Iron status and cellular immune competence

Iron Status and Cellular Immune Competence

M. F. Good, L. W. Powell, J. W. Halliday

S UM M A R Y. There is increasing evidence that both iron overload and iron deficiency are associated with sign&ant abnormalities of immune function. In diseases associated with iron overload there is increased susceptibility to both infection and neoplasia. The precise mechanisms are still being unravelled but iron overload has been shown to impair antigen-specifk immune responses and to reduce the number of functional helper precursor cells. Similarly, iron in vitro in concentrations reported to be present in the serum of patients with iron overload impairs the generation of cytotoxic T-cells, enhances suppressor T-cell activity and reduces the proliferative capacity of helper T-cells. The predominant tumor seen in iron overload is primary hepatocellular carcinoma; however other aektogical factors appear to be involved in addition to iron overload, especially hepatic cirrhosis. Never&less, primary liver cancer occurs much more frequently in hemochromatosis than in other forms of cirrhosis. Iron deticiency is associated with an altered response to infection but the relationship is again a complex one. The celhtbu mechanisms involved have yet to be clearly defined, although impaired T and B cell function have been demonstrated.

Iron overload and iron deficiency are not usually thought of clinically as being associated with immunological disturbances. However, abnormalities of immune function have been reported with increasing frequency in states of altered body iron stores (see below). In this review, we have re-examined the association between iron overload, infection, and tumorigenesis and the effects of iron deficiency on the immune system. The review also focuses on recent experimental studies which have examined directly the effects of excess iron and iron deficiency on immune competence.

M. F. Goad, The Laboratory of Parasitic Diseases, National Institute of Allergy and Infectious Diseases, NIH, Bethesda, Maryland, 20892, USA. L. W. Powell, J. W. HaUidsy, The Liver Unit, Department of Medicine, University of Queensland, Royal Brisbane Hospital, Brisbane, 4029, Australia. Bhd Reviews (I 988) 2, 4349 Q 1988 Longman Group UK Ltd

Iron Overload Iron Overload and Infection Any effect that iron has on the host parasite relationship depends on both the effect of iron on microbial growth, and its effect on host immune and nonspecific defence mechanisms. There are many reports of the concurrence of clinical iron overload and increased susceptibility to infection [reviewed by Weinberg and Gross & Newberne’*‘]. Rarely has an immunological basis for the concurrence been sought; however, impaired phagocytic function has been demonstrated. Thus, Waterlot et al3 studied hemodialysis patients and noted that those who had the highest serum ferritin concentrations had significantly more infectious episodes and had severely reduced leucocyte phagocytosis and myeloperoxidase activity when compared to those with normal concentrations of serum ferritin. Treatment with the iron chelating

44

IRON STATUS AND CELLULAR IMMUNE COMPETENCE

agent desferrioxamine resulted in a return to normal of the above parameters. In addition, van Asbeck et al4 noted a case of Listeria monocytogenes meningitis in an iron-overloaded patient with hemochromatosis. This organism is often a commensal and rarely causes pathology except in immunosuppressed patients. The authors observed that serum from the patient inhibited phagocytic cell function and, furthermore, reduction of iron stores by therapeutic phlebotomy reduced the inhibitory effect of serum on phagocyte function. In a subsequent report, van Asbeck et al5 studied several patients with a variety of iron-overload disorders and noted decreased phagocytic function in a majority of patients. There are many other examples of an association between iron overload and susceptibility to infection where a direct enhancing effect of serum on microbial growth has been found. 6*7 In many cases, however, the possibility that iron also has a direct effect on immunological and phagocytic function has not been excluded. Iron Overload and Neoplasia

