Sortilin-related VPS10 domain containing receptor 1 and Alzheimer's disease-associated allelic variations preferentially exist in female or type 2 diabetes mellitus patients in southern Han Chinese

Authors


Dr Guran Yu MD, Department of Neurology, Jiangsu Traditional Chinese Medicine Hospital, the Affiliated Hospital of Nanjing University of Traditional Chinese Medicine, Nanjing, Jiangsu Province 210029, China. Email: yushengzh@126.com

Abstract

Background:  Both in vitro and in vivo, overexpression of the sortilin-related VPS10 domain containing receptor 1 (SORCS1) protein lowers amyloid-β generation. Recent studies have shown that SORCS1 variations in intron 1 are associated with sporadic Alzheimer's disease (SAD), but the results remain inconsistent.

Methods:  In order to clarify the role of the SORCS1 gene in southern Han Chinese, we genotyped eight single nucleotide polymorphisms (SNP) of SORCS1 in 128 SAD patients and 92 healthy controls.

Results:  By dividing patients and controls according to apolipoprotein status, sex and whether they had type 2 diabetes mellitus, we found that rs7907690 C allele frequencies were significantly higher in the Alzheimer's disease (AD) patients with type 2 diabetes mellitus than in the controls (P= 0.041). Also, the rs600879 GG genotype and G allele worked as protective factors of SAD in women (GG genotype, P= 0.007; G allele, P= 0.009). In multilocus analysis, the frequency of an eight-single nucleotide polymorphism rs601883/rs7907690/rs600879/rs17277986/rs2900717/rs10884399/rs11193170/rs4918280 CCGGACGG haplotype was significantly higher in AD patients (6.3%), especially in female AD patients (9.5%), than in the controls (0.5%) (P= 0.003; P= 0.0002). However, the CTGGACGG haplotype was significantly lower in AD patients (9.3%) than in controls (20.3%) (P= 0.001). The association remained significant even after Bonferroni correction for the number of haplotypes.

Conclusion:  This study provides evidence that variations in the SORCS1 gene influence susceptibility to SAD in southern Han Chinese. The genetic link between AD and SORCS1 gene variations are influenced by ethnic background, sex and whether an individual has type 2 diabetes mellitus.

INTRODUCTION

Alzheimer's disease (AD) is the most common form of dementia. Mutations in amyloid precursor protein, presenilin 1 and presenilin 2 genes are involved in familial early-onset AD, which represents less than 1% of all AD cases.1 The great majority of AD cases are sporadic and late onset (age at onset >65 years). Only the apolipoprotein (APOE)ε4 allele is conclusively associated with increased genetic susceptibility to sporadic Alzheimer's disease (SAD).

As the APOEε4 allele only accounts for 20–25% of late onset AD, identification of other risk factors is of great importance. The sortilin-related VPS10 domain containing receptor 1 (SORCS1) belongs to the sortilin family of vacuolar protein sorting-10 domain containing proteins. It is a substrate of γ-secretase, which cuts amyloid precursor protein and generates amyloid-β (Aβ) peptide. It has been proved that in vitro and in vivo overexpression of SORCS1 protein lowers Aβ generation.2,3

Additionally, genetic studies suggest that variations in SORCS1 affect the risk of AD, especially in women. Among the single nucleotide polymorphisms (SNP) studies, rs600879 is the initial SNP that has been found to be associated with AD.4 Similarly, rs601883 and rs7907690 have also been found to affect AD risk,5 and other SNP, including rs17277986, rs2900717, rs10884399, rs11193170 and rs4918280 polymorphisms, have been found to significantly change the risk of AD in women. All the SNP affecting AD risk are in intron 1 of SORCS1, which is thought to be an important functional element in the gene.6 However, the results of association studies between the SNP in intron 1 and SAD have not been consistent.7–9

Furthermore, SORCS1 resides at a quantitative trait locus for type 2 diabetes mellitus (T2DM) in mice and rats.10,11 The SNP of SORCS1 also affect the risk of T2DM in humans, especially in women.12 Therefore, SORCS1 is genetically linked to both diabetes and AD.

To our knowledge, there have been no reports about SORCS1 gene polymorphism in Chinese patients with AD until now. In present study, we genotyped eight SNP (rs601883, rs7907690, rs600879, rs17277986, rs2900717, rs10884399, rs11193170 and rs4918280) in a southern Han Chinese population and studied the association between the variations and SAD. By dividing patients and healthy controls according to APOE status, sex and whether they had T2DM, we found that rs7907690 C allele frequencies were significantly higher in the AD with T2DM patients than in the controls. The rs600879 GG genotype and G allele worked as protective factors of SAD in women.

