Effects of insulin like growth factors on early embryonic chick limb myogenesis

Limb muscles derive from pax3 expressing precursor cells that migrate from the hypaxial somite into the developing limb bud. Once there they begin to differentiate and express muscle determination genes such as MyoD. This process is regulated by a combination of inductive or inhibitory signals including Fgf18, retinoic acid, HGF, Notch and IGFs. IGFs are well known to affect late stages of muscle development and to promote both proliferation and differentiation. We examined their roles in early stage limb bud myogenesis using chicken embryos as an experimental model. Grafting beads soaked in purified recombinant IGF-I, IGF-II or small molecule inhibitors of specific signaling pathways into developing chick embryo limbs showed that both IGF-I and IGF-II induce expression of the early stage myogenic markers pax3 and MyoD as well as myogenin. Their effects on pax3 and MyoD expression were blocked by inhibitors of both the IGF type I receptor (picropodophyllotoxin, PPP) and MEK (U0126). The PI3K inhibitor LY294002 blocked IGF-II, but not IGF-I, induction of pax3 mRNA as well as the IGF-I, but not IGF-II, induction of MyoD mRNA. In addition SU5402, an FGFR/ VEGFR inhibitor, blocked the induction of MyoD by both IGFs but had no effect on pax3 induction, suggesting a role for FGF or VEGF signaling in their induction of MyoD. This was confirmed by in situ hybridization showing that FGF18, a known regulator of MyoD in limb myoblasts, was induced by IGF-I. In addition to their well-known effects on later stages of myogenesis via their induction of myogenin expression, both IGF-I and IGF-II induced pax3 and MyoD expression in developing chick embryos, indicating that they also regulate early stages of myogenesis. The data suggests that the IGFs may have slightly different effects on IGF1R signal transduction via PI3K and that their stimulatory effects on MyoD expression may be indirect, possibly via induction of FGF18 expression.


Introduction
During development the limb muscles are derived from pax3 expressing cells from the hypaxial region of somites. These cells delaminate and migrate into the limb buds where they begin to differentiate and express muscle specific markers such as members of the Myogenic Regulatory Factor (MRF) family of transcription factors [1][2][3][4][5]. The migration of these cells is induced by CXCR4 [6,7] and HGF [8][9][10], which also acts to prevent premature differentiation of these cells. The majority of the migratory cells will contribute to muscle although some will also become endothelial cells [11]. Once in the limb, the myogenic precursors form the dorsal and ventral muscle masses and begin to differentiate, a process regulated by the induction the MRFs; first myoblasts express Myf5, then MyoD, myogenin and finally MRF4 [12].
Numerous signaling molecules regulate the differentiation of the limb myoblasts. Their differentiation is inhibited by sonic hedgehog [13] and BMP [14], promoted by FGFs, such as FGF18 [15,16], while other molecules can act to either block or induce myogenic genes depending on the stage of development and concentration, such as retinoic acid [16,17].
The insulin like growth factors, IGF-I and IGF-II, are well characterized promoters of muscle growth in development [18], including in chicken embryos [19]. They act through the IGF type 1 receptor in muscle growth and regeneration [20] primarily by promoting the AKT/ mTOR and MAPK signaling pathways [21][22][23].
During limb development several components of the IGF signaling machinery are expressed [24] and IGF signaling regulates the formation of the limb skeleton [25]. Retroviral overexpression of IGF-I in limbs also increases muscle size by promoting myoblast proliferation, leading to increased numbers of muscle fibres [19], and in ovo injection of IGF-I can have effects lasting into adulthood [26]. However, as well as promoting proliferation, IGFs can also induce myogenin expression [27] and it is clear that they have a complex role in developing muscle.
To try and understand the effects of IGFs during early embryonic myogenesis we used the chicken embryo limb bud as a model [28,29] by grafting beads soaked in purified growth factors or other signaling inhibitory molecules at defined stages of embryogenesis to determine their effects on myogenesis. Here we show that grafting IGF beads into early developing chicken embryo limbs induces the expression of pax3, a marker of proliferative muscle precursor cells, while later grafting also induces both MyoD and myogenin, which are associated with the early and late stages of myogenic differentiation respectively. Using various inhibitors we show that the effects on both pax3 and MyoD require MEK signaling while MyoD induction is dependent on secondary signaling through either FGFs or VEGF; in addition we show that IGF-I can induce FGF18 expression in limb buds. A PI3K inhibitor produced a more complex picture with different effects depending on whether the limbs were treated with IGF-I or-II.

