Effect of Cigarette Smoking on Pulmonary Function in Each Phenotype M of α-1-Protease Inhibitor

Effect of Cigarette Smoking on Pulmonary Function in Each Phenotype M of a-1-Protease Inhibitor* Takeshi Matsuse MD; Yoshinosuke Fukuchi , MD, FCCP; Hirotoshi Matsui, MD; Eiichi Sudo, MD; Takahide Nagase, MD; and Hajime Orimo, MD Human a-1-protease inhibitor (a-1-Pi) has been known to be a highly polymorphic protein. We hypothesized that antiprotease activity of each phenotype M of a1-protease inhibitor (PiM) might be different among smokers and that a variation of decrease in pulmonary function for a given amount of cigarette smoking might be associated with PiM phenotypes. To test this, we investigated the effect of cigarette smoking on pulmonary function in each PiM phenotype. The serum level of a1-Pi was measured by the turbidimetric immunoassay and the distribution of PiM phenotypes was determined using isoelectric focusing technique in 247 healthy subjects and 20 COPD patients. Serum levels of a-1-antitrypsin of healthy and COPD subjects were 205.1 ± 31.1 and 179.2 ± 44.4 ( ± SD) mg/ dL, respectively (p>0.01). The frequency of each PiM phenotype in healthy subjects was shown as follows: Mh 0.555; M 1Mz, 0.328; M 2, 0.041; M 1M 3, 0.057; M 2M 3, 0.016; M 3 , 0.004. The difference in the distribution of PiM phenotypes between healthy and COPD subjects was not significant. Single- and multiple-regression analyses showed that the ratio of FEV 1 to forced vital capacity (FVC), in

which FEV 1 is expressed as percentage of FVC, the maximum flow rate at 50% of FVC divided by measured body height CV5o/Ht), and the maximum flow rate at 25% of FVC divided by body height ('V 25 /Ht) were closely related to age and that Vz5/Ht also was related to smoking index. However, PiM phenotype was unrelated to those pulmonary function variables. We conclude that PiM phenotype is not a major determinant of difference in magnitude of pulmonary impairments caused by cigarette smoking in each individual. (Chest 1995; 107:395-400)

Human a-1-protease inhibitor (a-1-Pi) , also known as a-1-antitrypsin (a-1-A T) is a major inhibitor of serine proteases in plasma, and it is a principal inhibitor of proteolytic enzyme produced by inflammatory cells in the peripheral airways of cigarette smokers. 1 The association of the deficient alleles (Z and S) of protease inhibitor (Pi) with pulmonary emphysema and liver cirrhosis has been well known.2-4 Human a-1-Pi is a highly polymorphic plasma protein. The introduction of isoelectric focusing (IEF) in narrow-range polyacrylamide gel has considerably helped sorting out the genetic heterogeneity of phenotype M of a-1-protease inhibitor (PiM) allele into several subtypes, including M1(Val 213 ), M2, and M3. These PiM variants differ from M 1(Val 213) by one or two base changes. 1·5 The M1(Ala 213), which was recently recognized, cannot

be distinguished from M1(VaP 13 ) by IEF. 6 The M1(Ala 213 ) is only identified by sequencing at the DNA level or restriction fragment length polymorphism analysis. It also was reported that a-1-AT serum levels of individuals inheriting the M1(Ala213) gene in a homozygous fashion were in the same range as those for homozygous M1(Val 213 ) and that there was no significance in the function of M1(Val 213 ) and M1(Ala 213 ).6 All of these PiM phenotypes (a homozygous form with each other PiM variant) migrate in the M-region of the standard IEF gels (pH 4.2-4.9) . When inherited in a homozygous form, or heterozygous form with each other of PiM , the a-1-AT serum levels have been supposed to be normaP However, small differences were reported among the PiM phenotypes. 7 Three M allele heterozygotes (M1M2, M1M3, M2M3) as a group have higher means of a-1-AT serum level and lower variances than the group of three M allele homozygotes (M1 , M2, Ms). 7 In addition, there is a report of small differences in lung function between the heterozygotes and homozygotes.8 Although PiM phenotypes have not been thought to be related to any disease state,1·9 it was reported that the subtypes combined with an M2-al-