There is a strong association between local or generalized iron-overload and the development of neoplasia. Tumors have been reported with an increased incidence at the site of iron dextran injections, in the lungs of mine workers who inhale iron oxide, and in hemochromatosis patients, as documented below. Sarcomas have been reported at the site of iron dextran injections both in animals and humans.8*9 Langvad,’ for example, noted that mice often developed tumors at sites distant from that of the injection. Similarly, Bergeron et al” noted that when mice were given iron intraperitoneally at levels comparable to clinical doses for humans, they were much more sensitive to tumor cells which had been previously inoculated. There are only a limited number of human cases reported, but it is possible that tumors developing at sites other than that of the iron injection may have been missed. The association between iron overload and tumorigenesis is not restricted to injectable forms of iron. In an animal model, inhalation of iron oxide resulted in neoplasia at the site of accumulation of the iron 0xide.l’ Similarly, mine workers exposed to oxide dust have a much increased risk of developing neoplasms of the lower respiratory tract.‘* The major causes of iron overload are listed in the Table. The most clinically significant association with respect to iron overload and tumorigenesis occurs with the disease genetic hemochromatosis (HC). HC is inherited as an autosomal recessive trait. The HC allele, which is in strong linkage disequilibrium with HLA-A3, occurs at a frequency of 0.05 or greater in Caucasian populations. An association between HC and neoplasia was reported in 1971, when Powell et alI3 reviewed many cases of cirrhosis. They noted that

Tabk

Causes of iron overload

1. Genetic hemochromatosis 2. Secondary (acquired) iron overload (a) Iron-loading anemias, e.g. thalassemia (b) Transfusional iron overload (c)* Alcoholic cirrhosis (d)* Dietary iron overload (e)* Porta-systemic shunting (f)* Porphyria cutanea tarda *There is evidence that iron overload in these disorders is associated with at least one allele for hemochromatosis but this is currently under review

when HC was the cause of cirrhosis, primary liver cancer accounted for 36.6% of deaths, compared with O-10% of deaths when cirrhosis was the result of other diseases. More recent studies14*” have reported that the relative risk for hepatocellular cancer is greater than 200; however, neither study confirmed an increased risk for extra-hepatic neoplasia as suggested and adequate control population studies are lacking in these earlier reports. While iron overload would clearly appear to be related to risk for hepatocellular cancer, it is probably not the sole cause, since in HC, hepatic neoplasia has not been documented prior to the development of cirrhosis. Hepatic cirrhosis in the absence of iron overload has a much lower association with hepatic neoplasia (of the order of 510%). It is possible that cirrhotic livers are more prone to neoplastic events but that developing malignant clones are controlled by cellular immune mechanisms. In the presence of iron overload, however, immune function is compromised and malignant clones are unimpeded. Such a hypothesis would be consistent with the data presented below and with the data of Bergeron et al” using a mouse model. Another possibility is that an oncogene is inherited in linkage disequilibrium with the HC allele. Excess body iron may result in activation of such an oncogene. Hemosiderosis as a result of excess dietary intake (e.g. Bantu Siderosis) is also associated with primary hepatocellular carcinomas. Over 50% of all tumors that develop in black South African males have been reported to be primary hepatomas.*‘**’ However, the relative roles of cirrhosis, iron loading and the Hepatitis B virus have yet to be determined. previously'6~'7,'8,'9

The Efect

of Iron on Cellular Immune Function

Both iron deficiency and iron excess have been shown to alter cellular immune and non-specific defence functions. During the last 10 years, the effect on cellular immune function of excess iron in vitro has been studied by a number of investigators. Some of the earliest studies demonstrated that iron, either as ferric citrate or in the form of iron-saturated transferrin, inhibited human T-cell rosette formation.22*23 The first demonstration that iron inhibited functional lymphocyte activity was made by Bryan et a1.24 They demonstrated that ferric citrate at

BLOOD REVIEWS

concentrations of 100 uM or greater could inhibit a human mixed lymphocyte reaction. Iron was shown to act on the responder cells rather than the stimulator cells. The authors also observed that cells from HLA-A2 donors were significantly less susceptible to iron. They speculated that the survival advantage that HLA-A2 confers to sufferers of acute lymphocytic leukemia” and acute myeloid leukemia26 -conditions in which patients have high serum iron levels-may be the result of less inhibition of immune function. The development of infection and poor prognosis in leukemia has already been directly related to high serum iron levels.27 Keown & Descamps-Latscha2* reported that iron, in a variety of different forms, could inhibit antigeninduced lymphocyte proliferation and cytotoxic T lymphocyte (CTL) sensitization, but not CTL effector function. In their studies they used varying concentrations of ferric iron as ferric citrate, ferric nitrate, and also ferric chloride at pH 7.2 where solubility may be a problem. They also reported that Fe3’ did not affect mitogen-presentation by monocytes. Esfects of Ferritin on Immune Function A high serum ferritin concentration is characteristic of iron-overload. In vitro, ferritin has been shown to exert certain modulatory effects on T-cell function. Keown and Descamps-Latscha2* used ferritin in concentrations of 1.Og/L but the concentration of ferritin found even in patients with acute inflammatory conditions rarely exceeds 10 mg/L. Matzner et a129 demonstrated that ferritin at a concentration of 0.25-5 ug/ml caused a suppression of PHA and concanavalin A-induced blastogenesis; however, we have demonstrated that ferritin at concentrations found in the serum of patients with iron-overload does not inhibit human antigen-secific T-cell proliferation in vitro.30 It has been claimed that acidic isoferritins which may be present in certain tumors and which contain excess H-ferritin subunits, can inhibit granulocytemacrophage colony formation in vitro31 and multipotential and erythroid colony formation.32 This may be relevant to the pancytopenia associated with leukemia, since acidic isoferritins have been reported at much greater concentrations in leukemic patients.33 However, this view has been questioned.34

