Cellular effects of pulsatile GH release

Patterned GH release is observed in mammals, and continuously released GH has been shown to have little effect on growth, or on total hepatic IGF-1 levels when administered

below a threshold value (Gevers et a l, 1996; Wells et a l, 1994). At the cellular level the

GH signalling pathway which activates STAT5b is required for the pattern of dimorphic

growth seen in rodents (Bohl et al, 1997), and this pathway is less responsive to

continuous GH exposure - compared with intermittent incubation with GH. Indeed for GH to exert its full effect on the JAK2/STAT5b pathway, the withdrawal o f GH is required to reset the activation cascade, to enable negative regulatory pathways which are also activated by GH to dephosphorylate JAK2 and STAT5b (Ram and Waxman,

1999). This partly explains why GH administered as a pulse is more effective at inducing growth in rodents than the equivalent dose administered over a longer period of

time (Gevers et a l, 1996).

5.7 Regulated transcription using gene therapy vectors

As discussed in Chapter 1, GH is released from the pituitary in response to the integrated action o f GHRH and somatostatin, and gene transfer vectors have also been designed to regulate the release of hormones from muscle - using small molecules to control

transcription (reviewed in Harvey and Caskey, 1998; MacColl et a l, 1999). For

example, serum Epo levels can be modulated in muscle following administration of the tetracycline analogue doxycycline, to nude mice implantated with myoblasts encoding

erythropoietin (Bohl et a l, 1997).

The controlled release of growth hormone from striated muscle has also been achieved following implantation of engineered human fibrosarcoma cells into nude mice (Rivera

et a/., 1996) (Figure 5.1). The system used chemical inducers o f dimérisation to mediate hGH production in fibroblast cells. The human proteins FK506 binding protein (FKBP12) and rapamycin associated protein (FRP) form a high affinity complex with the immunosuppressant rapamycin, and the implanted fibroblast cells used in the study

by Rivera et a l, (1996) expressed two hybrid proteins. These consisted of the ligand

binding domains o f FKBP12 and FRP, fused with the zinc finger DNA binding domain of ZFBPl and the transcriptional activation domain of NF-kB p65 proteins respectively. Thus following administration of rapamycin, the fusion derivatives o f FKBP12 and FRP expressed in implanted fibroblasts formed a tripartite complex, activating GH expression from a ZFHDl-dependent promoter. A dose-dependent rise in serum growth hormone was subsequently observed in response to rapamycin administration and this is shown in Figure 5.1.

This method of inducible transcription has also been incorporated into AAV delivery vectors where hGH production could be repeatedly induced with rapamycin over several months in immunocompetent mice. However, levels o f hGH remained elevated for 1.5 days, and elements which reduced the period of hGH induction would have to be

incorporated into this system to mimic the physiological GH secretion pattern (Rivera et

al, 1999). This type of regulated GH production described above has yet to be

incorporated into the design of plasmid based DNA delivery methods.

5.8 Applications for GH and IGF-1 gene transfer into muscle

This section considers a number o f other applications using gene transfer of GH and

DNA binding domain / / - _hCMV ËWZFHDI I FKBP ("FKBP I FK8P Activation domain Target gen e hCMV /

. X

C l

#

Separate DNA-binding and activation domains

Artificially clustered transcriprioft com plex

10 9 8 7 6 5 sz 4 3 2 1 0 10 1 0.1 Rapamycin (mg/kg)

Figure 5.1 Rapamycin induction of growth hormone (GH) production in muscle.

Modified fibrosarcoma cells, implanted into the muscles o f nude mice, were induced to release GH in to the serum following administration o f rapamycin. (a) DNA sequences for expression o f human growth hormone (hGH) cDNA in human fibrosarcoma cells; DNA binding domain:hCMV - human cytomegalovirus immediate early promoter, E - epitope tag, N - SV40 viral T antigen nuclear localisation sequence, composite of ZFH Dl DNA binding domain with FKBP. Activation domain: FRP and transcription activation domain o f p65. Target gene ZFH D l/m in hCMV - tandem ZFHD 1 DNA binding sites/minimal CMV promoter for hGH expression, (b) Formation o f FRP- rapamycin-FKBP tripartite complex, and activation o f hGH transcription, (c) Rapamycin dose - dependent release o f hGH into the serum o f nude mice. Values are expressed as means ± S.E.M. From Rivera et al., (1996). With permission.

