary: the glycoprotein hormone alpha-subunit gene is regulated by a gonadotrope-specific protein. Molecular Cell Biology 12 2143–2153.
Jaquet P, Gunz G & Grisoli F 1985 Hormonal regulation of prolactin release by human prolactinoma cells cultured in serum free conditions. Hormone Research 22 153–163.
Jin Y, Surabhi RM, Fresnoza A, Lytras A & Cattini PA 1999 A role for A/T-Rich sequences and Pit-GHF-1 in a distal enhancer located in the human growth hormone locus region with preferential pituitary activity in culture and transgenic mice. Molecular Endocrinology 13 1249–1266.
Kafri T, Van Praag H, Gage FH & Verma IM 2000 Lentiviral vectors: regulated gene expression. Molecular Therapeutics 1 516–521.
Kajiwara K, Byrnes AP, Ohmoto Y, Charlton HM, Wood MJA & Wood KJ 2000 Humoral immune responses to adenovirus vectors in the brain. Journal of Neuroimmunology 103 8–15.
Kontogeorgos G, Asa SL, Kovacs K, Smyth HS & Singer W 1993 Production of alpha-subunit of glycoprotein hormones by pituitary somatotroph adenomas in vitro. Acta Endocrinologica 129 565–572.
Kumar M, Keller B, Makalou N & Sutton RE 2001 Systematic detremination of the packaging limit of lentiviral vectors. Human Gene Therapy 12 1893–1905.
Lee EJ & Jameson L 2002 Cell-specific CRE-mediated activation of the diphtheria toxin gene in pituitary tumor cells: potential for cytotoxic gene therapy. Human Gene Therapy 13 533–542.
Lee EJ, Anderson LM, Thimmapaya B, & Jameson JL 1999 Targeted expression of toxic genes directed by pituitary hormone promoters: a potential strategy for adenovirus-mediated gene therapy of pituitary tumors. Journal of Clinical Endocrinology and Metabolism 84 786–794.
Lee EJ, Thimmapaya B & Jameson JL 2000 Stereotactic injection of adenoviral vectors that target gene expression to specific pituitary cell types: implications for gene therapy. Neurosurgery 46 1461–1469.
Lemaigre FP, Peers B, Lafontaine DA, Mathy-Hartet M, Rousseau GG, Belayew A & Martial JA 1991 Transcriptional induction of human prolactin gene by cAMP requires two cis-acting elements and at least the pituitary specific factor Pit-1. Journal of Biological Chemistry 266 18127–18134.
Le Pechon-Vallée C, Magalon K, Rasolonjanahary R, Enjalbert A & Gérard C 2000 Vasoactive intestinal polypeptide and pituitary adenylate-activating polypeptide activate mitogen-activated protein kinase in the pituitary cell line GH4C1 by a 3′,5′-cyclic adenosine monophosphate pathway. Neuroendocrinology 72 46–56.
Limon A, Briones J, Puig T, Carmona M, Fornas O, Cancelas JA, Nadal M, Garcia J, Rueda F & Barquinero J 1997 High-titer retroviral vectors containing the enhanced green fluorescent protein gene for efficient expression in hematopoietic cells. Blood 90 3316–3321.
Liu M-L, Winther BL & Kay MA 1996 Pseudotransduction of hepatocytes by using concentrated pseudotyped vesicular stomatis virus G Glycoprotein (VSV-G)-moloney murine leukemia virus-derived retrovirus vectors: comparison of VSV-G and amphotropic vectors for hepatic gene transfer. Journal of Virology 70 2497–2502.
Melmed S, Ho K, Klibanski A, Reichlin S & Thorner M 1995 Recent advances in pathogenesis, diagnosis, and management of acromegaly. Journal of Clinical Endocrinology and Metabolism 80 3395–3402.
Moreau C, Rasolojanahary R, Zamora AJ, Enjalbert A, Kordon C & Llorens-Cortes C 1997 Expression of angiotensin II receptor subtypes AT1A and AT1B in enriched fractions of dispersed rat pituitary cells. Neuroendocrinology 66 416–425.
Naldini L, Blömer U, Gallay P, Ory D, Mulligan R, Gage FH, Verma IM & Trono D 1996a In vivo gene delivery and stable transduction of non-dividing cells by a lentiviral vector. Science 272 263–267.
Naldini L, Blömer U, Gallay P, Ory D, Mulligan R, Gage FH, Verma IM & Trono D 1996b Efficient transfer, integration, and sustained long-term expression of the transgene in adult rat brains injected with a lentiviral vector. PNAS 93 11382–11388.
Quentien MH, Pitoia F, Gunz G, Guillet MP, Enjalbert A & Pellegrini I 2002 Regulation of prolactin, GH, and Pit-1 gene expression in anterior pituitary by Pitx2: an approach using Pitx2 mutants. Endocrinology 143 2839–2851.
Rasmussen AK, Kayser L, Perrild H, Brandt M, Bech K & Feldt-Rasmussen U 1996 Human thyroid epithelial cells cultured in monolayers. Decreased thyroglobulin and cAMP response to TSH in 12-week-old secondary and tertiary cultures. Molecular and Cellular Endocrinology 116 165–172.
Saveanu A, Lavaque E, Gunz G, Barlier A, Taylor JE, Culler MD, Enjalbert A & Jaquet P 2002 Demonstration of enhanced potency of a chimeric somatostatin-dopamine molecule, BIM-23A387, in suppressing growth hormone and prolactin secretion from human pituitary somatotroph adenoma cells. Journal of Clinical Endocrinology and Metabolism 87 5545–5552.
Schoderbek WE, Kim KE, Ridgway EC, Mellon PL & Maurer RA 1992 Analysis of DNA sequences required for pituitary-specific expression of the glycoprotein hormone alpha-subunit gene. Molecular Endocrinology 6 893–903.
Southgate TD, Windeatt S, Smith-Erica J, Gerdes CA, Perone MJ, Morris I, Davis JRE, Klatzmann D, Lowenstein PR & Castro MG 2000 Transcriptional targeting to anterior pituitary lactotrophic cells using recombinant adenovirus vectors in vitro and in vivo in normal and estrogen/sulpiride-induced hyperplasic anterior pituitaries. Endocrinology 141 3493–3505.
Southgate TD, Stone D, Williams JC, Lowenstein PR & Castro MG 2001 Long-term transgene expression within the anterior pituitary gland in situ: impact on circulating hormone levels, cellular and antibody-mediated immune responses. Endocrinology 142 464–476.
Thomas CE, Schiedner G, Kochnek S, Castro MG & Löwenstein PR 2000 Peripheral infection with adenovirus causes unexpected long term brain inflammation in animals injected intracranially with first-generation, but not with high capacity, adenovirus vectors: toward realistic long term neurological gene therapy for chronic diseases. PNAS 97 7482–7487.
Turgeon JL, Kimura Y, Waring DW & Mellon PL 1996 Steroid and pulsatile gonadotropin-releasing hormone (GnRH) regulation of luteinizing hormone and GnRH receptor in novel gonadotrope cell line. Molecular Endocrinology 10 439–450.
Windle JJ, Weiner RI & Mellon PL 1990 Cell lines of the pituitary gonadotrope lineage derived by targeted oncogenesis in transgenic mice. Molecular Endocrinology 4 597–603.
Zennou V, Petit C, Guetard D, Nerhass U, Montagnier L & Charneau P 2000 HIV-1 genome nuclear import is mediated by a central DNA flap. Cell 85 173–185.
