Glucocerebrosidase Gene Mutations and Parkinson Disease

Stuart K. Shapira, M.D., Ph.D.
National Center on Birth Defects and Developmental Disabilities
Centers for Disease Control and Prevention

In November, 2004, Aharon-Peretz et al. reported that mutations in the glucocerebrosidase (GBA) gene increased the susceptibility for developing Parkinson Disease.

Aharon-Peretz J et al. Mutations in the glucocerebrosidase gene and Parkinson Disease in Ashkenazi Jews.
N Engl J Med. 2004;351:1972-7.

These investigators performed a case-control study of 99 patients with Parkinson Disease, 74 patients with Alzheimer's disease, and 1543 healthy control subjects for 6 mutations in the GBA gene. All patients and controls were of Ashkenazi Jewish ancestry by family history. Parkinson Disease was diagnosed according to the United Kingdom brain-bank criteria; patients underwent a physical, neurobehavioral, and neurologic examination that incorporated the Unified Parkinson Disease Rating Scale. Alzheimer's disease patients, serving as a comparison group, met the criteria for dementia of the Alzheimer's type according to the Diagnostic and Statistical Manual of Mental Disorders, 4th edition, and the criteria for probable Alzheimer's disease of the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer's Disease and Related Disorders Association. Healthy Ashkenazi Jews serving as a control group were from the same geographic area; the healthy controls were undergoing testing to identify heterozygosity for certain recessive diseases. The Parkinson Disease group included 55 men and 44 women, and the Alzheimer's disease group included 42 men and 32 women. No further information was provided about the cases or controls.

 

Question 1: What other information about the cases and controls would be important to know? Are there any specific concerns about the healthy controls selected for this study?

 

The N370S and 84GG alleles are the most frequent mutations in the GBA gene among Ashkenazi Jews, with carrier rates of 1 in 17.5 for N370S and 1 in 400 for 84GG; these alleles are associated with mild and severe Gaucher disease, respectively. Four other rare GBA mutations identified in Ashkenazi Jews are L444P, IVS2+1G-->A, V394L, and R496H. All 99 patients with Parkinson Disease, all 74 patients with Alzheimer's disease, and all 1543 healthy controls were evaluated for these 6 mutations. Among the 1543 control subjects, 95 were heterozygous for a Gaucher disease-causing mutation (6.2%). None of the control subjects were homozygous or compound heterozygous for a mutation. Of the carriers, 92 were heterozygous for N370S and 3 were heterozygous for 84GG; these results were consistent with a carrier rate of 1 in 16.7 for the N370S mutation and 1 in 514 for the 84GG mutation.

 

Question 2: Are the observed carrier frequencies for the N370S and 84GG mutations significantly different from the carrier estimates reported in the literature?

The authors defined a “carrier" for Gaucher disease as a person who was either heterozygous for a GBA mutation, homozygous for a mutation, or a compound heterozygote for two different GBA mutations. The genotypes that were found for the “carriers” in each study group are shown:

GBA genotypes in Parkinson Disease patients and controls*

Genotype	Parkinson Disease	Alzheimer's Disease	Healthy Controls
+/N370S	23	2	92
N370S/N370S	3	0	0
+/84GG	4	1	3
+/R496H	1	0	0
Total “Carriers”	31	3	95

*Adapted from Aharon-Peretz et al.
+ = wild type or non-mutant allele

 

Question 3: Is the definition that the authors proposed for a “carrier” for Gaucher disease appropriate for this analysis?

The rates of “carriage of Gaucher disease” (heterozygotes plus homozygotes for any mutation) were calculated for patients with Parkinson Disease, Alzheimer's disease, and control subjects:

Rates of carriage of GBA mutations among patients with Parkinson Disease and Controls*

Patient Populations	Number Tested	Number of “Carriers” (%)	95% Confidence Interval
Parkinson Disease	99	31 (31.3%)	22.2-40.4
Alzheimer's Disease	74	3 (4.1%)	0.0-8.5
Healthy Controls	1543	95 (6.2%)	5.0-7.4

*Adapted from Aharon-Peretz et al.

 

Question 4: Based on these findings, what is the odds ratio (estimated relative risk) for Parkinson Disease among “carriers” of Gaucher disease? Is the rate of “carriage” of a GBA mutation among patients with Alzheimer's disease increased compared with healthy controls?

The authors' overall conclusions from the study were (1) the prevalence of GBA mutations in the population of Ashkenazi Jews with Parkinson Disease by far outweighs the reported prevalence of mutations in other susceptibility genes, such as parkin and SNCA, and (2) mutations in the GBA gene emerge as a strong genetic determinant predisposing people to Parkinson Disease.
 

