Chromogranin B P413L variant as risk factor and modifier of disease onset for amyotrophic lateral sclerosis

We initially screened the CHGB gene for sequence variations in the entire ORF and the flanking intron/exon borders in 194 individuals, of French or French Canadian origin, with either familial or sporadic ALS and 258 healthy controls of similar age, geography, and ethnicity (Table 1). A total of 21 single nucleotide variations were detected in the CHGB gene, including eight previously unrecorded sequence variations that were not present in the SNP database (Table 2). Fifteen of these 21 sequence variations occurred within coding regions, five were found within introns, and one in the 3′ untranslated region of the gene. Interestingly, all of the coding variations were located in exon 4 and 11 of them led to amino acid changes. Two different bioinformatics protein prediction software (PolyPhen and Pmut) were used to assess whether the 11 missense variations are potentially pathological or lead to protein structure defect (Table S1). Among these 11 nonsynonymous variations, three (R101K, H230R, and R258Q) were not found in 516 control chromosomes. The R101K variation leading to a conservative amino acid change was present in only one ALS patient. The patient had no familial history of neurological disorder and he first presented classic ALS symptoms in the lower-limb at the age of 42 years. The H230R missense variation is predicted to have a functional effect on protein structure (Polyphen: probably damaging, Position-Specific Independent Count or PSIC score of 2.696) and likely a pathological effect (Pmut: prediction score of 0.5510) according to the protein-prediction software. This variation was found in one FALS patient and in one SALS patient, but was absent in control individuals. The non-SOD1 FALS patient first presented with classic ALS in the upper-limb at the age of 56 years. She had another affected brother who had died of ALS at the age of 83 years and, unfortunately, no DNA was available to confirm whether the affected brother was also carrying the variation. No clinical information was available for the SALS patient. The R258Q found in two different FALS cases represents a nonconserved amino acid change leading to the substitution of a polar amino acid for a polar uncharged amino acid. Unfortunately, no additional family members from the two French families were available for further study. The synonymous E169E silent variation was found in two ALS patients and not found in controls.

Of particular interest was the finding of a common variant (P413L) that was significantly increased in ALS patient chromosomes when compared to the control population (P < 0.0001). This variation is predicted to have a functional effect and to alter the protein structure (Polyphen: possibly damaging, PSIC score of 1.741) and also to have a pathological effect (Pmut: prediction score of 0.8075). To explore the significance of the association of the P413L variant with ALS, we screened 95 additional SALS patients and 190 additional nonneurological healthy individuals for this particular variant. Overall, the frequency of this particular SNP was of 17% in ALS patients as compared to 5% in controls [odds ratio (OR) of 3.7 with 95% confidence interval (CI) of 2.2–6.4], conferring a relative risk of 3.3 times greater to develop disease.

To validate our initial results with the French and French Canadian population, we further investigated the occurrence of the R101K, E169E, H230R, R258Q, and P413L sequence variations in the Swedish population (Table 3). The R101K substitution, previously found in one French Canadian SALS case, was not detected in either 832 patient chromosomes or 606 control chromosomes from Sweden. The H230R substitution was initially found in one FALS case from France and one French Canadian SALS patient. It was also present in two other SALS cases from Sweden, one FALS case homozygous for the D90A SOD1 mutation, and one Swedish control individual. The first SALS patient from Sweden carrying the H230R variation suffered from a sporadic progressive bulbar paresis variant of ALS. He developed his first symptom at the age of 61 years. The second SALS case from Sweden with the same variation had typical leg-onset ALS, which he developed at the age of 68 years. Interestingly, this variation was also found in one FALS patient homozygous for the D90A SOD1 mutation. The patient was a Finnish-born man with slowly progressing ALS with leg-onset at the age of 46 years. He also presented some ataxia symptoms, which is not uncommon in patients bearing the D90A SOD1 mutation. This patient had two unaffected older brothers in their sixties, as well as an unaffected son. He therefore was the only affected individual in his family but he is considered as a familial case because of the presence of the SOD1 mutation. Although the H230R variant was also found in one control individual from Sweden, an increased allele frequency is still observed for the substitution in total ALS cases we examined (OR 5.5; 95% CI of 0.6–47.3) (Table 4). The R258Q substitution, found initially in two French FALS cases, was also found in one FALS case from Sweden, and was not detected in control individuals. However, the R258Q variation is unlikely to be pathogenic, as this variation was not detected in three other affected family members from the same Swedish FALS pedigree. The E169E synonymous variation was not detected in Swedish patients or in controls. Although the frequency of the P413L variation in ALS patients from Sweden (5.2%) was not as high as in the French and French Canadian patients (17%), Swedish ALS patients carrying the P413L variation also had a greater risk for ALS (OR 1.5; 95% CI of 0.7–3.1).

