1. Introduction
Alpha-1 antitrypsin (AAT) is a major serum protease inhibitor produced predominantly by hepatocytes [
1,
2]. After its synthesis, AAT is translocated into the endoplasmic reticulum (ER) where it is folded to be secreted into the bloodstream [
2]. The most important, disease-causing AAT variant, termed ‘Pi*Z’, interferes with AAT secretion, thereby leading to AAT accumulation in hepatocytes [
2,
3]. The resulting condition is termed AAT deficiency (AATD) and represents one of the most common genetic disorders potentially leading to death [
1,
3]. The pathogenic consequences of AAT retention are reproduced in transgenic animals overexpressing the Pi*Z variant (Pi*Z mice; [
3]). Similarly to humans, these mice display periodic acid-Schiff-diastase (
PAS-
D)-positive AAT globules as well as chronic liver injury that progresses to liver fibrosis [
3,
4].
In humans, severe AATD is caused mainly by the homozygous Pi*Z mutation (termed ‘Pi*ZZ’ genotype), that is seen in ~1:2000 Caucasians. Pi*ZZ-associated liver disease displays a biphasic pattern and results in clinically relevant liver injury in up to 10% of children and significant liver fibrosis in 20–36% of Pi*ZZ adults [
5,
6,
7,
8,
9]. Pi*ZZ adults are also strongly susceptible to early-onset emphysema and lung affection constitutes the leading cause of Pi*ZZ-related mortality [
2]. Heterozygous Pi*Z carriage (‘Pi*MZ’ genotype) is detected in ~1:40 Caucasians and strongly predisposes to liver cirrhosis in presence of a second hit such as alcoholic and non-alcoholic liver disease [
9,
10,
11].
Although AATD is a monogenic disease, its course is highly heterogeneous and the factors leading to progression of liver disease are largely unknown [
8,
12]. Several lines of evidence suggested that iron overload may promote liver injury in AATD. In particular, a subset of AATD individuals displayed a moderate to severe hepatic iron overload [
13,
14]. As a possible explanation, a direct interaction between AAT and hepcidin, the master regulator of iron metabolism, has been described [
15]. Additionally, two studies suggested that presence of heterozygous Pi*Z mutation promotes the development of liver fibrosis in individuals with the most common form of hereditary hemochromatosis, that is caused by the homozygous
C282Y mutation of the Homeostatic Iron Regulator gene (
HFE) [
16,
17,
18]. Besides affecting iron metabolism,
HFE mutations, including
C282Y and the somewhat less pathogenic
H63D variant, were suggested to lead to ER stress and thereby to increase the proteotoxic injury caused by Pi*Z [
19,
20].
Similarly, altered iron metabolism was also described in multiple pulmonary diseases including chronic obstructive pulmonary disease (COPD). In the latter one, levels of iron and iron-binding proteins in the lung are increased with normal to reduced systemic iron availability [
21,
22,
23,
24]. Moreover, elevated levels of systemic iron are toxic to the lungs and correlate with disease severity and worsening lung function [
25,
26]. Notably, a genetic variant in iron responsive element binding protein 2 (IREB2), a protein regulating iron levels in the cells, was associated with COPD phenotype in Pi*ZZ individuals [
27].
Despite these multiple links, iron metabolism in individuals with severe AATD, i.e., the Pi*ZZ genotype, was never systematically examined. To address this, we analyzed a large international cohort of Pi*ZZ adults for parameters of iron metabolism as well as the presence of HFE mutations and directly studied the interaction between mild iron overload and AATD by crossbreeding Pi*Z mice with HFE knockouts.
