Parenteral iron therapy exacerbates experimental sepsis Rapid Communication

Kidney International, Vol. 65 (2004), pp. 2108–2112

Parenteral iron therapy exacerbates experimental sepsis Rapid Communication RICHARD A. ZAGER, ALI C.M. JOHNSON, and SHERRY Y. HANSON Department of Medicine, the University of Washington, and the Fred Hutchinson Cancer Research Center, Seattle, Washington

Parenteral iron therapy exacerbates experimental sepsis. Background. Catalytic iron can potentiate systemic inflammation via its pro-oxidant effects. This raises the possibility that parenteral iron administration might exacerbate a concomitant septic state. This study sought to experimentally test this hypothesis. Methods. Male CD-1 mice were subjected to experimental sepsis via intraperitoneal injection of heat-killed Escherichia coli ± concomitant intravenous iron sucrose (Venofer; 2 mg). Nonseptic mice ± iron therapy served as controls. Plasma tumor necrosis factor-a (TNF-a) levels were assessed 2 hours postinjections (serving as an inflammatory marker). Oxidative stress was gauged in heart or kidney tissue (at either 4 or 24 hours) by heme oxygenase-1 (HO-1) mRNA or protein levels. Overall sepsis severity was assessed by morbidity/mortality rates (at 24 hours). Results. Iron alone or sepsis alone each induced oxidant stress in heart and kidney (HO-1 mRNA/protein increases). When iron and E. coli were coadministered, additive or synergistic HO-1 mRNA/protein increments resulted. Iron injection alone only slightly raised TNF-a levels (from 0 to 2.3 pg/mL; P = 0.01). However, iron approximately doubled the TNF-a increments which arose from the septic state (1400 → 2600 pg/mL). Neither sepsis alone, nor iron alone, induced any mortality and no mice became moribund (0/24 mice). However, when iron + sepsis were combined, ∼60% of mice either died (5/12) or developed a moribund (2/12) state (P = 0.005). Conclusion. Parenteral iron administration can induce systemic oxidative stress and modest TNF-a release. However, when iron is given during experimental sepsis, profound increases in both processes, and ∼60% mortality, result. Given that renal failure patients have decreased antioxidant defenses and intermittently develop bacteremia, the potential for parenteral iron therapy to exacerbate clinical sepsis needs to be addressed.

Systemic inflammation is believed to be an important determinant of morbidity and mortality in patients with end-stage renal disease (ESRD). This concept is

Key words: iron sucrose, TNF-a, heme oxygenase 1, acute renal failure. Received for publication February 4, 2004 Accepted for publication March 2, 2004
C

2004 by the International Society of Nephrology

underscored by observations that serum C-reactive protein (CRP) levels, a surrogate marker of systemic inflammation, directly correlate with dialysis-associated complications (e.g., adverse cardiovascular events [1–5]). Oxidative stress, which can be driven by catalytic (free) iron, can promote inflammation via a number of potential pathways. For example, redox-sensitive transcription factors, such as nuclear factor-kappaB (NF-jB), cFos, c-Jun, activated protein-1 (AP-1), and hypoxia-inducible factor-1a (HIF-1a), are up-regulated by oxidative states [6–9]. Once activated, these molecules can recruit a variety of “downstream” inflammatory mediators, such as tumor necrosis factor-a (TNF-a), cytokine receptors, and adhesion molecules [10]. That antioxidants can attenuate NF-jB and AP-1 expression, and thus, suppress inflammation, supports this pathophysiologic link [11–15]. Parenteral iron formulations are routinely administered to hemodialysis patients to help correct anemia. However, currently employed iron preparations can each exert potent pro-oxidant effects [16–22]. Given these considerations, it seems feasible that their administration might predispose to life-threatening acute systemic inflammation, most notably as induced by sepsis. This study was undertaken to test this hypothesis. METHODS Sepsis model For this study, an experimental sepsis model was employed which involves the intraperitoneal injection of nonviable (boiled) Escherichia coli. This model induces the manifestations of sepsis, while avoiding the confounding variables which may result from in vivo bacterial replication [23–25]. Boiling does not denature endotoxin, which remains intact in the bacterial cell wall. Thus, the administration of heat-killed organisms should more accurately simulate a bacterial-induced endotoxin challenge than would soluble, purified, endotoxin injection. Male CD-1 mice (25 to 35 g) (Charles River Laboratories, Wilmington, MA, USA) were used for all experiments. They were placed in restrainers and divided into one