Growing and staging of experimental animals
Fertilized white leghorn chicken (Gallus gallus) eggs were purchased from Henry Stewart Limited (Norwich, UK). Eggs were incubated at 15˚C for up to 5 days until the day of use then transferred to 38˚C (Forma scientific CO2 water incubator) until they reached the required stages of development. Embryos were staged according to Hamburger and Hamilton [30].

IGF and pharmacological inhibitor beads
Heparin beads (Sigma H-5263) were soaked in recombinant human IGF-I or IGF-II (Peprotech) at 1mg/ml in phosphate buffered saline (PBS) with 0.1% Bovine Serum Albumin (BSA). AG 1-X2 beads (BioRad) were incubated in Picropodophyllotoxin (PPP, Tocris Bioscience), U0126 (Cell Signaling), LY 294002 (Calbiochem) or SU5402 (Calbiochem), all reconstituted in DMSO at 10mM. Beads were incubated for at least one hour in the dark before being washed briefly in 2% phenol red and rinsed in PBS before grafting. Beads were grafted into limb buds with a sharpened tungsten needle, resealed with sellotape and reincubated for 18-48h as described previously [31].

Paraffin embedding and sectioning of chick embryos
Embryos were washed twice in 1X PBS at RT and in sterile distilled water then Dehydrated through an ethanol series of 25%, 50%, 70%, 90% and 100% ethanol. Embryos were cleared in xylene then transferred into hot paraffin wax at 65˚C for 2 hours. Embryos were orientated in paraffin wax solution and cooled for 2 hours at 4˚C. Sections were cut at 7 μm thickness on a HM 355 microtome, placed onto glass microscope slides then mounted in Omnimount (Histological Mounting Medium HS-110). Slides were photographed on an Olympus BH-2 microscope.

