ELSEVIER Neuroscience Letters 177 (1994) 151-154

NEUROSClENC[ IEITIRS

Transforming growth factor-beta-2 (TGF-fl2) is associated with mature rat neuromuscular junctions

Ian. S. McLennan*, Kyoko Koishi

Department of Anatomy and Structural Biology. University of Otago, PO Box 913, Dunedin, New Zealand.

Received 19 April 1994; Revised version received 8 June 1994; Accepted 30 June 1994

Abstract The localisation of transforming growth factor-beta-2 (TGF-fl2) in skeletal muscles was investigated with immunohistochemistry.

In neonates, TGF-fl2 was distributed throughout the muscle fibres but as the fibres matured TGF-fl2 became restricted to the circumference of a small subpopulation of nuclei. These nuclei were judged to be the subsynaptic nuclei as they lay beneath the plasmalemma and were associated with ~-bungarotoxin-labelled neuromuscular junctions. These observations point to TGF-fl2 being either atrophic factor for mature motoneurones or an autocrine regulator of synaptic protein production.

Key words: TGF-fl2; Motoneuron; Neuromuscular junction; Schwann cell; Neurotrophic factor; Acetylcholine receptor

Transforming growth factor-beta-2 (TGF-fl2) is a member of a family of multifunctional polypeptide growth factors that includes TGF-fll and TGF-fl3 (70- 80% amino acid sequence homology) and the more dis- tantly related activins, inhibins and bone morphogenetic proteins (30-40% homology) [6]. Despite the similarity of their primary structure, each riaember of the TGF-fl family appears to mediate different cellular functions in vivo [2]. The TGF-fls are best known for their ability to induce the synthesis of extracellular matrix proteins and to inhibit cell proliferation. However, the TGF-fls are present in the central and peripheral nervous systems [2,11] and affect both neurones and non-neuronal cells in vitro. For instance, TGF-fll is a potent survival factor for cultured embryonic rat motoneurons [5] and can stimulate the proliferation of Schwann cells [8]. In this paper, we report that TGF-fl2 is associated with the subsynaptic nuclei of mature rat neuromuscular junc- tions (nmj).

Adult Wistar rats (15 weeks or older) were killed by cervical dislocation and their tibialis anterior muscles snap frozen in melting isopentane. The muscles were

*Corresponding author. Fax: (64) (3) 4797254. E-mail: I Mclennan@Gandalf.Otago.AC.NZ

0304-3940194157.00 © 1994 Elsevier Science Ireland Ltd. All rights reserved S S D I 0304-3940(94)00520-6

stored under liquid nitrogen and subsequently processed for immunohistochemistry as previously described [7]. The key steps of the procedure were that 5/.zm sections were cut with a cryostat and mounted onto gelatin- coated glass slides. After air-drying, the sections were fixed in neutral-buffered 4% paraformaldehyde at 4 °C for 30 min, incubated overnight at 4 °C with either an affinity purified rabbit polyclonal anti-TGF-fl2 (Santa Cruz Biotechnology) or an equivalent amount of non- immunised rabbit IgG (Sigma). The immunoreactivity was then developed using biotinylated-anti-rabbit IgG (Amersham), streptavidin biotinylated-horseradish per- oxidase complex (Amersham) and 3-amino-9-ethylcar- bamide (AEC) as the chromogen. The slides were not counterstained and were examined using brightfield and phase contrast microscopy with a Zeiss Axioplan photomicroscope.

Most of the muscle fibres contained very low levels of TGF-fl2 immunoreactivity but in some sections a few fibres contained intense immunoreactivity around some of their nuclei (Fig. l a). Similar staining was not ob- served with non-immune IgG (Fig. lc) or after incuba- tion with an anti-TGF-fl3 antibody (Fig. lb). The in- tensely positive fibres were often clustered together (Fig. la) and their distribution pointed to the immunoreactiv- ity being associated with the nmj. This was investigated

152 I.S. McLennan. K. Koishi/Ncuro.~ciencc Letters 177 119941 151 154

~ f

A

t

C

Fig. I. Photomicrographs of transverse sections of skeletal muscles, a: adult tibialis anterior staincd with anti-TGF-fl2. Tile arrows point to intense immunoreactivity associated with a subpopulation of nuclei. The arrowhead points to trace levels of immunoreactivity inside the circumference of a fibre, b: adult tibialis anterior stained with anti-TGF-fl3. The fibres arc weakly immunoreactive and the nmj are not stained, c: adult tibialis anterior stained with rabbit IgG (control antibody). The immunoreactivity was developed and photographed using the same conditions as for figures lal arid (b). d: tibialis anterior from a 2-day-old rat stained with anti-TGF-fl2. Magnification 250 ~ .