MATERIALS AND METHODS

Subjects

The study group consisted of 128 SAD patients (77 men and 51 women; mean age ± SD, 76.34 ± 7.49 years; mean age at onset ± SD, 74.86 ± 8.00 years). The patients were recruited from the Jiangsu Traditional Chinese Medicine Hospital (Nanjing, China). Probable AD was clinically diagnosed according to the criteria of National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer's Disease and Related Disorders Association. All patients with evidence of an autosomal dominant AD trait were excluded. The control group consisted of 92 healthy subjects (58 men and 34 women; mean age ± SD, 74.81 ± 6.87 years). Controls were recruited from the outpatient department of Jiangsu Traditional Chinese Medicine Hospital, where they underwent regular health examinations, and were confirmed to be healthy and neurologically fit by Mini-Mental State Examination, Hasegawa's Dementia Scale and general examinations. All subjects were representative of southern Han Chinese populations (‘southern’ defined as from south of the Yellow River). Informed consent was obtained from each subject or the subject's guardian. The protocol of this study was approved by the Institutional Ethics Committee at the Jiangsu Traditional Chinese Medicine Hospital.

Genotype analysis

Genomic DNA was extracted via routine blood sample from each subject. Direct sequencing wase performed in 20 individuals. The SNP rs601883, rs7907690, rs600879, rs17277986, rs2900717, rs10884399, rs11193170 and rs4918280 in the SORCS1 gene and rs429358 and rs7412 in the APOE gene were genotyped by ligase detection reaction (Amersham Biosciences, Piscataway, NJ, USA).13 In brief, a DNA fragment containing the polymorphic site was amplified by polymerase chain reaction in a PerkinElmer GeneAmp 9600 machine (Foster City, CA, USA), with primer pairs listed in Table 1. The amplicon size was 231 bp for rs601883, 230 bp for rs7907690, 197 bp for rs600879, 230 bp for rs17277986, 216 bp for rs2900717, 234 bp for rs10884399, 224 bp for rs11193170, 180 bp for rs4918280, 250 bp for rs429358 and 358 bp for rs7412. Then ligase detection reaction (35 cycles of 95°C 2 min, 94°C 30 s, 50°C 2 min) was performed in a final volume of 10 µL, which contained 1-µL 1 × New England Biolabs buffer, 1-µL 12.5 pmol/µL of each of the ligase detection reaction mixed probes, 1-µL 100-ng/µL polymerase chain reaction product, 0.05-µL (2-U/µL) Taq DNA Ligase (New England Biolabs, Ipswich, MA, USA) and 6.95-µL double-distilled water. The fluorescent products of the ligase detection reaction were differentiated by an ABI 377 sequencer (PerkinElmer). The result was analyzed by Genemapper Analysis software 3.7 (Applied Biosystems, Foster City, CA, USA).

Table 1. The primer and probe sequences
SNPPrimer sequence (5′–3′)Probe sequence (5′–3′)
  1. SNP, single nucleotide polymorphisms.

