Long-term human growth hormone expression and partial phenotypic correction by plasmid-based gene therapy in an animal model of isolated growth hormone deficiency

Sustained secretion of human growth hormone from TheraCyte devices encapsulated with PiggyBac-engineered retinal pigment epithelium cells

Growth hormone (GH) deficiency is characterized by impaired growth and development, and is currently treated by repeated administration of recombinant human GH (hGH). Encapsulated cell therapy (ECT) may offer a less demanding treatment-strategy for long-term production and release of GH into circulation. We used PiggyBac-based (PB) transposon delivery for engineering retinal pigment epithelial cells (ARPE-19), and tested a series of viral and non-viral promoters as well as codon-optimization to enhance transgene expression. Engineered cells were loaded into TheraCyte macrocapsules and secretion was followed in vitro and in vivo. The cytomegalovirus (CMV) promoter supports strong and persistent transgene expression, and we achieved clonal cell lines secreting over 6 µg hGH/10⁶ cells/day. Codon-optimization of the hGH gene did not improve secretion. ARPE-19 cells endured encapsulation in TheraCyte devices, and resulted in steady hormone release for at least 60 days in vitro. A short-term pilot experiment in immunodeficient SCID mice demonstrated low systemic levels of hGH from a single 40 µL capsule implanted subcutaneously. No significant increase in weight increase or systemic hGH was detected after 23 days in the GH-deficient lit/SCID mouse model using 4.5 µL capsules loaded with the highest secreting clone of ARPE-19 cells. Our results demonstrate that PB-mediated engineering of ARPE-19 is an efficient way to generate hormone secreting cell lines compatible with macroencapsulation, and our CMV-driven expression cassette allows for identification of clones with high level and long-term secretory activity without addition of insulator elements. Our results pave the way for further in vivo studies of encapsulated cell therapy.

Optimization of Mouse Growth Hormone Plasmid DNA Electrotransfer into Tibialis Cranialis Muscle of "Little" Mice

Previous non-viral gene therapy was directed towards two animal models of dwarfism: Immunodeficient (lit/scid) and immunocompetent (lit/lit) dwarf mice. The former, based on hGH DNA administration into muscle, performed better, while the latter, a homologous model based on mGH DNA, was less efficient, though recommended as useful for pre-clinical assays. We have now improved the growth parameters aiming at a complete recovery of the lit/lit phenotype. Electrotransfer was based on three pulses of 375 V/cm of 25 ms each, after mGH-DNA administration into two sites of each non-exposed tibialis cranialis muscle. A 36-day bioassay, performed using 60-day old lit/lit mice, provided the highest GH circulatory levels we have ever obtained for GH non-viral gene therapy: 14.7 ± 3.7 ng mGH/mL. These levels, at the end of the experiment, were 8.5 ± 2.3 ng/mL, i.e., significantly higher than those of the positive control (4.5 ± 1.5 ng/mL). The catch-up growth reached 40.9% for body weight, 38.2% for body length and 82.6%–76.9% for femur length. The catch-up in terms of the mIGF-1 levels remained low, increasing from the previous value of 5.9% to the actual 8.5%. Although a complete phenotypic recovery was not obtained, it should be possible starting with much younger animals and/or increasing the number of injection sites.

Partial correction of the dwarf phenotype by non-viral transfer of the growth hormone gene in mice: Treatment age is critical

Non-viral transfer of the growth hormone gene to different muscles of immunodeficient dwarf (lit/scid) mice is under study with the objective of improving phenotypic correction via this particular gene therapy approach. Plasmid DNA was administered into the exposed quadriceps or non-exposed tibialis cranialis muscle of lit/scid mice followed by electroporation, monitoring several growth parameters. In a 6-month bioassay, 50μg DNA were injected three times into the quadriceps muscle of 80-day old mice. A 50% weight increase, with a catch-up growth of 21%, together with a 16% increase for nose-to-tail and tail lengths (catch-up=19-21%) and a 24-28% increase for femur length (catch-up=53-60%), were obtained. mIGF1 serum levels were ~7-fold higher than the basal levels for untreated mice, but still ~2-fold lower than in non-dwarf scid mice. Since treatment age was found to be particularly important in a second bioassay utilizing 40-day old mice, these pubertal mice were compared in a third bioassay with adult (80-day old) mice, all treated twice with 50μg DNA injected into each tibialis cranialis muscle, via a less invasive approach. mIGF1 concentrations at the same level as co-aged scid mice were obtained 15days after administration in pubertal mice. Catch-up growth, based on femur length (77%), nose-to-tail (36%) and tail length (39%) increases was 40 to 95% higher than those obtained upon treating adult mice. These data pave the way for the development of more effective pre-clinical assays in pubertal dwarf mice for the treatment of GH deficiency via plasmid-DNA muscular administration.

Inhibitory effect of memantine, an NMDA-receptor antagonist, on electroporation-induced inward currents in pituitary GH3 cells

The membrane electroporation-induced inward current (IMEP) in pituitary tumor (GH3) cells was characterized. This current emerges irregularly when membrane hyperpolarizations to -200 mV with a holding potential of -80 mV were elicited. Neither E-4031 (10 μM), glibenclamide (30 μM), nor ZD7288 (30 μM) caused any effects on IMEP. The single-channel conductance and pore radius were estimated to be around 1.12 nS and 1.7 nm, respectively. LaCl3- and memantidine (MEM)-induced block of this current was also examined. The IC50 value for LaCl3- and MEM-induced inhibition of IMEP was 35 and 75 μM, respectively. However, unlike LaCl3, MEM (300 μM) did not exert any effect on voltage-gated Ca2+ current. In inside-out configuration, MEM applied to either external or internal surface of the excised patch did not suppress the activity of ATP-sensitive K+ channels expressed in GH3 cells, although glibenclamide significantly suppressed channel activity. This study provides the first evidence to show that MEM, a non-competitive antagonist of N-methyl D-aspartate receptors, directly inhibits the amplitude of IMEP in pituitary GH3 cells. MEM-mediated block of IMEP in these cells is unlinked to its inhibition of glutamate-induced currents or ATP-sensitive K+ currents. The channel-suppressing properties of MEM might contribute to the underlying mechanisms by which it and its structurally related compounds affect neuronal or neuroendocrine function.