Parenteral iron compounds sensitize mice to injury-initiated TNF- mRNA production and TNF- release
http://www.100md.com
《美国生理学杂志》
The University of Washington and the Fred Hutchinson Cancer Reseach Center, Seattle, Washington
ABSTRACT
Intravenous Fe is widely used to treat anemia in renal disease patients. However, concerns of potential Fe toxicity exist. To more fully define its spectrum, this study tested Fe's impact on systemic inflammation following either endotoxemia or the induction of direct tissue damage (glycerol-mediated rhabdomyolysis). The inflammatory response was gauged by tissue TNF- message expression and plasma TNF- levels. CD-1 mice received either intravenous Fe sucrose, -gluconate, or -dextran (FeS, FeG, or FeD, respectively; 2 mg), followed by either endotoxin (LPS) or glycerol injection 0–48 h later. Plasma TNF- was assessed by ELISA 2–3 h after the LPS or glycerol challenge. TNF- mRNA expression (RT-PCR) was measured in the kidney, heart, liver, lung, and spleen with Fe ± LPS treatment. Finally, the relative impacts of intramuscular vs. intravenous Fe and of glutathione (GSH) on Fe/LPS- induced TNF- generation were assessed. Each Fe preparation significantly enhanced LPS- or muscle injury-mediated TNF- generation. This effect was observed for at least 48 h post-Fe injection, a time at which plasma iron levels were increased by levels insufficient to fully saturate transferrin. Fe did not independently increase plasma TNF- or tissue mRNA. However, it potentiated postinjury-induced TNF- mRNA increments and did so in an organ-specific fashion (kidney, heart, and lung; but not in liver or spleen). Intramuscular administration, but not GSH treatment, negated Fe's ability to synergize LPS-mediated TNF- release. We conclude 1) intravenous Fe can enhance TNF- generation during LPS- or glycerol-induced tissue damage; 2) increased TNF- gene transcription in the kidney, heart, and lung may contribute to this result; and 3) intramuscular administration, but not GSH, might potentially mitigate some of Fe's systemic toxic effects.
endotoxemia; rhabdomyolysis; end-stage renal disease; iron deficiency; glutathione
PARENTERAL IRON THERAPY HAS become a mainstay in anemia management in patients with progressive or end-stage renal disease (ESRD) (15, 17, 18). Its need stems from a variety of factors, including low iron intake or absorption, dialysis-dependent blood loss, and/or "functional" deficiency, defined as an insufficient amount of bioavailable iron to support maximal erythropoietin-stimulated erythropoiesis. Fe therapy, with the resultant correction of anemia, is believed to improve quality of life and possibly decrease mortality (38). Nevertheless, concerns of potential Fe toxicity exist, based on the fact that all currently employed parenteral Fe preparations [Fe sucrose (FeS), Fe gluconate (FeG), and Fe dextran (FeD)] can exert marked prooxidant effects. This conclusion is based on results from in vivo (e.g., Refs. 19, 47, 49) and in vitro (e.g., Refs. 5, 30, 34, 47, 49) experiments and from clinical investigations (e.g., Refs. 1, 6, 14, 20, 26–28, 32). Oxidative stress is believed to play an important pathogenic role in atherogenesis (7, 8, 11, 24) and in renal disease progression (3, 10, 22, 31). These considerations raise further concerns regarding the safety of intravenous (iv) Fe treatment in patients with renal disease (47). The recent demonstration that iv FeS transiently exacerbates proteinuria and induces proximal tubular enzymuria in pre-ESRD patients supports these concerns (1).
An additional pathway by which iv Fe might adversely impact renal disease patients is via a potentiation of systemic inflammation, a correlate of ESRD (33, 42, 52). That levels of serum C-reactive protein, a surrogate marker of inflammation, correlate with dialysis-associated complications, e.g., adverse cardiovascular events (16, 23, 45), underscores the clinical relevance of this inflammatory state. Oxidative stress can promote inflammation via a number of potential pathways. For example, redox-sensitive transcription factors, most notably NF-B, cFos, c-Jun, activator protein 1, and hypoxia-inducible factor-1, are upregulated by oxidative stress (9, 13, 25, 35, 44). These can then recruit "downstream" mediators, such as TNF-, which can potentiate systemic inflammation (2, 4, 40, 41), or trigger apoptotic cell death (37, 39).
Given the above, it is reasonable to question whether parenteral iron-induced oxidative stress can potentiate a proinflammatory state. A recent brief report from this laboratory provides initial support for this concept (48). Experimental sepsis was induced in mice by injecting nonviable Escherichia coli ± iv FeS. Despite the fact that FeS alone evoked no discernible tissue injury/inflammation, it profoundly potentiated the E. coli-induced septic state. For example, FeS-treated mice manifested a doubling of E. coli-initiated TNF- generation, culminating in an 50% mortality rate (48). Clearly, then, whereas iv FeS appeared to be well tolerated, in concert with E. coli, it had profound toxic effects.
The present study was undertaken to gain further insights into the above-defined phenomenon, and by so doing better define the spectrum and mechanisms of parenteral Fe toxicity. Specifically, the following questions have been addressed. 1) Can FeS potentiate TNF- generation following purified endotoxin administration, or is this response only seen with intact bacterial injections 2) Given that our prior studies have found that FeS has by far the greatest cytotoxic potential of the currently employed Fe preparations (47, 49), is FeS unique in its ability to potentiate TNF- generation, or can other Fe compounds reproduce this result 3) What is the mechanism by which iv Fe potentiates TNF- generation in the setting of sepsis Specifically, does Fe increase TNF- generation via transcriptional or posttranscriptional events 4) What are the target organs that lead to exaggerated TNF- generation in response to an Fe-loading/septic state 5) Can iv Fe potentiate TNF- generation with non-LPS-driven tissue damage Experiments to gain insights into these, and other issues, form the basis of this report.
METHODS
General Methods
Male CD-1 mice (25–35 g; Charles River Laboratories, Wilmington, MA) were used for all experiments. All animal experiments were in accordance with, and approval of, Fred Hutchinson Cancer Research Center institutional animal care and and use committee guidelines. They were maintained under routine vivarium conditions with free food and water access throughout. Tail vein Fe injections were used and encompassed the following test agents: 1) FeS (Venofer, American Regent, Shirley, NY); 2) FeD (INFeD, Watson Pharmaceuticals, Morristown, NJ); or 3) FeG (Ferrlecit, Watson Pharmaceuticals). An equal amount of elemental Fe was used with each agent (2 mg: equaling 100 μl of FeS, 160 μl of FeG, or 40 μl of FeD of their respective stock solutions). Non-Fe-treated mice received a tail vein injection of 100 μl 0.9% saline as a sham for Fe treatment. Tail vein injections were performed by placing individual mice into cylindrical restraining holders, cannulating the vein with a 26-gauge needle, and injecting the test agents over 30 s.
Purified endotoxin (LPS) was obtained from Sigma (E. coli 0111:B4; catalogue no. L-2630; stock solution, 0.4 mg/ml saline). It was administered intraperitoneally (ip) in a dose of either 2 or 10 mg/kg. Mice not receiving LPS received an equal amount of the LPS vehicle (saline ip). As a second model of tissue injury, the glycerol model of muscle necrosis was employed. For this purpose, mice were lightly anesthetized with isoflurane and injected with 6 ml/kg of 50% glycerol, administered intramuscularly (im) into one hindlimb. The specific Fe treatments employed in the LPS or glycerol experiments are discussed in the individual protocols presented below.
Following the above interventions, the mice were allowed to recover from anesthesia and provided with free food and water access. Subsequent animal death was conducted under deep pentobarbital sodium anesthesia (2–4 mg/kg). The abdominal cavity of each mouse was opened with a midline incision, and a terminal phlebotomy was conducted from the inferior vena cava into a heparinized syringe. Plasma samples were stored at –70°C until assay for TNF- levels conducted by ELISA (R&D Systems; Minneapolis, MN; Ref. 49).
Specific Experiments
Immediate effect of different Fe compounds on LPS-induced TNF- increments. This experiment was undertaken to ascertain whether the above-noted Fe preparations impact LPS-mediated TNF- generation, and if so, ascertain whether differences exist among these Fe-containing polymers. Mice (n = 47) were subjected to one of the following tail vein injections: 1) 0.1 ml normal saline (sham for Fe treatment; n = 17); 2) FeS (n = 11), FeG (n = 8), or FeD (n = 11). Immediately after Fe or saline administration, the mice received an injection of LPS (2 mg/kg ip). Two hours post-LPS injection, the mice were killed and the terminal plasma samples were collected for TNF- assay.
Delayed effect of different Fe compounds on LPS-induced TNF- increments. The above experiment indicated that when Fe is administered concomitantly with LPS, a potentiation of plasma TNF- increments occurs (see RESULTS). This experiment was undertaken to ascertain whether Fe can also potentiate LPS-mediated TNF- generation after a 24-h delay, allowing for in vivo drug distribution/drug processing to occur. Potential differences among the test Fe compounds were also assessed. Mice (n = 20) received tail vein injections of either saline (n = 6) or Fe (FeS, n = 6; FeG, n = 4; or FeD, n = 4). Then, they were returned to their cages. Twenty-four hours later, each mouse received an LPS injection (10 mg/kg ip). Three hours post-LPS treatment, terminal plasma samples were obtained for TNF- assay.
Parenteral Fe effect on muscle injury-initiated TNF- levels. The following experiment was undertaken to ascertain whether parenteral Fe can potentiate TNF- generation when the latter is initiated by direct, focal tissue damage, as opposed to a systemic endotoxemic state. Toward this end, muscle injury was induced by im glycerol injection ± iv Fe therapy, as noted above (General Methods). Mice were divided into the following experimental groups: 1) iv FeS, followed immediately by glycerol injection (n = 10); 2) iv FeG+glycerol (n = 5); 3) iv FeD+glycerol (n = 7); and 4) sham iv Fe (0.1 ml of saline)+glycerol. Three hours later, the mice were killed and terminal plasma samples were analyzed for TNF- levels, as noted above.
Independent effect of parenteral Fe administration on TNF- levels. The following experiment was undertaken to ascertain the possible impact of parenteral Fe administration on circulating TNF- levels independent of concomitant forms of tissue stress (i.e., LPS or glycerol injection). To this end, 12 mice were divided into 4 equal groups, as follows: 1) sham Fe injection; 2) iv FeS; 3) iv FeG; or 4) iv FeD. Three hours postinjection, terminal plasma samples were collected and assayed for TNF- levels.