45

bound. It has been shown that in iron-overload, nontransferrin bound iron can be present at concentrations of up to 30 PM. 35*36In recent studies, we have examined the effect of non-transferrin bound Fe3+ at concentrations as low as 30 uM on the cloning efficiency and rate of clone growth of individual human tetanus toxoid-specific memory helper T-cell~.~~ We have shown that iron at the lowest concentrations tested reduced by more than 50% both the cloning efficiency of the helper T-cells and the rate of clone growth of the T-cells that did proliferate (Fig. 1). In contrast, human liver ferritin over a wide range of concentrations had no effect on helper T-cell growth. To examine the effects of low concentrations of non-transferrin bound iron on other T-lymphocyte subsets, we used a murine model and studied by clonal techniques the various cells involved in the generation of CTL.37 Thus, CTL-precursors (CTL-P), helper T-lymphocytes (HTL) and their precursors (HTL-P) and suppressor T-lymphocytes (STL) and their precursors (STL-P) were studied. Fe3 + at concentrations of 10 uM or greater directly inhibited the cloning frequency of CTL-P (Fig. 2). This effect was independent of its inhibitory effect on the growth rate of antigen-specific HTL. Iron at low concentrations also significantly increased the suppressive effect of concanavalin A-induced STL on the

6

/

T

Eflects of Non-Transferrin-Bound Iron on Immune Function In clinical iron overload, most of the excess iron is stored as ferritin or hemosiderin, while in the serum it is complexed to transferrin. Each transferrin molecule can bind two molecules of iron. In normal iron status, the transferrin is approximately one third saturated whereas in iron-overload, the transferrin may be more than 90% saturated. In such situations, a small, but significant amount of iron is ‘free’ or non-transferrin

0

0.03

0.10

0.30

1.0

Ferric Citrate (mM) Fig. 1 Dose response study of the effect of ferric citrate on the cloning efficiency (estimated HTL-P frequency) and rate of clone growth of human HTL. The horizontal bars represent the upper limit of the 95% confidence interval.

46

IRON STATUS AND CELLULAR IMMUNE COMPETENCE

Cells Per Well 750

140

1500

130 3g

. [I

..

0

.

0 .

120-

0

_. :

Y

: .

d F

110-

.

$

loo-

: -9-Q

I; 8

.

l

go-:

.

* o 0

Ferric Citrate (390 PM)

NHLF (10,000 pgil)

90% Iron Saturated Transferrin (2 g/l)

Fig. 3 NK activity of peripheral blood lymphocytes from normal subjects (0) and patients with hemochromatosis (0) expressed as per cent of control lytic unit values (LU,,) in the presence of ferric citrate (300 PM), human liver ferritin (10 000 pg/L) and transferrin at 90% iron saturation (2 g/L).

