Mutation in neurofilament transgene implicates RNA processing in the pathogenesis of neurodegenerative disease

Abstract

A mouse neurofilament light subunit (NF-L) transgene with a 36 bp c-myc insert at the end of the coding region was found to have neuropathic effects on enteric and motor neurons of transgenic mice. The severity of phenotype was related directly to the levels of transgenic mRNA expression. High levels of transgene expression were lethal to newborn pups, causing profound alterations in the development of the enteric nervous system and extensive vacuolar changes in motor neurons. Lower levels of transgene expression led to a transient stunting of growth and focal alterations of enteric and motor neurons. Because the positioning of the c-myc insert coincided with the location of the major stability determinant of the NF-L mRNA (Cañete-Soler et al., 1998a,b), additional studies were undertaken. These studies showed that the c-myc insert alters the ribonucleoprotein (RNP) complexes that bind to the stability determinant and disrupts their ability to regulate the stability of the transcripts. The findings indicate that expression of an NF-L transgene with a mutant mRNA stability determinant is highly disruptive to enteric and motor neurons and implicate alterations in RNA processing in the pathogenesis of a neurodegenerative condition.

Fig. 1.

Newborn founder mice A (a) and B (b) bearing the NF-L transgene withc-myc mutation have markedly distended abdomens.c, The intestine of pup A reveals a dilated and shortened small intestine (SI) between the stomach (S) and ileum (I). The distal end of the esophagus (arrow) and discontinuous ends of ileum (*) and colon (C) are identified also. Immunohistochemistry for PGP9.5 (Karaosmanoglu et al., 1996) delineates brown reactive products in the enteric ganglia in the walls of small intestine from pup A (d) and a nontransgenic littermate (e) that are everted and folded back on themselves and around the abdominal cavity (x). The wall of the control is studded with immunoreactive enteric ganglia in the myenteric plexus immediately below the serosal surface. In contrast, there is a marked reduction of immunoreactive enteric ganglia in the wall of the small intestine of pup A. Scale bars: d, e, 100 μm.

Fig. 2.

a, A cluster of anterior horn cells in the lumbar spinal cord of pup A is enlarged (b) to reveal the cytoplasmic vacuolar degeneration of motor neurons. The nuclei of vacuolated neurons contain prominent nucleoli. c, Similar vacuolated changes are present in a cluster of three motor neurons of founder mouse D. Stained with hemotoxylin alone (a, b) or with eosin (c). Scale bars: a, 75 μm;b, c, 25 μm.

Fig. 3.

a, Cross section of skeletal muscle at the level of the distal tibia from transgenic pup A showing large myofibers with hyperchromatic central nuclei and perinuclear vacuoles (arrow) scattered among numerous small cells without myofibrils. b, Cross section of skeletal muscle at the level of the distal tibia of a nontransgenic littermate control showing a uniform population of myofibers with central and peripheral nuclei. H&E stain. Scale bars: a, b, 50 μm.

Fig. 4.

a, b, Stunted growth in an 18 d F1 transgenic pup (arrow) as compared with two nontransgenic littermates. c, The transgenic pup (arrow) displays an abnormal reflex of flexing the hind- and forelimbs when held by the tail, as compared with extension of the limbs and writhing movements of a nontransgenic littermate.

Fig. 5.

RNA protection assay showing protected fragments of 322 and 212 nt from transgenic and endogenous NF-L mRNAs, respectively, in the brains of pup A (lane 1), pub B (lane 2), nontransgenic littermates of pups A and B (lanes 3, 4), founder mouse C (lane 7), two of her transgenic pups (lanes 5, 6), and founder mouse D (lane 8).

Fig. 6.

Protection assay (a) and quantitation (b) of NF-L and β-actin mRNA levels in Neuro 2a cells doubly transfected with tTA transactivator expression vector and target gene in which a Tn-10 tetracycline-inducible promoter drives the expression of wild-type (NF-L/wt) or mutant (NF-L/del and NF-L/c-myc) NF-L cDNA. The mutant transgenes either were deleted of 23 bp of distal coding region and 45 bp of proximal 3′-UTR (NF-L/del) or had a 36 bpc-myc insert and stop codon appended to the end of the coding region (NF-L/c-myc). The NF-L target genes were activated for 72 hr by withdrawal of tetracycline and then inactivated by the readdition of 0.5 μg/ml tetracycline (0 time point). An RNA protection assay detected radioactive fragments of 125 nt (filled arrow) and 83 nt (open arrow) from NF-L and β-actin mRNAs. NF-L/β-actin mRNA levels were averaged from three experiments. c, mRNA levels from mutant NF-L cDNAs with 36 bp c-myc mutation in the BglII site of exon 1 (NF-L/c-myc/BglII), in theEcoRI sites of the distal 3′-UTR (NF-L/c-myc/EcoRI), and at the end of the coding region (NF-L/c-myc). Mutant and wild-type NF-L cDNAs with tetracycline-inducible promoters were cotransfected in Neuro 2a cells, and mRNA levels were assayed by RT-PCR at 0, 24, and 48 hr after inactivation of transcription by the addition of tetracycline. Levels of mRNA from the mutant NF-L cDNAs are expressed as the percentage of mRNA level from the cotransfected wild-type NF-L cDNA.

Fig. 7.

Gel-shift (a) and cross-linking (b, c) assays of RNP complexes that form when brain extracts are incubated with probe A (23 nt of distal coding region and 45 nt of proximal 3′-UTR of NF-L) and probe A/c-myc (same probe with c-myc tag inserted between the coding region and 3′-UTR). Gel-shift assay (a) shows a faster migrating complex (solid arrows) competed by poly(C), enhanced by poly(U), and referred to as the C-binding complex. Whereas the C-binding complex primarily is composed of a single band (lower solid arrow) on probe A (lanes 2, 3), an additional slower migrating band (upper solid arrow) forms on probe A/c-myc (lanes 6, 7). A slower migrating complex (open arrows), referred to as the U/A-binding complex, also forms preferentially on the mutant probe. Cross-linking assay (b, c) shows that radioactivity from probes A and A/c-myc is cross-linked to a major 39 kDa polypeptide (solid arrow) and to a minor 80 kDa polypeptide (open arrow) and that cross-linkage to the 39 and 80 kDa polypeptides is competed by poly(C) and poly(U), respectively. Cross-linkage of radioactivity from probe A to the 39 kDa polypeptide is enhanced in the presence of poly(U) or poly(A⁺) (compare lanes 5 and6 with lane 2 in Fig. 1c), but not from probe A/c-myc (compare lanes 12 and 13 with lane 9 in Fig.1c). Likewise, cross-linkage of radioactivity from probe A to the 80 kDa polypeptide is enhanced in the presence of low (+) or high (++) levels of poly(C) (compare lanes 3 and4 with lane 2), but not from probe A/c-myc (compare lanes 10 and11 with lane 9). Incubations were conducted with or without extract (160 μg) and at low (20 ng) or high (1 μg) levels of poly(C), poly(U), or poly(A⁺) homoribopolymer competitors. A 5% nondenaturing acrylamide gel was used for the gel-shift assay; 10% PAGE was used for cross-linking studies. Molecular weights were estimated by comigration with prestained standards. Figure 1c is a briefly exposed autoradiogram of cross-linkage in the 39 kDa polypeptide.