Parenteral Iron Formulations: A Comparative Toxicologic Analysis
`and Mechanisms of Cell Injury
`
`Richard A. Zager, MD, Ali C.M. Johnson, BS, Sherry Y. Hanson, BS, and Haimanot Wasse, MD
`
`● Background: Multiple parenteral iron (Fe) formulations exist for administration to patients with end-stage renal
`disease. Although there are concerns regarding their potential toxicities, no direct in vitro comparisons of these
`agents exist. Thus, the present study contrasted pro-oxidant and cytotoxic potentials of four available Fe
`preparations: Fe dextran (Fe dext), Fe sucrose (Fe sucr), Fe gluconate (Fe gluc), and Fe oligosaccharide (Fe OS).
`Methods: Differing dosages (0.06 to 1 mg/mL) of each compound were added to either (1) isolated mouse proximal
`tubule segments, (2) renal cortical homogenates, or (3) cultured human proximal tubule (HK-2) cells (0.5- to 72-hour
`incubations). Oxidant injury (malondialdehyde generation) and lethal cell injury (percentage of lactate dehydroge-
`nase release; tetrazolium dye uptake) were assessed. Effects of selected antioxidants (glutathione [GSH], catalase,
`dimethylthiourea (DMTU), and sodium benzoate also were assessed. Results: Each test agent induced massive and
`similar degrees of lipid peroxidation. Nevertheless, marked differences in cell death resulted (Fe sucr >> Fe gluc >
`Fe dext ⯝ Fe OS). This relative toxicity profile also was observed in cultured aortic endothelial cells. Catalase,
`DMTU, and sodium benzoate conferred no protection. However, GSH and its constituent amino acid glycine blocked
`Fe sucr–mediated cell death. The latter was mediated by mitochondrial blockade, causing free radical generation
`and a severe adenosine triphosphate depletion state. Conclusions: (1) parenteral Fes are highly potent pro-
`oxidants and capable of inducing tubular and endothelial cell death, (2) markedly different toxicity profiles exist
`among these agents, and (3) GSH can exert protective effects. However, the latter stems from GSH’s glycine
`content, rather than from a direct antioxidant effect.
`© 2002 by the National Kidney Foundation, Inc.
`
`INDEX WORDS: Iron (Fe); lipid peroxidation; human proximal tubule (HK-2) cells; mitochondria; proximal tubules.
`
`PARENTERAL IRON (Fe) formulations have
`
`become mainstays for the treatment of Fe
`deficiency anemia, particularly in patients with
`end-stage renal disease who are on erythropoi-
`etin therapy.1 Although clearly effective in cor-
`recting Fe deficiency, concerns continue to sur-
`face regarding potential toxicities of these agents
`(eg,1-4). Two issues seem particularly relevant in
`this regard: first, the ability of parenteral Fe to
`initiate oxidative reactions, potentially increas-
`ing the risk for atherosclerosis5-8 and direct tissue
`injury9; and second, the risk for anaphylaxis,
`believed to arise from immunologic reactions to
`the carbohydrate moieties of these preparations,
`
`From the Department of Medicine, University of Washing-
`ton; and the Fred Hutchinson Cancer Research Center,
`Seattle, WA.
`Received November 6, 2001; accepted in revised form
`January 8, 2002.
`Supported in part by research grants no. RO1 DK-54200,
`RO-1 DK-38432,and RO1 DK53765 from The National
`Institutes of Health; and Abbott Laboratories, North Chi-
`cago, IL.
`Address reprint requests to Richard A. Zager, MD, Fred
`Hutchinson Cancer Research Center, 1100 Fairview
`Ave N, Rm D2 190, Seattle, WA 98109-1024. E-mail:
`dzager@fhcrc.org
`© 2002 by the National Kidney Foundation, Inc.
`0272-6386/02/4001-0012$35.00/0
`doi:10.1053/ajkd.2002.33917
`
`which serve as Fe envelops or carriers.10-12 The
`latter reactions have spurred the development of
`short-chain carbohydrate-based Fe compounds
`(eg, sucrose complexes or other oligosaccha-
`rides) believed to be less immunogenic. How-
`ever, the impact of these different carbohydrate
`carriers on the ability of Fe to induce oxidative
`injury has not been clearly defined. This is a
`highly relevant issue, given that Fe-mediated
`oxidative stress is critically dependent on its
`carrier (eg,
`increased injury with adenosine
`diphosphate (ADP)13 and decreased injury with
`ferritin or desferrioxamine [DFO]13-17).
