Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Feb 18.
Published in final edited form as: Regul Toxicol Pharmacol. 2018 May 29;97:17–23. doi: 10.1016/j.yrtph.2018.05.014

In vitro and in vivo DFO-chelatable labile iron release profiles among commercially available intravenous iron nanoparticle formulations

Amy Barton Pai a,*, Manjunath P Pai a, Dan E Meyer b, Brian C Bales b, Victoria E Cotero b, Nan Zheng c, Wenlei Jiang c
PMCID: PMC8855972  NIHMSID: NIHMS1776412  PMID: 29857115

Abstract

Intravenous (IV) iron formulations are complex colloidal suspensions of iron oxide nanoparticles. Small changes in formulation can allow more labile iron to be released after injection causing toxicity. Thus, bioequivalence (BE) evaluation of generic IV iron formulations remains challenging. We evaluated labile iron release in vitro and in vivo using a high performance liquid chromatography chelatable iron assay to develop a relational model to support BE. In vitro labile iron release and in vivo labile iron pharmacokinetics were evaluated for Venofer®, Ferrlecit®, generic sodium ferric gluconate complex, InFeD®, Feraheme® and a pre-clinical formulation GE121333. Labile iron release profiles were studied in vitro in 150mM saline and a biorelevant matrix (rat serum) at 0.952 mgFe/mL. In vivo plasma labile iron concentration-time profiles (t0-240 min) were studied in rats after a 40 mgFe/kg IV dose. In vitro labile iron release in saline was significantly higher compared to rat serum, especially with InFeD®. An in vitro release constant (iKr) was calculated which correlated well with maximal plasma concentrations in the in vivo rat PK model (R2 = 0.711). These data suggest an in vitro to in vivo correlation model of labile iron release kinetics could be applied to BE. Other generic IV iron formulations need to be studied to validate this model.

Keywords: Intravenous iron, Labile iron, Bioequivalence, Pharmacokinetics

1. Introduction

Anemia secondary to iron deficiency is a common clinical condition that impacts patients with stage 3 and 4 chronic kidney disease (CKD), end-stage renal disease (ESRD), cancer, inflammatory bowel disease, post-partum, and post-bariatric surgery. Oral iron replacement therapy is often ineffective in these populations due to poor tolerability and/or limited oral bioavailability. Over the past half-century, a multitude of intravenous iron formulations have been developed to treat iron-deficiency anemia. The highly oxidative properties of iron in its free, labile form require that its rate of iron release from the administered complex be controlled in vivo. Current commercially available intravenous (IV) iron formulations consist of iron oxyhydroxide cores surrounded by carbohydrate shells of various sizes and polysaccharide branch characteristics which varies by product (Danielson., 2004). The particle size of commercially available intravenous iron-carbohydrate complexes range from 5 to 100 nm, and thus all of these agents are nanoparticles or nanomedicines (Danielson., 2004; Jahn et al., 2011). IV iron nanoparticle formulations were approved and entered clinical practice in the late 1950’s, preceding the current era exploring nanomedicine scientific and regulatory frontiers (Duncan and Gaspar, 2011).

Commercially available IV iron agents are now projected to have sales of $3 billion by 2024, in large part due to the increasing number of patients with CKD (Global Industry Analysts, Inc). Patients receiving dialysis treatment, represent a population who receive repeated IV iron doses for long periods of time with typical annual cumulative doses ranging from 2.5 to 5 g of elemental iron (Bailie et al., 2015). Safety concerns with erythropoiesis-stimulating agents coupled with the capitated reimbursement structure for dialysis has driven production and approval of generic IV iron formulations (Molony et al., 2016; Charytan et al., 2015). A plethora of generic IV iron products (iron sucrose similar products) exist in the global market, while only one generic product exists on the United States market (i.e. generic sodium ferric gluconate complex). Reports of lower efficacy and higher adverse event rates with non-US iron sucrose similar formulations have been increasing in the literature (Martin-Malo et al., 2012; Stein et al., 2012; Lee et al., 2013). Specifically, a higher incidence of hypotension has been reported with some non-US iron sucrose similar formulations, including different lots of the same formulation (Martin-Malo et al., 2012; Stein et al., 2012). Association of labile iron release as a fundamental cause of these adverse drug events (excluding immunogenetic reactions) has strong biological plausibility and is supported by translational research evaluating several of the reference listed drug (RLD) products (Lee et al., 2013; Pai et al., 2007, 2011). Collectively, these factors present significant challenges in reproducible manufacturing that can impact advancement of generic bioequivalent intravenous products in the US (Brissot et al., 2012).