Fig. 2 Limit--dilution analysis of C57 spleen cells generating cytotoxicity against Balb/c allo determinants in the presence of either 1 mM ferric citrate (A) or 1 mM sodium citrate (0).

generation of CTL from CTL-P. The overall inhibitory effect of Fe3 + on the generation of CTL was thus shown to be due to an effect of iron on all the major T-lymphocyte subsets. Another lymphocyte subset, having a tumor surveillance function38 as well as a role in resistance to virus infection3’ is the natural killer (NK) cell. It has been reported that lymphocytes from iron-overloaded poly-transfused thalassemic patients have a reduced NK cell activity. Pre-incubation of the patients’ cells with iron-chelating agents was able to restore normal NK activity suggesting that the defect resulted from excess iron4’ Work in our laboratory, however, has shown that non-transfused HC patients with moderate to severe iron overload have a normal level of NK activity. Furthermore, NK cell activity from normal individuals was not significantly reduced by the addition in vitro of high concentrations of iron or ironbinding proteins (transferrin, ferritin) (Fig. 3).41 Thus, it appears that iron per se does not affect NK activity and the reduction observed in thalassemic patients may be transfusion-related and not due to iron alone. Hence, transfused donor lymphocytes may directly or indirectly alter NK cell activity. In another study of thalassemic patients, the investigators have examined both phagocytic and lytic activities of peripheral blood monocytes.42 They found that whereas the phagocytic activity did not differ

from that of controls, lytic activity (using C. pseudotropicalis as the target) was significantly decreased. They also found significant inverse correlations between lytic activity and both the age of the patients and serum ferritin levels. No correlation was observed with either liver damage or treatment with desferrioxamine. They therefore suggested that iron overload, per se, resulted in decreased microbicidal function. Their results support those of van Asbeck et al4 (see above). Although excess iron can inhibit lymphocyte function, these cells require iron for normal cytochrome and enzyme function, including ribonucleoside reductase, an enzyme required for DNA synthesis. Transferrin-bound iron appears to be the form in which iron must be supplied. 43 Optimally, lymphocyte transformation specific for a given antigen occurs in the presence of transferrin which is between 30% and 70% saturated. At higher saturation, lymphocyte transformation is markedly reduced,44 possibly because at such high saturation, a significant fraction of the iron will be ‘free’ or non-transferrin bound and this non-transferrin bound iron (up to 30 PM) is capable of inhibiting lymphocyte transformation.” The E#ect of Experimental Iron Overload on Lymphocyte Function

To examine the effect of in vivo iron overload on lymphocyte function we used a murine model where animals were iron loaded either by adding carbonyl iron to their diet or by parenteral administration of iron dextran.45 Histological examination of liver sections revealed that the former method resulted in hepatocyte loading (similar to HC) and the latter method resulted in Kupffer cell loading (similar to transfusional siderosis). Spleen cells from these ani-

BLOOD REVIEWS

mals were examined in vitro in a mixed lymphocyte culture in the presence of normal concentrations of iron. Spleen cells from iron-loaded animals had markedly reduced ability to generate allo-specific CTL; However, a frequency analysis of CTL-precursors (CTL-P) revealed that these same iron-loaded animals had normal numbers of allo-specific CTL-P, suggesting that a regulatory cell defect was responsible for the reduced ability to generate effector cells in the mixed lymphocyte culture. This was confirmed by demonstrating that an interleukin-2-containing supernatant from activated lymphocytes could restore the functional capacity of lymphocytes from the ironloaded animals in a mixed lymphocyte culture. A frequency analysis of helper T-cells in the spleens of the iron-loaded animals confirmed that these cells were indeed present at a reduced frequency, and mixing experiments between cells from iron-loaded and control animals provided no evidence for an involvement of suppressor T-cells. Iron overload can thus place a limit on the amount of available T-cell helper activity in vivo apparently by a reduction in the number of functional helper Tcells. The mechanism for this is not known, but it would be interesting to examine the possibility as to whether iron or an iron-binding protein interacts with the helper cell-specific CD4 molecule. Bryan et al46 found that iron (Fe3+) suppressed expression of the human helper T-cell CD4 molecule, on mitogenactivated cells, resulting in a significantly reduced helper: suppressor ratio. They showed this to be a specific defect, since iron did not alter expression of other markers on activated cells, including T8, Ia, and the sheep thermostable red blood cell receptor. Unfortunately, only high concentrations of iron were used in their study (300 uM), and such high concentrations of serum iron are not observed clinically even in severe iron overload. No mention has been made of how iron overload might directly exert its immunomodulatory effects. Unfortunately, this is not well understood. Certainly, non-protein bound iron within the lymphocyte would be toxic, promoting free radical generation and causing lipid peroxidation;47*4* However, toxicity could not explain the differential sensitivity of various lymphocyte subsets to iron. Intracellular iron may also modulate the expression and function of iron-dependent enzymes. For example, ribonucleoside reductase, which is involved in the reduction of 4-ribonucleoside diphosphates to deoxyribonucleoside diphosphates is iron dependent. 49 The iron-chelating agent desferrioxamine has been shown to inhibit this enzyme within lymphocytes by 90%50 and was also found to suppress mitogen-induced proliferation, mixed lymphocyte cultures, and the generation of CTL.” In summary, excess iron, at levels comparable to those found in the serum of patients with iron overload, has been shown to impede the generation of antigen-specific immune responses in humans and animals. In addition, iron overload in animals results