Comparison of effects of iv vs. im Fe on LPS-mediated TNF- generation.
iv vs. im FeG with a 24-h latency period. The following experiment was undertaken to ascertain whether use of different administration routes alters the extent to which Fe can potentiate LPS-mediated TNF- generation. Six mice received 2 mg of elemental Fe (FeG) via either the iv or the im route (n = 3 each). After allowing a 24-h Fe equilibration period, each of the mice received the 10 mg/kg ip LPS challenge. Three hours post-LPS injection, terminal plasma samples were obtained for TNF- assay. The results were compared with those obtained in six mice subjected to ip LPS injection without prior Fe administration.
iv vs. im FeG with a 48-h latency period. The above experiment was repeated, but with a 48-h latency period as follows: 1) iv FeG+0.1 ml im saline (n = 6); 2) im FeG+0.1 ml iv saline (n = 6); and 3) controls (0.1 ml iv saline+0.1 ml im saline; n = 7). (Note that each mouse received both an iv and im injection, thereby controlling for this variable among the groups.) The mice were then returned to their cages for a 48-h drug distribution/processing period. After completing this period, each of the mice was subjected to the 10 mg/kg ip LPS challenge. Three hours later, plasma samples were obtained for TNF- assay.
Plasma Fe and Fe binding capacity 48 h after iv and im FeG administration. The following experiment was performed to ascertain the effect of the above iv and im Fe treatments on plasma Fe and %saturation of total iron binding capacity (TIBC). Six mice were divided into three treatment groups of two mice each: 1) tail vein saline injection (controls); 2) tail vein FeG injection; and 3) im FeG administration. After a 48-h delay, terminal plasma samples were obtained and assayed for Fe, TIBC, and %TIBC saturation. These determinations were performed with a Beckman Synchrom LX20 autoanalyzer (University of Washington Medical Center Clinical Laboratory; Seattle, WA).
To gauge Fe absorption following im administration, three mice received a 2-mg Fe injection (as FeG) in one upper thigh. Two days later, the mice were anesthetized and subjected to a repeat im FeG injection, this time administered in the contralateral thigh. Five minutes later, both thighs were dissected and the relative amount of black Fe staining in each thigh was visually contrasted. In addition, 1-cm-long pieces of injected muscle tissue were obtained immediately after, and 48 h after, FeG injection. The muscle samples were fixed in 10% formalin and embedded in paraffin. Serial sections were prepared throughout the muscle length (15 sections/muscle), and these were stained with hematoxylin and eosin. Fe deposition and muscle injury were gauged by light microscopy.
Glutathione addition to FeS: impact on LPS-mediated TNF- generation. As an alternative approach to im Fe injection as a possible means to mitigate Fe-stimulated TNF- generation, the antioxidant glutathione (GSH) was added to FeS before its injection. To this end, 5 mg of GSH (50 μl of a 100 mg/ml stock saline solution) were added to 2 mg of FeS. FeS/GSH was then administered via tail vein injection (into 3 mice). Three control mice received the same 2-mg FeS injection, substituting 50 μl of saline for the GSH addition. Immediately thereafter, each of the six mice received a 2-mg/kg ip LPS injection. Three hours post-LPS injection, the mice were anesthetized and terminal plasma samples were obtained and assayed for TNF- levels.
Tissue TNF- mRNA level and responses to iv Fe, LPS, or Fe+LPS treatment. The following experiment had three objectives: 1) ascertain the effect of iv Fe on tissue TNF- mRNA levels; 2) test the impact of iv Fe on LPS-stimulated TNF- mRNA expression; and 3) ascertain whether different tissues have different TNF- mRNA responses to Fe, LPS, or Fe+LPS injections. Four groups of mice (n = 5 each) were established: 1) controls (iv saline); 2) iv FeS; 3) LPS (10 mg/kg ip); and 4) iv FeS+LPS. Three hours later, the mice were anesthetized, and samples of the following tissues were obtained: kidney cortex, liver, cardiac apex, lung, and spleen.
Tissue samples were immediately placed into TRIzol reagent (Invitrogen Life Technologies, Carlsbad, CA), and total RNA was extracted according to the manufacturer's instructions. The final RNA pellet was brought up in RNase-free water to a concentration of 0.5–2 μg/ml. RNA integrity was ascertained on an ethidium bromide-agarose gel (41). RT-PCR was performed using a 1st Strand Synthesis Kit for RT-PCR (Ambion, Austin, TX), as previously described in detail (51). The specific primers for TNF- and GAPDH (housekeeping gene) were designed with 50–60% GC composition (see Table 1). Multiplex PCR was possible due to the primers' similar annealing temperatures but dissimilar sizes of the two PCR products. PCR conditions were optimized for each tissue (37 cycles for TNF- in all tissues; 18, 22, 19, 25, and 25 GAPDH cycles for kidney, liver, heart, spleen, and lung, respectively). PCR products were analyzed by agarose gel electrophoresis and ethidium bromide staining. The PCR products were quantified by densitometry with a Typhoon 8600 scanner (Amersham Phamacia Biosciences, Piscataway, NJ). TNF- bands were expressed as ratios to the simultaneously obtained GAPDH bands. Results in each organ were compared between the four treatment groups. The %changes in TNF- mRNA (controls vs. therapies) between organs were also compared.
View this table:
Calculations and Statistics
All values are presented as means ± SE. Statistical comparisons of paired and unpaired data were performed by paired and unpaired Student's t-test, respectively. In addition, comparisons of TNF- mRNA levels between two treatments were assessed by a test of binomial proportions. If multiple comparisons were made, the Bonferroni correction was applied. Significance was judged by a P value of <0.05.
RESULTS
Comparative, Immediate Effect of iv Fe on LPS-Mediated TNF- Generation
Normal mouse plasma has a baseline TNF- level that is at or below the level of accurate detection by the employed ELISA (<1–2 pg/ml; Ref. 48). As shown in Fig. 1, LPS administration induced dramatic plasma TNF- increments. Mice that had been treated concomitantly with FeS manifested an approximate doubling of these LPS-induced TNF- increases. [Of note, these results were essentially identical to those previously reported by this group using Fe+E. coli (48)]. Unlike FeS, neither FeG nor FeD caused any potentiation of LPS-initiated plasma TNF- elevations.
Delayed Effect of Fe Compounds on LPS-Induced TNF- Increments
When non-Fe-treated mice were subjected to LPS injection, plasma TNF- levels rose to 760 ± 80 pg/ml. The upper limit of the 95% confidence band for these mice, 1,000 pg/ml, is presented as a horizontal line in Fig. 2. Each mouse in each of the three Fe pretreatment groups (i.e., iv Fe administration 24 h earlier) developed a post-LPS TNF- level that exceeded this 95% confidence range (i.e., in each mouse, Fe pretreatment sensitized to LPS-mediated TNF- generation). FeD and FeG appeared to exert comparable effects on post-LPS TNF- plasma levels. The FeS group had 60% greater elevations in post-LPS TNF- levels compared with either the FeD or the FeG group.
Fe Effects on TNF- Levels Following Hypertonic Glycerol-Mediated Muscle Injury
Glycerol injection alone caused only minimal TNF- generation, rising from control values of 1 to 2.5 ± 0.6 pg/ml (P < 0.01; Fig. 3). In each case, concomitant Fe administration raised postglycerol TNF- levels, and with statistically significant differences in degrees of elevations being observed (FeS > FeG > FeD). Of note, although Fe potentiated postglycerol TNF- increases (e.g., 6-fold increase with FeS, compared with glycerol alone), the absolute degrees of elevation were quite modest compared with those that followed LPS injection (note the different y-axis scales in Fig. 3 vs. Figs. 1 and 2).
TNF- Levels Following Fe Treatment Alone
Treatment of normal mice with any of the Fe compounds caused no significant increase in plasma TNF- levels compared with non-Fe-injected control mice (all values 1 pg/ml). This indicates that the TNF- differences observed in the above experiments could not be explained by an independent Fe effect on TNF- levels.
Comparison of Effects of im vs. iv Fe Injections on LPS-Mediated TNF- Levels
As shown in Fig. 4, left, mice that received iv FeG 24 h before the LPS challenge had a dramatic increase in plasma TNF- levels compared with non-Fe-treated, LPS-injected controls (P < 0.005). In striking contrast, im Fe pretreatment did not predispose mice to LPS-mediated TNF- release. When this same experiment was repeated, but with a 48-h lag time allowed between iv/im FeG treatment and LPS injection, a marked potentiation of LPS-mediated TNF- generation was still observed in the mice that had received prior iv FeG treatment (Fig. 4, right). Once again, im FeG did not potentiate LPS-mediated TNF- release (Fig. 4, right).
Plasma Fe and TIBC Assessments Following iv and im FeG Injections
Plasma Fe, TIBC, and %TIBC saturation are presented in Table 2. Plasma Fe concentrations were increased by 45% at 48 h post-iv FeG injection. This corresponded with a modest increase in both TIBC and %TIBC saturation, the latter rising from 66 to 86% with iv FeG treatment. In contrast, im FeG injection failed to raise plasma Fe levels. Rather, there tended to be a slight decrease in the Fe, TIBC, and %TIBC saturation with this treatment protocol.
View this table:
Muscle evaluation. By visual inspection, obvious black deposits were observed in muscle immediately following im injection. In contrast, muscle injection sites obtained after a 48-h lag time revealed no visual evidence for persistent Fe deposition. Representative histological sections through upper leg muscle immediately following, and 48 h following, FeG injection are presented in Fig. 5. As shown in Fig. 5, A and B, sections obtained from muscle 5 min after FeG injection demonstrated marked interstitial fluid accumulation in muscle ("lakes"), with wide separation of muscle fibers. On higher power, particulate material, consisting of the injected FeG, was apparent (Fig. 5B). In contrast, muscle obtained 48 h post-FeG injection appeared completely normal: there was no longer evidence of fluid-induced separation of muscle fibers or of particulate Fe deposition. Thus these findings supported the visual findings: that the injected FeG had been absorbed from muscle by 48 h postinjection. Of further note, FeG injection failed to induce either muscle fiber necrosis or an inflammatory response, suggesting that it had caused little, if any, muscle damage.