*From the Department of Geriatrics, Faculty of Medicine, University of Tokyo, Bunkyo-ku, Tokyo, Japan. Supported by research funds from the Smoking Research Foundation of Japan and in part by grants from the Ministry of Education, Science and Culture of Japan (grant 05770394). Manuscript received February 7, 1994;revision accepted June 8. Reprint requests: Dr. Matsuse, Dept . of Geriatrics, University of Tokyo , 7-3-1 Hongo , Bunkyo-ku , Tokyo 113, japan .

a-1-AT=a-1-antitrypsin; a-1-Pi=a-1-protease inhibitor; FVC= forced vital capacity; IEF=isoelectric focusing; Pi= protease inhibitor; PiM=phenotype M of a-1-protease inhibitor; Vso=the maximum flow rate at 50% of FVC; Vso/ Ht=the Vso divided by the measured body height; V2s=the maximum flow rate at the 25% of FVC; V2s/ Ht=the V25 divided by the measured body height

Key words: a-1-protease inhibitor; cigarette smokers; isoelectric focusing technique; phenotype

CHEST / 107 / 2 / FEBRUARY, 1995

395

lele (M1M2, Mz) showed higher prevalence of COPD.l 0 The M2 homozygote seems to have occurred more frequently in the asthmatic group.l 1 Furthermore, association between M1M2 phenotype and pulmonary fibrosis in rheumatoid arthritis was reported.l 2 Thus, a number of workers have showed findings linking PiM phenotypes to some pulmonary disease. As far as we know, however, little is known about the effect of the interaction of PiM phenotypes with cigarette smoking on pulmonary function impairments among healthy subjects and on the development of COPD. It is possible that anti protease effect of each a-1-Pi phenotype on neutrophil elastase in the site of inflammation caused by cigarette smoking might be different from each other. Thus, we hypothesized that antiprotease activity of each PiM phenotype might be different among smokers and that variation of decrease in pulmonary function for a given amount of cigarette smoking might be associated with PiM phenotypes. The aim of the present study was therefore to investigate the effect of cigarette smoking on pulmonary function in each PiM phenotype among healthy subjects. METHODS

Subjects and Pulmonary Function Test Two hundred forty-seven healthy subjects 21 to 61 years of age were recruited from the Health Check Clinic at the Tokyo Kenbikyoin Hospital (age, 49.4 ± 7.9 [SD]); 156 were men and 91 women). All of the subjects were active regular workers and free from pulmonary disease. Details of smoking histories were confirmed by the interviewer. Smoking histories of those subjects were as follows: 114 nonsmokers, 52 exsmokers, and 81 current smokers. Pulmonary function tests were performed on the same day of blood sampling. Pulmonary function was assessed with the spirometry and flow-volume curve. The FEV 1 and the forced vital capacity (FVC) were obtained from spirometry. The FEV t! FVC ratio, in which FEV 1 is expressed as percentage of the FVC. The maximum flow rate at the 50% and 25% of FVC levels ('V5o, V25) were obtained from the flow-volume curve. The Vso/ Ht and V25/ Ht are these flow rates divided by the measured body height, respectively. We adapted the following parameters for further analysis; FEVt/FVC ratio, Yso/ Ht, and Vzs/ Ht. We also evaluated these parameters as percentages of normal values among Japanese, which were obtained from the sources of the Japan Society of Chest Disease 13 (%FEVJ/FVC, %V5o/ Ht, %V25/ Ht). Twenty COPD subjects also were studied for a-1-Pi phenotyping. These COPD patients, aged 55 to 85 years, were selected randomly, and the diagnosis of COPD was confirmed by their medical histories and physical examination findings as well as by previous pulmonary function test results. They all had expiratory airflow limitation (a FEVt/FVC ratio below 70%; 43.9 ± 14.5% [SD]).

a-1-Protease Inhibitor Phenotyping and the Serum Level of a-1-Antitrypsin Serum samples were obtained from 247 healthy subjects and 20 patients with COPD. The PiM phenotyping was performed by the IEF procedure as reported by Yuasa et aP 4 with minor modifications. Gels with dimensions of 125X260X0.5 mm were made of