47

in a reduction of the numbers of functional helper precursor cells. If a similar event occurs in humans, iron overload would have a compounding effect in vivo on the generation of an immune response. Such an effect may in part explain the strong association between clinical iron overload and susceptibility to infection and neoplasia.

Iron Deficiency Iron deficiency has been reported to affect the incidence of infection, although the results are conflicting. The effect of iron supplementation on iron-deficient neonates is an area of considerable controversy. In a randomized, placebo-controlled study, Oppenheimer et al52 reported that iron-deficient infants from Papua New Guinea, developed fewer malarial infections than infants treated with iron-dextran injections. Other studies have demonstrated that ‘iron-stressed’ neonates have a lowered resistance to bacterial infections.53*54 A 7-fold increase in gram-negative meningitis and septicaemia occurred in babies within 1 week of injection of iron dextran. Similarly, iron stress to protein-deficient patients results in fully ironsaturated serum transferrin. Administering iron to these patients, without prior protein feeding to increase transferrin levels, results in increased susceptibility to infectionss5 Thus it could be argued that marginal iron deficiency is beneficial. Van Heerden et al 56 for example, demonstrated that in iron-deficient children with anemia, T-cell responses to mitogens were significantly reduced. In the absence of anemia, however, iron deficiency was not associated with a reduced mitogen response, Walter et al57 observed that the bactericidal activity of neutrophils from iron-deficient children was significantly reduced. Iron therapy resulted in a return to normal bactericidal activity, but this took 2 weeks to occur. The authors speculated that iron deficiency during the critical time of neutrophil development in the bone marrow resulted in a permanent impairment of their function. Animal studies have revealed that iron-deficient, anemic mice have impaired blastogenic responses to T and B cell-specific mitogens,” as well as a reduced capacity of generate allo-specific cytotoxic T lymphocytes. 59 The reduced responses to mitogens could be restored by iron repletion. In contrast, however, others have not found a reduced mitogenic response in lymphocytes from iron-deficient mice.60 While the mice in the latter study were anemic (Hb: i1.4g/dl compared with 13.5 g/d1 in control animals), they were not as severely anemic as the mice in the former study (Hb: 4.5 g/d1 compared with 12.6 g/d1 in the control group), which may explain the observed differences. A different study, in rats, however, has shown that while thymocytes from iron-deficient anemic rats exhibit reduced reactivity to mitogens, splenocytes from these rats have increased activity.‘j’ Antibody synthesis following immunization has

48

IRON STATUS AND CELLULAR IMMUNE COMPETENCE

been reported to be depressed in iron-deficient rat pups. Kochanowski & Sherman62 reported that in pups rendered iron deficient by feeding their mothers hypoferremic diets during gestation and lactation, antibody synthesis post-immunization was decreased by at least 50% compared with controls. Iron repletion post-weaning resulted in only slight improvement. The effect of iron deficiency on antibody formation could have been due to an effect on development or function of either B-cells or regulatory T-cells. Thus, iron deficiency is associated with an altered response to infection, although the relationship is complex. On the one hand, iron-deficient infants appear to be more resistant to some infections, e.g. malaria and iron supplementation may lead to lowered resistance to bacterial infection. On the other hand animal studies have demonstrated impaired T and B cell function and impaired antibody synthesis following immunization in iron-deficiency states. The subject is clearly one of considerable clinical significance but further detailed studies of immune function in iron deficiency and after iron supplementation are required before firm recommendations can be made. Meanwhile, iron deficiency in infancy should be corrected with some caution, especially in tropical countries.

12.

13.

14.

15.

16.

17.

18. 19.

20. 21.

Acknowledgements

22.

These studies were supported in part by the National Health and Medical Research Council of Australia and the University of Queensland Mayne Bequest.

