�9 1988 by the Humana Press Inc. All rights of any nature whatsoever reserved. 01634984188/1501-0167502.20

Abnormalities of Blood Selenium and Glutathione Peroxidase ActMty

in Patients with Acquired Immunodeficiency Syndrome

and Aids-Related Complex

BRAD M. DWORKIN, *'l WILLIAM S. ROSENTHAL, 1 GARY P. WORMSER, 2 LISA WEISS, ~ MIGUEL NUNEZ, 2

CAROL JOLINE, 2 AND ANTHONY HERP 3

'The Sarah C Upham Division of Gastroenterology, the Division of Infectious Diseases, the Department of Medicine, and 3the

Department of Biochemistry, New York/Vledical College, Valhalla, NY 10595

Received May 4, 1987; Accepted June 15, 1987

ABSTRACT

Severe protein-calorie malnutr i t ion is co m m o n in patients with AIDS and could contr ibute to the progressive deteriorat ion character- istic of that disease. Selenium deficiency could also have a negative impact on immune funct ion and other organ functions vital for recov- ery f rom infectious diseases. Therefore, to assess any role for sele- n ium in AIDS, we de te rmined plasma and erythrocyte se lenium lev- els and glutathione peroxidase activity in 13 patients with AIDS compared to 8 patients with AIDS-related complex (ARC) and 14 heal thy controls. Plasma selenium levels were significantly reduced in AIDS patients compared to controls (p < .0001) and to ARC (p < .02). Erythrocyte selenium levels in both AIDS and ARC were also reduced compared to controls (p < .02), but not to each other. Gluta thione peroxidase activity in AIDS was 28.9 _+ 1.4 U/g Hb vs 38.4 + 6.9 in ARC (p = NS) and 52.3 + 1.7 in controls (p < .0001 vs

*Author to whom all correspondence and reprint requests should be addressed.

Biological Trace Element Research ] 67 Vol. 15, 1988

168 Dworkin et al.

AIDS; p < .02 vs ARC). When all groups were combined, there were significant correlations between total lymphocyte count and both plasma selenium (r = .53; p < .002) and erythrocyte glutathione per- oxidase activity (r = .65; p < .0001). In addition, strong correlations were noted between plasma selenium and serum albumin (r = .68; p < .0001), plasma selenium and glutathione peroxidase (r = .77; p < .0001), and glutathione peroxidase and hematocrit (r = .66; p < .0001). In AIDS or ARC, no correlations between selenium with dis- ease duration or weight loss were present. We conclude that, in comparison to normals, patients manifesting infection with human immunodeficiency virus have evidence of selenium deficiency as de- termined by diminished plasma and erythrocyte levels and glutathione peroxidase activity. These abnormalities are most marked in patients with AIDS, but are also present in patients with AIDS- related complex. Selenium deficiency has important implications for the progression and pathogenesis of clinical disease in AIDS.

Index Entries: Blood selenium, abnormalities of; glutathione per- oxidase activity, in AIDS and AIDS-related complex patients.

INTRODUCTION

Protein-calorie malnutr i t ion and cachexia are c o m m o n in AIDS (ac- quired immunodef ic i ency syndrome) (1-3). Since deficiencies of mul t ip le nutr ients could fur ther impair immunolog ic host defense mechan i sms and other organ functions, it is possible that progressive malnutr i t ion contributes to the morbidi ty or mortal i ty of some AIDS pat ients (4,5).

Selenium deficiency has been associated with impaired cell-me- diated cytotoxicity (6-7). In se len ium deficient animals, oral candidiasis is seen (7,8). This infection is also frequent ly found in AIDS (3). Therefore, we initially invest igated blood se len ium status in 12 men with AIDS com- pared to controls (9). We found 59% lower plasma (p < .001) and 24% lower red b lood cell (p < .005) se len ium levels in AIDS as compared to normal values (9). A good correlation be tween plasma se lenium and se- r um a lbumin (r -- .77; p < .0001) was also no ted (9). The purpose of the present s tudy was to fur ther investigate and define the relat ionship be- tween se len ium and gluta thione peroxidase activity in pat ients with AIDS compared to those with "pre-AIDS" or AIDS-related complex (ARC) and normal controls (10,11).