`In view of these considerations, the present
`study was undertaken to provide what we believe
`is the first direct comparison of potential Fe
`toxicity induced by three commercially available
`(Fe dextran [Fe dext], Fe gluconate [Fe gluc], Fe
`sucrose [Fe sucr]) and one preclinical parenteral
`(Fe oligosaccharide [Fe OS]) Fe preparations. To
`experimentally address this issue, renal tubular
`cells and renal cortical homogenates were used
`as molecular targets. The rationale for their use is
`that the kidney’s high lipid content makes it an
`excellent substrate to gauge oxidative tissue dam-
`age.13 Furthermore, there is the possibility that
`the administration of parenteral Fe, if cytotoxic,
`might induce renal damage. If so, such a result
`
`90
`
`American Journal of Kidney Diseases, Vol 40, No 1 (July), 2002: pp 90-103
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1090 - Page 1
`
`

`

`PARENTERAL Fe TOXICITY
`
`91
`
`could potentially accelerate residual nephron loss
`in predialysis or dialysis patients.
`
`MATERIALS AND METHODS
`
`Isolated Proximal Tubular Segment Experiments:
`Determination of Fe Cytotoxicity
`
`Male CD-1 mice (25 to 35 g; Charles River Laboratories,
`Wilmington, MA), maintained under routine vivarium condi-
`tions with free food and water access, were used for all
`animal experiments. Cortical proximal tubular segments
`(PTSs) were harvested from these mice by previously de-
`scribed methods.18 In brief, mice were anesthetized using
`pentobarbital (⬃2 mg intraperitoneally), kidneys were imme-
`diately removed through a midline abdominal incision, and
`cortices were recovered by dissection with a razor blade on
`an iced plate. Cortical tissues were minced with a razor
`blade, digested with collagenase, passed through a stainless
`steel sieve, and pelleted by centrifugation (4°C). Viable
`PTSs were recovered by centrifugation through 32% Percoll
`(Pharmacia; Piscataway, NJ). After multiple washings in
`iced buffer, PTSs were suspended (⬃2 to 4 mg of PTS
`protein/mL) in experimentation buffer (NaCl, 100 mmol/L;
`KCl, 2.1 mmol/L; NaHCO3, 25 mmol/L; KH2PO4, 2.4
`mmol/L; MgSO4, 1.2 mmol/L; MgCl2, 1.2 mmol/L; CaCl2,
`1.2 mmol/L; glucose, 5 mmol/L; alanine, 1 mmol/L; Na
`lactate, 4 mmol/L; Na butyrate, 10 mmol/L; and 36-kd
`dextran, 0.6%) and gassed with 95% oxygen/5% carbon
`dioxide (final pH, 7.4). Finally, they were rewarmed to 37°C
`in a heated shaking water bath over 15 minutes; then they
`were ready for experimentation, described next.
`Six PTS preparations were each divided into five equal
`aliquots (1.25 mL of PTS suspension placed into 10-mL
`Erlenmeyer flasks; 2 to 4 mg of tubule protein/mL buffer)
`and incubated for 60 minutes at 37°C, as follows: (1) control
`incubation conditions (buffer only), (2) Fe dext (InFed;
`Schein Pharmaceuticals; Florham Park, NJ), (3): sodium
`ferric gluconate complex in sucrose (Fe gluc; Ferrlecit;
`Schein Pharmaceuticals), (4) Fe OS (Abbott Laboratories,
`North Chicago, IL), and (5) Fe sucr (Venofer; American
`Regent Laboratories; Shirley, NY). Each compound was
`added in a dose of 1 mg/mL of elemental Fe. After complet-
`ing 60-minute incubations, Fe cytotoxicity was assessed by
`determining the percentage of tubular lactate dehydrogenase
`(LDH) release.18 Also, the extent of Fe-mediated lipid peroxi-
`dation, gauged by tubular malondialdehyde (MDA) concen-
`trations, was assessed (thiobarbituric acid method13).
`
`Potential Sucrose Effects on Fe Toxicity in PTSs
`
`Sucrose has been implicated as a potential nephrotoxic
`agent,19-21 possibly by its action as an osmotic agent. Given
`that two of the agents (Fe sucr, Fe gluc) contain sucrose (300
`and 195 mg/mL, respectively), the following experiment
`was undertaken to ascertain whether sucrose could influence
`Fe’s cytotoxic effects. To this end, four sets of PTSs were
`prepared, and each was divided into four aliquots as follows:
`(1) control incubation conditions; (2) incubation with su-
`crose, 14.88 mg/mL (equaling the sucrose concentration in
`the previously described Venofer experiment); (3) 1 mg/mL
`of Fe (either Fe OS [n ⫽ 2] or Fe dext [n ⫽ 2]); and (4) Fe
`
`OS or Fe dext plus 14.88 mg/mL of sucrose (n ⫽ 2 each).