4. Discussion
In our study, we systematically assessed iron parameters in a large cohort of Pi*ZZ individuals as well as Pi*Z-overexpressing mice crossbred with
HFE-KOs. While the observed changes in iron parameters were small, we detected a consistent increase in serum iron levels in Pi*ZZ individuals vs. non-carriers as well as in DTg mice vs. PiZ animals. As hepcidin is the primary negative regulator of serum iron, diminished hepcidin production might be responsible for this observation. In fact, AAT induces hepcidin expression and the decreased serum AAT levels seen in Pi*ZZ individuals may therefore result in lower hepcidin values [
14]. In addition, decreased hepcidin production was seen in multiple liver disorders [
39] and this stress-induced suppression may translate to Pi*ZZ livers. This scenario was further supported by the fact that LSM ≥ 7.1kPa, suggestive of significant liver fibrosis, was associated with a further increase in serum iron levels in Pi*ZZ subjects. A similar, stress-driven mechanism may also apply to serum transferrin, that was decreased in sera from Pi*ZZ adults vs. non-carriers. Notably, transferrin is a negative acute phase reactant that is suppressed in various stress conditions [
3]. Moreover, transferrin transcription is driven via
HNF4 signaling and this pathway is down-regulated in Pi*Z-overexpressing mice [
32,
41,
42].
Pi*ZZ adults also had somewhat higher serum ferritin values than non-carriers. This finding might be both because of a higher iron load and/or increased liver injury since ferritin is produced predominantly in the liver and becomes released during liver damage [
43]. The latter mechanism might be particularly relevant in Pi*ZZ individuals with LSM ≥ 7.1kPa, that display the highest ferritin values. While clinically relevant ferritin increase (>1000 ng/mL), that is associated with an advanced risk of liver fibrosis in hereditary hemochromatosis [
18], was uncommon (2% of Pi*ZZ subjects), it was associated with elevated LSM values. Therefore, a combination of both variants might be detrimental in a small subset of affected people and these individuals need a close clinical follow-up [
14]. Notably, our findings cannot delineate whether increased iron is a cause or consequence of liver fibrosis in these Pi*ZZ individuals.
While the presence of significant liver fibrosis was associated with clear alterations in iron parameters, the occurrence of advanced lung disease did not significantly affect the parameters of iron metabolism. This finding is somewhat surprising since hypoxia is a well-established suppressor of hepcidin production [
44] and since genetic variants in
IREB2 associate with the COPD phenotype in Pi*ZZ individuals. In that respect, two important limitations of our study must be stressed: (i) our cohort provides only a limited lung phenotyping and (ii) our work did not assess genes regulating iron levels and distribution within the cells. Therefore, further studies are needed to uncover more subtle phenotypic consequences in the lung as well as to fully explore the wider iron pathways.
Given the above described findings, we decided to use
HFE-KO mice as a model of mild iron overload. Notably, these animals are well-suited for this goal, since they, similarly to Pi*ZZ individuals, display a mild increase in serum iron and ferritin values as well as an increased transferrin saturation [
37]. In the resulting DTg mice, we did not observe any signs of increased injury/fibrosis nor any obvious alterations in AAT metabolism. These data suggest that the minor alterations in iron metabolism occurring in the majority of Pi*ZZ individuals do not contribute to liver disease progression. However, we cannot exclude potentially detrimental consequences of severe iron overload that arises in a small subset of AATD individuals. This may also apply to individuals with a homozygous
HFE C282Y or compound heterozygous
C282Y/H63D genotype, that were not or only rarely found in our study, respectively.
Apart from iron overload, the presence of an
HFE mutation leads to ER stress [
19,
45] and this type of injury is particularly relevant in Pi*ZZ individuals, that display an accumulation of polymerized AAT in their ERs [
3]. To test the interaction between both hits, we genotyped our Pi*ZZ cohort for the
H63D and
C282Y variants of the
HFE gene, but the detected haplotypes did not show obvious differences in their LSM values. However, the haplotypes with higher risk of liver fibrosis development did either not occur (
C282Y homozygosity) or were rare (four individuals with
C282Y/
H63D compound heterozygosity) and further studies are needed to clarify their impact. Nevertheless, the presence of heterozygous
H63D and
C282Y variants, either alone or in combination does not seem to relevantly exacerbate the Pi*ZZ-associated liver injury. However, since we analyzed an adult cohort, we cannot exclude an effect of such variants on pediatric liver disease, that has been suggested in a previously published study [
20]. Moreover, we cannot evaluate the impact of
C282Y homozygosity, that occurs in <1% of the Caucasian population. Collectively, these data indicate that a mild iron overload can be tolerated in Pi*ZZ individuals, while markedly increased values of iron metabolism should trigger a corresponding clinical workup.