rs601883Forward: TCTTAACGGAACATCTGGCATACommon probe: P-AGAAACAGAAACAACATCATGAATGTTTTTTTTTTTTTTTT-FAM
 Reverse: GAGTTGCTGACAGGGATGCTOligo C: TTTTTTTTTTTTTTTTAACTGGACCGAGAGAACCTAAATGG
  Oligo G: TTTTTTTTTTTTTTTTTTAACTGGACCGAGAGAACCTAAATGC
rs7907690Forward: CACTGTGGAGCTGGTTCTGACommon probe: P-ATCCAGAGGTCGTCTCAAATCCAGCTTTTTTTTTTTTTTTTTTTTTT-FAM
 Reverse: ATCTTGCCACTGCTGGAAACOligo A: TTTTTTTTTTTTTTTTTTTTTTTTGACATTTTGCATGACTTCCCTAT
  Oligo G: TTTTTTTTTTTTTTTTTTTTTTTTTTGACATTTTGCATGACTTCCCTAC
rs600879Forward: GGGAGAACTGCAGTGGAGAGCommon probe: P-GCTGTTAGAAGATGCTGGAGAGTTCTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT-FAM
 Reverse: GATCTTCCTGGGACTGACCAOligo A: TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTAGTTCTCAAAACACCCTCGCTTAGT
  Oligo G: TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTAGTTCTCAAAACACCCTCGCTTAGC
rs17277986Forward: GCAAAGACAAGAAATGAGCAACommon probe: P-CCAAACTTCTGTAGTCAGCAACAAATTTTTTTTTTTTTTTTTT-FAM
 Reverse: TAATGCAGGCTCTTGGAAGGOligo C: TTTTTTTTTTTTTTTTTTCTCAGATTCCAAGAATTATTCAGCG
  Oligo T: TTTTTTTTTTTTTTTTTTTTCTCAGATTCCAAGAATTATTCAGCA
rs2900717Forward: AAAAGACAGCCCTGGGTCACommon probe: P-TTTGGCCTTGCTCCTATTACACAGATTTTTTTTTTTTTTTTTTTTTTTTTT-FAM
 Reverse: TGTTAAAATTGGCTCTGTAACCAOligo C: TTTTTTTTTTTTTTTTTTTTTTTTTTTCACCCTGCCACCATACATACTCTG
  Oligo T: TTTTTTTTTTTTTTTTTTTTTTTTTTTTTCACCCTGCCACCATACATACTCTA
rs10884399Forward: AGGTCGAGGAAACAGCATTCCommon probe: P-TGCCAAAAGCCAAGGGGCAGTGTTCTTTTTTTTTTTTTTTTTTTTTTTTTTTT-FAM
 Reverse: TGTGACCCACATTACCCAACOligo A: TTTTTTTTTTTTTTTTTTTTTTTTTTTTGAATTTTAGTTTTGTGAAGGGAGCT
  Oligo G: TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTGAATTTTAGTTTTGTGAAGGGAGCC
rs11193170Forward: TCAGGTAGGCTTAGTCTTGATACCCommon probe: P-TGAAAAAGTTTGAAATTTCATTGAATTTTTTTTTTTTTTTTTTTTTTTT-FAM
 Reverse: CAGCTTCCACCAATCACGTAOligo A: TTTTTTTTTTTTTTTTTTTTTTTTCCATGTAAATAATTGATCAGGCTAT
  Oligo C: TTTTTTTTTTTTTTTTTTTTTTTTTTCCATGTAAATAATTGATCAGGCTAG
rs4918280Forward: GGAGATGCCAACTCTTCTGCCommon probe: P-ATGGTTTCTCTAAGAACGGTAATGGTTTTTTTTTTTTTTTTTTTT-FAM
 Reverse: CTGCATGAATACCCTCCACAOligo C: TTTTTTTTTTTTTTTTTTTTACTGTGAACAGTAGTTTTTAATCTG
  Oligo T: TTTTTTTTTTTTTTTTTTTTTTACTGTGAACAGTAGTTTTTAATCTA
rs429358Forward: CGGAACTGGAGGAACAACTGCommon probe: P-CACGTCCTCCATGTCCGCGCCCAGCTTTTTTTTTTTTTTTTTTTT-FAM
 Reverse: GCGCTTCTGCAGGTCATCOligo C: TTTTTTTTTTTTTTTTTTTTCGCGGTACTGCACCAGGCGGCCGCG
  Oligo T: TTTTTTTTTTTTTTTTTTTTTTCGCGGTACTGCACCAGGCGGCCGCA
rs7412Forward: AGGGTGCTGATGGACGAGACCATGAAGGAGCommon probe: P-CTTCTGCAGGTCATCGGCATCGCGGTTTTTTTTTTTTTTTTTTTTTT-FAM
 Reverse: GCTCACGGATGGCGCTGAGGCCGCGCTCGGOligo C: TTTTTTTTTTTTTTTTTTTTTTCCCCGGCCTGGTACACTGCCAGGCG
  Oligo T: TTTTTTTTTTTTTTTTTTTTTTTTCCCCGGCCTGGTACACTGCCAGGCA

Statistical analysis

Hardy–Weinberg equilibrium was tested using the program in http://analysis.bio-x.cn/myAnalysis.php (Bio-X Center, Shanghai Jiao Tong University, Shanghai, China).14,15 Allele and genotype distributions in patients and controls were compared using χ2 test in SPSS v. 17.0 (SPSS Inc., Chicago, IL, USA). Linkage disequilibrium values (D' and r2) and estimation of haplotypes were performed on http://analysis.bio-x.cn/myAnalysis.php. The strength of association between alleles or genotypes and SAD was evaluated with the odds ratio (OR) presented with 95% confidence intervals (CI). P < 0.05 was considered to be statistically significant in all tests.

RESULT

We detected eight SNP in a southern Han Chinese sample (128 cases and 92 controls). For each polymorphism analyzed, the genotype distribution accorded with the Hardy–Weinberg equilibrium. In the single-locus analysis, none of the SNP showed a significant difference in either genotype or allele frequencies between the AD patients and healthy controls. When these data were stratified by APOEε4, no significant difference was observed either. In the AD with T2DM patients, rs7907690 C allele frequencies were significantly higher than in the controls (P= 0.041). (Table 2). Some previous reports have shown a strong effect of SORCS1 in female patients, so we analyzed the SNP in female AD patients and male controls. rs600879 GG genotype and G allele were significantly lower in female AD patients than in male controls (P= 0.007; P= 0.009) (Table 3). When we compared the female AD patients with all controls, rs600879 GG genotype and G allele still showed a lower trend in female AD patients (GG genotype, P= 0.067; G allele, P= 0.077) (Table 4).