Overall, the P413L (g.11748C>T) genetic variation was significantly overrepresented in ALS cases versus controls in the combined populations (OR 2.4; 95% CI 1.5–3.6), conferring a 2.2-fold greater relative risk to develop the disease (P < 0.0001) (see Table 4). It is noteworthy that the frequency of the “T” allele among control individuals reported here (4%) is identical to the published frequency of National Center for Biotechnology Information (NCBI) SNP database obtained from Utah individuals with Northern and Western European ancestry (for more details, see reference number “rs742710” in the NCBI SNP database).

Because the other identified genetic variations found in ALS cases are rare, it remains uncertain that they confer susceptibility to ALS. Nevertheless, when pooled together with the P413L variant, the combined CHGB variants significantly increased the risk of developing ALS by about 2.7 times (95% CI, 1.7–4.4), with 11.2% of total ALS patients carrying one of these variations compared to 4.1% in controls (OR 2.9; 95% CI of 1.7–4.9) (see Table 4).

The effect of the P413L variation on the age of onset was also assessed by deriving Kaplan–Mayer curves. ALS patients carrying the P413L variant were more likely to have an earlier age of onset compared to those without this variation. SALS patients carrying the P413L variation had a median age of onset of 55 years, compared to 62 years for SALS patients without the variation (P = 0.0001) (Fig. 1A). Similarly, FALS patients with the P413L substitution had a median age of onset of 43 years compared to 55 years for FALS patients without the variation (P = 0.01) (Fig. 1B). The mean age of onset from patients with the P413L variation was also significantly lower when compared to patients not carrying this variation (Fig. 1C). SALS patients carrying the P413L variation had a mean age of onset of 53 years, compared to 60 years for SALS patients not carrying the variation (P = 0.008); the mean age of onset from FALS patients bearing the P413L variation was 50 years, compared to 57 years for FALS patients not bearing it (P = 0.05). We also investigated the cosegregation of the P413L variation in two families with other affected individuals for which DNA and clinical information were available (Table 5). The two affected individuals carrying the P413L variation from both families developed disease at the age of 45 and 41 years, as compared to 64 and 63 years for the other affected individuals not carrying the variation in their respective family. These results support the notion that the P413L variation is associated with an earlier age of onset.

To investigate the effect of the H230R and P413L CHGB gene variants, human neuroblastoma cells SHSY-5Y were transiently transfected with expression vectors encoding CHGB WT, CHGB H230R, or CHGB P413L variants fused to enhanced green fluorescent protein (EGFP). The subcellular localization of CHGB-EGFP protein was examined together with the Golgi-marker plasmid DsRed-Golgi, which carries the Golgi-targeting sequence of the human gene encoding 1,4-galactosyl transferase. Confocal laser microscopy revealed that after 48 h only a fraction of the WT CHGB-EGFP colocalized (about 40%) with the Golgi marker (Fig. 2). This result is consistent with the normal sorting of the protein to the secretory granules after leaving the endoplasmic reticulum (ER) transGolgi network. In contrast, the H230R and P413L CHGB-EGFP variants were almost entirely sequestered in the ER/transGolgi network (see Fig. 2), suggesting a defective sorting and maturation of these ALS-linked CHGB variants into secretory granules.

All patients were assessed by expert clinicians and gave written informed consent. Diagnosis of ALS was made according to El Escorial ALS diagnostic criteria (23). Peripheral blood from each patient from France and Canada/Quebec (non-SOD1 FALS = 40; SALS = 58) and healthy controls matched for age and ethnicity (n = 258) were collected for the initial screen (see Table 1). Another set of samples from 191 nonrelated SALS cases from France and 190 matched controls were also examined when it was not possible to confirm or exclude association between CHGB and ALS. Blood from patients and controls of Swedish origin were also collected and used to validate our initial results (non-SOD1 FALS = 61; SOD1 FALS = 40; SALS = 315; Control = 303) (see Table 1). Overall, the non-SOD1 FALS cohort exhibited a mean age of onset of 56 years, a disease duration of 4.4 years, and a male-to-female ratio of 1:1. The mean age of onset of the SOD1-linked FALS cohort was 49 years, exhibited a disease duration of 11.3 years and a male-to-female ratio of 1.5:1. The SALS cohort exhibited a mean age of onset of 59 years, a disease duration of 3.4 years, and a male-to-female ratio of about 1.5:1.