`After completing a 60-minute 37°C incubation, the extent of
`cell injury and lipid peroxidation was assessed as noted
`previously. For statistical analysis, results for the two Fe
`preparations were treated as a single group because no
`discernible differences were apparent between them.
`
`Renal Cortical Homogenate Experiments:
`Determination of Fe-Induced Lipid Peroxidation
`
`Previous observations from this laboratory indicated that
`mitochondrial electron transport is critical to Fe-mediated
`lipid peroxidation in living cells because blockade of site 2
`or 3 of the respiratory chain with chemical inhibitors blocks
`the Fe-driven lipid peroxidative process.22 The following
`experiment was conducted to determine the relative capaci-
`ties of the test Fe preparations to induce lipid peroxidation in
`the absence of intact mitochondrial function. To this end, a
`nonviable homogenate of renal cortex was used as the lipid
`peroxidation target. For each experiment, renal cortex was
`dissected from two mice and homogenized in 3 mL of the
`previously described tubule buffer. Samples of this homoge-
`nate (200 ␮L) were added to 800 ␮L of PTS buffer (PTS
`protein concentration, 3.0 mg/mL), followed by the addition
`of one of the four test Fe compounds. Each compound was
`tested simultaneously using elemental Fe concentrations of
`0.2 and 0.8 mg/mL. Samples were gassed with 95% oxy-
`gen/5% carbon dioxide for 5 minutes, then incubated for 30
`minutes in a rotating 37°C water bath. At the completion of
`incubations, the extent of lipid peroxidation was determined
`by MDA assay. This experiment was repeated in its entirety
`on four separate occasions. MDA concentrations were com-
`pared between groups by analysis of variance, with aftertest-
`ing performed by unpaired Student’s t-test.
`
`Potential Sucrose Effects on Renal Cortical
`Homogenate Lipid Peroxidation
`
`To test whether sucrose might act as an antioxidant,
`potentially altering lipid peroxidation induced by Fe sucr,
`the previous experiment was repeated using non–sucrose-
`containing Fe preparations (Fe dext, Fe OS) with or without
`the equivalent amount of sucrose in the Fe sucr (Venofer)
`preparation (300 mg/mL of stock solution). After 30-minute
`incubations, as noted previously, the extent of MDA genera-
`tion was determined (n ⫽ 4 separate experiments).
`
`Fe Dext Versus FeCl3-Mediated Lipid Peroxidation
`of Cortical Homogenates
`
`Given that carbohydrate moieties are used as shields to
`prevent Fe-mediated oxidative reactions, the following ex-
`periment assessed the relative degree to which this pharma-
`cological strategy is effective, at least as assessed by an in
`vitro experiment. Four sets of fresh renal cortical homoge-
`nates were prepared and incubated for 30 minutes with
`either 0, 50, or 100 ␮g/mL of either Fe dext (InFed) or FeCl3.
`The latter was chosen because each of the parenteral Fe
`preparations contains ferric (Fe3⫹) rather than ferrous Fe.
`After completing the incubations, conducted at 37°C, MDA
`concentrations were determined.
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1090 - Page 2
`
`

`

`92
`
`ZAGER ET AL
`
`HK-2 Proximal Tubule Cell Toxicity Experiments:
`Assessments of Fe Toxicity
`
`HK-2 cells, an immortalized cell line established from
`normal human kidney,23 were maintained under routine
`culture conditions in T-75 flasks with keratinocyte serum-
`free medium, as previously described.23 At near confluence,
`cells in each flask were detached by trypsinization and
`seeded into 24-well cluster plates (four plates per flask).
`Eight hours after seeding, each plate was divided into six
`groups (4 wells/group) as follows: (1) continued control
`incubation, and (2 to 6) incubation with one of the test Fe
`preparations in doses of 1, 0.5, 0.25, 0.126, or 0.06 mg/mL
`of elemental Fe. One plate was used for each Fe preparation.
`After a 16-hour incubation, the extent of cell injury was
`assessed by determining the percentage of LDH release in
`each well, as previously described.24 In addition, total LDH
`per well, an index of cell numbers/cell proliferation,24 was
`used as an additional index of relative Fe toxicity for the
`different Fe preparations. This experiment was repeated on
`three separate occasions.
`
`HK-2 Cell Proximal Tubular Cell Proliferation Using
`the MTT Assay
`
`As a second marker of Fe toxicity, the extent of cellular
`outgrowth/proliferation/viability was determined by the tet-
`razolium dye (MTT) assay method.24 To this end, plates of
`HK-2 cells containing dose titrations of each of the four test
`Fe preparations were established as noted previously. How-
`ever, they were allowed to remain in culture for 72 hours
`after the Fe additions. At the completion of this period,
`viable cell numbers were assessed by the MTT assay, as
`previously performed in this laboratory24 (n ⫽ 4 with each
`test agent).