Table 2. Distributions of the polymorphisms in Alzheimer's patients and controls
 GenotypeAllele
TotalGG (%)CG (%)CC (%) P-valueG (%)C (%) P-value
rs601883        
 AD12867 (52.3)52 (40.6)9 (7.0)0.565186 (72.7)70 (27.3)0.479
 AD with T2DM4824 (50.0)18 (37.5)6 (12.5)0.33266 (68.8)30 (31.2)0.889
 Control9242 (45.7)44 (47.8)6 (6.5) 128 (69.6)56 (30.4) 
APOE (−)        
 AD9447 (50.0)41 (43.6)6 (6.4)0.809135 (71.8)53 (28.2)0.551
 Control8237 (45.1)39 (47.6)6 (7.3) 113 (68.9)51 (31.1) 
APOE (+)        
 AD3420 (58.8)11 (32.4)3 (8.8)0.43851 (75.0)17 (25.0)1.000
 Control 105 (50.0)5 (50.0)0 (0) 15 (75.0)5 (25.0) 
 GenotypeAllele
TotalTT (%)TC (%)CC (%) P-valueT (%)C (%) P-value
rs7907690        
 AD12895 (74.2)29 (22.7)4 (3.1)0.496219 (85.5)37 (14.5)0.270
 AD with T2DM4830 (62.5)17 (35.4)1 (2.1)0.10077 (80.2)19 (19.8) 0.041
 Control9273 (79.3)18 (19.6)1 (1.1) 164 (89.1)20 (10.9) 
APOE (−)        
 AD9470 (74.5)23 (24.5)1 (1.1)0.839163 (86.7)25 (13.3)0.628
 Control8264 (78.0)17 (20.7)1 (1.2) 145 (88.4)19 (11.6) 
APOE (+)        
 AD3425 (73.5)6 (17.6)3 (8.8)0.48756 (82.4)12 (17.6)0.161
 Control 109 (90.0)1 (10.0)0 (0) 19 (95.0)1 (5.0) 
 GenotypeAllele
TotalGG (%)GA (%)AA (%) P-valueG (%)A (%) P-value
rs600879        
 AD128108 (84.4)20 (15.6)0 (0)0.143236 (92.2)20 (7.8)0.158
 AD with T2DM4841 (85.4)7 (14.6)0 (0)0.30689 (92.7)7 (7.3)0.320
 Control9282 (91.1)8 (8.9)0 (0) 172 (95.6)8 (4.4) 
APOE (−)        
 AD9482 (87.2)12 (12.8)0 (0)0.382176 (93.6)12 (6.4)0.396
 Control8274 (91.4)7 (8.6)0 (0) 155 (95.7)7 (4.3) 
APOE (+)        
 AD3426 (76.5)8 (23.5)0 (0)0.41560 (88.2)8 (11.8)0.444
 Control 108 (88.9)1 (11.1)0 (0) 17 (94.4)1 (5.6) 
 GenotypeAllele
TotalGG (%)GA (%)AA (%) P-valueG (%)A (%) P-value
rs17277986        
 AD12890 (70.3)35 (27.3)3 (2.3)0.786215 (84.0)41 (16.0)0.701
 AD with T2DM4834 (70.8)13 (27.1)1 (2.1)0.89481 (84.4)15 (15.6)0.832
 Control9266 (71.7)25 (27.2)1 (1.1) 157 (85.3)27 (14.7) 
APOE (−)        
 AD9466 (70.2)25 (26.6)3 (3.2)0.675157 (83.5)31 (16.5)0.750
 Control8258 (70.7)23 (28.0)1 (1.2) 139 (84.8)25 (15.2) 
APOE (+)        
 AD3424 (70.6)10 (29.4)0 (0)0.55758 (85.3)10 (14.7)0.590
 Control 108 (80.0)2 (20.0)0 (0) 18 (90.0)2 (10.0) 
 GenotypeAllele
TotalAAGAGG P-valueAG P-value
rs2900717        
 AD12878 (60.9)45 (35.2)5 (3.9)0.967201 (78.5)55 (21.5)0.941
 AD with T2DM4829 (60.4)18 (37.5)1 (2.1)0.74276 (79.2)20 (20.8)0.943
 Control9257 (62.0)31 (33.7)4 (4.3) 145 (78.8)39 (21.2) 
APOE (−)        
 AD9455 (58.5)34 (36.2)5 (5.3)0.812144 (76.6)44 (23.4)0.547
 Control8251 (62.2)28 (34.1)3 (3.7) 130 (79.3)34 (20.7) 
APOE (+)        
 AD3423 (67.6)11 (32.4)0 (0.00)0.17557 (83.8)11 (16.2)0.368
 Control 106 (60.0)3 (30.0)1 (1.00) 15 (75.0)5 (25.0) 
 GenotypeAllele
TotalCCTCTT P-valueCT P-value
rs10884399        
 AD12863 (49.2)53 (41.4)12 (9.4)0.318179 (69.9)77 (30.1)0.591
 AD with T2DM4824 (50.0)19 (39.6)5 (10.4)0.63467 (69.8)29 (30.2)0.661
 Control9252 (56.5)29 (31.5)11 (12.0) 133 (72.3)51 (27.7) 
APOE (−)        
 AD9448 (51.1)34 (36.2)12 (12.8)0.823130 (69.1)58 (30.9)0.747
 Control8245 (54.9)26 (31.7)11 (13.4) 116 (70.7)48 (29.3) 
APOE (+)        
 AD3415 (44.1)19 (55.9)0 (0)0.15049 (72.1)19 (27.9)0.240
 Control 107 (70.0)3 (30.0)0 (0) 17 (85.0)3 (15.0) 
 GenotypeAllele
TotalGGGTTT P-valueGT P-value
rs11193170        
 AD12863 (49.2)52 (40.6)13 (10.2)0.356178 (69.5)78 (30.5)0.544
 AD with T2DM4824 (50.0)19 (39.6)5 (10.4)0.60567 (69.8)29 (30.2)0.671
 Control9251 (56.7)28 (31.1)11 (12.2) 130 (72.2)50 (27.8) 
APOE (−)        
 AD9448 (51.1)33 (35.1)13 (13.8)0.817129 (68.6)59 (31.4)0.630
 Control8245 (55.6)25 (30.9)11 (13.6) 115 (71.0)47 (29.0) 
APOE (+)        
 AD3415 (44.1)19 (55.9)0 (0)0.22949 (72.1)19 (27.9)0.330
 Control 106 (66.7)3 (33.3)0 (0) 15 (83.3)3 (16.7) 
 GenotypeAllele
TotalGGGAAA P-valueGA P-value
  1. A, adenine; AD, Alzheimer's disease; APOE, apolipoprotein; APOE (+), at least one APOEε4 allele carrier; APOE (−), non-APOEε4 allele carrier; C, cytosine; G, guanine; T, thymine; T2DM, type 2 diabetes mellitus.