A common pharmaceutical science approach to evaluation of a generic formulation includes assessment and benchmarking of physicochemical characteristics (PCC) of the available innovator (i.e. the RLD). The PCC of RLD IV iron-carbohydrate complex formulations have been well-described in the literature (Danielson., 2004; Jahn et al., 2011). These iron-carbohydrate complex formulations differ with regard to molecular weight, carbohydrate shell chemistry, shell and particle diameter and osmolality (Jahn et al., 2011; Tinkle et al., 2014). These iron-carbohydrate formulations are sensitive to pH, temperature and other conditions during the manufacturing process, presenting major technical challenges to reproducible manufacturing (Duncan and Gaspar, 2011). However, it has been shown that complexes of similar molecular weight can be synthesized using multiple different manufacturing procedures, suggesting that the iron complex may be thermodynamically stable (Yang et al., 2010). The variance in coating of IV iron complex formulations is known to impact relative labile iron release. In general, agents with smaller carbohydrate coatings (e.g. iron sucrose) release labile iron at higher rates and extents compared to agents surrounded with larger carbohydrate coatings. (e.g. ferumoxytol) (Pai et al., 2007; Balakrishnan et al., 2009; Shi et al., 2013).

Assessment of both in vitro and in vivo labile iron release profiles from IV iron formulations can augment data provided by PCC to better understand how the disposition of generic complex drugs compares to the RLD. As a consequence, the primary objective of this study was to systematically evaluate labile iron release profiles of six IV iron formulations both in vitro and in vivo. We also sought to use a systematic approach to evaluate the potential for development of an in vitro to in vivo correlation model to support potential generic formulation bioequivalence applications.

2. Materials and methods

2.1. In vitro study design

To evaluate labile iron release in vitro, single lots of commercially available IV iron formulations (Venofer®) (lot 3101), Ferrlecit® (lot D2C250A), generic sodium ferric gluconate complex (NDC 00591-0149-87, Watson Laboratories, Inc, lot 132095.1), InFeD® (lot 13W08A) and Feraheme® (lot AA7895) were incubated in 150mM saline and in a bio-relevant matrix (rat serum). The IV iron formulations were diluted into 150mM saline or rat serum as a biorelevant matrix at an iron concentration of 0.952 mg Fe/mL (17.3 mM Fe). Rat serum was pooled from whole blood drawn from rats who were housed in the vivarium. This concentration was selected to represent the predicted maximum plasma concentration of the study formulations following a 40 mg elemental iron/kg body weight injection in a rat. The 40 mg Fe/kg dose was selected to limit the need for dilution of the agents, which could impact formulation stability and labile iron release profiles prior to injection, and to yield a detectable free-iron signal in vivo as determined by pilot study 1. To expand the classes of agents evaluated beyond carbohydrate-based coating chemistries, we additionally tested GEH121333 (GE Global Research, Niskayuna, NY, lot 120401), which is a research-stage iron oxide nanoparticle formulation with a PEG-based coating (Shi et al., 2013). For sample preparation, ex vivo serum samples were warmed to room temperature, vortexed, and centrifuged prior to addition of reagents. The ex vivo samples were incubated for 3 h at ambient temperature prior the addition of desferroximine to measure chelatable labile iron by high-performance liquid chromatography (Tesoro et al., 2005). (See Labile Iron Measurement below.)

2.2. In vivo study design

Given that no prior studies were available that measured labile iron release from IV iron formulations in rats, we utilized an adaptive study design to identify the optimal sampling method to quantify labile iron across different IV iron formulations. The in vivo study was designed in an iterative three stage process to ensure that animal model experiments (n = 10 per formulation) were refined to maximize the information gained while minimizing the number of animals sacrificed for the experiment. All animal studies were conducted under an Animal Care and Use protocol approved by the General Electric (GE) Global Research Institutional Animal Care and Use Committee. Male Sprague-Dawley rats, 200–250 g in body weight, were purchased from Charles River Laboratories with a pre-placed jugular vein catheter. Rats were acclimated upon arrival for at least 2 days prior to study. Prior to study initiation, the indwelling catheters were flushed and lock solution replaced within the first day of arrival and at least once every 72 h to maintain patency. All blood samples (nominally 125 μL) were withdrawn via the jugular catheter immediately before and at various time points after injection (described below) of the test substance via the tail vein. After the collection of each blood sample, a volume of sterile saline equal to the blood volume withdrawn administered as a flush through the jugular catheter.

2.2.1. Optimal sample collection time strategy

Stage 1.