`
`Mechanisms for Fe Sucr Toxicity: Isolated PTS
`Experiments
`
`Results obtained from these experiments indicate that Fe
`sucr exerted the greatest cytotoxic effect (see Results). Thus,
`it was chosen as the test compound to ascertain potential
`cytotoxic mechanism(s) of these Fe preparations, as follows.
`
`Impact of Mitochondrial Respiration on
`Fe Sucr–Mediated Lipid Peroxidation
`As noted, intact mitochondrial site 2/site 3 electron trans-
`port sites are required for intracellular Fe-mediated oxidant
`stress.22 The following experiment was conducted to ascer-
`tain whether the addition of Fe sucr to the extracellular
`(tubule buffer) space also requires intact electron transport
`to induce its lipid peroxidative effects. To this end, four sets
`of isolated tubules were prepared, and each was divided into
`four equal aliquots and incubated under one of the following
`conditions for 30 minutes: (1) control incubation, (2) with
`100 ␮mol/L of antimycin A (AA; a site 2 respiratory chain
`inhibitor), (3) with 1 mg/mL of elemental Fe (Fe sucr), and
`(4) with Fe sucr plus AA. At the completion of 30-minute
`incubations, the percentage of LDH release and MDA con-
`centrations (per milligram of tubule protein) were deter-
`mined.
`
`To assess whether the effect of AA on renal tubular lipid
`peroxidation was specific for the Fe sucr preparation, its
`impact on both Fe OS and Fe dext was assessed. Four sets of
`isolated tubules were each divided into five aliquots, as
`follows: (1) control incubation, (2) incubation with Fe OS (1
`mg/mL of elemental Fe), (3) incubation with Fe dext (1
`mg/mL of elemental Fe), (4) Fe OS plus AA, and (5) Fe dext
`plus AA. After completing 30-minute incubations, the per-
`centage of LDH release and MDA concentrations were
`assessed.
`
`Confirmation of Fe as the Cytotoxic Moiety of
`the Fe Sucr Preparation
`The following experiment was conducted to prove that the
`elemental Fe in the Fe sucr (Venofer) preparation is the
`cytotoxic moiety. To this end, four sets of tubules were
`prepared, and each was divided into four equal aliquots as
`follows: (1) control incubation, (2) incubation with 2 mmol/L
`of DFO, (3) incubation with 1 mg/mL of elemental Fe
`(Venofer), and (4) DFO plus Venofer. After completing
`30-minute incubations, the percentage of LDH release and
`MDA concentrations were determined.
`
`Trials of Selected Antioxidants in the Setting of
`Fe Sucr Cytotoxicity
`Hydroxyl radical scavengers. Because the hydroxyl radi-
`cal (OH•), generated through the Fenton/Haber-Weiss reac-
`tion, is widely implicated in Fe cytotoxicity,13,25,26 the follow-
`ing experiment was undertaken to ascertain its role in Fe
`sucr (Venofer)-mediated tubular cell death. Four tubule
`preparations were each divided into five equal aliquots as
`follows: (1) control incubation, (2, 3) incubation with 1
`mg/mL of elemental Fe (as Fe sucr), (4) incubation with Fe
`sucr plus 20 mmol/L of dimethylthiourea (DMTU), and (5)
`incubation with Fe sucr plus 20 mmol/L of sodium benzoate.
`In this regard, both DMTU and benzoate are potent hydroxyl
`radical scavengers,13,25,26 and neither exerts an independent
`effect on tubule viability.13 After completing 30-minute
`incubations, cytotoxicity was assessed by determining the
`percentage of LDH release and MDA concentrations.
`Glutathione and catalase. The previous experiment was
`repeated in four PTS preparations, substituting glutathione
`(GSH; 5 mmol/L) or catalase (5,000 U/mL) for the DMTU/
`benzoate additions.13
`Impact of glycine on Fe sucr–mediated tubular cytotoxic-
`ity. GSH may confer cytoprotection either through its
`antioxidant properties or by serving as a source of glycine.27
`The latter protection is more prominently expressed against
`adenosine triphosphate (ATP) depletion–mediated, rather
`than free radical–mediated, tubular cell injury.28 The follow-
`ing experiments were undertaken to explore these issues.
`1. Impact of glycine on Fe sucr toxicity. Isolated tubule
`preparations were divided into 3 equal aliquots and main-
`tained under: (1) control incubation, (2) incubation with Fe
`sucr (1 mg/mL of elemental Fe), or (3) incubation with Fe
`sucr plus 5 mmol/L of glycine (n ⫽ 4 with each treatment).
