Iron Mining for Erythropoiesis.

Iron is necessary for essential processes in every cell of the body, but the erythropoietic compartment is a privileged iron consumer. In fact, as a necessary component of hemoglobin and myoglobin, iron assures oxygen distribution; therefore, a considerable amount of iron is required daily for hemoglobin synthesis and erythroid cell proliferation. Therefore, a tight link exists between iron metabolism and erythropoiesis. The liver-derived hormone hepcidin, which controls iron homeostasis via its interaction with the iron exporter ferroportin, coordinates erythropoietic activity and iron homeostasis. When erythropoiesis is enhanced, iron availability to the erythron is mainly ensured by inhibiting hepcidin expression, thereby increasing ferroportin-mediated iron export from both duodenal absorptive cells and reticuloendothelial cells that process old and/or damaged red blood cells. Erythroferrone, a factor produced and secreted by erythroid precursors in response to erythropoietin, has been identified and characterized as a suppressor of hepcidin synthesis to allow iron mobilization and facilitate erythropoiesis.

Induction of erythroferrone in healthy humans by micro-dose recombinant erythropoietin or high-altitude exposure.

The erythropoietin (Epo)-erythroferrone (ERFE)-hepcidin axis coordinates erythropoiesis and iron homeostasis. While mouse studies have established that Epo-induced ERFE production represses hepcidin synthesis by inhibiting hepatic BMP/SMAD signaling, evidence for the role of ERFE in humans is limited. To investigate the role of ERFE as a physiological erythroid regulator in humans, we conducted two studies: first, 24 males received six injections of saline (placebo), recombinant Epo (rhEpo) 20 UI kg-1 (micro-dose) or 50 UI kg-1 (low-dose). Second, we quantified ERFE in 22 subjects exposed to high altitude (3800 m) for 15 hours. In the first study, total hemoglobin mass (Hbmass) increased after low- but not after micro-dose injections, when compared to placebo. Serum ERFE levels were enhanced by rhEpo, remaining higher than after placebo for 48 (micro-dose) or 72 hours (low-dose) post-injections. Conversely, hepcidin levels decreased when Epo and ERFE arose, before any changes in serum iron parameters occurred. In the second study, serum Epo and ERFE increased at high altitude. The present results demonstrate that in healthy humans ERFE responds to slightly increased Epo levels not associated with Hbmass expansion and down-regulates hepcidin in an apparently iron-independent way. Notably, ERFE flags micro-dose Epo, thus holding promise as novel anti-doping biomarker.

Iron Availability in Tissue Microenvironment: The Key Role of Ferroportin.

Body iron levels are regulated by hepcidin, a liver-derived peptide that exerts its function by controlling the presence of ferroportin (FPN), the sole cellular iron exporter, on the cell surface. Hepcidin binding leads to FPN internalization and degradation, thereby inhibiting iron release, in particular from iron-absorbing duodenal cells and macrophages involved in iron recycling. Disruption in this regulatory mechanism results in a variety of disorders associated with iron-deficiency or overload. In recent years, increasing evidence has emerged to indicate that, in addition to its role in systemic iron metabolism, FPN may play an important function in local iron control, such that its dysregulation may lead to tissue damage despite unaltered systemic iron homeostasis. In this review, we focus on recent discoveries to discuss the role of FPN-mediated iron export in the microenvironment under both physiological and pathological conditions.

Macrophage ferroportin is essential for stromal cell proliferation in wound healing.

Iron recycling by macrophages is essential for erythropoiesis, but may also be relevant for iron redistribution to neighboring cells at the local tissue level. Using mice with iron retention in macrophages due to targeted inactivation of the iron exporter ferroportin, we investigated the role of macrophage iron release in hair follicle cycling and wound healing, a complex process leading to major clinical problems, if impaired. Genetic deletion of ferroportin in macrophages resulted in iron deficiency and decreased proliferation in epithelial cells, which consequently impaired hair follicle growth and caused transient alopecia. Hair loss was not related to systemic iron deficiency or anemia, thus indicating the necessity of local iron release from macrophages. Inactivation of macrophage ferroportin also led to delayed skin wound healing with defective granulation tissue formation and diminished fibroplasia. Iron retention in macrophages had no impact on the inflammatory processes accompanying wound healing, but affected stromal cell proliferation, blood and lymphatic vessel formation, and fibrogenesis. Our findings reveal that iron/ferroportin plays a largely underestimated role in macrophage trophic function in skin homeostasis and repair.