DNA was extracted from whole blood and transformed lymphoblastoid cell lines by using standard procedures. Respectively, 12 and 9 PCR primer pairs were designed from genomic DNA to amplify each exon of the CHGB genes, including the flanking splice sites (Table S2). Products were PCR-amplified, checked on agarose gels, and then sequenced by using the forward primer for all of the amplicons. The PCR fragment containing variations were also sequenced on the reverse strands. Each of the variants found were also genotyped on our control populations by allele-specific oligomerization, as previously described (24). Variants that were not found in control individuals or found to be increased in ALS patients were also sequenced in our extra set of SALS samples, as well as in our set of Swedish patients and controls to validate our results. The P413L variation for all extra patients and controls were genotyped by restriction enzyme digestion. The P413L variation abolished an MspI restriction site. PCR fragments amplified from DNA of extra patients and controls were digested with MspI (New England BioLabs) and run on agarose gel to confirm the presence or absence of the P413L variation. PCR primer pair were designed to amplify a 500-bp PCR product in which the P413L variation is located in the middle of the PCR fragment (5′-aacgtcagcatggccagtttag-3′ and 3′-gagggtcgtagtatgggttgaaca-5′).

Deviations from Hardy–Weinberg equilibrium were tested by using the observed allele frequency by the χ² contingency test (1 degree of freedom). We assessed differences in distribution of alleles and genotypes between the groups (SALS, FALS, and control) by calculating the χ² contingency test using the Fisher's exact test, from which we derived P-values with one degree of freedom. P-values less than 0.04 were considered statistically significant. Odds ratios were also derived only when significant P-values were found. Median age of onset and P-values derived from the survival curves among patients were calculated by using the log-rank Mantel–Cox test with Graph Pad Prim5 software. Unpaired t test (mean ± standard error) was used to calculate and compare the mean age of onset for each group.

The PolyPhen program (http://genetics.bwh.harvard.edu/pph) and Pmut (http://mmb2.pcb.ub.es:8080/PMut) were used to predict the functional significance of nonsynonymous cSNPs. The PolyPhen method uses sequence, phylogenetic, and structural information to predict the effect of a particular amino acid change. Sequence alignment is used to generate a quantitative PSIC score and this is combined with any available sequence annotation or structural information to generate a qualitative prediction of a SNP as “benign” (PSIC score <0.5), “possibly damaging” (0.5 <PSIC score interval <2.0) and “probably damaging” (PSIC score >2.0) (25). A PSIC score > 0.5 associated with a benign prediction indicates that the studied substitution is rarely or never observed in the protein family. Pmut exploits a computational approach to the prediction of disease-associated amino acid mutations, using only sequence-based information (amino acid properties, evolutionary information, secondary structure and accessibility predictions, and database annotations). Mutations are predicted to be either pathological (score >0.5) or neutral (score <0.5) (26).

Mammalian expression vector for human chromogranin B tagged with EGFP (pEGFP-N3-CHGB) was generated as previously described (8). Two CHGB H230R and P413L sequence variations were introduced within the pEGFP-N3-CHGB plasmid by site-directed mutagenesis according to the manufacturer protocol (Agilent technologies, Stratagene division). PCR products obtained from WT pEGFP-N3-CHGB were digested by DpnI and subsequently transformed into Escherichia coli DH5alpha competent cells. The PCR primer pairs used for the site directed mutagenesis were: 5′-ttctcaggagaagacacgtagccgagagaagagta-3′ and 5′-tactcttctctcggctacgtgtcttctcctgagaa-3′ for H230R; 5′-ggggccttgagctgggaaagggacg-3′and 5′-cgtccctttcccagctcaaggcccc-3′ for P413L. For cell transfections, human neuroblastoma cell-line SHSY-5Y was cultured in DMEM containing 15% FBS, nonessential amino acids (Invitrogen) and antibiotic (penicillin/streptomycin; Invitrogen) at 37 °C in 5% CO₂ and 100% humidified condition. Cells were transfected by using Lipofectamine plus (Invitrogen) according to the manufacturer's protocol and were counted under epifluorescence microscope. Statistical analysis was done by using one-way ANOVA test and P values were then derived.