Recent studies have proposed that insulin like growth factors can function as neurotrophic factors, and reverse the peripheral neuropathy that is associated with diabetes. For example, intramuscular plasmid transfer o f IGF-1 coding sequence has been used to reduce the progressive motor neuropathy observed in diabetic mice, as a 4 week treatment using plasmid encoded IGF-1 increased nerve and muscle activity to the

level o f non diabetic controls (Coleman et a l, 1995).

It was not clear from this study by Coleman et a/.,(1995) where the IGF-1 sequence was

being expressed, and several studies demonstrate that both viral and plasmid vectors can be transported retrogradely to motorneurones innervating the muscle injection site - with

subsequent expression o f the introduced gene in the neuronal cell body (Haase et a l,

1997; Johnson I, personal communication). Although vector uptake by motorneurones might be beneficial in treating motor neurone diseases, the unregulated release of protein could also result in further damage occurring to the peripheral and central nervous systems and therefore must be investigated further.

5.9 IGF-1, GH and muscle atrophy

In addition to the methods described above, transfer of IGF-1 and GH into muscle have been proposed as a means for decreasing the muscle atrophy which is associated with ageing. Growth hormone has an anabolic effect on muscle tissue, and myoblast gene

transfer of GH into muscle tissue has also been shown to decrease atrophy in vivo

(Vandenburgh et a l, 1998). However, expression o f GH in muscle could induce insulin

resistance, by effectively decreasing glucose utilisation by this tissue. The long-term effects of this type of approach would have to be carefully considered.

IGF-1 gene transfer has also been used to decrease muscle atrophy. For example, an AAV vector encoding IGF-1 was used to prevent the age-related loss in type IIB muscle fibres in one study (described in Chapter 3). Based on these observations the authors suggested that this approach could, ‘form the basis for gene therapy for both ageing- related loss of muscle function and impairments associated with muscle disease’,

(Barton Davis et a l, 1998). Furthermore, injection of plasmid-IGF-1 into muscle

increases calcineurin activity and lactate levels, and the number o f myofibres with central nucleii - consistent with new muscle formation due to satellite cell activation by IGF-1 (Semsarian e/ûf/., 1999).

For these methods to have a significant beneficial effect, the vector would have to injected into many sites in several different weight bearing muscles and this could raise circulating IGF-1 to an unsafe level. As described earlier, there is a proportionality

between circulating IGF-1 levels and cancer risk (Holly et a l, 1999) and this type of

therapy would therefore have to be fully justified were it to be used in humans. However, one solution to this problem may be the use of a non-systemic isoform of IGF-1, which works specifically in muscle by autocrine/paracrine mechanisms. For example, both skeletal and cardiac muscle tissue produce a number of isoforms o f IGF-1 in response to overload - and these may be involved in the local repair mechanisms in both types of muscle tissue (Reviewed in Goldspink, 1999).

5.10 Immunological effects of plasmid gene transfer vectors

In Chapter 4, the injection o f a typical gene transfer vector into MRL/MpJ mice was used to model the potential effects of repeated injection of plasmid DNA in individuals with a predisposition towards autoantibody production. The increase in dsDNA

antibodies in MRL/MpJ mice injected with pcDNA3E, and additional numbers of mice within this group with antibodies to components of the cell nucleus, indicate that this

method could induce autoantibody production in susceptible individuals (MacColl et al,

2000b). Additional studies on larger numbers should be carried out to determine the significance of these results. Although repeated injection of plasmid into immunologically tolerant Balb/C mice did not induce or augment autoantibody production, these results suggest that the gradual development of autoimmune responses should be monitored in subgroups of patients given genetic vaccines or other types of therapy based on dehvery methods using plasmids or other types o f vector.