Abstract

Alcohol metabolism is one of the biological determinants that could significantly be influenced by genetic polymorphisms in alcohol-metabolism genes. Alcohol dehydrogenase (ADH) converts alcohol to acetaldehyde, and aldehyde dehydrogenase (ALDH) converts acetaldehyde to acetate. The well-known genetic polymorphisms in ADH1B(His47Arg) and ALDH2(Glu487Lys) have dramatic effects on the rate of metabolizing alcohol and acetaldehyde, respectively. The protective allele of ADH1B (ADH1B*47His) encodes for a rapid ethanol-metabolizing enzyme, and the susceptible allele of the ALDH2 (ALDH2*487Lys) is strongly associated with decreased rate of metabolizing acetaldehyde. However, the combined genetic effects of both functional polymorphisms have not been clarified. The combined analysis of two polymorphisms among a Korean population (n = 1,032) revealed dramatic genetic effects on the risk of alcoholism. Individuals bearing susceptible alleles at both loci have 91 times greater risk for alcoholism [odds ratio (OR) = 91.43, P = 1.4 × 10−32] and individuals bearing one susceptible and one protective allele at either loci have 11 times greater risk (OR = 11.40, P = 3.5 × 10−15) compared with subjects who have both protective alleles. The attributable fraction of those genetic factors, calculated based on population controls, indicates that alcoholism in 86.5% of alcoholic patients can be attributed to the detrimental effect of ADH1B*47Arg and/or ALDH2*487Glu in Korean population.

INTRODUCTION

Alcohol metabolism occurs in two major steps: oxidation of alcohol to acetaldehyde by the alcohol dehydrogenases (ADHs) enzymes, especially by ADH1B, and further oxidation of acetaldehyde into acetate by aldehyde dehydrogenase enzymes (ALDHs), mainly by ALDH2. Encoding genes for these two representative alcohol-metabolizing enzymes display polymorphisms (ADH1B His47Arg and ALDH2 Glu487Lys) that show different alcohol/acetaldehyde oxidizing capability among individuals (1–4). The ADH1B*47His allele represents a much higher activity of ADH1B with ∼40 times higher Vmax than the homozygotes for the ADH1B*47Arg form, which enables increased alcohol elimination from the blood after alcohol consumption (1,5). The ALDH2*487Lys allele encodes a catalytically inactive subunit (1,5), which causes alcohol-related adverse reactions including flushing, palpitation, nausea, headache, drowsiness, breathlessness and general discomfort (6). These adverse reactions in subjects with ALDH2*487Lys, as a result of excessive acetaldehyde accumulation, tend to reduce alcohol consumption, subsequently reducing the risk of alcoholism.

Many previous studies have reported genetic associations of ADH1B His47Arg and ALDH2 Glu487Lys with alcoholism, especially in Asian populations (7–12). However, the combined genetic effects of these two loci have not yet been clarified. In the current study, the combined genetic effects of ADH1B and ALDH2 genotypes were analyzed in a Korean population (n = 1032).

RESULTS

In the Korean population, the ADH1B His47Arg polymorphism showed, as expected, a dramatic genetic association with the risk of alcoholism. Referent analysis of ADH1B His47Arg revealed that its genetic mode for the risk of alcoholism is apparently susceptible and recessive, when comparing the strength and magnitude of associations of heterozygotes and homozygotes for ADH1B*47Arg. Only 6.2% of normal controls had the ADH1B*47Arg/Arg genotype, compared with 34.1% in alcoholics [recessive mode; odds ratio (OR) = 7.80 95% confidential interval (CI), 5.18–11.75, P= 8.2 × 10−23]. In addition, the genotype distribution of ADH1B His47Arg showed severe deviation from Hardy–Weinberg equilibrium (HWE) (P = 3.17 × 10−28) in alcoholics, whereas no deviation was detected among normal controls. The deviation from HWE occurring only in alcoholics strongly suggests that a selection bias (or pressure) is involved, which would be additional direct evidence of association of this polymorphism with the risk of alcoholism (Table 1).

Table 1.

