Reduced C9orf72 protein levels in frontal cortex of amyotrophic lateral sclerosis and frontotemporal degeneration brain with the C9ORF72 hexanucleotide repeat expansion

Abstract

An intronic G(4)C(2) hexanucleotide repeat expansion in C9ORF72 is a major cause of amyotrophic lateral sclerosis and frontotemporal lobar degeneration. Several mechanisms including RNA toxicity, repeat-associated non-AUG translation mediated dipeptide protein aggregates, and haploinsufficiency of C9orf72 have been implicated in the molecular pathogenesis of this disorder. The aims of this study were to compare the use of two different Southern blot probes for detection of repeat expansions in an amyotrophic lateral sclerosis and frontotemporal lobar degeneration pathological cohort and to determine the levels of C9orf72 transcript variants and protein isoforms in patients versus control subjects. Our Southern blot studies identified smaller repeat expansions (250-1800 bp) that were only detectable with the flanking probe highlighting the potential for divergent results using different Southern blotting protocols that could complicate genotype-phenotype correlation studies. Further, we characterize a new C9orf72 antibody and show for the first time decreased C9orf72 protein levels in the frontal cortex from patients with a pathological hexanucleotide repeat expansion. These data suggest that a reduction in C9orf72 protein may be a consequence of the disease.

Keywords: Amyotrophic lateral sclerosis; C9orf72; Frontotemporal dementia; Frontotemporal lobar degeneration; Repeat expansion; Southern blotting.

Supplementary Fig. 1

Supplementary Fig. 2

Supplementary Fig. 3

Supplementary Fig. 4

Supplementary Fig. 5

Supplementary Fig. 6

Fig. 1

C9ORF72 repeat expansion analysis of frontal cortex and cerebellar postmortem tissue using 2 Southern blot protocols. Genomic DNA extracted from frontal cortex (A) and cerebellar cortex (B) were digested with AluI/DdeI and analyzed using the published hexanucleotide Southern blot probe. Cerebellar cortex gDNA were also analyzed with a novel flanking Southern blot probe following EcoRI/BamHI restriction digest (C). The flanking probe allows detection of normal range alleles (white arrowhead) that matched the genotypes determined by flanking PCR (Supplementary Table 1). Analysis of path cases (ALS1-15) with the hexanucleotide probe demonstrated the presence of large repeat expansion products (>4000 bp) compared with controls (CON1-6). The expansion products were weaker for ALS5, ALS7, and ALS14. NDO5077 (unaffected control) and ALS15 LCL (expansion positive case) samples were used as internal controls for Southern blots. Analysis of cerebellar tissue with the flanking probe indicated that ALS7 had a reduced modal repeat size of 3000–4200 bp (indicated by black arrowhead). ALS5 had a weaker expansion signal compared with other samples. Abbreviations: PCR, polymerase chain reaction; LCL, lymphoblastoid cell line.

Fig. 2

Detection of an LCL specific product in the Gwent kindred affected case NDO6769 (A). Southern blot of Gwent kindred DNA panel using the DIG-labeled C9ORF72 flanking probe identifies an approximate 60 repeat product in NDO6769 LCL gDNA (indicated with a black arrowhead with normal range alleles marked with a white arrowhead) but not blood extracted gDNA. This product was undetectable with the hexanucleotide probe (B). Abbreviations: CB, cerebellum; DIG, digoxigenin; gDNA, genomic DNA; LCL, lymphoblastoid cell line.

Fig. 3

Quantification of C9ORF72 transcripts in different brain regions. (A) Reported genomic architecture of C9ORF72 with the location of SYBR green qPCR assays. (B) qPCR analysis of frontal cortex samples (10 cases and 5 controls). Data points show linearized and scaled ΔCt values, error bars show standard deviation with mean expression values as horizontal bars. Statistical significance was determined using 1-tailed t tests assuming unequal variance. Cases show a trend of reduced expression across all samples and assays although this is most apparent with exon 1b-containing variant (NM_018325.3). (C) qPCR for exon 1b containing variant 1 (NM_018325.3) shows significantly reduced expression in frontal cortex and cerebellum of expansion carriers (n = 5) compared to expansion negative (n = 8) control cases, as well as a trend toward reduced expression in hippocampus and spinal cord. (D) RT-PCR with forward primers in exon 1a and 1b, respectively, and reverse primer in exon 2 for the detection of variants NM_001256054.1, NM_145005.5, and NM_018325.3 in all examined regions. While NM_018325.3 appears to be more abundant overall, no qualitative difference in the splice variant ratio is apparent between genotypes. Abbreviations: RT-PCR, reverse transcriptase polymerase chain reaction; qPCR, quantitative polymerase chain reaction.

Fig. 4

Analysis of C9orf72 protein levels in postmortem tissue using the custom rabbit polyclonal antibody C9-3721. (A) Immunoblotting of HEK293T lysates detects approximate 48 kDa and 50 kDa endogenous protein species. Knockdown of C9orf72 using a panel of siRNA duplexes (siC91-4) leads reduction of the 48-kDa species (black arrowhead, endo C9). β-actin was used a loading control. (B) Initial immunoblots of postmortem frontal cortex tissue detected variable lower molecular weight (white arrowhead) and approximate 48-kDa protein species (black arrowhead) that obscured the endogenous C9orf72 bands. These suspected GFAP-related products were removed using the pre-absorption method. (C) Immunoblot analysis of HEK293Ts transfected with increasing amounts of a GFAP cDNA construct. C9-3721 showed evidence of cross-reactivity with recombinant GFAP that is removed following pre-absorption using a recombinant GFAP column. (D) Example immunoblots of protein extractions from frontal cortex (11 cases and 6 controls) and cerebellar cortex (12 cases and 6 controls) with recombinant myc-tagged C9orf72 long form blot control (myc-C9 LF, black arrowheads) used for LI-COR quantification. Endogenous C9orf72 long form migrates at approximately 48 kDa (endo C9, white arrowheads). (E) Quantification of C9orf72 protein levels using β-actin as a normalizer for total protein. Cases show a significant reduction in the long form compared with controls in frontal cortex using a 1-tailed t test assuming unequal variance. This was not observed in cerebellar cortex samples. Abbreviations: cDNA, complementary DNA; GFAP, glial fibrillary acidic protein; siRNA, small interfering RNA.