2108

Zager et al: Iron sucrose worsens sepsis

of four treatment groups: (1) Controls: 100 lL tail vein injection of 0.9% saline (sham for IV iron treatment); (2) IV Fe injection: 2 mg elemental Fe3+ , given in the form of Fe sucrose (Venofer; American Regent, Shirley, NY, USA); (3) Sepsis syndrome: IP injection of 2 mL of a heat killed E. coli slurry, containing ∼1 × 109 organisms per mL of saline [23]; additionally, 100 lL of saline was injected via the tail vein as a sham for IV iron treatment (see group 4); and (4) Sepsis + Fe injection: IV Fe sucrose followed immediately by IP E. coli injection. The mice were then returned to their cages and given free food and water access. Assessments were made at 2, 4, or 24 hours later, as described below. Survival, renal cortical heme oxygenase-1 (HO-1) protein levels: 24-hour assessments Mice subjected to each of the above treatments (N = 12 per each group) were evaluated using one of the following criteria: (1) survival; (2) survival, but in an obvious moribund state (no spontaneous movement, decreased body temperature to touch, decreased respiration); or (3) death. All surviving animals were deeply anesthetized with pentobarbital (3 to 4 mg intraperitoneally), followed by a midline laparotomy, terminal phlebotomy [from the inferior vena cava (IVC)], and bilateral renal resection. Renal cortical protein extracts were probed by Western blot for HO-1 protein [26] as a marker of oxidant tissue stress [27, 28]. Renal insufficiency was quantified by blood urea nitrogen (BUN) and plasma creatinine concentrations. Renal cortical and cardiac oxidant stress: 4-hour assessments The above mouse protocols were repeated (N = 6 animals per group), followed 4 hours later by induction of pentobarbital anesthesia. The kidneys and heart were removed through abdominal and thoracic incisions, respectively. Total RNA was extracted from renal cortical and cardiac apex tissue samples, followed by reverse transcription-polymerase chain reaction (RT-PCR) for quantifying HO-1 mRNA and glyceraldehyde-3phosphate dehydrogenase (GAPDH) mRNA levels, as previously described [26]. The ratio of HO-1 mRNA/ GAPDH mRNA, assessed by densitometry, served as a semiquantitative marker of oxidative tissue stress. Systemic inflammatory response: TNF-a levels at 2 hours posttreatment The four above-described protocols were repeated (N = 5 mice per group). Two hours later, the mice were anesthetized and subjected to terminal phlebotomy from

2109

the IVC. The severity of the systemic inflammatory response was gauged by plasma TNF-a concentrations [by enzyme-linked immunosorbent assay (ELISA)] (R&D Systems, Minneapolis, MN, USA). Statistics All values are presented as means ± 1 SEM. Statistical comparisons were performed by unpaired Student t test, unless stated otherwise. Significance was judged by a P value of <0.05. RESULTS 24-hour analyses Survival rates. All of the mice which received either E. coli alone (N = 12) or intravenous iron alone (N = 12) survived the protocols and none in either group appeared moribund. In contrast, in the iron + sepsis group, five mice died and two were moribund. Thus, 7/12 of mice in the iron + sepsis group were either moribund or dead vs. 0/12 in the sepsis only group (P = 0.005 by Fisher’s exact test). Renal function. Neither iron alone nor sepsis alone raised plasma creatinine concentrations above control levels. However, when these two challenges were combined, significant plasma creatinine elevations resulted (Fig. 1A). Sepsis raised BUN concentrations to 55 ± 10 mg/dL (P < 0.02 vs. control values) (Fig. 1B). Sepsis + iron raised BUNs to 81 ± 13 mg/dL. While higher than sepsis alone, this degree of BUN elevation did not achieve statistical significance [Note: The BUN and creatinine levels could not be determined for the five mice which died from the combined sepsis + iron protocol. Given that these animals were obviously the sickest, the exclusion of their BUN/creatinine values most likely caused an underestimate of the severity of renal insufficiency in the iron + sepsis group]. HO-1 Western blotting. Fe sucrose alone caused a modest increase in HO-1 protein expression (Fig. 2, lane 3) (P < 0.02 vs. controls). Sepsis alone did not significantly increase HO-1 levels (lane 4). However, when iron and sepsis were combined, a striking increase in HO1 expression resulted (P < 0.001 vs. iron alone; consistent with synergistic oxidative stress). To assess the relative degree of HO-1 protein induction by the iron + sepsis protocol, it was contrasted with that observed in a renal cortical protein extract prepared from a mouse 24 hours postinduction of severe, glycerol- induced, rhabdomyolysis acute renal failure (BUN 212 mg/dL), an insult which is known to profoundly increase HO-1 protein levels [27, 28]. The level of HO-1 protein observed post glycerol was comparable to that seen following the sepsis + iron challenge (lane 5). This confirms the severity of HO-1 induction in the iron + sepsis group.