23.

References

24.

4.

5.

6.

I.

8. 9. 10. 11

Weinberg E D 1978 Iron and infection. Microbiological Reviews 42: 45-66 Gross R L, Newbeme P M 1980 Role of nutrition in immunologic function. Physiological Reviews 60: 188-302 Waterlot Y, Cantinieaux B, Hariga-Muller C, De Maertelaere-Laurent E, Vanherweghem J L, Fondu P 1985 Impaired phagocytic activity of neutrophils in patients receiving haemodialysis: the critical role of iron overload. British Medical Journal 291: 501-504 Van Asbeck B S, Marx J J M, Struyuenberg A, Verhoef J 1984 Functional defects in phagocytic cells from patients with iron overload. Journal of Infection 8: 232-240 Van Asbeck B S, Verbrugh H A, Van Oost B A, Marx J J M, Imhof H W, Verhoef J 1982 Lisreriu monocyrogenes meningitis and decreased phagocytosis associated with iron overload. British Medical Journal 284: 542-544 Schade A L, Caroline L 1944 Raw hen egg white and the role of iron in growth inhibition of Shigellu dysenteriae. Staphylococcus aureus, Escherichia coli, and Saccharomyces cerevisiae. Science 100: 14-15 Gordeuk V R, Brittenham G M, McLaren G D, Spagnuolo P J 1986 Hyperferremia in immunosuppressed patients with acute nonlymphocytic leukemia, and the risk of infection. Journal of Laboratory and Clinical Medicine 1081466-472 Weinbren K, Salm R, Greenberg G 1978 Intramuscular injections of iron compounds and oncogenesis in man. British Medical Journal i: 683-685 Langvad E 1968 Iron-dextran induction of distant tumors in mice. International Journal of Cancer 3: 415423 Bergeron R J, Streiff R R, Elliott G T 1985 Influence of iron on in vivo proliferation and lethality of L 1210 cells. Journal of Nutrition 115: 369-374 Nettesheim P, Creasia D A, Mitchell T J 1975 Carcinogenic and cocarcinogenic effects of inhaled synthetic smog and

25.

26.

27.

28.

29.

30.

31

32.

33.