Impact of GSH Addition to FeS on LPS-Mediated TNF- Release
Although the GSH stock solution was colorless, when it was added to FeS, it caused a modest color change (from black to reddish brown; consistent with a change of in Fe valence). GSH addition to the FeS did not diminish its ability to facilitate LPS-mediated TNF- generation: rather, it tended to increase, not decrease, this response: (FeS/GSH+LPS: TNF-, 621 ± 39 pg/ml; FeS alone+LPS: TNF-, 506 ± 42 pg/ml; P < 0.035 by paired t-test analysis for 3 sets of paired mice).
TNF- mRNA Tissue Levels Following FeS, LPS, or FeS+LPS Injections
TNF- mRNA levels in the kidney, liver, lung, heart, and spleen are presented in Table 3. By itself, FeS injection failed to cause a significant increase in message levels in any of the tissues examined. LPS caused massive (>10–20x) TNF- mRNA increases in kidney, liver, heart, and lung (compared with basal values). The spleen manifested the smallest TNF- increase of all test organs (about twice the increase with LPS treatment).
View this table:
Despite the fact that FeS administered alone failed to raise TNF- mRNA levels, it did augment TNF- mRNA increases in response to concomitant LPS injection (Table 3). However, this change occurred in an organ-specific fashion: increases in TNF- message levels were observed in the the kidney, heart, and lung, but not in the liver or spleen. However, even when a paired analysis of all organs together (LPS vs. LPS+FeS) is done, Fe treatment still caused a significant increase in LPS-driven TNF- mRNA increments (all organs, LPS alone: 2.5 ± 0.2; all organs, LPS+FeS: 3.1 ± 0.3; P < 0.025). Thus in composite, these experiments indicate that 1) FeS by itself does not alter TNF- message levels; 2) despite no apparent independent effect, Fe can augment LPS-mediated TNF- mRNA expression in the setting of a concomitant stimulus (in this case, LPS); and 3) these increases are not generalized to all organs but appear to occur in an organ-specific fashion.
DISCUSSION
We previously demonstrated that iv FeS injection dramatically sensitizes mice to E. coli-initiated TNF- generation without independently raising TNF- levels (48). Because whole E. coli were used in those experiments, the exact bacterial component(s) with which Fe interacted to produce this response could not be ascertained. Hence, the first goal of this study was to test whether these prior results were mediated by the endotoxin content of E. coli. The data obtained indicate that this is the case: a purified LPS injection completely recapitulated the previously observed FeS/E. coli results (twice plasma TNF- increases). Given these findings, it seems highly unlikely that additional bacterial components, e.g., proteases, help to mediate this response.
In prior publications, we suggested that different commercially available iron preparations exert differential cytotoxic effects (47, 49). In those experiments, performed in vitro and in vivo, FeS was found to be the most toxic, followed by FeG, and then FeD. While the exact explanation for this differential toxicity remains incompletely resolved, the degree to which these compounds undergo cellular uptake (FeS > FeG > FeD) is almost certainly involved (47). Each of our prior comparative studies defined toxicity primarily by Fe compound-induced degrees of mitochondrial dysfunction (ATP/ADP ratio reductions) or by tubular cell death. Therefore, the second goal of the current investigation was to ascertain whether consistent differential toxicity can also be observed when a fundamentally different toxicity end point is employed: plasma TNF- levels as an index of a proinflammatory state. The results obtained further support the concept that these compounds can, indeed, exert differential toxic effects, and in the same order as previously observed. When each of the three test agents (FeS, FeG, FeD) were administered concomitantly with LPS, only FeS potentiated the LPS-induced TNF- increment. During the glycerol-induced muscle injury protocol, each Fe compound increased TNF- levels, but to differing degrees, and in the same rank order (FeS > FeG > FeD) as previously observed with other toxicity end points (47, 49). Of note, none of the iron preparations independently raised plasma TNF- levels. This indicates that the iron compounds act by sensitizing mice to superimposed tissue injury/systemic stress rather than by exerting a direct inflammatory effect. That FeG injection into muscle induced no inflammatory reaction, as assessed histologically, helps to underscore this point.
In each of the above experiments, iron was injected simultaneously with LPS or with the glycerol challenge. Thus plasma Fe concentrations were at their maximum when the LPS or glycerol insult was imposed. This timing sequence allowed little opportunity for tissue Fe uptake, or for that matter, in vivo drug processing. Each of these two factors could clearly impact the results. Therefore, the third goal of this study was to ascertain whether a potentiation of TNF- generation might also occur after significant tissue Fe distribution/drug processing had occurred. To address this issue, mice were challenged with LPS 24 h following iv FeS, FeG, or FeD injection. As depicted in Fig. 2, each of these iron preparations sensitized the animals to subsequent LPS-mediated TNF- generation. Additional studies, conducted with FeG, demonstrated that synergistic toxicity could still be expressed even with a 48-h hiatus between Fe and LPS injection (Fig. 4). It is noteworthy that in this latter experiment, plasma iron concentrations were increased by 45%, insufficient to fully saturate transferrin. This indicates that although the employed iron dose (2 mg) is clearly "suprapharmacological" for humans, in the mouse it appears to have induced a clinically relevant result. While perhaps surprising, these results are consistent with the phenomenon of "scaling": i.e., that the drug dosage required to produce a given result in different species is inversely, and exponentially, related to species' mass or weight (e.g., Refs. 29 and 43). These considerations raise the possibility that the present findings might have clinical relevance, despite the Fe dose employed.
Oxidative stress is a well-defined activator of NF-B, resulting in a subsequent increase in TNF- gene transcription (35, 44). Hence, the next goal of this study was to test the hypothesis that parenteral iron exacerbates LPS-mediated TNF- generation by augmenting tissue levels of TNF- mRNA. Toward this goal, TNF- message levels were assessed in various organs either under control conditions or following FeS, LPS, or the combined FeS/LPS challenge. FeS injection alone failed to increase TNF- mRNA in any of the organs tested. (This is consistent with the observation that Fe injection by itself fails to raise plasma TNF-.) In contrast, LPS injection evoked striking TNF- mRNA increments, and in each test organ. Next to the liver, the kidney mounted the highest TNF- mRNA response (compared with its basal values). To our knowledge, this is the first direct comparison of relative degrees of LPS-triggered TNF- mRNA generation within different organs. Given the dramatic renal TNF- mRNA response, it seems plausible that the kidney is a prime contributor to plasma TNF- increases in endotoxemic states.
In contrast to the fact that FeS did not independently raise tissue TNF- mRNA, it did augment the LPS-initiated TNF- mRNA response. However, this occurred in an organ-specific fashion, being noted in the kidney, lung, and heart, but not in the liver or spleen. The mechanism by which FeS increased tissue TNF- mRNA and why these increments appeared in an organ-specific fashion remain unknown at that time. Nevertheless, these results support the concept that increased TNF- gene transcription was likely involved in the Fe-induced potentiation of TNF- production during the endotoxic state. However, it is important to note that TNF- production is regulated at both the transcriptional and translational level (e.g., Ref. 36). Thus it is premature to conclude that Fe impacts TNF- expression solely by transcriptional events. Also, it remains to be determined how FeS can increase TNF- message expression during endotoxemia and yet, when given alone, fails to produce this result.
Clearly, Fe therapy is required for anemia correction in patients with renal disease. Thus concerns of potential parenteral Fe toxicity, based on experimental data, are insufficient justification for withholding this therapy. Nevertheless, experimental results, such as those presented above, do provide an impetus for exploring alternative therapeutic approaches to mitigate potential iv Fe-induced adverse effects. Given that oxidative injury is central to Fe-mediated toxicity, one approach might be to administer Fe with antioxidant agent(s). However, a potential caveat to this approach is that antioxidants can facilitate Fe cycling between the ferrous and ferric state. Because this process can increase free radical generation and cytotoxicity (46, 50), and given that Fe2+ appears to be more cytotoxic than Fe3+ (46, 50), "antioxidant" therapy cannot simply be assumed to have a benefit. Indeed, the present findings that GSH worsened, rather than blocked, FeS's proinflammatory effect underscore this concern. Given these findings, we explored an alternative approach: delivering Fe via the im, rather than the iv, route. There are at least two theoretical benefits to this approach: first, the im route allows for slow Fe release, potentially mitigating toxicity; and second, im administration might allow the muscle, in a sense, to act as a tissue "buffer," absorbing some of Fe's toxic effects before systemic distribution. Initial experiments utilizing this approach appear to support these concepts. When FeG was administered im, it appeared to be well absorbed from muscle over a 48-h period, and yet it did not increase LPS-stimulated TNF- release. Despite these encouraging results, it is fully recognized that im Fe can exert its own focal toxic effects (pain, tissue pigmentation, and possible rhabdomyosarcoma) (12, 21). Whether such focal toxicity can itself be avoided (e.g., by coadministration of protective agents) and whether im Fe administration can adequately support erythropoietin-driven erythropoiesis remain unknown. Nevertheless, the current results suggest that further investigation of these issues may be justified.
In conclusion, the present results indicate that each of three clinically employed parenteral iron formulations (FeS, FeG, FeD) can sensitize mice to LPS-induced, or muscle injury-induced, TNF- release. However, degrees to which these agents sensitize mice to TNF- production under these conditions may differ, with an apparent rank toxicity order of FeS > FeG > FeD. Iron's ability to increase TNF- mRNA levels (most notably in the heart, lung, and kidney) during superimposed tissue stress may help mediate this response. Given a growing body of evidence that parenteral iron therapy may exert adverse effects, attempts to mitigate them appear warranted. Coadministration of an antioxidant is a theoretically attractive possibility. However, the ability of antioxidants to increase Fe cycling might render this approach unsuccessful, as noted above. Iron administration im is another potential alternative. However, whether injection site-related problems can be abrogated and whether this approach can maximally support erythropoiesis are important unresolved issues that will need to be addressed.
GRANTS
This work was supported by research grants from the National Institute of Diabetes and Digestive and Kidney Diseases (R01-DK-68520–01; R37-DK-038432–17).
FOOTNOTES
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
REFERENCES
Agarwal R, Vasavada N, Sachs NG, and Chase S. Oxidative stress, and renal injury with intravenous iron in patients with chronic kidney disease. Kidney Int 65: 2279–2289, 2004.
Antonicelli F, Brown D, Parmentier M, Drost EM, Hirani N, Rahman I, Donaldson K, and MacNee W. Regulation of LPS-mediated inflammation in vivo, and in vitro by the thiol antioxidant Nacystelyn. Am J Physiol Lung Cell Mol Physiol 286: L1319–L1327, 2004.