Results
To determine their effects on limb bud myogenesis, heparin beads soaked in either IGF-I or IGF-II were grafted into developing limb buds in ovo at HH stage 17 and incubated for 24 hours until they had reached HH stage 21/22. In situ hybridization showed a clear upregulation of pax3 mRNA in grafted limbs by both IGF-I (28/34 embryos, Fi g1 a,b) and IGF-II (22/ 30 embryos , Fig 1e and 1f). In these embryos MyoD was only rarely upregulated by either IGF-I (2/11 embryos, Fig 1c and 1d) or IGF-II (4/12 embryos, Fig 1g and 1h) after 24 hours. Control beads soaked in BSA had no effect on either pax3 (8/8 embryos, Fig 1I and 1j) or MyoD (6/6 embryos, Fig 1k and 1l) expression.
To see if these genes were induced by IGFs at later stages of development heparin beads soaked in either IGF-I or IGF-II were grafted into developing limb buds at HH stage 19 and incubated for 18 hours until they had reached HH stage 23. In these embryos IGF-I induced both pax3 ( To confirm that the induced expression of both pax3 and MyoD was in myogenic progenitors we cut transverse sections of manipulated embryos. These showed that upregulation of pax3 and MyoD mRNA was observed in the dorsal and ventral muscle masses (Fig 3a and 3b), the regions of the limb buds where myogenic differentiation normally occurs. In contrast the non-manipulated limbs and other expression domains, such as the dorsal neural tube and dorso-medial lip of the somite, which express pax3, and the myotome, which expresses MyoD, remained unaffected (Fig 3c and 3d).
To see if IGFs could also induce expression of later myogenic markers we incubated embryos following bead grafts to HH stage 24, the point at which myogenic cells begin to express markers of terminal differentiation such as myogenin. IGF-I beads grafted at HH stage 19 and incubated for 24h to reach HH stage 24 induced MyoD expression (8/18 embryos, Fig  4a and 4b), while beads grafted at HH stage 21/22 and incubated for 18h to reach HH stage 24 induced myogenin (7/9 embryos , Fig 4c and 4d). Similar effects were seen with IGF-II beads which induced both MyoD (10/19 embryos , Fig 4e and 4f) and myogenin (12/21 embryos, Fig  4g and 4h) at HH stage 24. In contrast BSA beads had no effect on either MyoD (6/6 embryos ,  Fig 4i and 4j) or myogenin (9/9 embryos , Fig 4k and 4l) expression.
To investigate the downstream signaling pathways mediating IGF induction of pax3 and MyoD we co-grafted IGF beads with beads soaked in small molecule signaling inhibitors. We tested these signal transduction inhibitor effects on pax3 induction following IGF beads grafted at HH stage 17 (Fig 5a, 5b, 5c and 5d) and MyoD induction after grafting at HH stage 19 (Fig 5e, 5f, 5g and 5h). Results are shown in Fig 5 and summarized in Table 1.
To confirm that the IGFs were acting through the IGF type 1 receptor we co-grafted IGFs with picropodophyllotoxin (PPP), a specific inhibitor of IGF1R autophosphorylation. PPP beads blocked IGF-I induction of both pax3 (12/18 embryos , Fig 5i and 5j) and MyoD (7/9  Fig 5o and 5p).
The ability of SU5402, an FGFR and VEGFR inhibitor, to block MyoD induction by IGFs was unexpected. One possible explanation was that this was an indirect effect caused by IGF induced upregulation of FGF. To test this we grafted IGF-I beads at HH stage 17, incubated embryos for 24h to HH stage 22 and then measured their effect on Fgf18 mRNA, which is known to induce MyoD expression in limb bud myogenic precursors. In 9/19 embryos we saw increased levels of Fgf18 in grafted limb buds (Fig 6).