Molecular regulation of cellular iron balance.

Handling a life-supporting yet redox-active metal like iron represents a significant challenge to cells and organisms that must not only tightly balance intra- and extracellular iron concentrations but also chaperone it during its journey from its point of entry to final destinations, to prevent inappropriate generation of damaging reactive oxygen species. Accordingly, regulatory mechanisms have been developed to maintain appropriate cellular and body iron levels. In intracellular compartments, about 95% of iron is protein-bound and the expression of the major proteins of iron metabolism is controlled by an integrated and dynamic system involving multilayered levels of regulation. However, dysregulation of iron homeostasis, which could result from both iron-related and unrelated effectors, may occur and have important pathological consequences in a number of human disorders. In this review, we describe the current understanding of the mechanisms that keep cellular iron balance and outline recent advances that increased our knowledge of the molecular physiology of iron metabolism. © 2017 IUBMB Life, 69(6):389-398, 2017.

Erythropoietin's inhibiting impact on hepcidin expression occurs indirectly.

Under conditions of accelerated erythropoiesis, elevated erythropoietin (Epo) levels are associated with inhibition of hepcidin synthesis, a response that ultimately increases iron availability to meet the enhanced iron needs of erythropoietic cells. In the search for erythroid regulators of hepcidin, many candidates have been proposed, including Epo itself. We aimed to test whether direct interaction between Epo and the liver is required to regulate hepcidin. We found that prolonged administration of high doses of Epo in mice leads to great inhibition of liver hepcidin mRNA levels, and concomitant induction of the hepcidin inhibitor erythroferrone (ERFE). Epo treatment also resulted in liver iron mobilization, mediated by increased ferroportin activity and accompanied by reduced ferritin levels and increased TfR1 expression. The same inhibitory effect was observed in mice that do not express the homodimeric Epo receptor (EpoR) in liver cells because EpoR expression is restricted to erythroid cells. Similarly, liver signaling pathways involved in hepcidin regulation were not influenced by the presence or absence of hepatic EpoR. Moreover, Epo analogs, possibly interacting with the postulated heterodimeric ß common EpoR, did not affect hepcidin expression. These findings were supported by the lack of inhibition on hepcidin found in hepatoma cells exposed to various concentrations of Epo for different periods of times. Our results demonstrate that hepcidin suppression does not require the direct binding of Epo to its liver receptors and rather suggest that the role of Epo is to stimulate the synthesis of the erythroid regulator ERFE in erythroblasts, which ultimately downregulates hepcidin.

Dietary iron overload induces visceral adipose tissue insulin resistance.

Increased iron stores associated with elevated levels of the iron hormone hepcidin are a frequent feature of the metabolic syndrome. The aim of this study was to assess the effect of dietary iron supplementation on insulin resistance and the role of hepcidin in C57Bl/6 male mice fed a standard or iron-enriched diet for 16 weeks. Iron supplementation increased hepatic iron and serum hepcidin fivefold and led to a 40% increase in fasting glucose due to insulin resistance, as confirmed by the insulin tolerance test, and to threefold higher levels of triglycerides. Iron supplemented mice had lower visceral adipose tissue mass estimated by epididymal fat pad, associated with iron accumulation in adipocytes. Decreased insulin signaling, evaluated by the phospho-Akt/Akt ratio, was detected in the visceral adipose tissue of iron overloaded mice, and gene expression analysis of visceral adipose tissue showed that an iron-enriched diet up-regulated iron-responsive genes and adipokines, favoring insulin resistance, whereas lipoprotein lipase was down-regulated. This resulted in hyperresistinemia and increased visceral adipose tissue expression of suppressor of cytokine signaling-3 (Socs3), a target of resistin and hepcidin implicated in insulin resistance. Acute hepcidin administration down-regulated lipoprotein lipase and up-regulated Socs3 in visceral adipose tissue. In conclusion, we characterized a model of dysmetabolic iron overload syndrome in which an iron-enriched diet induces insulin resistance and hypertriglyceridemia and affects visceral adipose tissue metabolism by a mechanism involving hepcidin up-regulation.

Liver iron modulates hepcidin expression during chronically elevated erythropoiesis in mice.