Baliga R, Ueda N, Walker PD, and Shah SV. Oxidant mechanisms in toxic acute renal failure. Am J Kidney Dis 29: 465–477, 1997.
Blackwell TS, Blackwell TR, Holden EP, Christman BW, and Christman JW. In vivo antioxidant treatment suppresses nuclear factor-B activation and neutrophilic lung inflammation. J Immunol 157: 1630–1637, 1996.
Breborowicz M, Polubinska A, Tam P, Wu G, and Breborowicz A. Effect of iron sucrose on human peritoneal mesothelial cells. Eur J Clin Invest 33: 1038–1044, 2003.
Cavdar C, Temiz A, Yenicerioglu Y, Caliskan S, Celik A, Sifil A, Onvural B, and Camsari T. The effects of intravenous iron treatment on oxidant stress, and erythrocyte deformability in hemodialysis patients. Scand J Urol Nephrol 37: 77–82, 2003.
Corti MC, Gaziano M, and Hennekens CH. Iron status and risk of cardiovascular disease. Ann Epidemiol 7: 62–68, 1997.
Daugherty A. Atherosclerosis: cell biology and lipoproteins. Curr Opin Lipidol 8: 11–12, 1997.
Esposito K, Nappo F, Marfella R, Giugliano G, Giugliano F, Ciotola M, Quagliaro L, Ceriello A, and Giugliano D. Inflammatory cytokine concentrations are acutely increased by hyperglycemia in humans: role of oxidative stress. Circulation 106: 2067–2072, 2002.
Fernandes-Real JM, Lopez-Bermejo A, and Ricart W. Cross-talk between iron metabolism, and diabetes. Diabetes 51: 2346–2354, 2002.
Gackowski D, Kruszewski M, Jawien A, Ciecierski M, and Olinski R. Further evidence that oxidative stress may be a risk factor responsible for the development of atherosclerosis. Free Radic Biol Med 31: 542–547, 2001.
Greenberg G. Sarcoma after intramuscular iron injection. Br Med J 1: 1508–1509, 1976.
Haddad JJ. Antioxidant, and prooxidant mechanisms in the regulation of redox(y)-sensitive transcription factors. Cell Signal 14: 879–897, 2002.
Handelman GJ, Walter MF, Adhikarla R, Gross J, Dallal GE, Levin NW, and Blumberg JB. Elevated plasma F2-isoprostanes in patients on long-term hemodialysis. Kidney Int 59: 1960–1966, 2001.
Horl WH. Adjunctive therapy in anaemia management. Nephrol Dial Transplant 17, Suppl 5: S56–S59, 2002.
Iseki K, Tozawa M, Yoshi S, and Fukiyama K. Serum C-reactive protein (CRP) and risk of death in chronic dialysis patients. Nephrol Dial Transplant 14: 1956–1960, 1999.
Kausz AT, Obrador GT, and Periera BJ. Anemia management in patients with chronic renal insufficiency. Am J Kidney Dis 36, Suppl 3: S39–S51, 2000.
Kosch M and Schaefer RM. Iron management in renal failure patients—how do we achieve the best results EDTNA ERCA J 28: 182–184, 2002.
Lim CS and Vaziri ND. The effects of iron dextran on the oxidative stress in cardiovascular tissues of rats with chronic renal failure. Kidney Int 65: 1802–1809, 2004.
Lim PS, Wei YH, Yu YL, and Kho B. Enhanced oxidative stress in haemodialysis patients receiving intravenous iron therapy. Nephrol Dial Transplant 14: 2680–2687, 1999.
Magnusson G, Flodh H, and Malmfors T. Oncological study in rats of Ferastral, an iron-poly-(sorbitol-gluconic acid) complex, after intramuscular administration. Scand J Haematol 32: S87–S88, 1977.
Nath KA, Vercellotti GM, Grande JP, Miyoshi H, Paya CV, Manivel JC, Haggard JJ, Croatt AJ, Payne WD, and Alam J. Heme protein-induced chronic renal inflammation: suppressive effect of induced heme oxygenase-1. Kidney Int 59: 106–117, 2001.
Owen WF and Lowrie EG. C-reactive protein as an outcome predictor for maintenance hemodialysis patients. Kidney Int 54: 627–636, 1998.
Ponraj D, Makjanic J, Thong PS, Tan BK, and Watt F. The onset of atherosclerotic lesion formation in hypercholesterolemic rabbits is delayed by iron depletion. FEBS Lett 459: 218–222, 1999.
Rahman I and MacNee W. Regulation of redox glutathione levels, and gene transcription in lung inflammation: therapeutic approaches. Free Radic Biol Med 28: 1405–1420, 2000.
Roob JM, Khoschsorur G, Tiran A, Horina JH, Holzer H, and Winklhofer-Roob BM. Vitamin E attenuates oxidative stress induced by intravenous iron in patients on hemodialysis. J Am Soc Nephrol 11: 539–549, 2000.
Rooyakkers TM, Stroes ES, Kooistra MP, van Faassen EE, Hider RC, Rabelink TJ, and Marx JJ. Ferric saccharate induces oxygen radical stress, and endothelial dysfunction in vivo. Eur J Clin Invest 32, Suppl 1: 9–16, 2002.
Salahudeen AK, Oliver B, Bower JD, and Oberts LJ. Increase in plasma esterified F-2 isoprostanes following intravenous iron infusion in patients on hemodialysis. Kidney Int 60: 1525–1531, 2001.
Schmidt-Nielsen K. Scaling: Why Is Animal Size So Important Cambridge, UK: Cambridge University Press, 1984.
Sengoelge G, Kletzmayr J, Ferrara I, Perschl A, Horl WH, and Sunder-Plassmann G. Impairment of trans-endothelial leukocyte migration by iron complexes. J Am Soc Nephrol 14: 2639–2644, 2003.
Shah SV. The role of reactive oxygen metabolites in glomerular disease. Annu Rev Physiol 57: 245–262, 1995.
Spittle MA, Hoenich NA, Handelman GJ, Adhikarla R, Homel P, and Levin NW. Oxidative stress, and inflammation in hemodialysis patients. Am J Kidney Dis 38: 1408–1413, 2001.
Stenvinkel P, Heimburger F, Paultre O, Diczfalusy U, Wang T, Berglund L, and Jogestrand T. Strong association between malnutrition, inflammation, and atherosclerosis in chronic renal failure. Kidney Int 55: 1899–1911, 1999.
Sturm B, Goldenberg H, and Scheiber-Mojdehkar B. Transient increase of the labile iron pool in HepG2 cells by intravenous iron preparations. Eur J Biochem 270: 3731–3738, 2003.
Sukamoto H. Iron regulation of hepatic macrophage TNF- expression. Free Radic Biol Med 32: 309–313, 2002.
Swantek JL, Cobb MH, and Geppert TD. Jun N terminal kinase/stress-activated kinase (JNK/SAPK) is required for lipopolysaccharide stimulation of tumor necrosis factor (TNF-) translation: glucocorticoids inhibit TNF- translation by blocking JNK/SAPK. Mol Cell Biol 17: 6274–6282, 1997.
Thorburn A. Death receptor-induced cell killing. Cell Signal 16: 139–144, 2004.
Valderrabano F, Jofre R, and Lopez-Gomez JM. Quality of life in end-stage renal disease patients. Am J Kidney Dis 38: 443–464, 2001.
Varfolomeev EE and Ashkenazi A. Tumor necrosis factor: an apoptosis JuNKie Cell 116: 491–497, 2004.
Villa P, Saccani A, Sica A, and Ghezzi P. Glutathione protects mice from lethal sepsis by limiting inflammation, and potentiating host defense. J Infect Dis 185: 1115–1120, 2002.
Visseren FL, Verkerk MS, van der Bruggen T, van der Bruggen T, Marx JJ, van Asbeck BS, and Diepersloot RJ. Iron chelation, and hydroxyl radical scavenging reduce the inflammatory response of endothelial cells after infection with Chlamydia pneumoniae or influenza A. Eur J Clin Invest 32, Suppl 1: 84–90, 2002.
Weiss G, Meusburger E, Radacher G, Garimorth K, Neyer U, and Mayer G. Effect of iron treatment on circulating cytokine levels in ESRD patients receiving recombinant human erythropoietin. Kidney Int 64: 572–578, 2003.
West GB, Woodruff WH, and Brown JH. Allometric scaling of metabolic rate from molecules, and mitochondria to cells and mammals. Proc Natl Acad Sci USA 99: 2473–2478, 2002.
Xiong S, She H, Takeuchi H, Han B, Engelhardt JF, Barton CH, Zandi E, Giulivi C, and Tsukamoto H. Signaling role of intracellular iron in NF-B activation. J Biol Chem 278: 17646–17654, 2003.
Yeun JY, Levine RA, Mantadilok V, and Kaysen GA. C reactive protein predicts all cause, and cardiovascular mortality in hemodialysis patients. Am J Kidney Dis 35: 469–476, 2000.
Zager RA and Foerder CA. Effects of inorganic iron, and myoglobin on in vitro proximal tubular lipid peroxidation and cytotoxicity. J Clin Invest 89: 989–995, 1992.
Zager RA, Johnson ACM, and Hanson SY. Parenteral iron nephrotoxicity: potential mechanisms, and consequences. Kidney Int 66: 144–156, 2004.
Zager RA, Johnson ACM, and Hanson SY. Parenteral iron therapy exacerbates experimental sepsis. Kidney Int 65: 2108–2112, 2004.
Zager RA, Johnson ACM, Hanson SY, and Wasse H. Parenteral iron formulations: a comparative toxicologic analysis, and mechanisms of cell injury. Am J Kidney Dis 40: 90–103, 2002.
Zager RA, Schimpf BA, Bredl CR, and Gmur DJ. Inorganic iron effects on in vitro hypoxic proximal renal tubular cell injury. J Clin Invest 91: 702–708, 1993.
Zager RA, Shah VO, Shah HV, Zager PG, Johnson AC, and Hanson S. The mevalonate pathway during acute tubular injury: selected determinants, and consequences. Am J Pathol 161: 681–692, 2002.