The first stage was designed to identify the dose of IV iron that would lead to a measurable profile of labile iron. Two rats were tested at each dose level (10-fold) of 4 mg Fe/kg and 40 mg Fe/kg, respectively, with a single IV iron formulation (Ferrlecit®). Initial sampling time points (n = 8) of 0 (pre-dose), 5, 15, 30, 60, 120, 240, and 360 min were selected to characterize the concentration-time profile of labile iron after intravenous dose administration of a 4 mg Fe/kg and 40 mg Fe/kg dose of Ferrlecit®.

Stage 2.

The second stage was designed to gain initial pharmacokinetic data of all six IV iron formulations in 3 animals using the defined dose of 40 mg Fe/kg selected based on Stage 1 results. The sampling time-points used in Stage 1 were refined to include a 90 and 150 min sampling time point. The sampling time-points were re-evaluated for each IV iron formulation after analysis of concentration-time data generated from this second stage.

Stage 3.

The third stage was designed to obtain the final pharmacokinetic data using the defined dose for all six IV iron formulations with the remaining animals with single doses of 40 mg Fe/kg. Based on the data from Stage 2, additional model-based predictions of the expected concentrations were performed to identify additional sampling time points. A total of 5 animals (3 from Stage 2 and an additional 2 animals) followed the sampling scheme noted above. The remaining 5 animals followed the modified sampling schema (0, 5, 15, 30, 60, 90, 120, and 150 min). This optimized sampling schema permitted acquisition of informative concentration-time data from 10 sampling time points spread through two sets of 5 animals.

2.3. Labile iron measurement

Quantification of labile iron concentrations was performed for the ex vivo samples by a validated HPLC-based chelatable iron assay using desferoximine (DFO) as the chelation agent (Tesoro et al., 2005). This assay was selected based on our previous work that demonstrated notable interference of detectable signal with other published assays in vitro. The HPLC-based chelatable iron assay produced robust results across in vitro and in vivo experiments (Pai et al., 2017). Chelatable iron was detected following chelation with 20 mM deferoxamine in Tris-HCl, pH 5.5 and quantified by integration of the colored ferioxamine peak following HPLC analysis. The HPLC analysis was performed using a Shimadzu HPLC system equipped with 2 LC-20AD pumps fitted with a DGU-20A3 solvent degasser, a SIL-20AHT autosampler, a SPD-M20A photo diode array detector, and a CBM-20A controller bus module. The HPLC system was controlled by a PC running Lab Solutions software v. 5.41.240. The HPLC analysis was run using a Waters Xterra C18 4.6 × 50 mm 5 μm column and 10 mM Tris-HCl, pH = 5.5 and acetonitrile as the mobile phase at a 1 mL/min flow rate using the time gradient from 0 to 7.1 min, the iron-DFO chelate eluted at 3.9 min and was detected at 427 nm.

2.4. Pharmacokinetic analysis

Exploratory data analyses were performed to evaluate the labile iron concentration-time profile. The labile iron concentration-time information was modeled using a parametric population pharmacokinetic analysis approach (ADAPT 5, BMSR, Los Angeles, CA). The initial structural model included two-compartments with bolus input with first-order linear rate of release (Kr), central compartment volume of distribution (Vc) with elimination from this compartment parameterized as clearance (CL). The bolus input was based on administration of the stage 1 defined-dose for each IV iron formulation that represented the total elemental iron dose in mg (transformed into millimoles). No published methods exist to simultaneously measure the bound iron within the parent formulation and the unbound iron (labile iron) from a serum sample with a background of endogenous iron. As a result, the fraction (F) of unbound iron in the administered dose was indeterminate and so the parameters Vc and CL represent apparent parameters of Vc/F and CL/F, respectively.

The base model was fit using the maximum likelihood expectation maximization algorithm (MLEM) using importance sampling (n = 2000 samples/iteration). Alternate models of higher complexity (inclusion of another compartment) were tested. The parameters CL/F and Vc/F were also tested with or without inter-individual variability to ascertain their influence on the estimation of Kr. The rationale for this approach was to ascertain whether the assumption of fixing Vc/F and CL/F as physiologically similar for labile iron across formulation would help to improve estimation of Kr that may be distinct by IV iron formulation. Discrimination between models was based on the Akaike Information Criterion (AIC). One-way Analysis of Variance (ANOVA) was used to compare the individual Kr estimates by IV iron formulation.