`At the completion of the 60-minute incubation, the percent-
`age of LDH release and MDA concentrations were deter-
`mined.
`2. Impact of Fe sucr on cellular ATP concentrations. The
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1090 - Page 3
`
`

`

`PARENTERAL Fe TOXICITY
`
`93
`
`previous experiment was repeated exactly as described ex-
`cept that incubations were conducted for 15 minutes. At the
`end of this time, the percentage of LDH release was deter-
`mined, then adenine nucleotides were immediately extracted
`in trichloroacetic acid (6.66%). ATP, ADP, and adenosine
`monophosphate content were determined by high-perfor-
`mance liquid chromatography, as previously described.29,30
`3. Effect of ouabain on Fe sucr–mediated ATP depletion.
`The previous experiment indicated that Fe sucr caused a
`marked tubule ATP depletion state. The following experi-
`ment was undertaken to ascertain whether this was caused
`by a primary disturbance in mitochondrial performance or
`increased ATP consumption (eg, induced by increased so-
`dium potassium adenosine triphosphatase (Na⫹,K⫹-ATPase)
`activity, possibly triggered by plasma membrane lipid peroxi-
`dation with a secondary increase in transmembrane sodium
`leak). To this end, four sets of tubules were treated as
`follows: (1) control incubation for 15 minutes, (2) incuba-
`tion with 1 mg/mL of Fe (Fe sucr) plus 5 mmol/L of glycine
`(to prevent lethal cell injury), and (3) Fe sucr plus glycine
`plus 2 mmol/L of ouabain (to inhibit Na⫹,K⫹-ATPase–
`driven ATP consumption). At the end of 15-minute incuba-
`tions, the percentage of LDH release and adenine nucleotide
`profiles were assessed.
`
`Assessment of Toxicity in Cultured Aortic
`Endothelial Cells
`
`The following experiment was undertaken to assess
`whether the test Fe compounds also exert toxicity in a
`nontubular cell target. Endothelial cells were chosen for this
`purpose because they have direct and immediate contact
`with these compounds after intravenous injection. Bovine
`aortic endothelial cells (passage 13), raised according to a
`previously established method,31 were provided by John
`Harlan, MD, University of Washington, Seattle, WA. They
`were cultured in T-75 flasks in Dulbecco’s Modified Eagle
`Medium (Sigma Chemicals, St Louis, MO) supplemented
`with 10% fetal calf serum (FCS) and containing penicillin
`(100 U/mL) and streptomycin (100 ␮g/mL). On reaching
`confluence, cells were trypsinized and used to seed two
`24-well cluster plates. Either 1 or 2 days after seeding, cells
`
`were challenged with either Fe OS, Fe dextr, Fe sucr, or Fe
`gluc (1 mg/mL of elemental Fe) in 5% FCS (n ⫽ 4 wells per
`plate). The remaining wells (n ⫽ 8 per plate) were main-
`tained in 5% FCS without an Fe challenge. Eighteen hours
`later, viable cell numbers were assessed by the MTT assay,
`as noted previously.
`
`Assessment of In Vivo Lipid Peroxidation With
`Fe Dext Administration
`
`The following experiment was undertaken to confirm that
`the previously obtained in vitro results have, at a minimum,
`some in vivo correlate. To this end, six mice were injected
`subcutaneously with 0.2 mL of Fe dext. Two hours later, mice
`were anesthetized, and a plasma sample was obtained from the
`inferior vena cava. These plasma samples and those obtained
`from six normal mice were assayed for MDA concentrations.
`
`Calculations and Statistics
`All values are presented as mean ⫾ 1 SEM. Individual
`pairs of data were compared by either paired or unpaired
`Student’s t-test, unless stated otherwise. If multiple sets of
`data were compared, the Bonferroni correction was applied.
`Multiple sets of data also were contrasted by analysis of
`variance after testing by Student’s t-test. Significance was
`considered for P less than 0.05.
`
`RESULTS
`
`Fe effects on Isolated PTS Viability
`
`Effects of 1 mg/mL of elemental Fe contained
`in each of the four test compounds are shown in
`Fig 1. Fe OS, Fe gluc (Ferrlicit), and Fe dext
`(InFed) failed to increase the percentage of LDH
`release compared with control incubated tubules
`(Fig 1, left panel). However, Fe sucr (Venofer)
`caused marked cytotoxicity, indicated by approxi-
`mately 50% LDH release (P ⬍ 0.002 compared
`with controls; Fig 1, left panel). Despite this
`
`Fig 1. Effects of the four
`test Fe compounds on
`mouse proximal tubular cell
`injury. (Left) Only Fe sucr in-
`duced cytotoxicity, assessed
`by LDH release. However,
`each of the four test Fe com-
`pounds induced striking lipid
`peroxidation, assessed by
`MDA generation. That the
`amount of lipid peroxidation
`was no greater with Fe sucr
`versus the other test agents
`suggests that lipid peroxida-
`tion was not the mechanism
`by which Fe sucr induced
`tubular cell death. Abbrevia-
`tion: Cont, control.