Analysis of genetic effect of ADH1B His47Arg and ALDH2Glu487Lys on the risk of alcoholism among Korean male subjects (n = 1032)

Loci Genotype Alcoholism, n (%) HWEa NC, n (%) HWEa OR (95% CI)b P-valueb OR (95% CI)c P-valuec Genetic effect 
ADH1B His47Arg His/His 217 (39.5)  298 (61.7)  – – Protective 
 His/Arg 145 (26.4) 3.17 × 10−28 155 (32.1) 0.110 1.29 (0.97–1.71) 0.09   Protective 
 Arg/Arg 187 (34.1)  30 (6.2)  8.56 (5.61–13.07) 2.8 × 0−23 7.80 (5.18–11.75) 8.2 × 0−23 Susceptible 
ALDH2 Glu487Lys Glu/Glu 530 (96.5)  346 (73.6)  – – Susceptible 
 Glu/Lys 19 (3.5) 0.680 122 (25.3) 0.297 0.10 (0.06–0.17) 4.5 × 10−19 0.10 (0.06–0.15) 3.6 × 10−21 Protective 
 Lys/Lys 0 (0.0)  15 (3.1)  – –   Protective 
Loci Genotype Alcoholism, n (%) HWEa NC, n (%) HWEa OR (95% CI)b P-valueb OR (95% CI)c P-valuec Genetic effect 
ADH1B His47Arg His/His 217 (39.5)  298 (61.7)  – – Protective 
 His/Arg 145 (26.4) 3.17 × 10−28 155 (32.1) 0.110 1.29 (0.97–1.71) 0.09   Protective 
 Arg/Arg 187 (34.1)  30 (6.2)  8.56 (5.61–13.07) 2.8 × 0−23 7.80 (5.18–11.75) 8.2 × 0−23 Susceptible 
ALDH2 Glu487Lys Glu/Glu 530 (96.5)  346 (73.6)  – – Susceptible 
 Glu/Lys 19 (3.5) 0.680 122 (25.3) 0.297 0.10 (0.06–0.17) 4.5 × 10−19 0.10 (0.06–0.15) 3.6 × 10−21 Protective 
 Lys/Lys 0 (0.0)  15 (3.1)  – –   Protective 

aP-value of genotype distribution deviation from Hardy–Weinberg equilibrium.

bLogistic regression models were used for calculating ORs (95% CI) and corresponding P-values. The common alleles were used as the referent genotype to the heterozygote and homozygote of the minor allele.

cOR and P-value of referent analysis between susceptible versus protective genotype groups.

The ALDH2Glu487Lys also showed apparent genetic effects on the risk of alcoholism [dominant mode; OR = 0.10 (95% CI, 0.06–0.15) and P= 3.6 × 10−21]. In contrast to ADH1B*47Arg, the genetic mode of ALDH2*487Lys clearly appeared to be protective and dominant in the referent analysis. Interestingly, subjects bearing ALDH2*487Lys/Lys appeared to have complete genetic protection against alcoholism (no homozygotes in alcoholics) (Table 1).

Next, in order to examine whether these two functional polymorphisms have any interaction effects on the risk of alcoholism, subgroup analysis according to genetic effects of each loci was performed. The results were that similar magnitudes of genetic effects with those of all subjects were shown for the two polymorphisms (Supplementary Materils, Table S1), suggesting clearly independent genetic effects on the risk of alcoholism.

To examine the combined genetic effect of ADH1B His47Arg and ALDH2 Glu487Lys, subjects were subdivided by genetic effects of both loci. Combined analysis revealed huge genetic effects on the risk of alcoholism, e.g. individuals bearing susceptible alleles at both loci were found to have 91 times greater risk for alcoholism (OR = 91.43, P= 1.4 × 10−32) and individuals bearing one susceptible and one protective allele at either loci had 11 times greater risk (OR = 11.40, P= 3.5 × 10−15) compared with subjects who have both protective alleles (Table 2). The results of this combined analysis are concordant with those of the separated analysis (independent effects; Supplementary Materils, Table S1) and clearly showed additive effects on the risk of alcoholism development.

Table 2.