rs4918280        
 AD12861 (47.7)54 (42.2)13 (10.2)0.442176 (68.8)80 (31.2)0.581
 AD with T2DM4823 (47.9)21 (43.8)4 (8.3)0.47467 (69.8)29 (30.2)0.806
 Control9250 (54.3)31 (33.7)11 (12.0) 131 (71.2)53 (28.8) 
APOE (−)        
 AD9447 (50.0)35 (37.2)12 (12.8)0.913129 (68.6)59 (31.4)0.856
 Control8243 (52.4)28 (34.1)11 (13.4) 114 (69.5)50 (30.5) 
APOE (+)        
 AD3414 (41.2)19 (55.9)1 (2.9)0.26247 (69.1)21 (30.9)0.161
 Control107 (70.0)3 (30.0)0 (0) 17 (85.0)3 (15.0) 
Table 3. Distributions of the polymorphisms in female Alzheimer's patients and male controls
 GenotypeAllele
TotalGG (%)CG (%)CC (%) P-valueG (%)C (%) P-value
rs601883        
 Female AD5121 (41.2)27 (52.9)3 (5.9)0.80269 (67.6)33 (32.4)0.627
 Male control 5826 (44.8)30 (51.7)2 (3.4) 82 (70.7)34 (29.3) 
 GenotypeAllele
TotalTT (%)TC (%)CC (%) P-valueT (%)C (%) P-value
rs7907690        
 Female AD5139 (76.5)10 (19.6)2 (3.9)0.78088 (86.3)14 (13.7)0.715
 Male control 5845 (77.6)12 (20.7)1 (1.7) 102 (87.9)14 (12.1) 
 GenotypeAllele
TotalGG (%)GA (%)AA (%) P-valueG (%)A (%) P-value
rs600879        
 Female AD5041 (80.4)10 (19.6)0 (0)0.00792 (90.2)10 (9.8)0.009
 Male control 5756 (96.6)2 (3.4)0 (0) 114 (98.3)2 (1.7) 
 GenotypeAllele
TotalGG (%)GA (%)AA (%) P-valueG (%)A (%) P-value
rs17277986        
 Female AD5136 (70.6)15 (29.4)0 (0)0.60287 (85.3)15 (14.7)0.992
 Male control 5842 (72.4)15 (25.9)1 (1.7) 99 (85.3)17 (14.7) 
 GenotypeAllele
TotalAA (%)GA (%)GG (%) P-valueA (%)G (%) P-value
rs2900717        
 Female AD5128 (54.9)21 (41.2)2 (3.9)0.54077 (75.5)25 (24.5)0.500
 Male control 5837 (63.8)18 (31.0)3 (5.2) 92 (79.3)24 (20.7) 
 GenotypeAllele
TotalCC (%)TC (%)TT (%) P-valueC (%)T (%) P-value
rs10884399        
 Female AD5128 (54.9)19 (37.3)4 (7.8)0.83575 (73.5)27 (26.5)0.966
 Male control 5833 (56.9)19 (32.8)6 (10.3) 85 (73.3)31 (26.7) 
 GenotypeAllele
TotalGG (%)GT (%)TT (%) P-valueG (%)T (%) P-value
rs11193170        
 Female AD5128 (54.9)18 (35.3)5 (9.8)0.96174 (72.5)28 (27.5)0.904
 Male control 5833 (56.9)19 (32.8)6 (10.3) 85 (73.3)31 (26.7) 
 GenotypeAllele
TotalGG (%)GA (%)AA (%) P-valueG (%)A (%) P-value
  1. A, adenine; AD, Alzheimer's disease; C, cytosine; G, guanine; T, thymine.