Zoccali C, Benedetto FA, Mallamaci F, Tripepi G, Fermo I, Foca A, Paroni R, and Malatino LS. Inflammation is associated with carotid atherosclerosis in hemodialysis patients. Creed Investigators. Cardiovascular risk extended evaluation in dialysis patients. J Hypertens 18: 1207–1213, 2000.(Richard A. Zager, A. C. M)
ABSTRACT
Intravenous Fe is widely used to treat anemia in renal disease patients. However, concerns of potential Fe toxicity exist. To more fully define its spectrum, this study tested Fe's impact on systemic inflammation following either endotoxemia or the induction of direct tissue damage (glycerol-mediated rhabdomyolysis). The inflammatory response was gauged by tissue TNF- message expression and plasma TNF- levels. CD-1 mice received either intravenous Fe sucrose, -gluconate, or -dextran (FeS, FeG, or FeD, respectively; 2 mg), followed by either endotoxin (LPS) or glycerol injection 0–48 h later. Plasma TNF- was assessed by ELISA 2–3 h after the LPS or glycerol challenge. TNF- mRNA expression (RT-PCR) was measured in the kidney, heart, liver, lung, and spleen with Fe ± LPS treatment. Finally, the relative impacts of intramuscular vs. intravenous Fe and of glutathione (GSH) on Fe/LPS- induced TNF- generation were assessed. Each Fe preparation significantly enhanced LPS- or muscle injury-mediated TNF- generation. This effect was observed for at least 48 h post-Fe injection, a time at which plasma iron levels were increased by levels insufficient to fully saturate transferrin. Fe did not independently increase plasma TNF- or tissue mRNA. However, it potentiated postinjury-induced TNF- mRNA increments and did so in an organ-specific fashion (kidney, heart, and lung; but not in liver or spleen). Intramuscular administration, but not GSH treatment, negated Fe's ability to synergize LPS-mediated TNF- release. We conclude 1) intravenous Fe can enhance TNF- generation during LPS- or glycerol-induced tissue damage; 2) increased TNF- gene transcription in the kidney, heart, and lung may contribute to this result; and 3) intramuscular administration, but not GSH, might potentially mitigate some of Fe's systemic toxic effects.
endotoxemia; rhabdomyolysis; end-stage renal disease; iron deficiency; glutathione
PARENTERAL IRON THERAPY HAS become a mainstay in anemia management in patients with progressive or end-stage renal disease (ESRD) (15, 17, 18). Its need stems from a variety of factors, including low iron intake or absorption, dialysis-dependent blood loss, and/or "functional" deficiency, defined as an insufficient amount of bioavailable iron to support maximal erythropoietin-stimulated erythropoiesis. Fe therapy, with the resultant correction of anemia, is believed to improve quality of life and possibly decrease mortality (38). Nevertheless, concerns of potential Fe toxicity exist, based on the fact that all currently employed parenteral Fe preparations [Fe sucrose (FeS), Fe gluconate (FeG), and Fe dextran (FeD)] can exert marked prooxidant effects. This conclusion is based on results from in vivo (e.g., Refs. 19, 47, 49) and in vitro (e.g., Refs. 5, 30, 34, 47, 49) experiments and from clinical investigations (e.g., Refs. 1, 6, 14, 20, 26–28, 32). Oxidative stress is believed to play an important pathogenic role in atherogenesis (7, 8, 11, 24) and in renal disease progression (3, 10, 22, 31). These considerations raise further concerns regarding the safety of intravenous (iv) Fe treatment in patients with renal disease (47). The recent demonstration that iv FeS transiently exacerbates proteinuria and induces proximal tubular enzymuria in pre-ESRD patients supports these concerns (1).
An additional pathway by which iv Fe might adversely impact renal disease patients is via a potentiation of systemic inflammation, a correlate of ESRD (33, 42, 52). That levels of serum C-reactive protein, a surrogate marker of inflammation, correlate with dialysis-associated complications, e.g., adverse cardiovascular events (16, 23, 45), underscores the clinical relevance of this inflammatory state. Oxidative stress can promote inflammation via a number of potential pathways. For example, redox-sensitive transcription factors, most notably NF-B, cFos, c-Jun, activator protein 1, and hypoxia-inducible factor-1, are upregulated by oxidative stress (9, 13, 25, 35, 44). These can then recruit "downstream" mediators, such as TNF-, which can potentiate systemic inflammation (2, 4, 40, 41), or trigger apoptotic cell death (37, 39).
Given the above, it is reasonable to question whether parenteral iron-induced oxidative stress can potentiate a proinflammatory state. A recent brief report from this laboratory provides initial support for this concept (48). Experimental sepsis was induced in mice by injecting nonviable Escherichia coli ± iv FeS. Despite the fact that FeS alone evoked no discernible tissue injury/inflammation, it profoundly potentiated the E. coli-induced septic state. For example, FeS-treated mice manifested a doubling of E. coli-initiated TNF- generation, culminating in an 50% mortality rate (48). Clearly, then, whereas iv FeS appeared to be well tolerated, in concert with E. coli, it had profound toxic effects.
The present study was undertaken to gain further insights into the above-defined phenomenon, and by so doing better define the spectrum and mechanisms of parenteral Fe toxicity. Specifically, the following questions have been addressed. 1) Can FeS potentiate TNF- generation following purified endotoxin administration, or is this response only seen with intact bacterial injections 2) Given that our prior studies have found that FeS has by far the greatest cytotoxic potential of the currently employed Fe preparations (47, 49), is FeS unique in its ability to potentiate TNF- generation, or can other Fe compounds reproduce this result 3) What is the mechanism by which iv Fe potentiates TNF- generation in the setting of sepsis Specifically, does Fe increase TNF- generation via transcriptional or posttranscriptional events 4) What are the target organs that lead to exaggerated TNF- generation in response to an Fe-loading/septic state 5) Can iv Fe potentiate TNF- generation with non-LPS-driven tissue damage Experiments to gain insights into these, and other issues, form the basis of this report.
METHODS
General Methods
Male CD-1 mice (25–35 g; Charles River Laboratories, Wilmington, MA) were used for all experiments. All animal experiments were in accordance with, and approval of, Fred Hutchinson Cancer Research Center institutional animal care and and use committee guidelines. They were maintained under routine vivarium conditions with free food and water access throughout. Tail vein Fe injections were used and encompassed the following test agents: 1) FeS (Venofer, American Regent, Shirley, NY); 2) FeD (INFeD, Watson Pharmaceuticals, Morristown, NJ); or 3) FeG (Ferrlecit, Watson Pharmaceuticals). An equal amount of elemental Fe was used with each agent (2 mg: equaling 100 μl of FeS, 160 μl of FeG, or 40 μl of FeD of their respective stock solutions). Non-Fe-treated mice received a tail vein injection of 100 μl 0.9% saline as a sham for Fe treatment. Tail vein injections were performed by placing individual mice into cylindrical restraining holders, cannulating the vein with a 26-gauge needle, and injecting the test agents over 30 s.
Purified endotoxin (LPS) was obtained from Sigma (E. coli 0111:B4; catalogue no. L-2630; stock solution, 0.4 mg/ml saline). It was administered intraperitoneally (ip) in a dose of either 2 or 10 mg/kg. Mice not receiving LPS received an equal amount of the LPS vehicle (saline ip). As a second model of tissue injury, the glycerol model of muscle necrosis was employed. For this purpose, mice were lightly anesthetized with isoflurane and injected with 6 ml/kg of 50% glycerol, administered intramuscularly (im) into one hindlimb. The specific Fe treatments employed in the LPS or glycerol experiments are discussed in the individual protocols presented below.
Following the above interventions, the mice were allowed to recover from anesthesia and provided with free food and water access. Subsequent animal death was conducted under deep pentobarbital sodium anesthesia (2–4 mg/kg). The abdominal cavity of each mouse was opened with a midline incision, and a terminal phlebotomy was conducted from the inferior vena cava into a heparinized syringe. Plasma samples were stored at –70°C until assay for TNF- levels conducted by ELISA (R&D Systems; Minneapolis, MN; Ref. 49).
Specific Experiments
Immediate effect of different Fe compounds on LPS-induced TNF- increments. This experiment was undertaken to ascertain whether the above-noted Fe preparations impact LPS-mediated TNF- generation, and if so, ascertain whether differences exist among these Fe-containing polymers. Mice (n = 47) were subjected to one of the following tail vein injections: 1) 0.1 ml normal saline (sham for Fe treatment; n = 17); 2) FeS (n = 11), FeG (n = 8), or FeD (n = 11). Immediately after Fe or saline administration, the mice received an injection of LPS (2 mg/kg ip). Two hours post-LPS injection, the mice were killed and the terminal plasma samples were collected for TNF- assay.
Delayed effect of different Fe compounds on LPS-induced TNF- increments. The above experiment indicated that when Fe is administered concomitantly with LPS, a potentiation of plasma TNF- increments occurs (see RESULTS). This experiment was undertaken to ascertain whether Fe can also potentiate LPS-mediated TNF- generation after a 24-h delay, allowing for in vivo drug distribution/drug processing to occur. Potential differences among the test Fe compounds were also assessed. Mice (n = 20) received tail vein injections of either saline (n = 6) or Fe (FeS, n = 6; FeG, n = 4; or FeD, n = 4). Then, they were returned to their cages. Twenty-four hours later, each mouse received an LPS injection (10 mg/kg ip). Three hours post-LPS treatment, terminal plasma samples were obtained for TNF- assay.
Parenteral Fe effect on muscle injury-initiated TNF- levels. The following experiment was undertaken to ascertain whether parenteral Fe can potentiate TNF- generation when the latter is initiated by direct, focal tissue damage, as opposed to a systemic endotoxemic state. Toward this end, muscle injury was induced by im glycerol injection ± iv Fe therapy, as noted above (General Methods). Mice were divided into the following experimental groups: 1) iv FeS, followed immediately by glycerol injection (n = 10); 2) iv FeG+glycerol (n = 5); 3) iv FeD+glycerol (n = 7); and 4) sham iv Fe (0.1 ml of saline)+glycerol. Three hours later, the mice were killed and terminal plasma samples were analyzed for TNF- levels, as noted above.
Independent effect of parenteral Fe administration on TNF- levels. The following experiment was undertaken to ascertain the possible impact of parenteral Fe administration on circulating TNF- levels independent of concomitant forms of tissue stress (i.e., LPS or glycerol injection). To this end, 12 mice were divided into 4 equal groups, as follows: 1) sham Fe injection; 2) iv FeS; 3) iv FeG; or 4) iv FeD. Three hours postinjection, terminal plasma samples were collected and assayed for TNF- levels.