Combined analysis of association of ADH1B His47Arg and ALDH2 Glu487Lys with the risk of alcoholism in Korean male subjects (n = 1032)

Genotype No. of subjects OR (95% CI)a P-valuea OR (95% CI)b P-valueb OR (95% CI)c P-valuec P-valued 
ADH1B His47Arg ALDH2 Glu487Lys Alcoholism, n (%) NC, n (%) PCe, n (%)        
Protective (His/His or His/Arg) Protective (Glu/Lys or Lys/Lys) 12 (2.2) 128 (26.5) 95 (24.8) –    
Protective (His/His or His/Arg) Susceptible (Glu/Glu) 350 (63.7) 325 (67.3) 263 (68.5) 11.49 (6.24–21.16) 4.7 × 10−15 11.40 (6.19–20.99) 3.5 × 10−15   2.2 × 10−31 
Susceptible (Arg/Arg) Protective (Glu/Lys or Lys/Lys) 7 (1.3) 9 (1.9) 9 (2.3) 8.30 (2.62–26.24) 0.0003   16.14 (8.80–29.60) 2.6 × 10−19  
Susceptible (Arg/Arg) Susceptible (Glu/Glu) 180 (33.8) 21 (4.3) 17 (4.4) 91.43 (43.42–192.50) 1.4 × 10−32 91.43 (43.42–192.50) 1.4 × 10−32    
Genotype No. of subjects OR (95% CI)a P-valuea OR (95% CI)b P-valueb OR (95% CI)c P-valuec P-valued 
ADH1B His47Arg ALDH2 Glu487Lys Alcoholism, n (%) NC, n (%) PCe, n (%)        
Protective (His/His or His/Arg) Protective (Glu/Lys or Lys/Lys) 12 (2.2) 128 (26.5) 95 (24.8) –    
Protective (His/His or His/Arg) Susceptible (Glu/Glu) 350 (63.7) 325 (67.3) 263 (68.5) 11.49 (6.24–21.16) 4.7 × 10−15 11.40 (6.19–20.99) 3.5 × 10−15   2.2 × 10−31 
Susceptible (Arg/Arg) Protective (Glu/Lys or Lys/Lys) 7 (1.3) 9 (1.9) 9 (2.3) 8.30 (2.62–26.24) 0.0003   16.14 (8.80–29.60) 2.6 × 10−19  
Susceptible (Arg/Arg) Susceptible (Glu/Glu) 180 (33.8) 21 (4.3) 17 (4.4) 91.43 (43.42–192.50) 1.4 × 10−32 91.43 (43.42–192.50) 1.4 × 10−32    

aOR and P-value for referent analysis of protective/protective genotype group.

bOR and P-value of genotype groups with one and two copy of susceptible alleles versus protective/protective genotype group.

cOR and P-value of genotype groups with one or two copies of susceptible allele versus protective/protective genotype group.

dGlobal P-value.

ePopulation controls (n = 384). Population controls were used for attributable fraction (AF, 87.7%) calculation [OR (10.53) and frequency (75.2% of PC) of subjects with one or two susceptible alleles for either loci)

The attributable fraction (AF) was calculated by OR (10.53) and frequency (subjects with one or two susceptible alleles for either loci; 75.2%) based on population controls (Table 2). AF indicates that alcoholism in 86.5% of alcoholic patients in the Korean population can be attributed to the detrimental effects of ADH1B*47Arg and/or ALDH2*487Glu.

DISCUSSION

Alcoholism [MIM# 103780] is a leading cause of morbidity and premature death. Several lines of evidence suggest a substantial genetic component to the risk for alcoholism. Alcoholism is believed to be a multifactorial and polygenic disorder involving complex gene-to-gene and gene-to-environment interactions. According to the National Council of Alcoholism and Drug Dependence (NCADD) and the American Society of Addiction Medicine (ASAM), ‘alcoholism is a primary, chronic disease with genetic, psychosocial and environmental factors influencing its development and manifestations. The disease is often progressive and fatal. It is characterized by continuous or periodic impaired control over drinking, preoccupation with the drug alcohol, use of alcohol despite adverse consequences and distortion in thinking, most notably denial’.

In this study, by analyzing the well-known genetic polymorphisms in ADH1B(His47Arg) and ALDH2(Glu487Lys) in a large Korean sample (n= 1032), we were able to show that the combined effect of the two alleles has a huge impact on disease phenotype. The employment of the large sample gives the study a great statistical power, and the evidences from this study might be, therefore, indisputable. In terms of combined analysis of two functional polymorphisms, <5% of people in the Korean population (disease and population controls) have susceptible/susceptible genotypes. In addition, although considerable allele frequency variation occurs even among East Asian populations, up to 25% of Asians have protective genotypes at both loci, whereas very low numbers of European or African populations have been reported to have these protective genotypes (13,14). The much higher frequencies of the protective ADH1B and ALDH2 alleles (ADH1B*47His and ALDH2*487Lys) in some Asian populations, therefore, clearly provide some East Asians, especially Koreans and Japanese (6), with protection against alcoholism, compared with most European or African populations who have susceptible alleles (ADH1B*47Arg and ALDH2*487Glu).