396

4 mL acrylamide solution (29.1% of acrylamide and 0.9% bis-acrylamide), 4 gglycerol, 1.6 mL Pharmalyte (pH 4.2 to 4.9) (Pharmacia), 40 mg ACES [n-(2-acetamido)-2-aminoethane sulfonic acid] (Dojin Chemical Laboratory Co. , Kumamoto, Japan) and 15.2 mL water. After addition of 240 tLL ammonium persulfate (15 mg / 500 tLL) and 10 tLL of tetramethyl-ethylenediamine (TEMED) gels were polymerized. Electrode strips were applied using 0.5 moi/L CH3 COOH as an anolyte and 0.2N NaOH as a catholyte. Cooling temperature was 10° C. Separation was carried out at a maximum voltage of 2,500 V, 20 rnA for 4.5 h, including prefocusing at 1,000 V. Sample-soaked filter papers were placed for 30 min at the cathodic side. After removing the filter, IEF was carried out for 3.5 h. At completion of the run, gels were fixed with 12% trichloroacetic acid solution for 30 min. After washing briefly with distilled water, gels were rinsed in a destaining solution (25% methanol, 4% acetic acid) for 10 min, followed by staining with 0.05% solution of Coomassie Brilliant Blue (CBB G-250; Nakarai, Japan) for 10 min at 60° C. After destaining again, gels were treated with 10% glycerol for 30 min. Control serum samples for each PiM phenotype were applied at each IEF run. The a-1-PiM phenotyping of each sample was performed by two investigators independently without prior knowledge of physiologic data. The serum level of a-1-AT was determined by the turbidimetric immunoassay 15 using Nittobo's kit (TIAa1-AT, Nittobo Medical, Tokyo, Japan). For the turbidometric assay of a-1-AT, human a-1-AT purified by the manufacturer (Nittobo Co. Medical Laboratory. Fukushima, Japan) was used to generate a standard reference. According to the manufacturer's laboratory, the coefficient of variation of these measurements using this kit is less than 5%, and the range of level of a-1-AT measured by this kit is 0 to 600 mg/ dL. In addition, the values obtained by this assay (y) and by the radial immunodiffusion assay (x) using the Hoechest's kit (NOR-Partigen-a1-antitrypsin, Hoechest Japan. Tokyo, Japan) correlated well; y=l.032X-2.032 (r=0.991 n=56).

Statistical Analysis Results are expressed as means± 1 SD. Statistical analyses were performed using the SAS program (SAS Institute Inc., Cary, NC). Continuous variables (such as age, smoking index, the serum level of a-1-AT) were compared between the COPD group and the healthy control subjects using the two-sided Student's t test. An analysis of variance was used for multiple groups of single variables. These variables were compared among PiM phenotypes using analysis of variance. Differences in the PiM phenotype distribution between groups were tested by means of Fisher's exact test. Single-and multiple-regression analyses were performed to assess the relationship between pulmonary function (FEV d FVC ratio, Yso/ Ht, V25 / Ht, %FEVt/FVC, %V50 / Ht, %V 25 / Ht) and variables such as smoking index, sex, age, serum level of a-1-AT, and subtypes of a-1-Pi. The influence of categoric variables (such as sex and PiM phenotype) was assessed using dummy variables. A probability value of 0.05 or less was accepted for statistical significance. RESULTS

The band patterns of each PiM phenotype (M 1, M1M2, M2, M1M3, M2M3, M3) obtained after IEF were consistent with those reported previously by Yuasa et aP 4 To assess the reproducibility of the a-1-PiM phenotyping, the control serum of each PiM phenotype was reexamined a total of 14 times to provide 56 repeat analyses. The results from the test of the reproducibility of the a-1-PiM phenotyping on the control serum of each PiM phenotype showed Effect of Smoking on Pulmonary Function in PiM (Matsuse et al)

Table !-Distribution of a-1-Protease Inhibitor Phenotypes Hea lthy Subjects (n=247)

Phenotypes

No. of Cases

Frequency

M1 M1M2 M2 M1M3 M2M3 M3

137 81 10 14 4 1

0.555 0.328 0.041 0.057 0.016 0.004

Table 3-Serum a-1-Antitrypsin Levels in Healthy Subjects by Phenotype Group*

COPD (n=20) No. of Cases 14 5

Frequency 0.700 0.250 0.050

0 0 0

Phenotypes

M1 MtM2 M2 M1M3 M2M3 M3

No. of Cases 42 17 3 3 0

% 63.6 25.8 4.6 4.6 1.5

Current Smokers With %V25/ Ht ~ 80% (n=15) No. o f Cases 6 7 0 2 0 0

All Subjects (n=247)