2110

Zager et al: Iron sucrose worsens sepsis

A

B

<0.01

0.5

125 NS

0.45 105

0.35

85 BUN, mg/dl

Creatinine, mg/dl

0.4

0.3 0.25 0.2

<0.02

65

45

0.15 0.1

25 0.05 0

5 C

Fe

No Fe

Fe

Sepsis

C

Fe

No Fe

Fe

Sepsis

Fig. 1. Plasma creatinine and blood urea nitrogen (BUN) concentrations 24 hours following iron sucrose (Fe) injection, induction of sepsis syndrome, or the combination thereof. (A) Neither iron nor sepsis alone induced significant plasma creatinine elevations, compared to the controls. Conversely, combined iron + sepsis caused an approximate doubling of the plasma creatinine concentration. (B) Sepsis alone, but not iron alone, caused a significant increase in BUN concentrations. Combination iron + sepsis was associated with higher BUNs than either treatment alone, but this difference did not achieve statistical significance [as noted in the Results section, five of 12 mice in the iron + sepsis group died prior to blood sampling; this likely caused an underestimate of the severity of BUN and creatinine elevations in this group].

4-hour analyses HO-1 mRNAs. Iron alone and sepsis alone each caused a significant increase in HO-1 mRNA in both renal cortex and cardiac apex (Table 1). When sepsis + iron were combined, significantly greater HO-1 mRNA increases were observed in both tissues, compared to values observed with either iron or sepsis alone. TNF-a assessments at 2 hours after iron/sepsis protocols TNF-a could not be detected in plasma from control animals. Iron injection caused modest TNF-a increments (mean, 2.29 pg/mL) (see Table 1). Sepsis caused striking TNF-a elevations (mean, 1446 pg/mL). The degree of this sepsis-induced TNF-a increase was approximately doubled by concomitant iron treatment (to 2600 pg/mL) (Table 1). DISCUSSION This brief report provides experimental support for the concept that intravenous iron administration can acutely trigger oxidative stress, and that by so doing, it can exacerbate a superimposed, sepsis-triggered, severe systemic inflammatory event. These principles are indicated by the

following pieces of evidence. First, within 4 hours of intravenous iron injection, HO-1 mRNA increments were noted in both myocardium and renal cortex, consistent with multiorgan-based oxidative stress [27, 28]. Second, at 24 hours, renal cortical HO-1 protein increments resulted, indicating the functional significance of the HO-1 mRNA induction. Third, with concomitant E. coli administration, iron induced either additive or synergistic HO-1 mRNA/protein increments. Finally, although intravenous iron injection induced seemingly trivial plasma TNF-a increments, it approximately doubled the TNF-a elevations which resulted from the experimental septic state. By itself, E. coli “sepsis” caused no mortality, with all animals surviving in a relatively healthy state. However, with iron coadministration, E. coli “sepsis” induced an approximate 50% to 60% mortality rate. Clearly then, the above evidence of additive/synergistic oxidative stress and cytokine (TNF-a) release, evinced by the combined sepsis + iron challenge, had marked biologic relevance as assessed at the whole animal level. Given that a suprapharmacologic iron dosage was used in these experiments, clinical relevance cannot be ascribed to the present observations. However, neither should this possibility simply be dismissed. In this regard, it is noteworthy that ESRD patients have

Zager et al: Iron sucrose worsens sepsis

500

2111

**

450 400

Density units

350 300 250 200

*

150 100 50

Sepsis + Fe

C

Fe

Sepsis

Gly

Table 1. Heme oxygenase (HO-1) mRNA and tumor necrosis factor-a (TNF-a) assessments following the induction of the experimental protocols End point

Controls

Iron sucrose

HO-1 mRNA 0.64 ± 0.4 1.34 ± 0.02a (kidney) HO-1 mRNA 0.30 ± 0.01 0.68 ± 0.07a (heart) Plasma TNF-a 0.00 2.29 ± 0.56c pg/mL