Interestingly, the genotype distribution of both ADH1B and ALDH2 differs greatly between Asian, especially East Asian, and other major populations. Recently, the global frequencies of ADH1B*47His have been well-defined. The frequency of the ADH1B*47His allele is particularly high in East/West Asian populations and moderate in southeast Asian and North African populations, but the allele is almost absent in sub-Saharan African, European and Native American populations (13,15). Similarly, ALDH2*487Lys has only been detected among East Asian populations (14). Although other social/cultural factors might be involved, these different genetic backgrounds regarding alcohol metabolism could very likely explain the lower incidence of alcohol abuse in some East Asian populations, including the Korean population (16).

Several plausible evidences for natural selection of this allele in Koreans, Japanese and Han Chinese have been suggested (13,17). The researchers proposed interesting hypotheses that might explain the high frequencies of ADH1B*47His and ALDH2*487Lys in East Asian populations. Among them, the hypothesis of selection pressure for higher ADH1B (mediated by ADH1B*47His) and lower ALDH2 (mediated by ALDH2*487Lys), against some endemic disease(s), including parasitism, in area with high frequencies of those alleles, might be the most plausible explanation, although the nature of the selective pressure and the time period during which it operates are still unknown. Interestingly, Li et al. (13) presented very convincing evidence that the distribution of ADH1B*47His increased in frequency independently in West and East Asia after humans had spread across Eurasia. Similarly, natural selection of ALDH2*487Lys in East Asian populations was also suggested (14).

In summary, we re-confirmed that the well-known ADH1B His47Arg and ALDH2 Glu487Lys have the dramatic genetic associations with the risk of alcoholism, and that their genetic effects are clearly independent among Korean population. When combined together, those two loci revealed huge genetic effects on the risk of alcoholism. In addition, AF indicates that alcoholism in 87.7% of alcoholic patients in the Korean population can be attributed to the detrimental effect of ADH1B*47Arg and/or ALDH2*487Glu, which suggesting that the two polymorphisms are major genetic components for alcoholism, at least, in Korean population.

MATERIALS AND METHODS

Study subjects and genotyping of polymorphisms

The patients used in this study (n= 549, all males; mean age = 46.1, range = 20–73), all of whom were alcohol-dependent according to the Diagnostic and Statistical Manual of Mental Disorders IV (DSM-IV) criteria (American Psychiatric Association, 1994), were recruited from Hangang Sacred Heart Hospital, Yong-In Mental Hospital, Hanmaum Hospital, Gumin Hospital, Hando Hospital (Hanllym University groups) and another network of multicenter mental hospitals in Korea (Holy Family Hospital, Chamsarang Mental Hospital, Chunchon Mental Hospital; Catholic University groups). The patients had neither major medical nor co-morbid psychiatric illnesses other than alcohol-related disorders and/or nicotine dependence (43%). The controls were unrelated healthy male employees of Hangang Sacred Heart Hospital (n = 483, all male; mean age = 33.2 years, range = 20–77). Most of the participating employees were non-drinkers; only some were occasional light drinkers, as revealed by a drinking habit questionnaire. Subjects who had first-degree relatives with major psychiatric disorders, such as schizophrenia, mood disorders or substance-use disorders other than nicotine dependence, were excluded. Additional population controls (n= 384, all males) were used for AF calculation. The Institutional Review Board of each hospital approved the study, and informed consents were obtained.

The ADH1B His47Arg and ALDH2 Glu487Lys polymorphisms were genotyped using the TaqMan method (18), which has been described in our previous work (9,19).