by labelling nmj with rhodamine-conjugated a-bungaro- toxin (Molecular Probes) and developing the TGF-/32 immunoreactivity with streptavidin-BOD1PY (Molecu- lar Probes). The BODIPY and rhodamine tags were then separately excited using either the 488 nm or 568 nm lines of a krypton-argon laser and the emission analysed using a BioRad 600 confocal microscope. There was a one-to- one correspondence between intensely stained TGF-/32 profiles and ct-bungarotoxin synaptic sites. When the BODIPY and rhodamine fluorescences were simultane- ously viewed, the TGF-/3-2 (BODIPY) lay beneath the c~-bungarotoxin (rhodamine) signal (Fig. 2), implying that it was associated with the muscle fibre rather than with Schwann cells or fibroblasts. This was further inves- tigated by labelling the plasmalemma of fibres with an anti-dystrophin (DYS2: Novocastra laboratories) linked to anti-mouse-IgG-BODIPY (Molecular Probes) and si- multaneously visualising TGF-/32 (linked to strep- tavidin-Texas Red). The TGF-/3-2 immunoreactivity sur- rounded the sub-synaptic nuclei and was not detected outside of the plasmalemma (Fig. 3).

The antigen used to make the anti-TGF-/32 was a 26 amino acid peptide corresponding to the carboxy termi-

nal region of human TGF-fl2. This peptide has partial homology to TGF-fll (17 amino acids out of 26) and TGF-fl3 (21 out of 26) but is claimed by the manufac- turer to be specific to TGF-fl2 (Santa Cruz Biotechnol- ogy). The specificity of the antibody was checked by determining its ability to detect TGF-fll (recombinant, human), TGF-fl2 (porcine) and TGF-fl3 (recombinant, chicken) on slot blots (all proteins were from R&D Sys- tems). The results obtained with these proteins are rele- vant to the present study as the TGF-fls are highly con- served between species. 0.1, 1 and 10 ng of each of the TGF-fls was applied to a nitrocellulose membrane and the reactivity of the proteins with the anti-TGF-fl2 visu- alised using the reagents described for the immunohisto- chemistry, except that diaminobenzidine was used as the chromogen. The anti-TGF-fl2 reacted with 10 ng of TGF-fl3 but with an intensity that was less than its reac- tion with 1 ng of TGF-fl2. There was no significant reactivity with 10 ng of TGF-fll (not shown). It is there- fore possible that the anti-TGF-fl2 will detect high con- centrations of TGF-fl3 in tissue sections. However, TGF-fl3 is not concentrated at the nmj as an anti-TGF- /33 antibody {Santa Cruz Biotechnology) weakly stained

I.S. McLennan, K. Koishi/Neuroscience Letters 177 (1994) 151 154

J , " .

153

2C 3C'

Fig. 2. Electronic micrograph of a transverse section of an adult tibialis anterior that was double-labelled with anti-TGF-fl2 (linked to streptavidin- BODIPY) and rhodamine-conjugated ct-bungarotoxin. The BODIPY and rhodamine tags were sequentially excited with the 488 nm and 568 n m lines of a krypton-argon laser and their emissions independently recorded with a BioRad 600 confocal microscope, a: TGF-fl2 immunoreactivity (BODIPY). The arrow points to the position of a sub-synaptic nucleus, b: a nmj labelled with rhodamine conjugate ~-bungarotoxin. c: the merged images of(a) and (b). Two branches o f a nmj (red) are clearly evident and lie either side o f a subsynaptic nucleus, which is causing a local protuberance of the surface membrane. The TGF-fl2 immunoreactivi ty is coloured green and is subjacent to the nmj (red). Magnification (a) and (b) 2,500 × ; (c) 5,000 x .

Fig. 3. Electronic micrograph of a transverse section of an adult tibialis anterior tha: was double-labelled with anti-TGF-fl2 (linked to streptavidin- Texas red) and anti-dystrophin (linked to ant i-mouse IgG-BODIPY). The BODIPY and Texas red tags were sequentially excited with the 488 nm and 568 nm lines of a krypton-argon laser and their emissions independently recorded with a BioRad 600 confocal microscope, a: TGF-fl2 immunoreactivity. The arrow points to the position of a sub-synaptic nucleus, b: dystrophin immunoreactivity. Dystrophin is a marker for the p lasmamembrane of muscle fibres [12]. c: the merged images of (a) and (b). Note that the TGF-fl2 immunoreactivity is coloured red and lies within the p lasmamembrane of a fibre (coloured green). Magnification (a) and (b) 2,500 x ; (c) 5,000 x .

154 LS. McLennan, K. Kotshi/Neuroscience Letters 177r1994) 151 154

the cytoplasm of muscle fibres but did not stain mature nmj (Fig. lb). This indicates that the TGF-fl isoform at the nmj is most likely TGF-fl2.