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1090 - Page 4
`
`

`

`94
`
`ZAGER ET AL
`
`Fig 2. Effects of sucrose addition on Fe-mediated cytotoxicity. Exogenous sucrose was added to either Fe dext
`or Fe OS to reproduce the concentration present in the Fe sucr preparation. (Results with the two Fe preparations
`are combined for presentation because they did not differ.) (Left) Sucrose addition did not render the Fe compounds
`cytotoxic (no increase in LDH release; <13% for all groups) or (right) alter the extent of lipid peroxidation. Thus,
`selective Venofer (Fe sucr) toxicity could not be ascribed to its sucrose content.
`
`marked increase in cell death with Fe sucr com-
`pared with the other Fe compounds, each com-
`pound induced striking and equal degrees of
`lipid peroxidation (Fig 1, right panel). This sug-
`gests that lipid peroxidation was not the mediator
`of Fe sucr–induced cell death.
`
`Effects of Sucrose on Fe Toxicity and Lipid
`Peroxidation in Isolated PTSs
`Given that Fe sucr contains a high concentration
`of sucrose (300 mg/mL in the stock solution), this
`experiment assessed whether sucrose could explain
`the preferential increase in cell death with this Fe
`preparation. However, as shown in Fig 2, this was
`not the case: an equivalent amount of sucrose
`added to either Fe OS or Fe dext did not render
`either of them cytotoxic (⬍13% LDH release;
`results with two compounds combined for presenta-
`tion). Furthermore, sucrose addition did not alter
`the extent of lipid peroxidation induced by the Fe
`compounds (Fig 2, right panel).
`
`Fe-Induced Lipid Peroxidation Using a Cortical
`Homogenate Lipid Target
`Paradoxically, each of the Fe preparations except
`Fe sucr (Venofer) induced significant lipid peroxi-
`dation of renal cortical homogenate at the 0.2-
`mg/mL elemental Fe dose (Fig 3, left panel). (Note
`that the dotted line in Fig 3 represents the upper
`limit of the 95% confidence range for normal MDA
`values.) At the 0.8-mg/mL Fe dose, Fe OS, Fe dext,
`and Fe gluc produced marked and similar lipid
`
`peroxidation (all P ⫽ not significant [NS] versus
`each other). As with the 0.2-mg/mL Fe dose, at the
`0.8-mg/mL Fe dose, Fe sucr still induced signifi-
`cantly less lipid peroxidation compared with the
`other three Fe preparations (*P ⬍ 0.01).
`
`Evaluation of Potential Sucrose Effects on Cortical
`Homogenate Lipid Peroxidation
`As shown in Fig 4, the addition of sucrose did
`not alter the extent of either Fe OS– or Fe
`dext–induced lipid peroxidation. Hence, the rela-
`tive blunting of lipid peroxidation in the previ-
`ously noted Fe sucr (Venofer) experiments can-
`not be explained by Venofer’s sucrose content.
`
`Comparison of Lipid Peroxidation With Fe Dext
`Versus FeCl3
`As shown in Fig 5, relatively small doses of
`FeCl3 (50 and 100 ␮g/mL) induced substantial
`lipid peroxidation, with MDA levels increasing
`three to four times greater than control values.
`Although Fe dext caused less lipid peroxidation
`than FeCl3, the percentage of reduction was only
`25% and 35% with the 50- and 100-␮g/mL Fe
`doses, respectively (Fig 5). Thus, the dextran
`moiety was relatively ineffective in blocking
`Fe’s oxidative effect.
`
`Assessment of Fe Toxicity Using HK-2 Cells
`Percentage of LDH Release
`Fe-mediated percentage of LDH release val-
`ues from cultured HK-2 cells after 16-hour Fe
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1090 - Page 5
`
`

`

`PARENTERAL Fe TOXICITY
`
`95
`
`Fig 3. Fe-mediated lipid
`peroxidation in renal corti-
`cal homogenates. The upper
`limit of the 95% confidence
`band for normal homoge-
`nate MDA concentrations is
`shown by the dotted line.
`(Left) Each Fe compound ex-
`cept Fe sucr caused signifi-
`cant lipid peroxidation at the
`0.2-mg/mL Fe concentration.