rs4918280        
 Female AD5127 (52.9)19 (37.3)5 (9.8)0.99173 (71.6)29 (28.4)0.998
 Male control5831 (53.4)21 (36.2)6 (10.3) 83 (71.6)33 (28.4) 
Table 4. Distributions of the polymorphisms in female Alzheimer's patients and all controls
 GenotypeAllele
TotalGG (%)CG (%)CC (%) P-valueG (%)C (%) P-value
rs601883        
 Female AD5121 (41.2)27 (52.9)3 (5.9)0.84269 (67.6)33 (32.4)0.737
 Control 9242 (45.7)44 (47.8)6 (6.5) 128 (69.6)56 (30.4) 
 GenotypeAllele
TotalTT (%)TC (%)CC (%) P-valueT (%)C (%) P-value
rs7907690        
 Female AD5139 (76.5)10 (19.6)2 (3.9)0.52488 (86.3)14 (13.7)0.475
 Control 9273 (79.3)18 (19.6)1 (1.1) 164 (89.1)20 (10.9) 
 GenotypeAllele
TotalGG (%)GA (%)AA (%) P-valueG (%)A (%) P-value
rs600879        
 Female AD5041 (80.4)10 (19.6)0 (0)0.06792 (90.2)10 (9.8)0.077
 Control 9282 (91.1)8 (8.9)0 (0) 172 (95.6)8 (4.4) 
 GenotypeAllele
TotalGG (%)GA (%)AA (%) P-valueG (%)A (%) P-value
rs17277986        
 Female AD5136 (70.6)15 (29.4)0 (0)0.73487 (85.3)15 (14.7)0.994
 Control 9266 (71.7)25 (27.2)1 (1.1) 157 (85.3)27 (14.7) 
 GenotypeAllele
TotalAA (%)GA (%)GG (%) P-valueA (%)G (%) P-value
rs2900717        
 Female AD5128 (54.9)21 (41.2)2 (3.9)0.67277 (75.5)25 (24.5)0.519
 Control 9257 (62.0)31 (33.7)4 (4.3) 145 (78.8)39 (21.2) 
 GenotypeAllele
TotalCC (%)TC (%)TT (%) P-valueC (%)T (%) P-value
rs10884399        
 Female AD5128 (54.9)19 (37.3)4 (7.8)0.64975 (73.5)27 (26.5)0.821
 Control 9252 (56.5)29 (31.5)11 (12.0) 133 (72.3)51 (27.7) 
 GenotypeAllele
TotalGG (%)GT (%)TT (%) P-valueG (%)T (%) P-value
rs11193170        
 Female AD5128 (54.9)18 (35.3)5 (9.8)0.83574 (72.5)28 (27.5)0.953
 Control 9251 (56.7)28 (31.1)11 (12.2) 130 (72.2)50 (27.8) 
 GenotypeAllele
TotalGG (%)GA (%)AA (%) P-valueG (%)A (%) P-value
  1. A, adenine; AD, Alzheimer's disease; C, cytosine; G, guanine; T, thymine.

rs4918280        
 Female AD5127 (52.9)19 (37.3)5 (9.8)0.87573 (71.6)29 (28.4)0.947
 Control9250 (54.3)31 (33.7)11 (12.0) 131 (71.2)53 (28.8) 