3. Results

3.1. In vitro labile iron release

HPLC analysis of the chelatable iron present in IV iron formulations revealed an increase in Fe-DFO peak area as a function of time on several replicate analyses. A similar increase in signal attributed to Fe-DFO was not observed after the 180 min incubation at ambient temperature when non-complexed FeCl3 was used as the iron source. Based on this observation, we infer that the continuing increase in Fe-DFO peak area is attributable to continued release of chelatable Fe from the IV iron formulations. Fig. 1 illustrates the linear fit and 95% confidence interval plots of the change in the labile iron concentration over time after a 180 min incubation period. The labile iron concentrations were consistently higher and more variable when formulations were diluted in 150 mM of saline compared to rat serum. These profiles permitted us to compute the slope for each formulation using the data generated from the rat serum as a representative in vitro rate of release constant (iKr). The predicted mean (95% confidence interval) in vitro maximum concentrations (iCmax) of labile Fe reflective of the intercept value in rat serum along with the iKr values are included in Table 1 by formulation.

Fig. 1.

Fig. 1.

Open in a new tab

Linear fit [95% confidence interval] plots of the natural logarithm labile iron-concentration (μM) over time after the 180 min incubation step in rat serum and saline, respectively.

Table 1.

Summary of in vitro rate of release constant (iKr) and the maximum labile iron concentration (iCmax) by formulation (estimate [95% confidence interval]).

Formulation iKr (h−1) iCmax (μM)
Venofer 0.0369 [0.0326, 0.0411] 138 [115, 161]
Ferrlecit 0.0318 [0.0256, 0.0381] 595 [572, 618]
SFG complex 0.0282 [0.0237, 0.0327] 411 [392, 430]
INFeD 0.0277 [0.0231, 0.0323] 155 [144, 166]
Feraheme 0.0263 [0.0229, 0.0296] 278 [254, 302]
GEH121333 0.0442 [0.0362, 0.0521] 174 [166, 182]
Open in a new tab

3.2. In vivo labile iron pharmacokinetic data

The defined optimal sampling schema generated measurable labile iron concentration-time information from all the animals utilized in the experimental procedure at a dosage of 40 mg Fe/kg. This dosage was not associated with any changes in vital sign parameters in the rat. As shown in Fig. 2, the concentration-time profile was highest at the earlier time points, as is expected with an intravenous bolus injection, and then declined in a monoexponential (Venofer®, Ferrlecit®, SFGC) or linear (InFed®, Feraheme®, GEH121333) manner. The labile iron concentration-time profile of GEH121333 was visibly different than other formulations tested in vivo and a summary of the central tendencies of exposure parameters are provided in Table 2. The Cmax and AUC of GEH121333 was higher than that observed with the commercially available formulations. Similar labile iron AUCs were measured among the commercially available formulations but an approximately 2–3 fold longer half-life was observed with the larger molecular weight compounds (Feraheme®, InFeD®) compared to the smaller molecular weight compounds (Venofer®, Ferrlecit®, SFGC). The labile concentration-time data were explained optimally by a 1-compartment model with no improvement in the model with the use of higher-order models. The model predictions were closely correlated to observed labile iron concentrations (Fig. 3). Table 3 includes a summary of the population pharmacokinetic system parameters. As shown, the relative standard error (%RSE) was below 30% for the parameter estimates of CL/F and V/F and close to 100% for Kr. Given the expectation that the distribution and clearance of labile iron should in theory not be related to the formulation, we attempted to fix these parameters to improve our population estimate of Kr. However, fixing the CL/F and V/F in a stepwise manner (Model 2 and 3) did not improve (increased AIC) the model fit and increased the %RSE of Kr implying intrinsic differences in this parameter by formulation. As a consequence, the population model was utilized to generate empiric Bayesian estimates of the system parameters by formulation and provided in Table 4. The Kr values generated from the larger molecular weight compounds (Feraheme®, InFeD®) were not significantly different from GEH121333 but were significantly (p < 0.003) lower than the smaller molecular weight compounds (Venofer®, Ferrlecit®, and SFGC®). No significant differences in system parameters were noted between the smaller molecular weight compounds.

Fig. 2.

Fig. 2.

Open in a new tab

Scatter and fitted plots of labile iron concentrations over time post administration by formulation.

Table 2.

Summary of the mean (% coefficient of variation) in vivo labile iron concentration-time parameters by formulation (n = 10 animals).

Formulation Cmax (μM) AUC0-4h (μM•h) Half-life (h)
Venofer 206 (18.8) 403 (20.1) 2.15 (37.5)
Ferrlecit 216 (38.0) 432 (31.2) 2.72 (50.6)
SFGC 227 (21.8) 438 (15.5) 3.20 (72.7)
InFeD 156 (35.0) 495 (29.3) 6.00 (50.1)
Feraheme 102 (38.3) 338 (35.1) 9.41 (48.6)
GEH121333 413 (21.6) 1311 (20.6) 10.4 (33.3)
Open in a new tab

Cmax, maximum concentration, AUC0-4h, area under the curve from time 0–4 h.