Discussion
The specification and differentiation of muscle cells provides an excellent paradigm to examine inductive events during development. As pax3 expressing precursor cells migrate into limb buds they begin to differentiate by expressing MyoD and, subsequently, myogenin in response to a range of signals [12,33].
Although IGFs are well known to regulate muscle formation in embryos and muscle growth in adult animals [34][35][36] there remain many unanswered questions about their roles; for example how they are able to promote both proliferation and differentiation of myoblasts, behaviours that should be mutually exclusive. Here we use the chicken embryo model to examine some of the signaling events that underlie these activities by grafting beads soaked in IGF-I or IGF-II along with specific signaling inhibitors.
In early limb buds (HH stage 17) both IGFs were able to induce upregulation of pax3 mRNA but not MyoD, while in slightly later limbs (HH stage 19) both pax3 and MyoD mRNA were induced. We have previously shown that these early myogenic progenitors are resistant to signals that promote differentiation and that this is mediated by retinoic acid [16]. The data we present here is consistent with the idea that signals in the early limb prevent premature differentiation of muscle cells, presumably to ensure that there are sufficient precursors produced to contribute to the muscles.  The ability of IGFs in older embryos to induce pax3 expression, a marker of proliferative precursors, and MyoD and myogenin, markers of early and late stages of differentiation, is less easy to explain although this is a commonly observed feature of these molecules [34]. There are several models that could explain this apparent paradox.
The increase in pax3 expression could be due to (i) increased proliferation of the migratory precursors, (ii) higher numbers of cells migrating from the hypaxial somite into the limbs or (iii) recruitment of additional cells from within the limb to the myogenic lineage. As limb bud mesenchyme cells do not contribute to muscle [37,38] or express pax3 [39,40] it is unlikely that these cells are being respecified and therefore the recruitment (iii) model is unlikely. It is also possible, given that in situ hybridization does not provide single cell resolution, that the increase in pax3 levels indicates higher transcription of pax3 mRNA within the same number of cells. It is also important to bear in mind that we cannot exclude indirect effects on pax3 expression; for example IGFs could be inducing expression of other signaling molecules, such as HGF, that are known to enhance myoblast proliferation [8][9][10].
The induction of MyoD and myogenin are harder to explain in the context of increased pax3 expression. It is possible that this is a stochastic effect because there are more progenitors in the limb; increased MyoD and myogenin levels are observed simply because there are more myogenic cells differentiating. An alternative explanation is that IGF signaling has different effects at different stages of development and differentiation. This is consistent with the differences seen at HH stages 17 and 19. At earlier stages myogenic precursors are not competent to induce MyoD, either because of their epigenetic state or because of high levels of inhibitors in the proximal limb, such as retinoic acid [16]. In later embryos RA levels have declined and the cells will also have moved further along the differentiation process, potentially making them able to respond to IGFs in a different way. One other possible explanation is that the IGFs will also affect the surrounding limb bud mesenchyme, changing the signaling environment of the cells. Support for this model comes from the surprising observation that MyoD induction by IGFs is blocked by an inhibitor of the FGF and VEGF receptors. In this model IGFs act indirectly by the induction of pro-myogenic signals, such as FGF18, in the surrounding tissues. We number of grafts. A summary diagram of which molecules inhibited IGF-I and/or IGF-II induction of pax3 and MyoD mRNA is also included. have shown previously that beads soaked in FGF18 can induce MyoD, but not pax3, expression in developing limbs at these stages of development [16] and the induction of Fgf18 mRNA by IGF-I, as well as the ability of SU5402 to block IGF induced MyoD expression, is consistent with this model. However, as this inhibitor also blocks signaling though VEGFR2, which is known to affect myogenesis, it is also possible that IGFs induce members of the VEGF family and these could also contribute to induction of MyoD. Further complexity in the responses of limb bud myoblasts to IGFs is apparent when they are exposed to a variety of signal transduction inhibitors. pax3 and MyoD induction in response to IGF-I and-II were both blocked by PPP, an IGF1R inhibitor, and U0126, which blocks MEK activity and so prevents ERK phosphorylation. Surprisingly, given the very well characterized links between IGFs and PI3K/AKT/mTOR signaling, the PI3K inhibitor LY294002 specifically blocked IGF-I induced MyoD expression and IGF-II induced pax3 expression. This could be because, in these embryos, the IGF-I and -II are interacting with other receptors, for example the insulin receptor to trigger this pathway [41]. However, this is hard to reconcile with the data showing that PPP can block all these responses.
In summary our data show limb bud muscle precursors at HH stage 17 respond to IGF signaling by upregulating pax3, a marker of early proliferating muscle precursor cells and that later, at HH stage 19, they also upregulate later markers of differentiation, MyoD and myogenin. All these events are controlled by the IGF1R receptor, which is expressed throughout the limb buds at these stages [24], involving signaling through the ERK MAPK pathway. IGF-I and-II appear to have differential effects through PI3K signaling at HH stages 17 and 19 while the induction of MyoD is, at least in part, dependent on FGF receptors, possibly through induction of FGF18 in the limb bud mesenchyme.

Ethical approval
All experiments were completed before 14 days of incubation, two thirds of the way through chicken embryo development. Therefore embryos used in this project does are not regulated under the UK Animals (Scientific Procedures) Act of 1986. All procedures were discussed and agreed with the University of Nottingham ethics officer. Fertilised eggs were purchased from reputable commercial suppliers (Henry Stewart) who specialise in providing eggs for research. Their farms are registered under DEFRA's Poultry Health Scheme which ensures disease control programmes that are up to the latest EU standards. They also comply with RSPCA's Freedom Food Code of Practice promoting the 'Five Freedoms' for best practice welfare.