`At the 0.8-mg/mL Fe concen-
`tration, each of the four test
`Fe compounds induced sig-
`nificant
`lipid peroxidation
`(mean values > 95% confi-
`dence
`band). However,
`again, Fe sucr induced sig-
`nificantly less (*P < 0.01)
`lipid peroxidation than the
`three other compounds.
`
`exposures are listed in Table 1. Using percentage
`of LDH release as the marker of overt toxicity/
`cell death, only Fe gluc and Fe sucr caused
`cytotoxicity. Fe sucr was substantially more toxic
`than Fe gluc, indicated by an approximately two
`times greater percentage of LDH release at the
`1-mg/mL concentration and because the former,
`but not the latter, caused significant LDH release
`at the 0.12-mg/mL Fe concentration.
`
`Total LDH content
`Table 2 lists total HK-2 cell LDH levels after
`the previously described 16-hour incubations with
`each of the test Fe compounds. Loss of total
`LDH reflects not lethal cell injury per se (eg, as
`does percentage of LDH release), but total num-
`
`ber of cells (serving as a proliferation marker24).
`As seen, a significant decrement in total LDH
`was observed with each test Fe compound (com-
`pared with controls, 0 mg/mL of Fe). The rank
`order of toxicity was Fe sucr ⬎ Fe gluc ⬎ Fe
`OS ⫽ Fe dext. (This is based on both LDH totals
`observed at the highest test Fe dose [1 mg/mL]
`and significant declines in total LDH [denoted by
`* in Table 2] at differing Fe concentrations [Fe
`sucr ⱖ 0.06 mg/mL, Fe gluc ⱖ 0.12 mg/mL, Fe
`dext and Fe OS, only at 1 mg/mL]).
`
`Assessment of Cell Viability/Proliferation With
`the MTT Assay
`Results of the 72-hour Fe incubation/MTT
`assay experiments are listed in Table 3 as the
`
`Fig 4. Potential effect of
`sucrose on Fe-mediated lipid
`peroxidation in renal cortical
`homogenates. The addition
`of sucrose in a dose equal to
`that present in Venofer (Fe
`sucr) did not blunt lipid per-
`oxidation induced by either
`(left) Fe OS or (right) Fe dext.
`Thus, the lesser amount of
`lipid peroxidation induced by
`Venofer (Fe sucr) in the renal
`cortical homogenate experi-
`ment (Fig 3) could not be ex-
`plained by a potential su-
`crose-mediated antioxidant
`effect. (Light colored bars, no
`Fe ⴞ sucrose; dark bars,
`Fe ⴞ sucrose).
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1090 - Page 6
`
`

`

`96
`
`ZAGER ET AL
`
`Fig 5. Comparison of lipid
`peroxidation induced by
`FeCl3 and Fe dext. Both FeCl3
`and Fe dext caused marked
`lipid peroxidation at both test
`concentrations. However,
`the shielded Fe (dext) caused
`only 25% to 35% less lipid
`peroxidation than unshielded
`FeCl3 (indicating an appar-
`ently trivial dextran effect on
`Fe-mediated oxidative stress).
`
`percentage of reduction in cellular MTT uptake.
`The latter was calculated using the control MTT
`uptake value ([Fe value ⫼ control value] ⫻
`100%). Results obtained largely mirrored those
`of the previously described total LDH experi-
`ment, as follows: (1) each compound caused a
`significant decrease in MTT uptake at
`the
`1-mg/mL concentration, and (2) the rank order of
`toxicity was Fe sucr ⬎ Fe gluc ⬎ Fe OS ⫽ Fe
`dext. A statistical comparison of results obtained
`with the 1-mg/mL dosage for each compound is
`shown in Fig 6.
`
`Mechanisms of Fe Sucr Toxicity
`Role of Mitochondrial Respiration in
`Fe Sucr–Mediated Lipid Peroxidation
`the
`As shown in the left panel of Fig 7,
`addition of AA did not affect MDA generation
`over a 30-minute incubation period in the ab-
`sence of Fe sucr. However, in the presence of Fe
`sucr, AA caused an approximately 40% reduc-
`
`tion in MDA concentrations (P ⬍ 0.02). Thus,
`site 2 respiratory chain blockade mitigated Fe
`sucr–mediated lipid peroxidation (indicating a
`partial mitochondrial basis).
`As expected, AA induced substantial lethal
`cell injury, reflected by an approximately 50%
`LDH release (Fig 7, right panel). The combina-
`tion of Fe sucr plus AA caused less than additive
`toxicity: for example, AA plus Fe sucr caused
`58% LDH release. However, the sum of AA-
`mediated LDH release (50%) plus Fe sucr–
`mediated LDH release (32%) was 82%. (Note
`that
`this is consistent with observations pre-
`sented next that Fe sucr–mediated cell death is
`caused in large part by ATP depletion; thus, it is
`not expected to be highly toxic in the presence of
`an already existing AA-induced ATP depletion
`state.)