METHODS

Thirty-five subjects seen at Universi ty Hospital , New York Medical College, Westchester Coun ty Medical Center were s tudied. All pat ients and controls represent new subjects, no t repor ted in our prior manu - script (9).

Biological Trace Element Research VoL 15, 1988

Selenium Deficiency in AIDS 169

This included 13 patients meeting the Centers for Disease Control definition of AIDS (10,11). There were 11 males and 2 females with a mean (_ SEM) age of 32.9 + 1.1 yr. Epidemiologic risks for AIDS in- cluded 4 homosexuals, 7 with intravenous drug use, 1 Haitian, and 2 sex- ual partners of high risk persons. One homosexual also used intravenous drugs. Eight patients were diagnosed as having AIDS-related complex (ARC) (11). There were 5 males and 3 females with a mean age of 42.3 --- 4.4 yr. Risk factors in this group included 1 homosexual, 6 with intrave- nous drug use, and 1 sexual partner of a high risk person. All patients with AIDS and ARC were seropositive for human immunodeficiency virus infection (HIV). These two groups were contrasted with 14 normal, healthy controls without any identifiable risks for AIDS. This group was comprised of 6 males and 8 females with a mean age of 33.1 + 2.5 yr.

All patients had blood drawn by standard venipuncture techniques for selenium and glutathione peroxidase. Plasma and whole blood sele- nium was determined by a spectrofluorometric method, as previously described (9,12-16). Red blood cell selenium was then calculated from those determinations, using the hematocrit value (9,14-16). All selenium and glutathione peroxidase assays were run blindly with respect to the clinical status of any subject.

Glutathione peroxidase activity was determined by a modification of the coupled assay procedure (17-19). Red blood cells (RBC) were sepa- rated from whole blood by centrifugation, then were washed three times with cold isotonic salines and frozen at -20~ All assays were per- formed within two weeks of initial freezing. Cells were lysed upon thawing. A RBC solution was made consisting of 50 I~L RBC mixed with 200 I~L Drabkins reagent and 1250 I~L .01M Tris Buffer containing .5% Triton X-100 (pH 7). In a semimicro cuvet were pipetted 600 I~L .005M Tris Buffer, .5 mM EDTA (pH 7), 100 I~L of .01M NaN3, 100 t~/L of glutathione (.1M GSH), 100 laL 2.4 U/mL glutathione reductase (Sigma) and 50 I~L of the diluted RBC solution. This was mixed well and incu- bated at 37~ for 10 min, followed by the addition of 50 I~L of NADPH, (Sigma), (3 mM, in .1% NaHCO3). The reaction was monitored for stabil- ity in a recording spectrophotometer (Shimadzu) for 3 min. Then 10 I~L of 1.2 mM hydrogen peroxide was added. A 3 min scan at 340 nm was made beginning at the time of the addition of the peroxide, and the en- zyme activity was calculated using a molar extinction coefficient of 6.22 X 10 3 for NADPH at 340 nm.

Hemoglobin concentration was measured on an aliquot of the same red blood cell samples using Sigma diagnostics total hemoglobin quanti- tative, colorimetric determination (procedure No. 525). All values are ex- pressed as units glutathione peroxidase activity/g of hemoglobin.

Clinical data including a history of weight loss, anorexia, diarrhea, Kaposi's sarcoma, and disease duration were recorded and correlated with selenium status. Laboratory values including hematocrit, serum al-

Biological Trace Element Research Vol. 15, 1988

170 Dworkin et aL

bumin , and total l y m p h o c y t e coun t were also noted. Reliable d ie tary his- tories were not available.

Statistical analysis was p e r f o r m e d us ing an Analysis of Variance (ANOVA), Fisher 's Exact, and S tuden t ' s t-test for n o n p a i r e d data (20). Correlat ion coefficients (Pearson r) were also de t e rmined , w h e r e appro- priate. All va lues are expressed as m e a n + s t anda rd error (SEM) (20).

The s tudy was a p p r o v e d by the N e w York Medical College Commi t - tee for the pro tec t ion of h u m a n subjects.