Sepsis

Sepsis + iron

1.08 ± 0.04a

1.73 ± 0.15b

0.51 ± 0.05a

0.83 ± 0.07b

1446 ± 161c

2600 ± 484d

HO-1 mRNAs in heart and kidney were assessed 4 hours posttreatment. All three treatment groups significantly differed from the controls (P < 0.01); The sepsis + iron sucrose group had significantly higher renal cortical and cardiac HO-1 mRNA values than either the iron alone (P < 0.05) or the sepsis alone (P < 0.001) group. TNF-a levels were assessed 2 hours after iron or Escherichia coli injections; c P < 0.01 by Wilcoxon rank sum test (controls 0 vs. iron treatment); d Sepsis vs. sepsis + iron, P < 0.05. a

b

diminished antioxidant defenses, presumably sensitizing them to pro-oxidative states [29–32]. Furthermore, following iron injection, clinical oxidative stress is known to result [17–22]. Iron is also a growth factor for many bacteria [33], enhancing in vivo proliferation. Parenteral iron formulations can also impair neutrophil migration [34], impairing host defenses. Thus, these considerations, along with the current findings, suggest multiple pathways by which intravenous iron therapy might potentiate sep-

C

Fig. 2. Heme oxygenase 1 (HO-1) in renal cortex, as assessed by Western blotting (densitometry readings). Samples were obtained 24 hours postinduction of the iron (Fe), sepsis, or iron + sepsis protocols, and the results were compared to control tissues (C). Iron caused a significant increase in HO-1 expression, compared to controls (∗ P < 0.02). Sepsis alone did not cause obvious HO-1 induction. However, sepsis + iron caused an approximate 3 × increase in HO-1, compared to values observed with iron treatment alone (∗∗ P < 0.001). Representative blots are presented under each bar. As a comparison for the iron + sepsis result, a Western blot of renal cortex obtained 24 hours postinduction of glycerol (gly) induced-rhabdomyolysis nephrotoxicity (BUN, 212 mg/dL) is presented. The postglycerol result and the iron + sepsis result were comparable (underscoring the severity of the HO-1 induction seen in the iron + sepsis group).

sis, and the “downstream” inflammatory consequences thereof. ACKNOWLDGMENT This work was supported by research grants from the National Institutes of Health (RO-1 DK54200 and RO-1 DK38432). Reprint requests to Richard A. Zager, M.D., Fred Hutchinson Cancer Research Center, 1100 Fairview Ave. N; Room D2-190, Seattle, WA 98109. E-mail: [email protected]

REFERENCES 1. YEUN JY, LEVINE RA, MANTADILOK V, KAYSEN GA: C reactive protein predicts all cause and cardiovascular mortality in hemodialysis patients. Am J Kid Dis 35:469–476, 2000 2. OWEN WF, LOWRIE EG: C reactive protein as an outcome predictor for maintenance hemodialysis patients. Kidney Int 54:627–636, 1998 3. ZOCCALI C, BENEDETTO FA, MALLAMACI F, et al: Inflammation is associated with carotid atherosclerosis in hemodialysis patients. Creed Investigators. Cardio-vascular Risk Extended Evaluation in Dialysis Patients. J Hyperten 18:1207–1213, 2000 4. ISEKI K, TOZAWA M, YOSHI S, FUKIYAMA K: Serum C-reactive protein (CRP) and risk of death in chronic dialysis patients. Nephrol Dial Transplant 14:1956–1960, 1999 5. STENVINKEL P, HEIMBURGER O, PAULTRE F, et al: Strong association between malnutrition, inflammation, and atherosclerosis in chronic renal failure. Kidney Int 55:1899–1911, 1999 6. RAHMAN I, MACNEE W: Regulation of redox glutathione levels and gene transcription in lung inflammation: therapeutic approaches. Free Rad Biol Med 28:1405–1420, 2000