UNLABELLED: The liver-derived peptide hepcidin controls the balance between iron demand and iron supply. By inhibiting the iron export activity of ferroportin, hepcidin modulates iron absorption and delivery from the body's stores. The regulation of hepcidin, however, is not completely understood and includes a variety of different signals. We studied iron metabolism and hepcidin expression in mice constitutively overexpressing erythropoietin (Epo) (Tg6 mice), which leads to excessive erythropoiesis. We observed a very strong down-regulation of hepcidin in Tg6 mice that was accompanied by a strong increase in duodenal expression of ferroportin and divalent metal tranporter-1, as well as enhanced duodenal iron absorption. Despite these compensatory mechanisms, Tg6 mice displayed marked circulating iron deficiency and low levels of iron in liver, spleen, and muscle. To elucidate the primary signal affecting hepcidin expression during chronically elevated erythropoiesis, we increased iron availability by either providing iron (thus further increasing the hematocrit) or reducing erythropoiesis-dependent iron consumption by means of splenectomy. Both treatments increased liver iron and up-regulated hepcidin expression and the BMP6/SMAD pathway despite continuously high plasma Epo levels and sustained erythropoiesis. This suggests that hepcidin expression is not controlled by erythropoietic signals directly in this setting. Rather, these results indicate that iron consumption for erythropoiesis modulates liver iron content, and ultimately BMP6 and hepcidin. Analysis of the BMP6/SMAD pathway targets showed that inhibitor of DNA binding 1 (ID1) and SMAD7, but not transmembrane serine protease 6 (TMPRSS6), were up-regulated by increased iron availability and thus may be involved in setting the upper limit of hepcidin. CONCLUSION: We provide evidence that under conditions of excessive and effective erythropoiesis, liver iron regulates hepcidin expression through the BMP6/SMAD pathway.

Iron trafficking and metabolism in macrophages: contribution to the polarized phenotype.

During inflammation, proinflammatory macrophages sequester iron as a well known bacteriostatic mechanism. Alternative activation of macrophages is linked to tissue repair, and during this process the expression pattern of genes important for iron homeostasis is distinct from that in proinflammatory macrophages. This leads to an increased capacity of the alternatively activated macrophages for heme uptake, via scavenger receptors, and for production of anti-inflammatory mediators via heme-oxygenase-dependent heme catabolism. Alternatively activated macrophages also release non-heme iron into tissues via ferroportin. Here, we propose that the iron-release-associated phenotype of alternatively activated macrophages significantly contributes to their role in various conditions, including tissue repair and tumor growth.

Iron Absorption: Molecular and Pathophysiological Aspects.

Iron is an essential nutrient for growth among all branches of life, but while iron is among the most common elements, bioavailable iron is a relatively scarce nutrient. Since iron is fundamental for several biological processes, iron deficiency can be deleterious. On the other hand, excess iron may lead to cell and tissue damage. Consequently, iron balance is strictly regulated. As iron excretion is not physiologically controlled, systemic iron homeostasis is maintained at the level of absorption, which is mainly influenced by the amount of iron stores and the level of erythropoietic activity, the major iron consumer. Here, we outline recent advances that increased our understanding of the molecular aspects of iron absorption. Moreover, we examine the impact of these recent insights on dietary strategies for maintaining iron balance.

Serum hepcidin levels and muscle iron proteins in humans injected with low- or high-dose erythropoietin.

Inhibition of hepcidin expression by erythropoietic signals is of great physiological importance; however, the inhibitory pathways remain poorly understood. To investigate (i) the direct effect of erythropoietin (Epo) and (ii) the contribution of putative mediators on hepcidin repression, healthy volunteers were injected with a single dose of Epo, either low (63 IU/kg, n = 8) or high (400 IU/kg, n = 6). Low-dose Epo provoked hepcidin down-modulation within 24 h; the effect was not immediate as hepcidin circadian variations were still present following injection. High-dose Epo induced no additional effect on the hepcidin response, that is hepcidin diurnal fluctuations were not abolished in spite of extremely high Epo levels. We did not find significant changes in putative mediators of hepcidin repression, such as transferrin saturation, soluble transferrin receptor, or growth differentiation factor 15. Furthermore, the potential hepcidin inhibitor, soluble hemojuvelin, was found unaltered by Epo stimulation. This finding was consistent with the absence of signs of iron deficiency observed at the level of skeletal muscle tissue. Our data suggest that hepcidin repression by erythropoietic signals in humans may not be controlled directly by Epo, but mediated by a still undefined factor.