RESULTS

W h e n c o m p a r e d to controls, pat ients wi th AIDS had significantly lower levels of plasma, who le blood, and red blood cell se len ium (Table 1). Patients wi th ARC also had r educed se len ium values compared to normals (Table 1). However , no significant dif ferences b e t w e e n AIDS and ARC were no t ed for who le blood or red blood cell se lenium. Plasma se len ium was significantly r educe d in AIDS vs ARC pat ients (p < .02; S tuden t ' s t-test) (Table 1).

There was also a significant decrease in g lu ta th ione peroxidase activ- ity in pat ients wi th AIDS vs controls (p < .0001; S tuden t ' s t-test) (Table 1; Fig. 1). A lesser, bu t significant decrease was p resen t b e t w e e n ARC and controls (p < .02; S tuden t ' s t-test) (Table 1). No significant difference in g lu ta th ione peroxidase was found b e t w e e n pat ients wi th AIDS and ARC (p = .1; S tuden t ' s t-test).

Table 1 Plasma, Whole Blood, and Red Blood Cell Selenium Levels and

Glutathione Peroxidase Activity in Patients with AIDS, AIDS-Related Complex (ARC) and Controls

Parameter AIDS ARC Controls Significance"

Plasma Selenium .035 _+ .002 .044 _+ .003 ~' .065 + .001' p < .0001 p.g/ml (N = 13) (N = 8) (N = 14)

Whole Blood .054 _+ .003 .060 _+ .004 .090 _+ .001' p < .0001 Selenium (N = 13) (N = 8) (N = 14) txg/ml

Red Blood Cell .099 _+ .009 .088 +_ .007 .122 + .002" p < .002 Selenium (N = 11) (N = 7) (N = 14) ixg/ml

Glutathione 28.9 + 1.4 38.4 _+ 6.9 52.3 _ 1.7" p < .0001 Peroxidase (N = 13) (N = 8) (N = 14) U/gm Hb

"ANOVA. ~'p < .02, vs AIDS, S t u d e n t ' s t - tes t (all M e a n _+ SEM). 'p < .0001, vs AIDS or ARC, S t u d e n t ' s t-test. "p < .02 vs AIDS, p < .0001 vs ARC, S t u d e n t ' s t- test . 'p < .0001 vs AIDS; p < .02 vs ARC, S t u d e n t ' s t- test .

Biological Trace Element Research VoL 15, 1988

Selenium Deficiency in AIDS 171

~ , 86"0~l~.

�9 "r 6 6 . 0 " - E

D 5 8 . 0 -

I.IJ r,~

5 0 . 0 - X 0 r I,I s_ 4 2 . 0 - w z 0 -r" 34.O -

<[

�9 -J 2 6 . 0 - (.9

1 8 . 0 -

@

e l m

41,

- - I -

I J , _ �9

i

II

A I D S ARC ( N = I 3 ) ( N = 8 )

C O N T R O L

( N : 14) Fig. 1. Glutathione peroxidase activity in patients with AIDS, AIDS-

related complex (ARC), and controls (p < .0001; ANOVA). All Mean + SEM.

In an effort to evaluate potential factors contributing to selenium deficiency in AIDS some clinical and historical factors were noted (Table 2). There were no significant differences between AIDS and ARC with regard to disease duration, weight loss, diarrhea, or epidemiologic back- grounds. Also, no significant correlations between plasma and red blood cell selenium levels and either disease duration or weight loss were ob- served. None of the patients had Kaposi's sarcoma. One AIDS patient had a non-Hodgkins B-cell lymphoma. Patients with AIDS tended to be younger than patients with ARC (p < .05; Student 's t-test). However , there were no significant differences in age be tween AIDS and controls (p = .94) or ARC and controls (p = .06; Student 's t-test).