Comparison of effects of iv vs. im Fe on LPS-mediated TNF- generation.
iv vs. im FeG with a 24-h latency period. The following experiment was undertaken to ascertain whether use of different administration routes alters the extent to which Fe can potentiate LPS-mediated TNF- generation. Six mice received 2 mg of elemental Fe (FeG) via either the iv or the im route (n = 3 each). After allowing a 24-h Fe equilibration period, each of the mice received the 10 mg/kg ip LPS challenge. Three hours post-LPS injection, terminal plasma samples were obtained for TNF- assay. The results were compared with those obtained in six mice subjected to ip LPS injection without prior Fe administration.
iv vs. im FeG with a 48-h latency period. The above experiment was repeated, but with a 48-h latency period as follows: 1) iv FeG+0.1 ml im saline (n = 6); 2) im FeG+0.1 ml iv saline (n = 6); and 3) controls (0.1 ml iv saline+0.1 ml im saline; n = 7). (Note that each mouse received both an iv and im injection, thereby controlling for this variable among the groups.) The mice were then returned to their cages for a 48-h drug distribution/processing period. After completing this period, each of the mice was subjected to the 10 mg/kg ip LPS challenge. Three hours later, plasma samples were obtained for TNF- assay.
Plasma Fe and Fe binding capacity 48 h after iv and im FeG administration. The following experiment was performed to ascertain the effect of the above iv and im Fe treatments on plasma Fe and %saturation of total iron binding capacity (TIBC). Six mice were divided into three treatment groups of two mice each: 1) tail vein saline injection (controls); 2) tail vein FeG injection; and 3) im FeG administration. After a 48-h delay, terminal plasma samples were obtained and assayed for Fe, TIBC, and %TIBC saturation. These determinations were performed with a Beckman Synchrom LX20 autoanalyzer (University of Washington Medical Center Clinical Laboratory; Seattle, WA).
To gauge Fe absorption following im administration, three mice received a 2-mg Fe injection (as FeG) in one upper thigh. Two days later, the mice were anesthetized and subjected to a repeat im FeG injection, this time administered in the contralateral thigh. Five minutes later, both thighs were dissected and the relative amount of black Fe staining in each thigh was visually contrasted. In addition, 1-cm-long pieces of injected muscle tissue were obtained immediately after, and 48 h after, FeG injection. The muscle samples were fixed in 10% formalin and embedded in paraffin. Serial sections were prepared throughout the muscle length (15 sections/muscle), and these were stained with hematoxylin and eosin. Fe deposition and muscle injury were gauged by light microscopy.
Glutathione addition to FeS: impact on LPS-mediated TNF- generation. As an alternative approach to im Fe injection as a possible means to mitigate Fe-stimulated TNF- generation, the antioxidant glutathione (GSH) was added to FeS before its injection. To this end, 5 mg of GSH (50 μl of a 100 mg/ml stock saline solution) were added to 2 mg of FeS. FeS/GSH was then administered via tail vein injection (into 3 mice). Three control mice received the same 2-mg FeS injection, substituting 50 μl of saline for the GSH addition. Immediately thereafter, each of the six mice received a 2-mg/kg ip LPS injection. Three hours post-LPS injection, the mice were anesthetized and terminal plasma samples were obtained and assayed for TNF- levels.
Tissue TNF- mRNA level and responses to iv Fe, LPS, or Fe+LPS treatment. The following experiment had three objectives: 1) ascertain the effect of iv Fe on tissue TNF- mRNA levels; 2) test the impact of iv Fe on LPS-stimulated TNF- mRNA expression; and 3) ascertain whether different tissues have different TNF- mRNA responses to Fe, LPS, or Fe+LPS injections. Four groups of mice (n = 5 each) were established: 1) controls (iv saline); 2) iv FeS; 3) LPS (10 mg/kg ip); and 4) iv FeS+LPS. Three hours later, the mice were anesthetized, and samples of the following tissues were obtained: kidney cortex, liver, cardiac apex, lung, and spleen.
Tissue samples were immediately placed into TRIzol reagent (Invitrogen Life Technologies, Carlsbad, CA), and total RNA was extracted according to the manufacturer's instructions. The final RNA pellet was brought up in RNase-free water to a concentration of 0.5–2 μg/ml. RNA integrity was ascertained on an ethidium bromide-agarose gel (41). RT-PCR was performed using a 1st Strand Synthesis Kit for RT-PCR (Ambion, Austin, TX), as previously described in detail (51). The specific primers for TNF- and GAPDH (housekeeping gene) were designed with 50–60% GC composition (see Table 1). Multiplex PCR was possible due to the primers' similar annealing temperatures but dissimilar sizes of the two PCR products. PCR conditions were optimized for each tissue (37 cycles for TNF- in all tissues; 18, 22, 19, 25, and 25 GAPDH cycles for kidney, liver, heart, spleen, and lung, respectively). PCR products were analyzed by agarose gel electrophoresis and ethidium bromide staining. The PCR products were quantified by densitometry with a Typhoon 8600 scanner (Amersham Phamacia Biosciences, Piscataway, NJ). TNF- bands were expressed as ratios to the simultaneously obtained GAPDH bands. Results in each organ were compared between the four treatment groups. The %changes in TNF- mRNA (controls vs. therapies) between organs were also compared.
View this table:
Calculations and Statistics
All values are presented as means ± SE. Statistical comparisons of paired and unpaired data were performed by paired and unpaired Student's t-test, respectively. In addition, comparisons of TNF- mRNA levels between two treatments were assessed by a test of binomial proportions. If multiple comparisons were made, the Bonferroni correction was applied. Significance was judged by a P value of <0.05.
RESULTS
Comparative, Immediate Effect of iv Fe on LPS-Mediated TNF- Generation
Normal mouse plasma has a baseline TNF- level that is at or below the level of accurate detection by the employed ELISA (<1–2 pg/ml; Ref. 48). As shown in Fig. 1, LPS administration induced dramatic plasma TNF- increments. Mice that had been treated concomitantly with FeS manifested an approximate doubling of these LPS-induced TNF- increases. [Of note, these results were essentially identical to those previously reported by this group using Fe+E. coli (48)]. Unlike FeS, neither FeG nor FeD caused any potentiation of LPS-initiated plasma TNF- elevations.
Delayed Effect of Fe Compounds on LPS-Induced TNF- Increments
When non-Fe-treated mice were subjected to LPS injection, plasma TNF- levels rose to 760 ± 80 pg/ml. The upper limit of the 95% confidence band for these mice, 1,000 pg/ml, is presented as a horizontal line in Fig. 2. Each mouse in each of the three Fe pretreatment groups (i.e., iv Fe administration 24 h earlier) developed a post-LPS TNF- level that exceeded this 95% confidence range (i.e., in each mouse, Fe pretreatment sensitized to LPS-mediated TNF- generation). FeD and FeG appeared to exert comparable effects on post-LPS TNF- plasma levels. The FeS group had 60% greater elevations in post-LPS TNF- levels compared with either the FeD or the FeG group.
Fe Effects on TNF- Levels Following Hypertonic Glycerol-Mediated Muscle Injury
Glycerol injection alone caused only minimal TNF- generation, rising from control values of 1 to 2.5 ± 0.6 pg/ml (P < 0.01; Fig. 3). In each case, concomitant Fe administration raised postglycerol TNF- levels, and with statistically significant differences in degrees of elevations being observed (FeS > FeG > FeD). Of note, although Fe potentiated postglycerol TNF- increases (e.g., 6-fold increase with FeS, compared with glycerol alone), the absolute degrees of elevation were quite modest compared with those that followed LPS injection (note the different y-axis scales in Fig. 3 vs. Figs. 1 and 2).
TNF- Levels Following Fe Treatment Alone
Treatment of normal mice with any of the Fe compounds caused no significant increase in plasma TNF- levels compared with non-Fe-injected control mice (all values 1 pg/ml). This indicates that the TNF- differences observed in the above experiments could not be explained by an independent Fe effect on TNF- levels.
Comparison of Effects of im vs. iv Fe Injections on LPS-Mediated TNF- Levels
As shown in Fig. 4, left, mice that received iv FeG 24 h before the LPS challenge had a dramatic increase in plasma TNF- levels compared with non-Fe-treated, LPS-injected controls (P < 0.005). In striking contrast, im Fe pretreatment did not predispose mice to LPS-mediated TNF- release. When this same experiment was repeated, but with a 48-h lag time allowed between iv/im FeG treatment and LPS injection, a marked potentiation of LPS-mediated TNF- generation was still observed in the mice that had received prior iv FeG treatment (Fig. 4, right). Once again, im FeG did not potentiate LPS-mediated TNF- release (Fig. 4, right).
Plasma Fe and TIBC Assessments Following iv and im FeG Injections
Plasma Fe, TIBC, and %TIBC saturation are presented in Table 2. Plasma Fe concentrations were increased by 45% at 48 h post-iv FeG injection. This corresponded with a modest increase in both TIBC and %TIBC saturation, the latter rising from 66 to 86% with iv FeG treatment. In contrast, im FeG injection failed to raise plasma Fe levels. Rather, there tended to be a slight decrease in the Fe, TIBC, and %TIBC saturation with this treatment protocol.
View this table:
Muscle evaluation. By visual inspection, obvious black deposits were observed in muscle immediately following im injection. In contrast, muscle injection sites obtained after a 48-h lag time revealed no visual evidence for persistent Fe deposition. Representative histological sections through upper leg muscle immediately following, and 48 h following, FeG injection are presented in Fig. 5. As shown in Fig. 5, A and B, sections obtained from muscle 5 min after FeG injection demonstrated marked interstitial fluid accumulation in muscle ("lakes"), with wide separation of muscle fibers. On higher power, particulate material, consisting of the injected FeG, was apparent (Fig. 5B). In contrast, muscle obtained 48 h post-FeG injection appeared completely normal: there was no longer evidence of fluid-induced separation of muscle fibers or of particulate Fe deposition. Thus these findings supported the visual findings: that the injected FeG had been absorbed from muscle by 48 h postinjection. Of further note, FeG injection failed to induce either muscle fiber necrosis or an inflammatory response, suggesting that it had caused little, if any, muscle damage.