Differential regulation of iron homeostasis during human macrophage polarized activation.

Iron metabolism in inflammation has been mostly characterized in macrophages exposed to pathogens or inflammatory conditions, mimicked by the combined action of LPS and IFN-gamma (M1 polarization). However, macrophages can undergo an alternative type of activation stimulated by Th2 cytokines, and acquire a role in cell growth and tissue repair control (M2 polarization). We characterized the expression of genes related to iron homeostasis in fully differentiated unpolarized (M0), M1 and M2 human macrophages. The molecular signature of the M1 macrophages showed changes in gene expression (ferroportin repression and H ferritin induction) that favour iron sequestration in the reticuloendothelial system, a hallmark of inflammatory disorders, whereas the M2 macrophages had an expression profile (ferroportin upregulation and the downregulation of H ferritin and heme oxygenase) that enhanced iron release. The conditioned media from M2 macrophages promoted cell proliferation more efficiently than those of M1 cells and the effect was blunted by iron chelation. The role of ferroportin-mediated iron release was demonstrated by the absence of differences from the media of macrophages of a patient with loss of function ferroportin mutation. The distinct regulation of iron homeostasis in M2 macrophages provides insights into their role under pathophysiological conditions.

Iron-dependent regulation of MDM2 influences p53 activity and hepatic carcinogenesis.

Iron overload is a risk factor for hepatocarcinoma, but the pathways involved are poorly characterized. Gene expression analysis in immortalized mouse hepatocytes exposed to iron or the iron chelator deferoxamine revealed that iron downregulated, whereas deferoxamine upregulated, mRNA levels of mouse double minute gene 2 (MDM2), the ubiquitin ligase involved in the degradation of the oncosuppressor p53. Regulation of MDM2 by iron status was observed at protein levels in mouse hepatocytes and rat liver, and was associated with specular changes in p53 expression. Iron dependent regulation of MDM2/p53 was confirmed ex-vivo in human monocytes, by manipulation of iron pool and in a genetic model of iron deficiency, leading to modulation of p53 target genes involved in the antioxidant response and apoptosis. Iron status influenced p53 ubiquitination and degradation rate, and the MDM2 inhibitor nutlin increased p53 levels in iron-depleted cells. Furthermore, nutlin enhanced the antiproliferative activity of deferoxamine in HepG2 hepatoblastoma cells. The MDM2 -309T > G promoter polymorphism, determining increased MDM2 and lower p53 activity, was associated with higher risk of hepatocarcinoma in cirrhotic patients with hemochromatosis, and with HFE mutations in patients with hepatocarcinoma without hemochromatosis, suggesting an interaction between MDM2 and iron in the pathogenesis of hepatocarcinoma. In conclusion, iron status influences p53 activity and antioxidant response by modulating MDM2 expression. MDM2 inhibitors may enhance the antiproliferative activity of iron chelators.

Alterations of systemic and muscle iron metabolism in human subjects treated with low-dose recombinant erythropoietin.

The high iron demand associated with enhanced erythropoiesis during high-altitude hypoxia leads to skeletal muscle iron mobilization and decrease in myoglobin protein levels. To investigate the effect of enhanced erythropoiesis on systemic and muscle iron metabolism under nonhypoxic conditions, 8 healthy volunteers were treated with recombinant erythropoietin (rhEpo) for 1 month. As expected, the treatment efficiently increased erythropoiesis and stimulated bone marrow iron use. It was also associated with a prompt and considerable decrease in urinary hepcidin and a slight transient increase in GDF-15. The increased iron use and reduced hepcidin levels suggested increased iron mobilization, but the treatment was associated with increased muscle iron and L ferritin levels. The muscle expression of transferrin receptor and ferroportin was up-regulated by rhEpo administration, whereas no appreciable change in myoglobin levels was observed, which suggests unaltered muscle oxygen homeostasis. In conclusion, under rhEpo stimulation, the changes in the expression of muscle iron proteins indicate the occurrence of skeletal muscle iron accumulation despite the remarkable hepcidin suppression that may be mediated by several factors, such as rhEpo or decreased transferrin saturation or both.