Question 5: What additional analysis of the data should be performed in order to determine the validity of the authors' second conclusion?

In order to determine the GBA genotypic contribution toward the development of Parkinson Disease, the Parkinson Disease and healthy control groups are compared as shown:

Genotype	Parkinson Disease	Healthy Controls	OR	95% Confidence Interval	p value
+/+	68	1448	ref
+/N370S	23	92	5.32	3.19-8.90	< 0.001
N370S/N370S	3	0
+/84GG	4	3	28.39	6.96-115.65	< 0.001
Total	98	1543

 

Question 6: Since there were no N370S homozygotes among the healthy controls, how can the odds ratio be calculated for this genotype?

The observed number of healthy controls with each genotype is compared to the expected number of controls, based on Hardy-Weinberg equilibrium:

Genotype	Observed	Expected
+/+	1448	1449.4622
+/N370S	92	89.1678
N370S/N370S	0	1.3713
+/84GG	3	2.9076
N370S/84GG	0	0.0894
84GG/84GG	0	0.0014

 

Question 7: Are the observed numbers for the genotypes that comprise the healthy controls significantly different from Hardy-Weinberg equilibrium? Is it valid to have used 1.3713 as the number of healthy controls with the N370S/N370S genotype in order to calculate the odds ratio?

The odds ratios for developing Parkinson Disease, the 95% confidence intervals, and the p values are shown for each GBA gene genotype:

Genotype	Parkinson Disease	Healthy Controls	OR	95% Confidence Interval	p value
+/+	68	1448	ref
+/N370S	23	92	5.32	3.19-8.90	< 0.001
N370S/N370S	3	0	46.59	7.74-279.09	< 0.001
+/84GG	4	3	28.39	6.96-115.65	< 0.001
Total	98	1543

 

Question 8: How strong is the association between each of the GBA gene genotypes and Parkinson Disease?

In order to quantify the contribution of the various GBA gene genotypes to the occurrence of Parkinson Disease at the population level, one can calculate the associated attributable fraction utilizing the formula of Miettinen:

Attributable fraction = f_c (R – 1) / R

Where f_c is the fraction of cases with the risk factor and R is the measure of relative risk (or odds ratio for rare diseases). The following table summarizes the calculation of attributable fraction for each genotype:

Genotype	fc	R	Attributable Fraction
+/N370S	23/98	5.32	19.1%
N370S/N370S	3/98	46.59	3.0%
+/84GG	4/98	28.39	3.9%

 

In order to estimate the attributable fraction from these data, it was necessary to assume that the cases and controls were comparable and population-based.

Question 9: Are the results of this analysis valid? If the result is correct, what is then suggested about the role of the GBA gene genotypes in Parkinson Disease at the population level?

The estimated absolute risk (incidence) of Parkinson Disease by genotype can be determined using the estimated frequencies of each genotype, and summing the genotype specific incidence rates in order to calculate the absolute risk. The allele and genotype frequencies are calculated as follows (allele frequencies derived from the population genetics expressions of p + q + r = 1 and [(p x p) + (2p x q) + (q x q) +

(2p x r) + (2q x r) + (r x r) = 1]; these formulas assume that both the N370S and 84GG alleles are in Hardy-Weinberg equilibrium, which was shown previously to be the case):

Allele Frequency	Mutation	N370S	84GG	None (+)
.02981	N370S	.00089	.00003	.02889
.00097	84GG	.00003	.0000009	.00094
.96922	None (+)	.02889	.00094	.93939

+ = wild type or non-mutant allele

Genotype Frequencies:
+/+ = 0.93939
+/N370S = 0.02889 x 2 = 0.05778
+/84GG = 0.00094 x 2 = 0.00188
N370S/N370S = 0.00089

 

Question 10: Based on these genotype frequencies, and assuming that the total population risk of Parkinson Disease is 1 per 100 people (1%), what is the estimated absolute risk of Parkinson Disease by genotype?

The calculation of absolute risk of Parkinson Disease by genotype showed that there are significant risks for developing Parkinson Disease, over the baseline population risk of 1%, for both N370S and 84GG carriers (4.0% and 21.1%, respectively), as well as for N370S homozygotes (34.7%). However, it is crucial to determine whether these are important risks from a population-based public health perspective.

Question 11: Would it beneficial to perform population-based screening of Ashkenazi Jews for the N370S mutation--the mutation with the highest incidence in the population--in order to identify those persons with an increased risk of developing Parkinson Disease?