Impact of GSH Addition to FeS on LPS-Mediated TNF- Release
Although the GSH stock solution was colorless, when it was added to FeS, it caused a modest color change (from black to reddish brown; consistent with a change of in Fe valence). GSH addition to the FeS did not diminish its ability to facilitate LPS-mediated TNF- generation: rather, it tended to increase, not decrease, this response: (FeS/GSH+LPS: TNF-, 621 ± 39 pg/ml; FeS alone+LPS: TNF-, 506 ± 42 pg/ml; P < 0.035 by paired t-test analysis for 3 sets of paired mice).
TNF- mRNA Tissue Levels Following FeS, LPS, or FeS+LPS Injections
TNF- mRNA levels in the kidney, liver, lung, heart, and spleen are presented in Table 3. By itself, FeS injection failed to cause a significant increase in message levels in any of the tissues examined. LPS caused massive (>10–20x) TNF- mRNA increases in kidney, liver, heart, and lung (compared with basal values). The spleen manifested the smallest TNF- increase of all test organs (about twice the increase with LPS treatment).
View this table:
Despite the fact that FeS administered alone failed to raise TNF- mRNA levels, it did augment TNF- mRNA increases in response to concomitant LPS injection (Table 3). However, this change occurred in an organ-specific fashion: increases in TNF- message levels were observed in the the kidney, heart, and lung, but not in the liver or spleen. However, even when a paired analysis of all organs together (LPS vs. LPS+FeS) is done, Fe treatment still caused a significant increase in LPS-driven TNF- mRNA increments (all organs, LPS alone: 2.5 ± 0.2; all organs, LPS+FeS: 3.1 ± 0.3; P < 0.025). Thus in composite, these experiments indicate that 1) FeS by itself does not alter TNF- message levels; 2) despite no apparent independent effect, Fe can augment LPS-mediated TNF- mRNA expression in the setting of a concomitant stimulus (in this case, LPS); and 3) these increases are not generalized to all organs but appear to occur in an organ-specific fashion.
DISCUSSION
We previously demonstrated that iv FeS injection dramatically sensitizes mice to E. coli-initiated TNF- generation without independently raising TNF- levels (48). Because whole E. coli were used in those experiments, the exact bacterial component(s) with which Fe interacted to produce this response could not be ascertained. Hence, the first goal of this study was to test whether these prior results were mediated by the endotoxin content of E. coli. The data obtained indicate that this is the case: a purified LPS injection completely recapitulated the previously observed FeS/E. coli results (twice plasma TNF- increases). Given these findings, it seems highly unlikely that additional bacterial components, e.g., proteases, help to mediate this response.
In prior publications, we suggested that different commercially available iron preparations exert differential cytotoxic effects (47, 49). In those experiments, performed in vitro and in vivo, FeS was found to be the most toxic, followed by FeG, and then FeD. While the exact explanation for this differential toxicity remains incompletely resolved, the degree to which these compounds undergo cellular uptake (FeS > FeG > FeD) is almost certainly involved (47). Each of our prior comparative studies defined toxicity primarily by Fe compound-induced degrees of mitochondrial dysfunction (ATP/ADP ratio reductions) or by tubular cell death. Therefore, the second goal of the current investigation was to ascertain whether consistent differential toxicity can also be observed when a fundamentally different toxicity end point is employed: plasma TNF- levels as an index of a proinflammatory state. The results obtained further support the concept that these compounds can, indeed, exert differential toxic effects, and in the same order as previously observed. When each of the three test agents (FeS, FeG, FeD) were administered concomitantly with LPS, only FeS potentiated the LPS-induced TNF- increment. During the glycerol-induced muscle injury protocol, each Fe compound increased TNF- levels, but to differing degrees, and in the same rank order (FeS > FeG > FeD) as previously observed with other toxicity end points (47, 49). Of note, none of the iron preparations independently raised plasma TNF- levels. This indicates that the iron compounds act by sensitizing mice to superimposed tissue injury/systemic stress rather than by exerting a direct inflammatory effect. That FeG injection into muscle induced no inflammatory reaction, as assessed histologically, helps to underscore this point.
In each of the above experiments, iron was injected simultaneously with LPS or with the glycerol challenge. Thus plasma Fe concentrations were at their maximum when the LPS or glycerol insult was imposed. This timing sequence allowed little opportunity for tissue Fe uptake, or for that matter, in vivo drug processing. Each of these two factors could clearly impact the results. Therefore, the third goal of this study was to ascertain whether a potentiation of TNF- generation might also occur after significant tissue Fe distribution/drug processing had occurred. To address this issue, mice were challenged with LPS 24 h following iv FeS, FeG, or FeD injection. As depicted in Fig. 2, each of these iron preparations sensitized the animals to subsequent LPS-mediated TNF- generation. Additional studies, conducted with FeG, demonstrated that synergistic toxicity could still be expressed even with a 48-h hiatus between Fe and LPS injection (Fig. 4). It is noteworthy that in this latter experiment, plasma iron concentrations were increased by 45%, insufficient to fully saturate transferrin. This indicates that although the employed iron dose (2 mg) is clearly "suprapharmacological" for humans, in the mouse it appears to have induced a clinically relevant result. While perhaps surprising, these results are consistent with the phenomenon of "scaling": i.e., that the drug dosage required to produce a given result in different species is inversely, and exponentially, related to species' mass or weight (e.g., Refs. 29 and 43). These considerations raise the possibility that the present findings might have clinical relevance, despite the Fe dose employed.
Oxidative stress is a well-defined activator of NF-B, resulting in a subsequent increase in TNF- gene transcription (35, 44). Hence, the next goal of this study was to test the hypothesis that parenteral iron exacerbates LPS-mediated TNF- generation by augmenting tissue levels of TNF- mRNA. Toward this goal, TNF- message levels were assessed in various organs either under control conditions or following FeS, LPS, or the combined FeS/LPS challenge. FeS injection alone failed to increase TNF- mRNA in any of the organs tested. (This is consistent with the observation that Fe injection by itself fails to raise plasma TNF-.) In contrast, LPS injection evoked striking TNF- mRNA increments, and in each test organ. Next to the liver, the kidney mounted the highest TNF- mRNA response (compared with its basal values). To our knowledge, this is the first direct comparison of relative degrees of LPS-triggered TNF- mRNA generation within different organs. Given the dramatic renal TNF- mRNA response, it seems plausible that the kidney is a prime contributor to plasma TNF- increases in endotoxemic states.
In contrast to the fact that FeS did not independently raise tissue TNF- mRNA, it did augment the LPS-initiated TNF- mRNA response. However, this occurred in an organ-specific fashion, being noted in the kidney, lung, and heart, but not in the liver or spleen. The mechanism by which FeS increased tissue TNF- mRNA and why these increments appeared in an organ-specific fashion remain unknown at that time. Nevertheless, these results support the concept that increased TNF- gene transcription was likely involved in the Fe-induced potentiation of TNF- production during the endotoxic state. However, it is important to note that TNF- production is regulated at both the transcriptional and translational level (e.g., Ref. 36). Thus it is premature to conclude that Fe impacts TNF- expression solely by transcriptional events. Also, it remains to be determined how FeS can increase TNF- message expression during endotoxemia and yet, when given alone, fails to produce this result.
Clearly, Fe therapy is required for anemia correction in patients with renal disease. Thus concerns of potential parenteral Fe toxicity, based on experimental data, are insufficient justification for withholding this therapy. Nevertheless, experimental results, such as those presented above, do provide an impetus for exploring alternative therapeutic approaches to mitigate potential iv Fe-induced adverse effects. Given that oxidative injury is central to Fe-mediated toxicity, one approach might be to administer Fe with antioxidant agent(s). However, a potential caveat to this approach is that antioxidants can facilitate Fe cycling between the ferrous and ferric state. Because this process can increase free radical generation and cytotoxicity (46, 50), and given that Fe2+ appears to be more cytotoxic than Fe3+ (46, 50), "antioxidant" therapy cannot simply be assumed to have a benefit. Indeed, the present findings that GSH worsened, rather than blocked, FeS's proinflammatory effect underscore this concern. Given these findings, we explored an alternative approach: delivering Fe via the im, rather than the iv, route. There are at least two theoretical benefits to this approach: first, the im route allows for slow Fe release, potentially mitigating toxicity; and second, im administration might allow the muscle, in a sense, to act as a tissue "buffer," absorbing some of Fe's toxic effects before systemic distribution. Initial experiments utilizing this approach appear to support these concepts. When FeG was administered im, it appeared to be well absorbed from muscle over a 48-h period, and yet it did not increase LPS-stimulated TNF- release. Despite these encouraging results, it is fully recognized that im Fe can exert its own focal toxic effects (pain, tissue pigmentation, and possible rhabdomyosarcoma) (12, 21). Whether such focal toxicity can itself be avoided (e.g., by coadministration of protective agents) and whether im Fe administration can adequately support erythropoietin-driven erythropoiesis remain unknown. Nevertheless, the current results suggest that further investigation of these issues may be justified.
In conclusion, the present results indicate that each of three clinically employed parenteral iron formulations (FeS, FeG, FeD) can sensitize mice to LPS-induced, or muscle injury-induced, TNF- release. However, degrees to which these agents sensitize mice to TNF- production under these conditions may differ, with an apparent rank toxicity order of FeS > FeG > FeD. Iron's ability to increase TNF- mRNA levels (most notably in the heart, lung, and kidney) during superimposed tissue stress may help mediate this response. Given a growing body of evidence that parenteral iron therapy may exert adverse effects, attempts to mitigate them appear warranted. Coadministration of an antioxidant is a theoretically attractive possibility. However, the ability of antioxidants to increase Fe cycling might render this approach unsuccessful, as noted above. Iron administration im is another potential alternative. However, whether injection site-related problems can be abrogated and whether this approach can maximally support erythropoiesis are important unresolved issues that will need to be addressed.
GRANTS
This work was supported by research grants from the National Institute of Diabetes and Digestive and Kidney Diseases (R01-DK-68520–01; R37-DK-038432–17).
FOOTNOTES
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
REFERENCES
Agarwal R, Vasavada N, Sachs NG, and Chase S. Oxidative stress, and renal injury with intravenous iron in patients with chronic kidney disease. Kidney Int 65: 2279–2289, 2004.
Antonicelli F, Brown D, Parmentier M, Drost EM, Hirani N, Rahman I, Donaldson K, and MacNee W. Regulation of LPS-mediated inflammation in vivo, and in vitro by the thiol antioxidant Nacystelyn. Am J Physiol Lung Cell Mol Physiol 286: L1319–L1327, 2004.