ferric oxide particles. Journal of the National Cancer Institute 55: 1599169 Anthoine D, Braun P, Cervoni P, Schwartz P, Lamy P 1979 Le cancer bronchique des mineurs de fer de Lorraine peut-il etre considere comme une maladie professionelle? A propos de 270 nouveaux cas observes de 1964 a 1978. Revue Francaise de Maladies Respiratoires 7: 63-65 Powell L W, Mortimer R, Harris 0 D 1971 Cirrhosis of the liver. A comparative study of the four major aetiological groups. Medical Journal of Australia 1: 941-950 Bradbear R A, Bain C, Siskind V, Scholfield F D, Webb S, Axelsen E, Halliday J W, Bassett M L, Powell L W 1985 Cohort study of internal malignancy in genetic hemochromatosis and other chronic non-alcoholic liver diseases. Journal of the National Cancer Institute 75: 81-84 Niederau C, Fischer R, Sonnenberg A, Stremmel W, Tramp&h H J, Strohmeyer G 1985 Survival and causes of death in cirrhotic and in non-cirrhotic patients with primary hemochromatosis. New England Journal of Medicine 313 (20): 12561262 Bomford A, Williams R 1976 Long term results of venesection therapy in idiopathic hemochromatosis. Quarterly Journal of Medicine 45: 61 l-623 Viteri A, Vemace S J, Schaffner F 1976 Extrahepatic malignancy in chronic liver disease: report of six cases. Gastroenterology 71: 1075-1078 Milder M S, Cook J D, Stray S, Finck C A 1980 Idiopathic hemochromatosis, an interim report. Medicine 59: 34-49 Ammann R W, Muller E, Bansky J, Schuler G, Hacki W H 1980 High incidence of extrahepatic carcinomas in idiopathic hemochromatosis. Scandinavian Journal of Gastroenterology 15: 733-736 Berman C 1958 Primary carcinoma of the liver. Adv. Cancer Research 5: 55-90 Robertson M A, Harington J S, Bradshaw E 1971 The cancer pattern in African gold miners. British Journal of Cancer 25: 395402 DeSousa M, Nishiya K 1978 Inhibition of E-rosette formation by two iron salts. Cellular Immunology 38: 203-208 Nishiya K, deSousa M, Tsoi E, Bognacki J J, de Harven E 1980 Regulation of expression of a human lymphoid cell surface marker by iron. Cellular Immunology 53: 71-83 Bryan C F, Nishiya K, Pollack M S, DuPont B, deSousa M 1981 Differential inhibition of the MLR by iron: association with HLA phenotype. Immunogenetics 12: 129-140 Rogentine G N, Trapani R J, Yankee R A, Henderson E S 1973 HL-A antigens and acute lymphoblastic leukemia. The nature of the HL-A2 association. Tissue Antigens 3: 47&476 Jeannet M, Fliedner V, Sultan-Kahan Z 1980 HLA-DRW antigens associated with acute leukemia. Proceedings of the VIIIth International Histocompatibility Workshop (Abstract) Caroline L, Rosner F, Koziner P J 1969 Elevated serum iron, low unbound transferrin and candidiasis in acute leukemia. Blood 34: 441451 Keown P, Descamps-Latscha B 1983 In vitro suppression of cell-mediated immunity by ferroproteins and ferric salts. Cellular Immunology 80: 257-266 Matzer Y, Hershko C, Polliack A, Konijn A M, Izak G 1979 Suppressive effect of ferritin on in vitro lymphocyte function. British Journal of Haematoloav 42: 345-353 Good M F, Chapman D E, PIdwell L W, Halliday J W 1986 The effect of iron (Fe3+) on the cloning efficiency of human memory T4+ lymphocytes. Clinical and Experimental Immunology 66: 340-347 Lu L, Broxmeyer H E, Meyers P A, Moore M A S, Thaler H T 1983 Association of cell cycle expression of Ia-like antigenic determinants on normal human multi-potential (CFUGEMM) and erythroid (BFU-E) progenitor cells with regulation by acidic isoferritins. Blood 61: 250-256 Broxmeyer H E, Bognacki J, Ralph P, Dorner M H, Lu L, Castro-Malaspina H 1982 Monocyte-macrophagederived acidic isoferritins: normal feedback regulators of granulocyte macrophage progenitor cells in vitro. Blood 60: 595-607 Broxmeyer H E, Bognacki J, Domer M H, deSousa M 1981 Identification of leukemia associated inhibitory activity as

BLOOD REVIEWS

34. 35.

36.

31.

38.

39.

40.

41.

42.

43.

44.

45.

46.

41.

acidic isoferritins. Journal of Experimental Medicine 153: 1426-1444 Jacobs A 1983 Do acidic isofertitins regulate haemopoiesis? British Journal of Haernatology 55: 199-202 Hershko C, Graham G, Bates G W, Rachmilewitz E A 1978 Non-specific serum iron on thalassaemia: An abnormal serum iron fraction of potential toxicity. British Journal of Haematology 40: 255-263 Batey R Cl, Lai Chung Fong P, Shamir S, Sherlock S 1980 A non-transfer&-bound serum iron in idiopathic hemochromatosis. Digestive Diseases and Sciences 25: 34&346 Good M F, Powell L W, Halliday J W 1987 The effect of non-transferrin bound iron on murine T lymphocyte subsets: Analysis by cloning techniques. Clinical and Experimental Immunology. In press Haller 0, Hansson M, Kiessling R, Wigzell H 1977 Role of nonconventional natural killer cells in resistance against syngeneic tumor cells in vivo. Nature 270: 609611 Blazer B, Patarroyo M, Klein E, Klein G 1980 Increased sensitivity of human lymphoid lines to natural killer cells after induction of the Epstein-Barr viral cycle by superinfection or sodium butyrate. Journal of Experimental Medicine 151: 614627 Akbar A N, Fitzgerald-Bocarsly PA, deSousa M, Giardinia P J, Hilgartner M W, Grady R W 1986 Decreased natural killer activity in Thalassemia Major: A possible consequence of iron overload. Journal of Immunology 136: 1635-1640 Chapman D E, Good M F, Powell L W, Halliday J W 1987 The effect of iron, iron-binding proteins and iron-overload on human natural killer cell activity. Journal of Gastroterology and Hepatology. In press Ballart I J, Estevez M E, Sen L, Diez R A, Giuntoli J, deMiani S A, Penalver J 1986 Progressive dysfunction of monocytes associated with iron overload and age in patients with thalassemia major. Blood 67: 105-109 Brock J H, Mainou-Fowler T 1983 The role of iron and transferrin in lymphocyte transformation. Immunology Today 4: 347-351 Brock J H 1981 The effect of iron and transferrin on the response of serum-free cultures of mouse lymphocytes to concanavalin A and lipopolysaccharide. Immunology 43: 387-392 Good M F, Chapman D E, Powell L W, Halliday J W 1987 The effect of experimental iron-overload on splenic T cell function: analysis using cloning techniques. Clinical and Experimental Immunology 68: 375-383 Bryan C F, Leech S H, Boxelka B 1986 The immunoregulatory nature of iron. II. Lymphocyte surface marker expression. Journal of Leukocyte Biology 40: 589-600 Halliwell B 1978 Superoxide-dependent formation of hydroxyl radicals in the presence of iron chelates. Is it a mechanism for hydroxyl radical production in biochemical systems? FEBS Letters 92: 321-326