2112

Zager et al: Iron sucrose worsens sepsis

7. HADDAD JJ: Antioxidant and prooxidant mechanisms in the regulation of redox(y)-sensitive transcription factors. Cell Sig 14:879–897, 2002 8. XIONG S, SHE H, TAKEUCHI H, et al: Signaling role of intracellular iron in NF-kappa B activation. J Biol Chem 278:17646–17654, 2003 9. SUKAMOTO H: Iron regulation of hepatic macrophage TNF alpha expression. Free Radical Biol Med 32:309–313, 2002 10. YOUNG JL, LIBBY P, SCHONBECK U: Cytokines in the pathogenesis of atherosclerosis. Thromb Haemostasis 88:554–567, 2002 11. ESPOSITO K, NAPPO F, MARFELLA R, et al: Inflammatory cytokine concentrations are acutely increased by hyperglycemia in humans: Role of oxidative stress. Circulation 106:2067–2072, 2002 12. VILLA P, SACCANI A, SICA A, GHEZZI P: Glutathione protects mice from lethal sepsis by limiting inflammation and potentiating host defense. J Infect Dis 185:1115–1120, 2002 13. RAHMAN I, MACNEE W: Regulation of redox glutathione levels and gene transcription in lung inflammation: therapeutic approaches. Free Rad Biol Med 28:1405–1420, 2000 14. BLACKWELL TS, BLACKWELL TR, HOLDEN EP, et al: In vivo antioxidant treatment suppresses nuclear factor-kappa B activation and neutrophilic lung inflammation. J Immunol 157:1630–1637, 1996 15. VISSEREN FL, VERKERK MS, VAN DER BRUGGEN T, et al: 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 16. ZAGER RA, JOHNSON ACM, HANSON SY, WASSE H: Parenteral iron formulations: A comparative toxicologic analysis and mechanisms of cell injury. Am J Kidney Dis 40:90–103, 2002 17. SALAHUDEEN AK, OLIVER B, BOWER JD, OBERTS LJ: Increase in plasma esterified F-2 isoprostanes following intravenous iron infusion in patients on hemodialysis. Kidney Int 60:1525–1531, 2001 18. ROOB JM, KHOSCHSORUR G, TIRAN A, et al: Vitamin E attenuates oxidative stress induced by intravenous iron in patients on hemodialysis. J Am Soc Nephrol 8:475–486, 1997 19. LIM PS, WEI YH, YU YL, KHO B: Enhanced oxidative stress in haemodialysis patients receiving intravenous iron therapy. Nephrol Dial Transplant 14:2680–2687, 1999 20. CAVDAR C, TEMIZ A, YENICERIOGLU Y, et al: The effects of intravenous iron treatment on oxidant stress and erythrocyte deformability in hemodialysis patients. Scand J Urol Nephrol 37:77–82, 2000

21. ROOYAKKERS TM, STROES ES, KOOISTRA MP, et al: Ferric saccharate induces oxygen radical stress and endothelial dysfunction in vivo. Eur J Clin Invest 32(Suppl 1): 9–16, 2002 22. AGARWAL R, VASAVADA N, SACHS NG, et al: Oxidative stress and renal injury with intravenous iron in patients with chronic kidney disease. Kidney Int (in press) 23. ZAGER RA, PRIOR RB: Gentamicin and Gram-negative bacteremia: A synergism for the development of experimental nephrotoxic acute renal failure. J Clin Invest 78:196–204, 1986 24. ZAGER RA: Escherichia coli endotoxin injections potentiate experimental ischemic renal injury. Am J Physiol 251:F988–F994, 1986 25. ZAGER RA, JOHNSON ACM, HANSON SY: Sepsis syndrome stimulates proximal tubule cholesterol synthesis and suppresses the SR-B1 cholesterol transporter. Kidney Int 63:123–133, 2003 26. ZAGER RA, JOHNSON ACM, HANSON SY: Parenteral iron mediated nephrotoxicity: Potential mechanisms and consequences. Kidney Int 65:2108–2112, 2004 27. NATH KA, BALLA G, VERCELLOTTI GM, et al: Induction of heme oxygenase is a rapid, protective response in rhabdomyolysis in the rat. J Clin Invest 90:267–270, 1992 28. NATH KA, VERCELLOTTI GM, GRANDE JP, et al: Heme protein- induced chronic renal inflammation: Suppressive effect of induced heme oxygenase-1. Kidney Int 59:106–117, 2001 29. ROXBOROUGH HE, MERCER C, MCMASTER D, et al: Plasma glutathione peroxidase activity is reduced in hemodialysis patients. Nephron 84:278–281, 2000 30. DE CAVANAGH EM, FERDER L, CARRASQUEDO F, et al: Higher levels of antioxidant defenses in enalapril-treated versus non-enalapriltreated hemodialysis patients. Am J Kid Dis 34:445–455, 1999 31. GALLI F, VARGA Z, BALLA J, et al: Vitamin D, lipid profile, and peroxidation in hemodialysis patients. Kidney Int 59(Suppl 78):S148– S154, 2001 32. BOAZ M, SMETANA S, WEINSTEIN T, et al: Secondary prevention with antioxidants of cardiovascular disease in end stage renal disease (SPACE): Randomized placebo-controlled trial. Lancet 356:1213– 1218, 2000 33. SENGOELGE G, KLETZMAYR J, FERRARA I, et al: Impairment of trans-endothelial leukocyte migration by iron complexes. J Am Soc Nephrol 14:2639–2644, 2002 34. BULLEN J, GRIFFITHS E, ROGERS H, WARD G: Sepsis: The critical role of iron. Microbes Infect 2:409–415, 2000