In multilocus analysis, the frequency of an eight-SNP rs601883/rs7907690/rs600879/rs17277986/rs2900717/rs10884399/rs11193170/rs4918280 CCGGACGG haplotype was significantly higher in AD patients (6.3%) than in controls (0.5%) (χ2= 8.791, P= 0.003, OR = 13.038, CI = 2.346–72.445). The frequency of the CTGGACGG haplotype was significantly lower in AD patients (9.3%) than in controls (20.3%) (χ2= 10.678, P= 0.001, OR = 0.394, CI = 0.222–0.696) (Table 5). After Bonferroni correction, the frequency of an eight-SNP rs601883/rs7907690/rs600879/rs17277986/rs2900717/rs10884399/rs11193170/rs4918280 CCGGACGG haplotype was significantly higher in AD patients than in controls (P= 0.045), and the frequency of the CTGGACGG haplotype was significantly lower in AD patients than in controls (P= 0.045). Because these two P-values resulted from calculations according to the Bonferroni correction formula and the original P-values (P= 0.003 and P= 0.001), they have not been included in Table 5. However, after Bonferroni correction, only the frequency of the CCGGACGG haplotype was significantly higher in female AD patients (P= 0.004) (Table 6).

Table 5. Haplotype frequencies in Alzheimer's patients and controls (frequency <0.05 in both control and patients have been excluded)
IIIIIIIVVVIVIIVIIIPatients (%)Controls (%)χ2 P-valueOR (95% CI)
  1. Global χ2= 19.591. Degrees of freedom = 5. Fisher's P-value is 0.002. I, rs601883; II, rs7907690; III, rs600879; IV, rs17277986; V, rs2900717; VI, rs10884399; VII, rs11193170; VIII, rs4918280; A, adenine; C, cytosine; CI, confidence interval; G, guanine; OR, odds ratio; T, thymine.

CCGGACGG6.30.58.7910.00313.038 (2.346∼72.445)
CTGGACGG9.320.310.6780.0010.394 (0.222∼0.696)
GCGGACGG5.47.20.5440.4610.742 (0.335∼1.643)
GTGAGTTA7.87.60.0250.8741.061 (0.513∼2.194)
GTGGACGG37.737.034.50.4921.169 (0.758∼1.802)
GTGGATTA12.69.61.0950.2951.398 (0.745∼2.624)
Table 6. Haplotype frequencies in female Alzheimer's patients and controls (frequency <0.05 in both control and patients have been excluded)
IIIIIIIVVVIVIIVIIIPatients (%)Controls (%)χ2 P-valueOR (95% CI)
  1. Global χ2= 21.577. Degrees of freedom = 6. Fisher's P-value is 0.0015. I, rs601883; II, rs7907690; III, rs600879; IV, rs17277986; V, rs2900717; VI, rs10884399; VII, rs11193170; VIII, rs4918280; A, adenine; C, cytosine; CI, confidence interval; G, guanine; OR, odds ratio; T, thymine.

CCGGACGG9.50.513.7080.000220.062 (3.426–117.473)
CTGGACGG13.220.32.0800.14930.603 (0.302–1.204)
GCGGACGG1.27.24.8340.02790.154 (0.023–1.031)
GTGAGTTA6.17.60.1840.66780.805 (0.298–2.172)
GTGGACGG33.134.50.0140.90710.968 (0.562–1.667)
GTGGATTA12.09.60.4700.49301.316 (0.599–2.894)
GTGGGCGG7.84.81.1800.27731.734 (0.636–4.732)

DISCUSSION

SORCS1, which belongs to the sortilin family of vacuolar protein sorting-10 domain containing proteins, expresses in the developing and mature murine central nervous system,16–19 and it is involved in intracellular sorting and trafficking functions.20 Furthermore, it is capable of influencing the production of Aβ. Several studies have shown that the variations in the SORCS1 gene affect AD risk, and these AD-associated allelic variants are located in intron 1 of SORCS1. However, different populations have different AD-associated allelic variations.3,6

Since previous studies have shown that possible AD-associated allelic variants in the SORCS1 gene are concentrated around eight SNP in intron 1, no other tag SNP in SORCS1 have been found to be associated with AD risk.3,4,6 As such, in this study, we only genotyped the eight variations and found that that none of the polymorphisms was associated with the disease in single-locus analysis. Haplotype analysis showed that rs601883/rs7907690/rs600879/rs17277986/rs2900717/rs10884399/rs11193170/rs4918280 CCGGACGG haplotype was associated with an increased risk of AD (P= 0.003), but that the CTGGACGG haplotype was associated with a decreased risk of AD (P= 0.001). After Bonferroni correction for the number of haplotypes, the association remained significant (P= 0.045 or 0.015). In six independent populations, Reitz et al. found that there were different AD-associated allelic variants in the SORCS1 gene in different populations, and no single SNP was associated with AD in all six independent data sets.3 In the southern Han Chinese population, we found two haplotypes were associated with AD risk. Similar to Reitz et al., we also attributed this result to ethnic difference.