Statistics

χ2 tests were used to determine if the individual variants were in HWE. Logistic regression analyses, controlling for age as covariate, were used to calculate ORs and the P-values for case–control analysis. The AF was calculated by the formula AF = f(R –1)/(1+f(R –1)), where f is the frequency of the risk factor in the population and R is the measure of the OR (20). In combined analysis (Table 2), several facts were considered when choosing a method of analysis, including (i) the two SNPs were not linked to each other, (ii) no significant interactions were detected through interaction analysis (P> 0.36, data not shown) and (iii) both SNPs have obvious effects and their genetic modes are apparent [susceptible/recessive (ADH1B*47Arg) and protective/dominant (ALDH2*487Lys) in this study of a Korean population]. Based on these considerations, we have adopted one of the simplest methods of analysis, e.g. susceptible versus protective at two loci (four subgroups). The protective/protective subgroup was used as the referent to the other subgroups. Logistic regression analyses, controlling for age as covariate, were used to calculate ORs and the P-values for case–control analysis. All statistical analyses were performed using SAS 9.1 (SAS Institute, Inc., Cary, NC, USA).

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG Online.

Conflict of Interest statement. None declared.

FUNDING

This work was supported by a grant from the National Research Lab Program as part of the National Research and Development Program of the Ministry of Commerce, Industry and Energy of Korea (M1-0302-00-0073). This study was also partly supported by a grant from Biomedical Brain Research Center (A040042) and a grant of the Korea Health 21 R&D Project, Ministry of Health & Welfare (03-PJ10-PG13-GD01-0002) funded by the Ministry of Health & Welfare, Republic of Korea.