Nonsmokers (n=ll4)

Current Smokers (n=81 )

M1 M1M2 M2 M1M3 M2M3 M3

205.7±34.0 204.6±28.3 213.4 ±30.1 205.2± 17.6 177.0±29.5 190

203.8±40.5 207.1±27.9 202.8±19.7 204.3± 18.7

211.2±28.5 207.9±28.4 248.0±23.1 207.2±20.6 187

190

*Values are expressed a s mg/ dL of a-1-antitrypsin.

that this phenotyping was good, reproducible , and that the rate of agreement of reproducibility was 55 out of 56 or 98.2%. Furthermore, interobserver agreement for PiM phenotyping was good (the rate of agreement was 95.8%). The results of the distribution of PiM phenotypes are summarized in Table 1. The frequency of each PiM phenotype in healthy subjects was shown as follows, M1, 0.555; M1M2, 0.328; M2, 0.041 ; M1M3, 0.057; M2M3 , 0.016; M3, 0.004. The allele frequencies of Pi•M1, p(M 2, and p(M3 were estimated at 0.7449, 0.2201, and 0.0382, respectively. Since the number of each PiM phenotype was small except for M1 and M1M2, we divided PiM phenotypes into three groups (M 1, M1M2, and the others); then we analyzed the difference in the distributions of M1, M1M2, and the others which consist of M2, M3, M1Ms, and M2Ms between healthy subjects and COPD patients using Fisher's exact test. The difference in the distribution of M1, M1M2, and the others between healthy subjects and COPD patients was not significant (p=0.548, Fisher's exact test) . We also compared the distributions of frequencies of PiM subtype allele heterozygotes (M 1M2, M1Ms, and M2Ms) and homozygotes (M 1, M2, and M3) in these two groups. However, there

Current Smokers With %V25/Ht• <80% (n=66)

Phenotypes

%

40.0 46.7 13.3

• The %V25 / Ht value is the V25 / Ht expressed as a p ercentage of the normal values for the Japanese population, a s d m e onstrated by the Japanese Society of Chest Disease.l 3

was no significant difference in frequency of heterozygotes and homozygotes as a group between healthy subjects and C 80%) revealed that the difference in the distribution between these two groups was not significant (p=0.209, Fisher's exact test) . There was also no significant difference in %FEY If FVC, %V50 / Ht, and %V25/ Ht between the group of PiM allele heterozygotes (M1M2, M1Ms, and M2Ms) and the group of homozygotes (M1 , M2, and Ms) . The serum level of a-1-A T in each PiM phenotype tested was identical (Table 3) . Serum levels of a-1-AT of healthy subjects and COPD patients were 205.1 ± 31.1 and 179.2±44.4 (±SD) mg/ dL, respectively (p>0.01 , Student's t test), as shown in Table 4. Single- and multiple-regression analyses showed that both V50/ Ht and V25/ Ht were closely related to age (p>0.01) and that V25/ Ht also was related to smoking index (p>0.05) . However, PiM phenotype was unrelated to those pulmonary function variables (Table 5) . The valid estimation of correlation between V25/ Ht and cigarette smoking was only possible for M1. and M1M2 phenotype with sufficient numbers o.f subJects. Thus, we analyzed the relationship of V25/ Ht to smoking index (the number of cigarettes smoked per Table 4-Serum Levels of a-1-Antitrypsin in Healthy Subjects and Chronic Obstructive Pulmonary Disease Patients Subjects

Level of a-1-Antitrypsin mg/dL

All healthv subjects (n=247 ) Nonsm ~kers (n=ll4) C urrent smokers (n= 8l) COPD patients (n=20)

205.1 ± 31.3 204.8± 34. 3 211.0±28.4 179.2 ± 44.4 *

*Probability value is ess l than 0.01 for comparison with the serum a-1-AT levels of all healthy subjects. CHEST / 107 / 2 / FEBRUARY, 1995

397

Table 5-Multiple-Regression Analyses: General Linear Models of SAS (Healthy Current Smoker [n=81]) Probability Value