Baliga R, Ueda N, Walker PD, and Shah SV. Oxidant mechanisms in toxic acute renal failure. Am J Kidney Dis 29: 465–477, 1997.
Blackwell TS, Blackwell TR, Holden EP, Christman BW, and Christman JW. In vivo antioxidant treatment suppresses nuclear factor-B activation and neutrophilic lung inflammation. J Immunol 157: 1630–1637, 1996.
Breborowicz M, Polubinska A, Tam P, Wu G, and Breborowicz A. Effect of iron sucrose on human peritoneal mesothelial cells. Eur J Clin Invest 33: 1038–1044, 2003.
Cavdar C, Temiz A, Yenicerioglu Y, Caliskan S, Celik A, Sifil A, Onvural B, and Camsari T. The effects of intravenous iron treatment on oxidant stress, and erythrocyte deformability in hemodialysis patients. Scand J Urol Nephrol 37: 77–82, 2003.
Corti MC, Gaziano M, and Hennekens CH. Iron status and risk of cardiovascular disease. Ann Epidemiol 7: 62–68, 1997.
Daugherty A. Atherosclerosis: cell biology and lipoproteins. Curr Opin Lipidol 8: 11–12, 1997.
Esposito K, Nappo F, Marfella R, Giugliano G, Giugliano F, Ciotola M, Quagliaro L, Ceriello A, and Giugliano D. Inflammatory cytokine concentrations are acutely increased by hyperglycemia in humans: role of oxidative stress. Circulation 106: 2067–2072, 2002.
Fernandes-Real JM, Lopez-Bermejo A, and Ricart W. Cross-talk between iron metabolism, and diabetes. Diabetes 51: 2346–2354, 2002.
Gackowski D, Kruszewski M, Jawien A, Ciecierski M, and Olinski R. Further evidence that oxidative stress may be a risk factor responsible for the development of atherosclerosis. Free Radic Biol Med 31: 542–547, 2001.
Greenberg G. Sarcoma after intramuscular iron injection. Br Med J 1: 1508–1509, 1976.
Haddad JJ. Antioxidant, and prooxidant mechanisms in the regulation of redox(y)-sensitive transcription factors. Cell Signal 14: 879–897, 2002.
Handelman GJ, Walter MF, Adhikarla R, Gross J, Dallal GE, Levin NW, and Blumberg JB. Elevated plasma F2-isoprostanes in patients on long-term hemodialysis. Kidney Int 59: 1960–1966, 2001.
Horl WH. Adjunctive therapy in anaemia management. Nephrol Dial Transplant 17, Suppl 5: S56–S59, 2002.
Iseki K, Tozawa M, Yoshi S, and Fukiyama K. Serum C-reactive protein (CRP) and risk of death in chronic dialysis patients. Nephrol Dial Transplant 14: 1956–1960, 1999.
Kausz AT, Obrador GT, and Periera BJ. Anemia management in patients with chronic renal insufficiency. Am J Kidney Dis 36, Suppl 3: S39–S51, 2000.
Kosch M and Schaefer RM. Iron management in renal failure patients—how do we achieve the best results EDTNA ERCA J 28: 182–184, 2002.
Lim CS and Vaziri ND. The effects of iron dextran on the oxidative stress in cardiovascular tissues of rats with chronic renal failure. Kidney Int 65: 1802–1809, 2004.
Lim PS, Wei YH, Yu YL, and Kho B. Enhanced oxidative stress in haemodialysis patients receiving intravenous iron therapy. Nephrol Dial Transplant 14: 2680–2687, 1999.
Magnusson G, Flodh H, and Malmfors T. Oncological study in rats of Ferastral, an iron-poly-(sorbitol-gluconic acid) complex, after intramuscular administration. Scand J Haematol 32: S87–S88, 1977.
Nath KA, Vercellotti GM, Grande JP, Miyoshi H, Paya CV, Manivel JC, Haggard JJ, Croatt AJ, Payne WD, and Alam J. Heme protein-induced chronic renal inflammation: suppressive effect of induced heme oxygenase-1. Kidney Int 59: 106–117, 2001.
Owen WF and Lowrie EG. C-reactive protein as an outcome predictor for maintenance hemodialysis patients. Kidney Int 54: 627–636, 1998.
Ponraj D, Makjanic J, Thong PS, Tan BK, and Watt F. The onset of atherosclerotic lesion formation in hypercholesterolemic rabbits is delayed by iron depletion. FEBS Lett 459: 218–222, 1999.
Rahman I and MacNee W. Regulation of redox glutathione levels, and gene transcription in lung inflammation: therapeutic approaches. Free Radic Biol Med 28: 1405–1420, 2000.
Roob JM, Khoschsorur G, Tiran A, Horina JH, Holzer H, and Winklhofer-Roob BM. Vitamin E attenuates oxidative stress induced by intravenous iron in patients on hemodialysis. J Am Soc Nephrol 11: 539–549, 2000.
Rooyakkers TM, Stroes ES, Kooistra MP, van Faassen EE, Hider RC, Rabelink TJ, and Marx JJ. Ferric saccharate induces oxygen radical stress, and endothelial dysfunction in vivo. Eur J Clin Invest 32, Suppl 1: 9–16, 2002.
Salahudeen AK, Oliver B, Bower JD, and Oberts LJ. Increase in plasma esterified F-2 isoprostanes following intravenous iron infusion in patients on hemodialysis. Kidney Int 60: 1525–1531, 2001.
Schmidt-Nielsen K. Scaling: Why Is Animal Size So Important Cambridge, UK: Cambridge University Press, 1984.
Sengoelge G, Kletzmayr J, Ferrara I, Perschl A, Horl WH, and Sunder-Plassmann G. Impairment of trans-endothelial leukocyte migration by iron complexes. J Am Soc Nephrol 14: 2639–2644, 2003.
Shah SV. The role of reactive oxygen metabolites in glomerular disease. Annu Rev Physiol 57: 245–262, 1995.
Spittle MA, Hoenich NA, Handelman GJ, Adhikarla R, Homel P, and Levin NW. Oxidative stress, and inflammation in hemodialysis patients. Am J Kidney Dis 38: 1408–1413, 2001.
Stenvinkel P, Heimburger F, Paultre O, Diczfalusy U, Wang T, Berglund L, and Jogestrand T. Strong association between malnutrition, inflammation, and atherosclerosis in chronic renal failure. Kidney Int 55: 1899–1911, 1999.
Sturm B, Goldenberg H, and Scheiber-Mojdehkar B. Transient increase of the labile iron pool in HepG2 cells by intravenous iron preparations. Eur J Biochem 270: 3731–3738, 2003.
Sukamoto H. Iron regulation of hepatic macrophage TNF- expression. Free Radic Biol Med 32: 309–313, 2002.
Swantek JL, Cobb MH, and Geppert TD. Jun N terminal kinase/stress-activated kinase (JNK/SAPK) is required for lipopolysaccharide stimulation of tumor necrosis factor (TNF-) translation: glucocorticoids inhibit TNF- translation by blocking JNK/SAPK. Mol Cell Biol 17: 6274–6282, 1997.
Thorburn A. Death receptor-induced cell killing. Cell Signal 16: 139–144, 2004.
Valderrabano F, Jofre R, and Lopez-Gomez JM. Quality of life in end-stage renal disease patients. Am J Kidney Dis 38: 443–464, 2001.
Varfolomeev EE and Ashkenazi A. Tumor necrosis factor: an apoptosis JuNKie Cell 116: 491–497, 2004.
Villa P, Saccani A, Sica A, and Ghezzi P. Glutathione protects mice from lethal sepsis by limiting inflammation, and potentiating host defense. J Infect Dis 185: 1115–1120, 2002.
Visseren FL, Verkerk MS, van der Bruggen T, van der Bruggen T, Marx JJ, van Asbeck BS, and Diepersloot RJ. Iron chelation, and hydroxyl radical scavenging reduce the inflammatory response of endothelial cells after infection with Chlamydia pneumoniae or influenza A. Eur J Clin Invest 32, Suppl 1: 84–90, 2002.
Weiss G, Meusburger E, Radacher G, Garimorth K, Neyer U, and Mayer G. Effect of iron treatment on circulating cytokine levels in ESRD patients receiving recombinant human erythropoietin. Kidney Int 64: 572–578, 2003.
West GB, Woodruff WH, and Brown JH. Allometric scaling of metabolic rate from molecules, and mitochondria to cells and mammals. Proc Natl Acad Sci USA 99: 2473–2478, 2002.
Xiong S, She H, Takeuchi H, Han B, Engelhardt JF, Barton CH, Zandi E, Giulivi C, and Tsukamoto H. Signaling role of intracellular iron in NF-B activation. J Biol Chem 278: 17646–17654, 2003.
Yeun JY, Levine RA, Mantadilok V, and Kaysen GA. C reactive protein predicts all cause, and cardiovascular mortality in hemodialysis patients. Am J Kidney Dis 35: 469–476, 2000.
Zager RA and Foerder CA. Effects of inorganic iron, and myoglobin on in vitro proximal tubular lipid peroxidation and cytotoxicity. J Clin Invest 89: 989–995, 1992.
Zager RA, Johnson ACM, and Hanson SY. Parenteral iron nephrotoxicity: potential mechanisms, and consequences. Kidney Int 66: 144–156, 2004.
Zager RA, Johnson ACM, and Hanson SY. Parenteral iron therapy exacerbates experimental sepsis. Kidney Int 65: 2108–2112, 2004.
Zager RA, Johnson ACM, Hanson SY, and Wasse H. Parenteral iron formulations: a comparative toxicologic analysis, and mechanisms of cell injury. Am J Kidney Dis 40: 90–103, 2002.
Zager RA, Schimpf BA, Bredl CR, and Gmur DJ. Inorganic iron effects on in vitro hypoxic proximal renal tubular cell injury. J Clin Invest 91: 702–708, 1993.
Zager RA, Shah VO, Shah HV, Zager PG, Johnson AC, and Hanson S. The mevalonate pathway during acute tubular injury: selected determinants, and consequences. Am J Pathol 161: 681–692, 2002.
Zoccali C, Benedetto FA, Mallamaci F, Tripepi G, Fermo I, Foca A, Paroni R, and Malatino LS. Inflammation is associated with carotid atherosclerosis in hemodialysis patients. Creed Investigators. Cardiovascular risk extended evaluation in dialysis patients. J Hypertens 18: 1207–1213, 2000.(Richard A. Zager, A. C. M)