Evaluation of laboratory data revealed that AIDS patients t ended to be more anemic than either controls or patients with ARC (Table 3). Con- sidering all three groups together, an excellent correlation was found be-

Biological Trace Element Research VoL 15, 1988

172 D w o r k i n e t al.

Table 2 Clinical and Historical Parameters in Patients with AIDS and AIDS-Related Complex (ARC)

Parameter AIDS ARC Significance

Durat ion of 44.5 + 7.2" 50.7 + 20.6 p = NS I .... disease, weeks (N = 13) (N -- 7)

Weight loss present 11/13 (85%) 6/7 (86%) p = NS' Weight loss, 7.0 _+ 1.7 18.8 _ 8.5 p = .09"

pounds (N = 11) (N = 6) Diarrhea present 5/13 (38%) 2/8 (25%) p = NS' Anorexia present 4/12 (33%) 1/6 (17%) p = NS' In t ravenous d rug 7/13 (54%) 6/8 (75%) p = NS'

use Homosexual 4/13 (31%) 1/8 (13%) p = NS'

"Student's t = test, all mean +_ "NS = not significant 'Fishers exact.

SEM..

t w e e n h e m a t o c r i t a n d p l a s m a s e l e n i u m (r = .78; p < .0001), b u t n o t w i t h r e d b l o o d cell s e l e n i u m (r = .26; p = .1). A s t r o n g c o r r e l a t i o n was p r e s - e n t b e t w e e n h e m a t o c r i t a n d g l u t a t h i o n e p e r o x i d a s e ac t iv i ty , (r = .66; p < .0001) (Fig. 2).

A s imi la r t r e n d was o b s e r v e d for s e r u m a l b u m i n (Table 3). A signif i- can t p o s i t i v e c o r r e l a t i o n ex i s t ed b e t w e e n p l a s m a s e l e n i u m a n d a l b u m i n (r = .68; p < .0001), b u t n o t b e t w e e n r ed b l o o d cell s e l e n i u m a n d a lbu- m i n leve ls (r = .16; p = .4) G l u t a t h i o n e p e r o x i d a s e ac t iv i ty , h o w e v e r , d i d c o r r e l a t e w i t h s e r u m a l b u m i n levels (r = .61; p < .0005).

As e x p e c t e d , total l y m p h o c y t e c o u n t s w e r e low in b o t h p a t i e n t s w i t h AIDS a n d A R C (Table 3). W h e n all g r o u p s w e r e c o m b i n e d t h e r e w e r e c o r r e l a t i o n s n o t e d b e t w e e n to ta l l y m p h o c y t e c o u n t a n d b o t h p l a s m a sele- n i u m (r = .53; p < .002) a n d g l u t a t h i o n e p e r o x i d a s e ac t iv i ty (r = .65; p < .0001).

Table 3 Laboratory Parameters in Patients with AIDS, AIDS-Related Complex

(ARC), and Controls

Parameter AIDS ARC Controls Significance"

Hematocri t , % 31.6 + 1.6 37.4 + 2.0" 44.7 _+ 1.1' p < .0001 (N = 11) (N = 7) (N = 14)

Serum albumin, 3.3 + .2 4.0 _+ .3 4.8 _+ .1 '1 p < .0001 g/dl (N = 9) (N = 5) (N = 14)

Total lymphocyte 962 +_ 202 1205 + 271 1894 + 219 '~ p < .01 count (N = 11) (N = 6) (N = 14)

~ ~p < .05, vs AIDS, Student's t-test (all Mean _+ SEM). 'p < .02, vs AIDS or ARC, Student's t-test. "p < .002 vs AIDS or ARC, Student's t-test.

Biological Trace Element Research Vol. 15, 1988

Selenium Deficiency in AIDS 173

1"

E

ILl 03 < n X 0 tlr ILl O_

I.LI z 0 "1-

I- ,< I- _1 (9

Fig. 2.

90.0

75.0

60.0

45.0 -

5 0 . 0 -

1 5 . 0 -

o AIDS (N=I I ) �9 ARC (N =7)

- A Control(N =14)

- y =-8.16 + 1.29x r =0.66 A

--_ p< O.O00l AZ~ A ~

0 O 0 0 0

- - / O0 �9

I L I I I I I I I I I I

0 I0.0 20.0 30.0 40.0 50.0 60.0

HEMATOCRIT % Glutathione peroxidase activity vs hematocrit in patients with

AIDS, AIDS-related complex (ARC), and controls.