Y/ Xi•

Cigarette X Years Smoking Indext

Age

a-1-AT Level

Sex

Pi Phenotype

SI*Pi Phenotypet

NS§ 1\S 0.0438 NS NS 0.0501

0.0006 0.0021 0.0001 NS NS 0.0292

NS NS NS NS NS NS

NS 0.0263 NS NS NS 0.0400

NS NS NS NS NS NS

NS NS NS NS NS NS

FEY t/FVC ratio Yso/ Ht v25/Ht

%FEVt/FVC %V50 /Ht %V25 / Ht

•y = ;t+ Effect[xJ]+ Effect[x2]+ Effect(x3]+ Effect[x4 ]+ Effect[xs]+ Effect[xt *xs]+<. tCigarettes X years represents the number of cigarettes smoked times the number of years smoked. tSI*Pi phenotype is the interaction of smoking index and Pi phenotype. §NS, not significant.

day times the number of years smoked) and compared the decline rates of V2s/ Ht in M1 and M1M2 phenotype groups with each other (Fig 1). Decline in V25/Ht with increasing cigarette smoking in both PiM phenotype groups was identical, and there was no difference between these two regression lines. DISCUSSION

a-1-Protease inhibitor is the major serine protease inhibitor in human serum and plays an important role in preventing lung tissue damage by the release of neutrophil elastase during inflammation. Cigarette smoking and severe deficiency of a-1-Pi are wellestablished as risk factors in the development of COPD. In addition, a-1-Pi has been known to be a highly polymorphic protein, and many subtypes and variant alleles have been identified at Pi locus by electrophoretic techniques.l The most common variant, known as type M, consists of at least six subtypes: Mt[Ala 213], Mt[Vai2 13], M2, Ms, M4, and Ms, characterized by normal serum values.1·9 .1 6 The S and Z

r=-0.38 p<0.059

0.5 r=-0.40 p
0.0

L__---'-------'-----'-----......l....-

0

500

1000

1500

2000

Cigarettes x Years FIGURE l. Relationship between V25 /Ht and cigarette smoking (smoking index). x axis: cigarettes X years means the number of cigarettes smoked per day times the number of years smoked. The solid line shows the regression line in the M 1 phenotype group (n=48, r=-0.40, p
variant results in deficient or low serum value of a-1-Pi. The Z variant is associated in its homozygous form with emphysema, probably due to the inadequacy of the antiprotease defense mechanisms in the lung. The heterozygous MZ (but not MS) phenotype also may be associated with obstructive lung disease, although this is extremely controversial. 17- 19 Furthermore, the effect of heterozygosity for each PiM phenotype on pulmonary function impairment due to cigarette smoking remains to be determined.1.1° Here, we present the distribution of PiM phenotypes in healthy Japanese subjects and assess the pulmonary function impairments in current smokers with these phenotypes. In this study, we identified both M1(Val 213 ) and Mt(Ala 213 ) as Mt because M1(Val 213 ) and M1(Ala 213 ) cannot be distinguished from each other by IEF. 6 To assess the reproducibility of the a-1-PiM phenotyping, the control serum of each PiM phenotype was reexamined. The results from the test of the reproducibility of the a-1-PiM phenotyping on the control serum of each PiM phenotype showed that this phenotyping was good, reproducible, and that the agreement of reproducibility was 55 out of 56 or 98.2% . We did not confirm this PiM phenotyping using IEF by independent measures such as genotyping. However, this IEF technique is a well established method to analyze the PiM phenotype. In addition, the PiM phenotyping of all samples was performed by two investigators independently without prior knowledge of physiologic data. Interobserver agreement for PiM phenotyping also was good (the rate of agreement was 95.8%). Furthermore, all cases of nonagreement were reexamined to determine the phenotype. In this study, the frequencies of pj"M allele were essentially identical to those reported in other Japanese populations14 and in other East Asian nations.20 However, the PiM2 allele frequency in the present study was relatively higher compared with that in literature reported in European nations. 21 -23 When inherited in a homozygous form, or het-