`In contrast to these results with Fe sucr, AA
`did not limit lipid peroxidation induced by either
`Fe dext or Fe OS (all MDA values for these
`
`Table 1. Percentage of LDH Release From HK-2 Cells After a 16-Hour Exposure to Fe Compounds
`
`Compound
`
`0 mg/mL
`
`0.06 mg/mL
`
`0.12 mg/mL
`
`0.25 mg/mL
`
`0.5 mg/mL
`
`1 mg/mL
`
`Fe dext
`Fe OS
`Fe gluc
`Fe sucr
`
`11 (1)
`11 (1)
`11 (1)
`9 (1)
`
`11 (2)
`12 (1)
`13 (1)
`13 (1)
`
`11 (1)
`11 (1)
`13 (1)
`15 (2)*
`
`11 (1)
`12 (1)
`12 (1)
`18 (2)†
`
`12 (1)
`12 (1)
`16 (1)*
`25 (2)†
`
`12 (1)
`13 (1)
`16 (1)*
`30 (2)†
`
`NOTE. Numbers in parentheses ⫽ 1 SEM.
`*P ⬍ 0.05.
`†P ⬍ 0.001 compared with control cells (0 mg/mL Fe concentration).
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1090 - Page 7
`
`

`

`PARENTERAL Fe TOXICITY
`
`97
`
`Table 2. Total HK-2 Cell LDH Content After 16-Hour Exposures to Fe Compounds
`
`Compound
`
`0 mg/mL
`
`0.06 mg/mL
`
`0.12 mg/mL
`
`0.25 mg/mL
`
`0.5 mg/mL
`
`1 mg/mL
`
`Fe dext
`Fe OS
`Fe gluc
`Fe sucr
`
`21 (3)
`21 (3)
`21 (3)
`21 (3)
`
`20 (3)
`21 (3)
`20 (3)
`19 (3)*
`
`20 (3)
`20 (3)
`19 (3)*
`18 (2)*
`
`20 (3)
`21 (3)
`19 (3)*
`17 (2)*
`
`19 (3)
`20 (3)
`18 (3)*
`16 (2)*
`
`19 (2)*
`19 (2)*
`17 (2)*
`14 (2)*
`
`NOTE. LDH contents expressed in units. Data correspond to that in Table 1.
`*P ⬍ 0.01 versus control cells (ie, 0 mg/mL Fe exposure).
`
`experiments ranged from 22 to 26 nmoles/mg of
`protein regardless of whether AA was present).
`Thus, these results indicate the relative specific-
`ity of AA effects on Fe sucr–driven MDA genera-
`tion. As noted, neither Fe OS nor Fe dext induced
`cell death (controls, 10% ⫾ 1%; Fe OS, 10% ⫾
`1%; Fe dext, 11% ⫾ 1%; all P ⫽ NS). In the
`presence of AA, marked cytotoxicity resulted
`that was unaffected by the presence or absence of
`the concomitant Fe challenge (56% to 58% LDH
`release).
`
`Role of Free Fe in Fe Sucr Cytotoxicity
`To prove that Fe sucr toxicity was Fe depen-
`dent, the ability of an Fe chelator to decrease Fe
`sucr toxicity was assessed. As shown in Fig 8,
`left panel, the addition of DFO was able to
`significantly blunt Fe sucr cytotoxicity, indicat-
`ing the Fe dependence of Fe sucr cytotoxicity.
`However, as shown in Fig 8, right panel, this
`cytoprotection was completely dissociated from
`a change in lipid peroxidation given that no
`DFO-mediated decrement in MDA generation
`was observed.
`
`Role of Hydroxyl Radical in Fe Sucr
`Cytotoxicity
`The Fe sucr challenge induced 48% ⫾ 3%
`LDH release. Neither DMTU nor sodium benzo-
`ate diminished this toxicity (52% ⫾ 1% and 45%
`⫾ 2% LDH release, respectively). Furthermore,
`
`neither agent decreased Fe sucr–mediated MDA
`generation (Fe sucr alone, 25 ⫾ 1; Fe ⫹ DMTU,
`26 ⫾ 1; Fe ⫹ sodium benzoate, 22 ⫾ 1
`nmoles/mg protein). For comparison, control tu-
`bule MDA concentrations were 1.0 ⫾ 0.1
`nmoles/mg protein.
`
`Role of Hydrogen Peroxide in Fe Sucr
`Cytotoxicity
`In stark contrast to the DMTU/benzoate re-
`sults, GSH conferred complete protection against
`Fe sucr–mediated LDH release (Fig 9, l