REFERENCES

1
Bosron
W.F.
Li
T.K.
Genetic polymorphism of human liver alcohol and aldehyde dehydrogenases, and their relationship to alcohol metabolism and alcoholism
Hepatology
 , 
1986
, vol. 
6
 (pg. 
502
-
510
)
2
Yokoyama
A.
Muramatsu
T.
Omori
T.
Yokoyama
T.
Matsushita
S.
Higuchi
S.
Maruyama
K.
Ishii
H.
Alcohol and aldehyde dehydrogenase gene polymorphisms and oropharyngolaryngeal, esophageal and stomach cancers in Japanese alcoholics
Carcinogenesis
 , 
2001
, vol. 
22
 (pg. 
433
-
439
)
3
Marchand
L.L.
Wilkinson
G.R.
Wilkens
L.R.
Genetic and dietary predictors of CYP2E1 activity: a phenotyping study in Hawaii Japanese using chlorzoxazone
Cancer Epidemiol. Biomarkers Prev.
 , 
1999
, vol. 
8
 (pg. 
495
-
500
)
4
Crabb
D.W.
Edenberg
H.J.
Bosron
W.F.
Li
T.K.
Genotypes for aldehyde dehydrogenase deficiency and alcohol sensitivity. The inactive ALDH2(2) allele is dominant
J. Clin. Invest.
 , 
1989
, vol. 
83
 (pg. 
314
-
316
)
5
Yoshida
A.
Hsu
L.C.
Yasunami
M.
Genetics of human alcohol-metabolizing enzymes
Prog. Nucleic Acid Res. Mol. Biol.
 , 
1991
, vol. 
40
 (pg. 
255
-
287
)
6
Matsuo
K.
Wakai
K.
Hirose
K.
Ito
H.
Saito
T.
Tajima
K.
Alcohol dehydrogenase 2 His47Arg polymorphism influences drinking habit independently of aldehyde dehydrogenase 2 Glu487Lys polymorphism: analysis of 2,299 Japanese subjects
Cancer Epidemiol. Biomarkers Prev.
 , 
2006
, vol. 
15
 (pg. 
1009
-
1013
)
7
Chen
C.C.
Lu
R.B.
Chen
Y.C.
Wang
M.F.
Chang
Y.C.
Li
T.K.
Yin
S.J.
Interaction between the functional polymorphisms of the alcohol-metabolism genes in protection against alcoholism
Am. J. Hum. Genet.
 , 
1999
, vol. 
65
 (pg. 
795
-
807
)
8
Chen
W.J.
Loh
E.W.
Hsu
Y.P.
Chen
C.C.
Yu
J.M.
Cheng
A.T.
Alcohol-metabolising genes and alcoholism among Taiwanese Han men: independent effect of ADH2, ADH3 and ALDH2
Br. J. Psychiatry
 , 
1996
, vol. 
168
 (pg. 
762
-
767
)
9
Choi
I.G.
Son
H.G.
Yang
B.H.
Kim
S.H.
Lee
J.S.
Chai
Y.G.
Son
B.K.
Kee
B.S.
Park
B.L.
Kim
L.H.
, et al.  . 
Scanning of genetic effects of alcohol metabolism gene (ADH1B and ADH1C) polymorphisms on the risk of alcoholism
Hum. Mutat.
 , 
2005
, vol. 
26
 (pg. 
224
-
234
)
10
Higuchi
S.
Polymorphisms of ethanol metabolizing enzyme genes and alcoholism
Alcohol Alcohol Suppl.
 , 
1994
, vol. 
2
 (pg. 
29
-
34
)
11
Whitfield
J.B.
Alcohol dehydrogenase and alcohol dependence: variation in genotype-associated risk between populations
Am. J. Hum. Genet.
 , 
2002
, vol. 
71
 (pg. 
1247
-
1250
author reply 1250–1241
12
Maezawa
Y.
Yamauchi
M.
Toda
G.
Suzuki
H.
Sakurai
S.
Alcohol-metabolizing enzyme polymorphisms and alcoholism in Japan
Alcohol Clin. Exp. Res.
 , 
1995
, vol. 
19
 (pg. 
951
-
954
)
13
Li
H.
Mukherjee
N.
Soundararajan
U.
Tarnok
Z.
Barta
C.
Khaliq
S.
Mohyuddin
A.
Kajuna
S.L.
Mehdi
S.Q.
Kidd
J.R.
, et al.  . 
Geographically separate increases in the frequency of the derived ADH1B*47His allele in eastern and western Asia
Am. J. Hum. Genet.
 , 
2007
, vol. 
81
 (pg. 
842
-
846
)
14
Oota
H.
Pakstis
A.J.
Bonne-Tamir
B.
Goldman
D.
Grigorenko
E.
Kajuna
S.L.
Karoma
N.J.
Kungulilo
S.
Lu
R.B.
Odunsi
K.
, et al.  . 
The evolution and population genetics of the ALDH2 locus: random genetic drift, selection, and low levels of recombination
Ann. Hum. Genet.
 , 
2004
, vol. 
68
 (pg. 
93
-
109
)
15
Osier
M.V.
Pakstis
A.J.
Soodyall
H.
Comas
D.
Goldman
D.
Odunsi
A.
Okonofua
F.
Parnas
J.
Schulz
L.O.
Bertranpetit
J.
, et al.  . 
A global perspective on genetic variation at the ADH genes reveals unusual patterns of linkage disequilibrium and diversity
Am. J. Hum. Genet.
 , 
2002
, vol. 
71
 (pg. 
84
-
99
)
16
Grant
B.F.
Dawson
D.A.
Stinson
F.S.
Chou
S.P.
Dufour
M.C.
Pickering
R.P.
The 12-month prevalence and trends in DSM-IV alcohol abuse and dependence: United States, 1991–1992 and 2001–2002
Drug Alcohol Depend.
 , 
2004
, vol. 
74
 (pg. 
223
-
234
)
17
Han
Y.
Gu
S.
Oota
H.
Osier
M.V.
Pakstis
A.J.
Speed
W.C.
Kidd
J.R.
Kidd
K.K.
Evidence of positive selection on a class I ADH locus
Am. J. Hum. Genet.
 , 
2007
, vol. 
80
 (pg. 
441
-
456
)
18
Livak
K.J.
Allelic discrimination using fluorogenic probes and the 5′ nuclease assay
Genet. Anal.
 , 
1999
, vol. 
14
 (pg. 
143
-
149
)
19
Kim
D.J.
Park
B.L.
Yoon
S.
Lee
H.K.
Joe
K.H.
Cheon
Y.H.
Gwon
D.H.
Cho
S.N.
Lee
H.W.
NamGung
S.
, et al.  . 
5′-UTR polymorphism of dopamine receptor D1 (DRD1) associated with severity and temperament of alcoholism
Biochem. Biophys. Res. Commun.
 , 
2007
, vol. 
357
 (pg. 
1135
-
1141
)
20
Adams
M.J.
Jr.
Khoury
M.J.
James
L.M.
The use of attributable fraction in the design and interpretation of epidemiologic studies
J. Clin. Epidemiol
 , 
1989
, vol. 
42
 (pg. 
659
-
662
)

Author notes

The authors wish it to be known that, in their opinion, the first two authors should be regarded as joint First Authors.