Additionally, we found that the rs600879 GG genotype and G allele decreased AD risk in women (P= 0.007; P= 0.009). The effect of the sexual dimorphism of the SORCS1 gene on AD has also been found in another study.6 Furthermore, Aβ42 levels were increased in the brains of female SORCS1 hypomorphic mice, but not in male mice.21 However, the exact mechanism remains elusive. There are several possible hypotheses for the sexual dimorphism of the SORCS1 gene's effect on AD. First, oestrogen may have some effect on AD onset. Several randomized clinical trials have reported positive effects of the oestrogen-replacement therapy on cognitive performance.22–28 Animal studies have also shows that oestrogens can increase neuronal connectivity and prevent or slow age-related cognitive decline.29,30 Second, oestrogen and the variations may interfere with cell type-specific or tissue-specific expression of SORCS1. Third, interaction with mitochondrial DNA mutations is thought to be involved in AD onset, as there is evidence of mitochondrial DNA in AD cases.31,32 Another possible mechanism is genomic imprinting, such as methylation, which is supported by the findings of the increased number of unmethylated sites in AD patients in comparison to controls.6,22 Oestrogen and the variant of SORCS1 may collectively interfere with unmethylated sites in female AD patients. All these explanations remain to be confirmed.

SORCS1, which belongs to the sortilin family of vacuolar protein sorting-10 domain containing proteins, is highly expressed in the brain, heart, kidneys, pancreatic islets and β-cell lines.6,20 It binds to platelet-derived growth factor and plays a role in maintaining or expanding pancreatic islet vasculature during islet growth and in compensation for insulin resistance. Impaired vascular structure may lead to disrupted islet architecture and directly affect insulin secretion to the blood.33–35 Studies in mice and humans all suggest that the effect of SORCS1 on insulin secretion is directly relevant to the development of T2DM.10,12

Previous studies found that SORCS1 resides at a quantitative trait locus for T2DM in mice and that the promoter, first exon and first intron of SORCS1 are the major loci influencing fasting plasma insulin levels. Genetic variations in these regions leads to decreased fasting insulin levels via reduced insulin secretion following a glucose challenge and pancreatic islet disruption.10,11 The study by Goodarzi et al. in Mexican Americans suggested that distinct allelic variants of the SORCS1 gene are associated with alterations in fasting insulin levels, insulin secretion and T2DM.12 Dysfunction of SORCS1 contributes to both the amyloid precursor protein/Aβ disturbance underlying AD and the insulin/glucose disturbance underlying diabetes.35 In our study, we found that the rs7907690 C allele, which is in intron 1 of SORCS1, increased AD risk in T2DM patients (P= 0.041). We speculate that the rs7907690 C allele may be the genetic mechanism that links SORCS1 to both diabetes and AD. Because the association between the rs7907690 C allele and AD patients with T2DM is marginal (P= 0.041) and the AD with T2DM sample size is small, the results need further study to be validated.

It is a difficult process to identify underlying susceptibility genes in diseases of complex genetic aetiology such as AD. Thus, replication is important in determining which genes have a substantial effect. In our study, we have replicated the result that AD-associated allelic variations of SORCS1 exist in women. In addition, we found that there is an association between SORCS1 variations and AD patients with T2DM. However, there were some limitations in our study. One limitation was small sample size, especially with regard to AD patients with T2DM. To avoid a chance positive result, enlarging the sample size, especially that of AD patients with T2DM, is necessary. Another limitation was the lack of functional studies about SORCS1 variations influence AD risk.

In summary, our results show that there are two haplotypes in intron 1 of the SORCS1 gene that affected AD risk in a southern Han Chinese population. Furthermore, the rs7907690 C allele increases AD risk in T2DM patients, whereas rs600879 GG genotype and G allele decrease AD risk in women. We conclude that, in addition to ethnic background, sex and T2DM status influence the genetic link between AD and the SORCS1 gene. Further studies with an enlarged sample size are still required to validate our findings. In addition, to get a better understanding of the roles of the gene variations of SORCS1 in AD, functional studies are necessary.

ACKNOWLEDGMENTS

We sincerely appreciate all the subjects for their participation in this study and all the medical staff involved in specimen collection. We also thank Shanghai BioWing Applied Biotechnology (China) for providing genotyping service. This work was supported by grants from the Jiangsu Administration Bureau of Traditional Chinese Medicine (Nanjing, China) (LZ09074). The first two authors (YH and ZF) contributed equally to this work.

Ancillary