A good correlation between glutathione peroxidase activity and plasma selenium was present (r = .77; p < .0001), (Fig. 3). Glutathione peroxidase activity also significantly correlated with red blood cell sele- nium levels (r = .37; p < .04).

D I S C U S S I O N

A relative state of selenium deficiency, as determined by diminished plasma, whole blood and erythrocyte levels, and glutathione peroxidase activity is common in patients manifesting infection with human immun- odeficiency virus (HIV) (Table 1, Fig. 1). These abnormalities are most marked in patients with AIDS, but are also present in patients with AIDS-related complex (Table 1) (10,11). The exact mechanism causing this deficiency remains unknown. Previously, we have shown dimin- ished blood selenium levels and other nutritional abnormalities in a dif- ferent group of AIDS patients regardless of the presence or absence of gastrointestinal malabsorption (2,9). Protein-calorie malnutrition is c o m -

Biological Trace Element Research Vol. 15, 1988

174 D w o r k i n e t aL

.1-

E

bJ 09 <

X 0 or" I.iJ 13_

I.IJ z 0 "1" F- < }.-

.d (.9

9 0 . 0 -

7 5 . 0 -

6 0 . 0

4 5 . 0 -

3 0 . 0 -

15.0 -

0

Fig. 3.

o A I D S (N =13) �9 ARC (N = 8) ,a Contro l (N =14)

y = 5 . 8 2 + 7 5 2 . 5 9 x r = 0 . 7 7 A

p< 0 . 0 0 0 1 A Az ~ zx

0 o 0 0

0

~,, 1 0 . 0 2

I I I I I I 0 .05 0 . 0 4 0 . 0 5 0 . 0 6 0 . 0 7 0 . 0 8

P L A S M A S E L E N I U M I~.g/ml

Glutathione peroxidase activity vs plasma selenium in patients with AIDS, AIDS-related complex (ARC), and controls.

mon in AIDS (1-4). Although there was an excellent correlation between serum albumin and plasma selenium, selenium is primarily transported in plasma bound to non-albumin proteins (21,22). Therefore, decreased plasma selenium levels are not a result of decreased albumin binding of selenium. One hypothesis that could explain this finding is that protein malnutrition results in decreased synthesis of both albumin and sele- nium binding proteins. Despite this, diminished red blood cell levels and glutathione peroxidase activity are further evidence of true selenium de- pletion in AIDS (23). Overall, there was good correlation between blood selenium values and glutathione peroxidase activity (Fig. 3). This finding is similar to that reported by others (23-26). Thus, one feature of the ca- chexia commonly seen in AIDS patients is selenium depletion.

The potential relevance of this deficiency as a contributor to clinical disease in AIDS is widespread. Although complex, immunologically AIDS is characterized by deficient effector T-cell response (27-30). Many of the immune defects described in the syndrome are caused by deple- tion and/or dysfunction of the T4 inducer/helper subset of T-lymphocytes (27-30). Although it is absolutely clear that selenium deficiency is not re-

Biological Trace Element Research VoL 15, 1988

Selenium Deficiency in AIDS 17..5

sponsible for the immune defects in AIDs, it is possible that selenium de- pletion could have a further negative impact on the remaining immune competence (6-8,31-33). In animals, selenium depletion results in im- paired fungicidal activity by neutrophils, resulting in yeast infections. In AIDS patients, infections with Candida albicans are common (3,29). Sele- nium deficient animals have reduced levels of thymic hormone, also seen in AIDS (29,31). AIDS patients have impaired serologic responses follow- ing de novo immunizations, which is also true of selenium deficient ani- mals (7,29,31). T-cells from selenium deficient animals show decreased proliferation when presented with antigen and decreased synthesis of interleukin-2 (32). These abnormalities commonly occur in AIDS (28-30). Natural killer cell and T-lymphocyte mediated cytotoxity are reduced in both selenium deficient animals and in patients with AIDS (6,28-30). Pa- tients receiving long term parenteral nutrition with very low blood sele- nium levels had reduced levels of helper T-cells (T4), which increased with selenium repletion (33). In our series, we have shown significant correlations between plasma selenium levels and erythrocyte glutathione peroxidase activity with the total lymphocyte count. These correlations do not prove causation. However, they lend support to the supposition that selenium deficiency superimposed on AIDS could further enhance immune suppression, thereby contributing to the development of oppor- tunistic infections and malignancies.