Setup for human sera MALDI profiling: the case of rhEPO treatment.

The implementation of high-throughput technologies based on qualitative and quantitative methodologies for the characterization of complex protein mixtures is increasingly required in clinical laboratories. MALDI profiling is a robust and sensitive technology although the serum high dynamic range imposes a major limitation hampering the identification of less abundant species decreasing the quality of MALDI profiling. A setup to improve these parameters has been performed for recombinant human erythropoietin (rhEPO) monitoring in serum, analyzing the effects of two commercially available columns (MARS Hu7 and Hu14) for immunodepletion, and two matrices (α-cyano-4-hydroxycinnamic acid and 2',4'-dihydroxyacetophenone) for peak quality improvement. The immunodepletion capability of both columns was determined by 2-D DIGE, which precisely revealed the efficacy of Hu14 in protein removal and the serum dynamic range decrement. In addition, the type of matrix, the sample dilution, and the efficacy of optimized parameters were used for serum profiling of ten healthy subjects before and after rhEPO treatment. The principal component analysis indicates that a combination of Hu14 column and 2',4'-dihydroxyacetophenone matrix increases data quality allowing the discrimination between treated and untreated samples, making serum MALDI profiling suitable for clinical monitoring of rhEPO.

Macrophages and Iron: A Special Relationship.

Macrophages perform a variety of different biological functions and are known for their essential role in the immune response. In this context, a principal function is phagocytic clearance of pathogens, apoptotic and senescent cells. However, the major targets of homeostatic phagocytosis by macrophages are old/damaged red blood cells. As such, macrophages play a crucial role in iron trafficking, as they recycle the large quantity of iron obtained by hemoglobin degradation. They also seem particularly adapted to handle and store amounts of iron that would be toxic to other cell types. Here, we examine the specific and peculiar iron metabolism of macrophages.

Unconventional endocytosis and trafficking of transferrin receptor induced by iron.

The posttranslational regulation of transferrin receptor (TfR1) is largely unknown. We investigated whether iron availability affects TfR1 endocytic cycle and protein stability in HepG2 hepatoma cells exposed to ferric ammonium citrate (FAC). NH4Cl and bafilomycin A1, but not the proteasomal inhibitor MG132, prevented the FAC-mediated decrease in TfR1 protein levels, thus indicating lysosomal involvement. Knockdown experiments showed that TfR1 lysosomal degradation is independent of 1) endocytosis mediated by the clathrin adaptor AP2; 2) Tf, which was suggested to facilitate TfR1 internalization; 3) H-ferritin; and 4) MARCH8, previously implicated in TfR1 degradation. Notably, FAC decreased the number of TfR1 molecules at the cell surface and increased the Tf endocytic rate. Colocalization experiments confirmed that, upon FAC treatment, TfR1 was endocytosed in an AP2- and Tf-independent pathway and trafficked to the lysosome for degradation. This unconventional endocytic regulatory mechanism aimed at reducing surface TfR1 may represent an additional posttranslational control to prevent iron overload. Our results show that iron is a key regulator of the trafficking of TfR1, which has been widely used to study endocytosis, often not considering its function in iron homeostasis.

Adenosine A(2)A receptor but not HIF-1 mediates Tyrosine hydroxylase induction in hypoxic PC12 cells.

Tyrosine hydroxylase (TH) is the rate-limiting enzyme in the biosynthesis of catecholamines released by oxygen-sensitive cells in response to hypoxic conditions. Adenosine is released in response to hypoxia in the central nervous system and CGS21680, an adenosine A(2)A receptor agonist, induces TH transcription. As we have previously demonstrated the A(2)A receptor-mediated induction of HIF-1 in macrophages and hepatocytes, we investigated the involvement of HIF-1 in the adenosine-mediated activation of TH expression. Exposure to adenosine or CGS21680 increased TH mRNA and protein levels in PC12 cells. Transcription of a reporter gene under the control of the wild type rat TH promoter was induced 3.5-fold in CGS21680-treated cells, but neither the mutation of the hypoxia responsive element in the TH promoter nor the co-transfection of a dominant negative of the HIF-1 beta subunit prevented the increase in transcription; furthermore, CGS21680 increased CREB binding activity but did not induce HIF-1 DNA binding activity or protein levels. To investigate whether HIF-1 was involved in the hypoxia-mediated induction of TH, PC12 cells were exposed to hypoxia in the presence of the A(2)A receptor antagonist ZM241385, which prevented hypoxia-dependent TH induction despite HIF-1 activation; in line with this finding, the inhibition of HIF-1 did not abolish TH induction in hypoxic PC12 cells. These results indicate that, under hypoxic conditions, TH (a key factor in systemic adaptation to reduced oxygen availability) is not regulated by HIF-1, the primary modulator of the response to hypoxia, but by the adenosine A(2)A receptor-mediated signalling pathway.