49

48. Heys A D, Dormandy T L 1981 Lipid peroxidation in ironoverloaded spleens. Chnical Science 66: 295-301 49. Brown N C, Eliasson R, Reichard P, Thelander L 1969 Spectrum and iron content of protein B2 from ribonucleoside diphosphate reductase. European Journal of Biochemistry 9: 5 12-5 18 50 Hoffbrand A V, Ganeshaguru K, Hooton J W L, Tattersall M H N 1976 Effect of iron deficiency and desferrioxamine on DNA synthesis in human cells. British Journal of Haematology 33: 5 17-526 51 Bowem N, Ramshaw I A, Badenoch-Jones P, Doherty P C 1984 Effect of an iron chelating agent on lymphocyte proliferation. Australian Journal Experimental Biology and Medical Science 62: 743-754 52 Oppenheimer S J, Gibson F D, Macfarlane S B, Moody J B, Harrison C, Spencer A, Bunari 0 1986 Iron supplementation increases prevalence and effects of malaria: Report on clinical studies in Papua New Guinea. Transactions of the Royal Society for Tropical Medicine and Hygiene 80: 603-6 I 2 53. Barry D M J, Reeve A W 1977 Increased incidence of gramnegative neonatal sepsis with intramuscular iron administration. Pediatrics 60: 908-912 54 Becroft D M 0, Dix M R, Farmer K 1977 Intramuscular iron-dextran and susceptibility of neonates to bacterial infection. Archives of Diseases of Children 62: 778-781 55. McFarlane H, Reddy S, Adcock K J, Adeshina H, Cooke A R, Akena J 1970 Immunity, transferrin and survival in kwashiorkor. British Medical Journal 4: 268-270 56. Van Heerden C, Oosthuizen R, van Wyk H, Prinsloo P, Anderson R 1981 Evaluation of neutrophil and lymphocyte function in subjects with iron deficiency. South African Medical Journal 59: 11l-1 13 57. Walter T, Arredondo S, Arevalo M, Stekel A 1986 Effect of iron therapy on phagocytosis and bactericidal activity in neutrophils of iron-deficient infants. American Journal of Clinical Nutrition 44: 877-882 58. Kuvibidila S, Nauss K M, Baliga B S, Suskind R M 1983a Impairment of blastogenic response of splenic lymphocytes from iron-deficient mice: in vivo repletion. American Journal of Clinical Nutrition 37: 15-25 59. Kuvibidila S R, Baliga B S, Suskind R M 1983b The effect of iron-deficiency anemia on cytolytic activity of mice spleen and peritoneal cells against allogeneic tumor cells. American Journal of Clinical Nutrition 38: 238-244 60. Mainou-Fowler T, Brock J H 1985 Effect of iron deficiency on the response of mouse lymphocytes to concanavalin A: The importance of transferrin-bound iron. Immunology 54: 325-332 61. Soyano A, Candellet D, Layrisse M 1982 Effect of iron deficiency on the mitogen-induced proliferative response of rat lymphocytes. International Archives of Allergy and Applied Immunology 69: 353-357 62. Kochanowski B A, Sherman A R 1985 Decreased antibody formation in iron-deficient rat pups--effect of iron repletion. American Journal of Clinical Nutrition 41: 278-284