The risk of a variety of tumors, including Kaposi's sarcoma is in- creased in patients with AIDS (3). A relative selenium deficiency may also predispose to an increased risk of malignancies (34-36). No patient in the present series had Kaposi's sarcoma. The absence of this tumor may reflect the preponderance of intravenous drug abusers, in whom Kaposi's sarcoma is less common than in homosexuals with AIDS (2,37). At this time, it is unclear whether selenium plays any role in cancer asso- ciated with AIDS.

Patients with AIDS are also predisposed to develop a diffuse, four- chamber congestive cardiomyopathy that is refractory to therapy (38,39). The clinical and pathological features of this cardiomyopathy are similar to the cardiomyopathy described in association with selenium deficiency (40-42). Although frank cardiomyopathy was not evident in the patients in this report, cardiac testing to detect subtle abnormalities was not per- formed (39). The mean whole blood selenium level in our AIDS patients was somewhat higher than that reported from endemic areas for Keshan disease or in a patient with selenium-related cardiomyopathy (23,41,42). However, given the range of low selenium levels, the progressive down- hill natural history of AIDS, and their underlying immunodeficiency, AIDS patients may still be at high risk. At this time, the relationship be- tween selenium deficiency and cardiomyopathy in AIDS patients re- mains speculative.

Finally, anemia is also common in AIDS patients, and was present in our series (Table 3). There was an excellent correlation between blood se-

Biological Trace Element Research Vol. 15, 1988

176 Dworkin et al.

l en ium levels, g luta thione peroxidase activity, and hematocri t (Fig. 2). Hemolyt ic anemia may occur in se len ium deficiency (22,23,43). Al though both anemia and se lenium deficiency were found in our AIDS pat ients and the two f indings were s t rongly correlated, a cause and effect rela- t ionship cannot be conc luded from the present data.

We conclude that se len ium deficiency is commonplace in pat ients wi th AIDS and AIDS-related complex. The clinical implications of this deficiency, as d iscussed above, are potentially of great impor tance to AIDS patients. Interactions with other nutr ients related to se lenium, such as vi tamin E, or related to o ther over lapping biochemical functions, such as zinc and i m m u n e funct ion or vi tamin A and cancer risk, could modu la te or amplify the impact of the se len ium deficit. More data are n e e d e d to proper ly define the role of se len ium as a factor in the patho- phys io logy of disease in pat ients with AIDS and AIDS-related complex.

REFERENCES

1. D. P. Kotler, J. Wang, and R. N. Pierson, Am. ]. Clin. Nutr. 42, 1255 (1985). 2. B. Dworkin, G. P. Wormser, W. S. Rosenthal, et al. Am. J. Gastro. 80, 774

(1985). 3. A. S. Fauci, A. M. Macher, D. L. Longo, et al., Ann. Int. Med. 100, 92 (1984). 4. R. H. Gray, A. J. Pub. Health 73, 1332 (1983). 5. R. T. Chelbowski, Nutr. Cancer 7, 85 (1985). 6. H. C. Meeker, M. L. Eskew, W. Schenchenzuber, J. Leuk. Biol. 38, 451

(1985). 7. J. E. Spallholz, "Selenium: What role in immunity and immune

cytotoxicity," in, Selenium in Biol. and Med., J. E. Spallholz, J C. Martin and H. E. Ganther, eds., Avi, Westport, CT, 1981, pp. 103-117.

8. R. Boyne and J. R. Arthur, J. Comp. Pathol. 89, 151 (1979). 9. B. M. Dworkin, W. S. Rosenthal, G. P. Wormser, et al., J. Parent. Ent. Nutr.

