Background
Iron deficiency is the most widespread nutrient deficiency, causing considerable developmental impairment in children [
1‐
4], and it is a major cause of anaemia in tropical and low-income countries [
5,
6]. However, iron deficiency has also been proposed to protect against malaria [
7‐
10], and there is some evidence that iron supplementation can increase malaria susceptibility [
11,
12]. This has made the concept of dietary iron supplementation in areas of stable transmission of malaria parasites somewhat controversial. The issue is complicated by the inherent difficulty in accurately assessing iron status in malaria patients, because all known markers of iron stores are increased by inflammation or are otherwise affected by malaria [
13]. This means that severely ill malaria patients might inadvertently be misclassified as iron replete. Concurrent helminth infections, haemoglobinopathies, glucose-6-phosphate-dehydrogenase, and molecular haptoglobin variants can all affect measurements, adding further complexity to epidemiological studies in malaria-endemic areas [
14].
To some extent, these difficulties can be overcome in studies of experimental animal models of malaria, as these allow full control over the cause, degree, and alleviation of iron deficiency, as well as control over the timing and clinical consequences of malaria infection in iron-deficient and iron-replete animals. Nevertheless, such studies have also shown conflicting results [
15‐
19].
In the present study, the effect of nutritional iron deficiency and iron supplementation on survival and infection parameters in a murine model of malaria is explored. The focus is on ferric carboxymaltose, which is an effective parenteral iron supplementation that can be administered easily and safely [
20‐
24]. This supplement offers a cost-effective and fast correction of iron deficiency that overcomes the defective absorption of oral iron among parasitaemic recipients, and is without the adverse effects of earlier intravenous preparations.
Methods
Experimental animal model
Pathogen-free male A/J mice were purchased from Harland Laboratories (Venray, Netherlands) or Jackson Laboratories (Sacramento, CA, USA). The animals were kept under standard conditions in a closed, ventilated rack system with food/water access ad libitum. Control animals were fed a standard, iron-replete diet (100 mg Fe/kg, Altromin 1319), whereas test animals were fed an iron-deficient diet (5 mg Fe/kg; Altromin C1038, Lage, Germany) to induce iron deficiency anaemia. Haemoglobin levels were monitored weekly until levels were 20% below those in control animals. This was usually achieved in 3–4 weeks.
Some animals were treated daily for 3 consecutive days with i.v. injection (200 μL) of either ferric carboxymaltose (equivalent of 600 μg Fe; Ferinject, Vifor Pharma, Glattbrugg, Switzerland) [
25] or ferrous sulfate p.o. (600 μg Fe in 200 μL water). In those experiments, control animals received the same volume of either carboxymaltose i.v. (900 μg) or saline p.o.
For malaria infection,
P. chabaudi AS-infected erythrocytes (IEs), obtained from frozen stock (originally a gift from David Walliker, University of Edinburgh, UK), were used. After thawing, the parasites were passaged in donor mice before infecting study animals i.p. [
26].
A body temperature below 30 °C was used as a humane endpoint for death, and such severely hypothermic animals were killed by cervical dislocation as required by the Inspectorate [
27].
Outcome measurements
Weight, body temperature, haemoglobin level, reticulocytaemia, and asexual parasitaemia were measured daily until malaria symptoms developed. Subsequently, the body temperature and clinical condition of the animals were checked thrice daily.
Body temperatures were measured with an infrared thermometer (845, Testo, Lenzkirch, Germany), with temperatures below 31 °C further assessed by a rectal probe (DM852, Ellab, Hillerød, Denmark), as described [
27]. Haemoglobin levels were measured in blood from tail nicks by alkaline haematin D–575 spectrophotometry, as described [
28]. Reticulocytaemias and asexual parasitaemias were determined by flow cytometry and assessment of DNA/RNA content of acridine orange-labelled blood as described [
21,
29]. FGF23 was measured using a C-terminal, homologous two-site enzyme-linked immunosorbent assay (ELISA), catalogue number #60-6300 (Immutopics Inc, CA, USA). Hepcidin (hepc-1) was detected with a solid-phase ELISA, based on the principles of competitive binding, catalogue number SKU# HMC-001, Intrinsic Lifesciences, CA, USA.
Histochemistry
In some experiments, mice were censored to allow harvesting of organs at different time points during infection. These animals were anesthetized using fentanyl/fluanisone (Hypnorm) and Midazolam i.p., immediately perfused with saline and subsequently perfusion fixated using formaldehyde 4%, pH 7.4, and killed in the process. Livers and spleens were initially incubated in formaldehyde (24 h) and then stored in 70% ethanol. All organs were embedded in paraffin, sectioned (2 µm), and stained using Haematoxylin–Eosin and Perls’ Prussian Blue iron staining. Sections were evaluated by light microscopy. The examiner was blinded to treatment group affiliations.
Statistics
Statistical analyses were done in SAS v. 9.4 (SAS Institute, NC, USA). The log-rank test was used to assess the statistical significance of survival effects, and continuous outcome measures were analysed using a mixed-effects model (Autoregressive, order 1) with individual mice and treatment groups included as repeated and fixed effects, respectively. Hepcidin and FGF23 levels were analysed using log-transformed values in Satterthwaite’s approximate t test and ANOVA. P values < 0.05 were considered statistically significant.
Discussion
Iron is an essential nutrient, but its redox properties makes it potentially cytotoxic. Levels of non-transferrin-bound iron ions are therefore tightly regulated by hormones such as hepcidin. In the face of infection, a balance must furthermore be struck between the host need for iron and the potential benefits of minimizing the availability of this nutrient to invading microorganisms [
37]. This has created substantial controversy regarding the role of iron status in host health versus susceptibility to important infectious diseases such as malaria, and the relative benefits and detriments of correcting iron deficiency by dietary supplementation in malaria-endemic areas [
38].