Low doses of recombinant human erythropoietin does not affect C-terminal FGF23 in healthy men.

Recombinant human erythropoietin (rhEpo) can improve human performance, but misuse remains difficult to detect. C-terminal fibroblast growth factor 23 (cFGF23) was recently demonstrated to increase following injection of a single high dose rhEpo, but the effect of more frequent low doses is unknown. Using a randomized double-blind placebo-controlled design, we investigated whether 2 weeks of subcutaneous injections three times a week of 50 IU/kg Eprex (low-dose) or 20 IU/kg Eprex (micro-dose) increase cFGF23 levels compared with saline (placebo) injections in 24 healthy males. Venous blood was sampled at day -3, 0, 1, 3, 11, 14, 18, and 25 of the treatment and analyzed for cFGF23 and erythropoietin concentration ([EPO]). The level of cFGF23 was similar at days -3, 0, 1, 3, 11, 14, 18, and 25 with the low-dose (23 ± 4, 26 ± 5, 23 ± 7, 27 ± 6, 25 ± 8, 24 ± 10, 22 ± 5, and 24 ± 7 RU/mL, respectively), micro-dose (23 ± 6, 25 ± 5, 23 ± 8, 28 ± 9, 27 ± 7, 25 ± 9, 25 ± 5, and 23 ± 6 RU/mL, respectively) and placebo (23 ± 6, 24 ± 6, 26 ± 7, 26 ± 6, 31 ± 6, 31 ± 7, 24 ± 4, and 27 ± 8 RU/mL, respectively) treatment. The correlation coefficient between plasma [EPO] and plasma cFGF23 levels was R2 = 0.01 and insignificant. The results demonstrate that cFGF23 is not sensitive to low doses of subcutaneous rhEpo injections in healthy males.

Adenosine-dependent activation of hypoxia-inducible factor-1 induces late preconditioning in liver cells.

UNLABELLED: The cellular mechanisms by which ischemic preconditioning increases liver tolerance to ischemia/reperfusion injury are still poorly understood. This study investigated the role of the hypoxia-inducible factor-1 (HIF-1) in the protection associated with the late phase of liver preconditioning. Late preconditioning was induced in primary cultured rat hepatocytes by a transient (10 minute) hypoxic stress or by 15 minutes incubation with the adenosine A(2A) receptors agonist CGS21680 24 hours before exposure to 90 minutes of hypoxia in a serum-free medium. Late preconditioning induced the nuclear translocation of HIF-1 and the expression of carbonic anhydrase IX (CAIX), a HIF-1-regulated transmembrane enzyme that catalyzes bicarbonate production. Such effects were associated with prevention of hepatocyte killing by hypoxia and the amelioration of intracellular acidosis and Na+ accumulation. The inhibition of PKC-mediated and PI3-kinase-mediated signals with, respectively, chelerythrine and wortmannin abolished HIF-1 activation and blocked both CAIX expression and the protective action of late preconditioning. CAIX expression was also prevented by interfering with the transcriptional activity of HIF-1 using a dominant negative HIF-1beta subunit. The inhibition of CAIX with acetazolamide or the block of bicarbonate influx with disodium-4-acetamido-4'-isothiocyanato-stilben-2,2'-disulfonate also reverted the protective effects of late preconditioning on intracellular acidosis and Na+ accumulation. CONCLUSION: The stimulation of adenosine A(2A) receptors induced late preconditioning in liver cells through the activation of HIF-1. HIF-1-induced expression of CAIX increases hepatocyte tolerance to ischemia by maintaining intracellular Na+ homeostasis. These observations along with the importance of HIF-1 in regulating cell survival indicates HIF-1 activation as a possible key event in liver protection by late preconditioning.