erozygous form with each other of the PiM, the a-1-AT serum levels have been thought to be normal.1 However, small differences were reported among the PiM phenotypes. 7 Three M allele heterozygotes (M1M2, M1M3, M2M3) as a group have higher means of a-1-AT serum level and lower variances than the group of three M allele homozygotes (M1, M2, Ms). 7 In addition, it is supposed that variants of a-1-Pi (including M2 and M3) might be more susceptible to local inactivation by oxidation or proteolytic cleavage.l 2 This probability is suggested by the findings of the decrease in half-life of other a-1-Pi variants such as the S variant in vivo. 24 Furthermore, variants might differ in effectiveness against neutrophil elastase. An additional possibility is supposed that variants may differ in their ability to enter lung tissue sites of inflammation.l 2 Thus, we hypothesized that antiprotease activity of each PiM phenotype might be different among current smokers and that variation of pulmonary function impairments for a given amount of cigarette smoking among healthy subjects might be associated with PiM phenotypes. In this study, however, there was no apparent association of variant types with pulmonary function impairments which were expressed by V2s/ Ht among current smokers and also among patients with COPD as a whole. Bencze and colleagues 10 reported that PiM2 was significantly more frequent in patients with COPD. They concluded that apart from the known a-1-PI deficiencies as in PiZZ phenotype, the PiMM subtype and especially the M2M2 and M1M2 phenotypes seem to be combined with a higher prevalence of COPD. However, they compared the prevalence of a-1-PiM phenotypes of COPD patients and healthy individuals without considering the smoking histories. On the other hand, Gaillard and colleagues 11 noted that there was an increase in the prevalence of the M2M2 phenotype in the asthmatic subject group in comparison with the general population; however, they believed that the number of M2M2 phenotype individuals identified was small and that the finding should be interpreted with caution. We found no increased prevalence of variant phenotypes in the COPD groups. Moreover, we found no relationship between the PiM phenotype and early damage of small airways due to cigarette smoking. There is a report of small differences in lung function between the heterozygotes and homozygotes.8 It has been described that lung function in women was greater in the M phenotype heterozygotes at the Pi locus than in the M phenotype homozygotes, although no effect of the Pi locus was found in men. 8 However, we found no difference in frequency of PiM heterozygotes and PiM homozygotes as a group between COPD and healthy subjects. Moreover, we could not find such a tendency in the

current smokers, in terms of %FEY I/FYC, %V so/ Ht, and %Y2s/ Ht. In our study, single- and multiple-regression analyses were performed to assess the relationship between pulmonary function (FEY I/FYC ratio, Yso/ Ht, Y2s/ Ht, %FEYI/FYC, %V5o/ Ht, and %V2s/ Ht) and variables such as smoking index, sex, age, serum level of a-1-Pi, and subtypes of a-1-PiM, which showed that pulmonary function impairment caused by cigarette smoking is unrelated to PiM phenotype. The present study showed that each phenotype of PiM has a comparable serum level of a-1-AT in healthy subjects irrelevant to smoking history. These findings were essentially similar when the analysis was performed only in current smokers. The mean serum level of a-1-A T of COPD patients was 179.2±44.4 (±SD) mg/ dL, and this was relatively low compared with that of healthy subjects. The possible reason for their relatively low a-1-A T levels may be related to the difference in the nutritional state between healthy and COPD subjects. Thus, we conclude that PiM phenotype is not a major determinant of difference in magnitude of pulmonary impairments caused by cigarette smoking in each individual. The lack of association between PiM variants and pulmonary function impairments in the smoker population suggests that other geneticenvironmental interactions, such as latent viral infection,25 are involved in the pathogenesis of COPD. Acknowledgements: The authors thank Dr. I. Yuasa (the Tottori University School of Medicine, Japan ) for providing control serum samples, Mr. K. Matsuno for his excellent technical assistance, and Dr. T. Hamada for his statistical advice. We also thank the Tokyo Kenbikyouin Hospital for their assistance and in particular Dr. Y. Hiraiwa and Ms. M. Saito. REFERENCES

1 Brantly M, Nukiwa T, Crystal RG. Molecular basis of alpha-1antitrypsin deficiency. Am J Med 1988; 84(suppl 6A):13-31 2 Laurell CB, Eriksson S. The electrophoretic alpha-1-globulin pattern of serum in alpha-1-antitrypsin deficiency. Scand J Clin Lab Invest 1963; 15:132-40 3 Hutchison DCS. Natural history of alpha-1-protease inhibitor deficiency. Am J Med 1988; 84(suppl 6A):3-12 4 Eriksson S, Carlson J, Velez R. Risk of cirrhosis and primary liver cancer in alpha 1-antitrypsin deficiency. N Eng! J Med 1986; 314:736-39 5 Frants RR , Eriksson AW. Reliable classification of six Pi M subtypes by separator isoelectric focusing. Hum Hered 1978; 28:201-09 6 Nukiwa T, Brantly M, Ogushi F, et al. Characterization of the M1(Aia 213 ) type of a1-antitrypsin, a newly recognized common "normal" a1-antitrypsin haplotype. Biochemistry 1987; 26: 5259-67 7 Oakeshott JG, Muir A, Clark P, et al. Effects of the protease inhibitor (Pi) polymorphism on alpha-1-antitrypsin concentration and elastase inhibitory capacity in human serum. Ann Hum Bioi 1985; 12:149-60 8 Gibson JB, Martin NG, Oakeshott JG, et al. Lung function in an Australian population: contributions of polygenic factors and the Pi locus to individual differences in lung function in a CHEST / 107 / 2 / FEBRUARY, 1995