10, 405 (1986). 10. Center for Disease Controls, Morbid. Mortal. Weekly 34, 373 (1985). 11. Center for Disease Controls, Morbid. Mortal. Weekly 35, 334 (1986). 12. P. A. Whetter, D. E. Ullrey, J. Assoc. Off. Anal. Chem. 61, 927 (1978). 13. O. E. Olson, J. Assoc. Off. Anal. Chem. 52, 627 (1969). 14. B. Dworkin, L. J. Newman, S. Berezin, et al., J. Parent. Ent. Nutr. 11, 38

(1987). 15. B. Dworkin, S. Weseley, W. S. Rosenthal, et al. Am. J. Med. Sci. 30, 6 (1987). 16. B. Dworkin, W. S. Rosenthal, G. G. Gordon, Alcohol. Clin. Exp. Res. 8, 535

(1984). 17. L. Floh6 and W. A. Guenzler, Methods in Enzymology 105, 114 (1984). 18. D. E. Paglia and W. N. Valentine, J. Lab. Clin. Med. 70, 158 (1967). 19. B. Faraji, H. K. Kang, and J. L. Valentine, Clin. Chem. 33, 539 (1987). 20. G. W. Snedcor and W. G. Cochran, Statistical Methods, Iowa State Univ.

Press, Ames, IA, 1980. 21. A. Motsenbocker and A. L. Tappel, Biochim. Biophys. Acta 719, 147 (1982). 22. R. F. Burk, "Selenium in Man," in, Trace Element in Human Health and Dis-

ease, vol. II, A. Prasad, ed., Academic, New York 1976, pp. 105-133. 23. O. A. Levander, Fed. Proc. 44, 2579 (1985). 24. C. D. Thomson and M. F. Robinson, Am. J. Clin. Nutr. 33, 303 (1980).

Biological Trace Element Research VoL 15, 1988

Selenium Deficiency in AIDS 177

25. H. W. Lane, S. Dudrick, and D. C. Warren, Proc. Soc. Exp. Biol. Med. 167, 383 (1981).

26. A. Kauppila, H. Korpela, U. M. Makila, et al., Brit. Med; J. 192, 150 (1987). 27. J. Lawrence, Scientif. Am. 253, 84 (1985). 28. F. P. Siegal, Semin. Onc. 11, 29 (1984). 30. G. V. Quinnan, J. P. Siegel, J. S. Epstein, et al., Ann. Int. Med. 103, 710

(1985). 31. R. K. Chandra and D. H. Dayton, Nutr. Res. 2, 721 (1982). 32. M. Fischback, S. Jenkinson, N. Talal, et al., Cin. Res. 32, 232 A (Abstract). 33. M. E. Shils, D. Jacobs, Cunningham-Rundles, et al., Am. Soc. Clin. Nutr.

(Abstract), 1983. 34. Committee of Diet, Nutrition and Cancer, Assembly of Life Sciences, Natl.

Res. Council: Minerals in Diet, Nutrition and Cancer. Natl. Acad. Press, Wash- ington, DC, 1982, pp. 162-201.

35. W. C. Willett, J. S. Morris, S. Pressel, et al. The Lancet 1983; 2, 130. 36. G. N. Schrauzer, "Selenium and cancer; Historical developments and per-

spectives," in J. E. Spallholz, J. L. Martin, and H. E. Ganther, eds., Selenium in Biology and Medicine, Avi, Westport, CT, 1981, pp. 98-102.

37. D. C. DeJarlais, M. Marmot, P. Thomas, et al., New Engl. J. Med. 210, 1119 (1984).

38. I. S. Cohen, D. W. Anderson, R. Virmani, et al. New Engl. J. Med. 315, 628 (1986).

39. L. Fink, N. Reichek, and M. G. Sutton, Am. J. Cardiol. 54, 1161 (1984). 40. R. ,A. Johnson, S. S. Baker, J. T. Fallon, et al., New Engl. J. Med. 304, 1210

(1981). 41. C. R. Fleming, J. T. Lie, J. T. McCall et al., Gastroenterology 83, 689 (1982). 42. Keshan Disease Research Group, Chinese Med. J., 92, 471 (1979). 43. S. Gross, Semin. Hematol. 3, 187 (1976).

Biological Trace Element Research VoL 15, 1988