In the present study, nutritional iron deficiency was found to seriously aggravate
P. chabaudi AS infection in A/J mice. Thus, an inoculum size that normally produced a self-limiting infection in iron-replete animals caused substantial mortality among iron-deficient mice. The obligation to kill animals with a body temperature below 30 °C (to minimize suffering) may have led to some overestimation of mortality, but these animals were all clinically assessed moribund, why the implementation of this humane endpoint did not markedly affect the outcome of the experiments or the conclusions that can be drawn. As such, the presented data are at variance with experimental animal studies [
15,
18] and human field studies that have indicated that iron deficiency can protect against malaria [
7,
9,
10,
39].
Both altered iron bioavailability and erythrocyte physiology have been implicated as mechanisms underlying the putative protective effect of iron deficiency [
19,
40]. Clark et al. speculated that iron deficiency might induce a reduction in the erythropoietic rate and the synthesis of microcytic, iron-deficient erythrocytes unsuitable as host cells for the parasites [
41]. In contrast, Matsuzaki-Moriya et al. found that iron deficiency anaemia did not in itself have an effect on parasite growth, and instead suggested that parasitized erythrocytes from malaria patients with iron deficiency anaemia are more susceptible to phagocytosis than control cells [
18]. However, there have also been studies challenging the protective effect of iron deficiency. Thus, Lelliott et al. found that mice carrying a missense mutation in the transferrin receptor 1 gene, which renders them anemic, showed higher
P. chabaudi parasitaemias and mortality than wildtype mice [
19], similar to the findings in the present study. Furthermore, iron is required for normal immune function, and iron deficiency anaemia is a risk factor in acute lower respiratory tract infection and in acute otitis media [
42,
43].
With respect to the malaria consequences of correcting iron deficiency, this study documents a markedly beneficial effect of short-term iron supplementation on the survival of iron-deficient,
P. chabaudi AS-infected A/J mice. The iron supplementation did not lead to increased parasitaemia, in agreement with our previous safety study [
21]. On the other hand, it also did not reduce parasitaemia, thus suggesting that the improved survival rates were not due to a direct effect on parasites but rather due to an improved host response. In the initial survival experiment, using a lethal inoculum (Fig.
2), the effect on survival might be ascribed to reversal of hypoglycaemia [
30,
31], since additional experiments showed a similar survival effect after administration of dextrose (unpublished data). However, in iron-deficient mice infected with low inoculum, the survival mediated by ferric carboxymaltose was not due to a glycemic effect, since carboxymaltose did not rescue the mice. Further studies are needed to dissect the host effects causing vulnerability to
P. chabaudi in iron deficient mice, and how these are affected by ferric carboxymaltose.
This study’s results also indicate that the temporal relationship between iron deficiency and parasitaemia is important. Thus, restoration of iron stores immediately prior to infection did not improve survival in our hands, but rather tended to hasten fatal outcomes. These findings are in line with those of Clark et al., which led them to suggest a period of vulnerability during the transition from iron deficiency to iron replete status [
41]. Thus, a ‘window of opportunity’ may exist, where adjunct intravenous iron supplementation would be beneficial. In that context, it may be worth noting that Chang et al. found that early reticulocytosis increased morbidity in a murine malaria model, whereas late reticulocytosis had the opposite effect [
44]. Furthermore, the study’s findings suggest that iron supplementation should be administered over as short a period as possible in order to reduce the period of vulnerability, and preferably while taking precautions against malaria and other infections. In this connection, it should be noticed that the protective effect of the long-acting partner drug (e.g. lumefantrine or piperaquine) may shield against a temporary risk of malaria if i.v. iron is administered concomitantly with artemisinin-based combination therapy.
The fact that this study was based on pilot experiments, and thus had clearly predefined hypotheses, supports the conclusions regarding an effect of iron deficiency as well as of ferric carboxymaltose therapy on severity of P. chabaudi infections. However, the data should be interpreted with caution.
Previous clinical studies of iron supplementation have shown conflicting results. While some findings suggest that supplementation may be unwise [
12,
45], others reported no negative effect on malaria susceptibility of iron supplementation [
46‐
48]. Recent meta-analyses indicate that iron supplementation does not increase the risk of malaria or malaria-related death, when regular surveillance and treatment services are provided [
49,
50].
This study’s results do not support the proposition by Braithwaite et al. that FGF23 can be used as a marker of iron status in patients exposed to malaria [
35], which would be an improvement over hepcidin, which is both affected by and itself affecting inflammation [
32,
34]. In particular, mice with nutritional iron deficiency could have either high or low FGF-23, indicating that this is not a useful marker of iron deficiency during malaria. Furthermore, all dying mice had severe anaemia, low hepcidin levels and high FGF23 levels. The data do not allow a conclusion on whether the high FGF23 levels in these mice were primarily caused by anaemia or inflammation [
36]. The combination of high FGF23 and low hepcidin levels could possibly serve as a marker of severity of malaria and a predictor for outcome. Further research on how iron-regulating hormones affect survival in complicated malaria is warranted.
Authors’ contributions
LM, JALK and FCC designed the study, and TS, CH, LH revised the study design and offered valuable comments. FCC, LM, TS and CH carried out the experiments. ECL analyzed and interpreted the histopathology. LH, FCC and JAKL were major contributors in drafting the manuscript. All authors read and approved the final manuscript.