399

sample of twins. Ann Hum Biol1983; 10:547-56 9 Crystal RG, Brantly ML, Hubbard RC, et al. The alpha1-antitrypsin gene and its mutations: clinical consequences and strategies for therapy. Chest 1989; 95:196-208 lO Bencze K, Sabatke L, Fruhmann G. Alpha 1-antitrypsin: the PiMM subtypes-do they play a role in development of chronic obstructive pulmonary disease? Chest 1980; 77:761-63 l l Gaillard MC, Kilroe-Smith TA, Nogueira C, et a!. Alpha-1protease inhibitor in bronchial asthma: phenotypes and biochemical characteristics. Am Rev Respir Dis 1992; 145:1311-15 12 Michalski JP, McCombs CC, Scopelitis E, et a!. Alpharantitrypsin phenotypes, including M subtypes, in pulmonary disease associated with rheumatoid arthritis and systemic sclerosis. Arthritis Rheum 1986; 29:586-91 13 The Japanese Society of Chest Disease. Japanese standard predicted values of pulmonary functions. Jap J Thoracic Dis 1993; 31(suppl) 14 Yuasa I, Suenaga K, Gotch Y, et a!. PI(a 1-antitrypsin) polymorphism in the Japanese: confirmation of PI.M4 and description of new PI variants. Hum Genet 1984; 67:209-12 15 Viedma JA, de Ia Iglesia A, Parera M, eta!. A new automated turbidimetric immunoassay for quantifying a 1-antitrypsin in serum. Clin Chern 1986; 32:1020-22 16 Weidinger S, Jahn W , Cujnik F, et al. Alpha-1-antitrypsin: evidence for a fifth PiM subtype and a new deficiency allele Pi*Zaugsberg· Hum Genet 1985; 71:27-29

400

17 Cox DW, Hoeppner VH, Levison H. Protease inhibitors in patients with chronic obstructive pulmonary disease: the alpha1antitrypsin heterozygote controversy. Am Rev Respir Dis 1976; 113:~1-06

18 Janus ED. Alpha ~-antitrypsin Pi types in COPD patients [letter]. Chest 1988; 94:446-47 19 Buist AS, Burrows B, Cohen A, et a!. Guidelines for the approach to the patient with severe hereditary alpha-1-antitrypsin deficiency. Am Rev Respir Dis 1989; 140:1494-97 20 SabaN. a1-protease inhibitor (PI) subtypes in seven populations of East Asia. Ann Hum Biol1990; 17:229-34 21 Klasen EC, Bos A, Simmelink HD. PI (arantitrypsin) subtypes: frequency of PI*M4 in several populations. Hum Genet 1982; 62:139-41 22 Weidinger S, Cleve H, Patutschnick W. Alpha 1-antitrypsin: evidence for a fourth PiM allele-distribution of the PiM subtypes in Southern Germany. Z Rechtsmed 1982; 88:203-11 23 Thymann M. Distribution of alpha-1-antitrypsin (Pi) phenotypes in Denmark determined by separator isoelectric focusing in agarose gel. Hum Hered 1986; 36:19-23 24 Jeppsson JO, Laurell CB, Nosslin B, et a!. Catabolic rate of alpha-1-antitrypsin of Pi types Sand M malton and asialylate M protein in man. Clin Sci Mol Med 1978; 55:103-07 25 Matsuse T, Hayashi S, Kuwano K, et a!. Latent adenoviral infection in the pathogenesis of chronic airways obstruction. Am Rev Respir Dis 1992; 146:177-84

Effect of Smoking on Pulmonary Function in PiM (Matsuse et alj