The association of TGF-fl2 with mature nmj is sugges- tive that TGF-fl2 may regulate some aspect(s) of neu- romuscular communication. There are several possibili- ties. First, the survival of motoneurones is contingent on them receiving trophic support from the cells that they are associated with, especially the muscle fibres. Several factors, including NT-3, NT-4 and FGF-5, have been recently implicated as motoneurone trophic factors but this does not preclude TGF-fl2 having a role in this area. Neurones are known to be influenced by combinations of growth factors and to have different trophic require- ments as they mature [1,9,10]. In this context, it is partic- ularly interesting that TGF-fl2 was concentrated at ma- ture nmj (Fig. la) but was widely distributed in immature fibres (from 2-day-old pups, Fig. ld). The TGF-fl2 asso- ciated with immature fibres may be available to imma- ture motoneurones although the presence of TGF-fl2 in muscle precursor cells (unpublished observations) points to the early expression of TGF-fl2 being related to some aspect of muscle fibre formation. Second, synapses have very specialised structures and there are over 30 proteins that are selectively localised to the synaptic region of a muscle fibre [4]. Rat muscle fibres contain about a thou- sand nuclei and some mechanism exists which results in the subsynaptic nuclei differentially producing mRNA for synaptic proteins [3,4]. TGF-fl2 is associated with subsynaptic nuclei and it is possible that TGF-fl2 is an autocrine regulator of the subsynaptic nuclei: that is, it is part of the mechanism that directs the subsynaptic nuclei to produce synaptic mRNAs. Third, terminal Schwann cells and fibroblasts associated with the neu- romuscular junction have different properties to their non-synaptic counterparts [4]. The mechanism which generates these differences is unknown but could involve synaptically-produced TGF-fl2. These alternatives are currently being explored by several laboratories associ- ated with the University of Otago.

Mr. Piotr Swierczynski is thanked for his technical assistance. This work was supported by a grant from The Health Research Council (NZ) and through the use of equipment funded by The Lottery Board (NZ), The Health Research Council (NZ) and The Otago Research Society.

[1] Eagleson, K.L. and Bennett, M.R., Motoneurone survival require- ments during development: the change from immature astrocyte dependence to myotube dependence, Dev. Brain Res., 29 (1986) 161-172.

[2] Flanders, K.C., Ludecke, G., Engels, S., Cissel, D.S., Roberts, A.B., Kondaiah, P., Lafyatis, R., Sporn, M.B. and Unsicker, K., Localization and actions of transforming growth factor-fls in the embryonic nervous system, Development, 113 (1991) 183-191.

[3] Fontaine, B., Sassoon, D., Buckingham, M. and Changeux, J.-P., Detection of the nicotinic acetylcholine receptor ~-subunit mRNA by in situ hybridisation at neuromuscular junctions of 15-day-old chick striated muscles, EMBO J., 7 (1988) 603~509.

[4] Hall, Z.W. and Sanes, J.R., Synaptic structure and development: the neuromuscular junction, Cell, 72 (Suppl.) (1993) 99-121.

[5] Martinou, J.C., Levanthai, A., Valette, A. and Weber, M.J.. Transforming growth factor-beta-1 is a potent survival factor for rat embryo motoneurons in culture, Dev. Brain Res., 52 (1990) 175-181.

[6] Massague, J., The transforming growth factor-beta family, Annu. Rev. Cell Biol., 6 (1990) 597-641.

[7] McLennan, I.S., Resident macrophages (ED2 and ED3 positive) do not phagocytose degenerating rat skeletal muscle fibres, Cell Tissue Res., 272 (1993) 193-196.

[8] Ridley, A.J., Davis, J.B., Stroobant, P. and Land, H., Transform- ing growth factor-beta-1 and factor-beta-2 are mitogens for rat Schwann cells, J. Cell Biol., 109 (1989) 3419-3424.

[9] Thoenen, H. and Barde, Y.A., Physiology of nerve growth factor, Physiol. Rev., 60 (1980) 1284-1335.

[10] Thoenen, H., Hughes, R.A. and Sendtner, M., Towards a compre- hensive understanding of the trophic support of motoneurons, C.R. Acad. Sci. Serie III. - Life Sci., 316 (1993) 1161-1163.

[11] Unsicker, K., Flanders, K.C., Cissel, D.S., Lafyatis, R. and Sporn, M.B., Transforming growth factor-beta isoforms in the adult rat central and peripheral nervous system, Neuroscience, 44 (1991) 613~525.

[12] Zubrzycka-Gaarn, E.E., Bulman, D.E., Karpati, G., Burghes, A.H.M., BelfaU, B., Klamut, H.J., Talbot, J., Hodges, R. S., Ray, P.N. and Worton, R.G., The Duchenne muscular dystrophy gene product is localized in sarcolemma of human skeletal muscle, Na- ture, 333 (1988) 466-469.