Fig. 3.

Fig. 3.

Open in a new tab

Scatter and fitted observed labile iron concentrations over individual model predicted labile iron concentrations illustrated by formulation.

Table 3.

Summary of system parameter estimates (%RSE) and model discrimination function, Akaike information criterion (AIC) generated as apparent clearance (CL/F) and apparent volume (V/F) were modeled with or without inter-individual variability.

Model CL/F (mL/min) V/F (mL) Kr (h−1) AIC
1 2.97 (19.3) 1020 (9.80) 1.09 (98.9) 4376
2 Fixed 1330 (11.0) 1.42 (232) 4859
3 Fixed Fixed 1.53 (343) 5158
Open in a new tab

Table 4.

Summary of mean (%CV) system parameter estimates by empiric Bayesian analyses for each formulation (N = 10 in each group).

Formulation CL/F (mL/min) V/F (mL) Kr (min−1)
Venofer 6.49 (39.9) 1041 (17.1) 2.22 (24.1)
Ferrlecit 5.43 (40.3) 1075 (33.4) 2.02 (33.9)
SFGC 4.86 (36.6) 987 (20.2) 2.07 (41.8)
InFeD 3.41 (47.0) 1245 (19.7) 1.07 (30.2)
Feraheme 3.59 (69.7) 1972 (35.6) 0.701 (66.1)
GEH121333 0.774 (46.2) 506 (21.1) 0.972 (28.7)
Open in a new tab

3.3. In vitro to In vivo correlation

The in vitro data lead to the generation of two parameters (iKr and iCmax) that were plotted against system parameters generated in vivo. No significant correlations were identified when these parameters were tested against system parameters. However, a significant correlation was observed between the iKr values and Cmax values observed in vivo. Fig. 4 illustrates the good correlation (R2 = 0.711) observed between the natural logarithm transformed Cmax values to the iKr values by formulation. The mean (standard error) is 56.5(18.0) and 3.46 (0.596) for the slope (p = 0.035) and intercept (p = 0.004), respectively. The R2 value is 0.711 or a correlation coefficient (R) of 0.84. We believe that it is reasonable to ascribe this relationship as “good”. However, this relationship could be driven by the high Cmax values.

Fig. 4.

Fig. 4.

Open in a new tab

Scatter and linear fit of observed mean natural logarithm maximum concentration in in vivo to in vitro rate of release constant (Kr).

4. Discussion

The widespread use of IV iron therapy particularly to manage anemia in patients with ESRD, has spurred the development of numerous formulations. Currently available IV iron-carbohydrate formulations are synthesized using co-precipitation techniques to coat iron hydroxide cores with carbohydrate moieties (Molday and MacKenzie, 1982). Using this approach, alterations in carbohydrate chemistries or manufacturing conditions in the co-precipitation reaction can lead to marked differences in physicochemical properties. This can include particle core size and carbohydrate content differences in the outer coating, both of which may affect labile iron release (Barrow et al., 2016). Use of a single contemporary physicochemical characterization technique (e.g. dynamic light scattering) may be unable to sufficiently distinguish subtle differences among iron-carbohydrate nanoparticles (Bhattacharjee, 2016). Alternatively, advanced characterization methods could also identify subtle differences that are not clinically relevant. Thus, methods that are predictive of in vivo performance could potential explain observations on non-US formulations that are reported to behave differently in patients (Lee et al., 2013; Elford et al., 2013).

To date, animal studies comparing non-US iron sucrose formulations have evaluated total serum iron concentrations and transferrin saturation (TSAT, serum iron/total iron binding capacity x 100) and found values to be higher in animals receiving the non-US iron sucrose similar preparations versus the RLD Venofer® (Toblli et al., 2011, 2012, 2015; Elford et al., 2013). However, TSAT is not a direct measurement of the reactive labile iron species and does not adequately represent the potential for deleterious redox reactions. Although TSAT values greater than 100% do strongly infer the presence of labile iron, we and others, have shown that labile iron is present at TSAT values less than 100%, limiting the utility of this parameter for the evaluation of potential toxicity related to labile iron (Pai et al., 2007; Slotki and Cabantchik, 2015).

This study systematically evaluated labile iron release profiles from commercially available IV iron formulations as an alternate approach to the current paradigm. Importantly, the series of test formulations we studied included the only generic product approved in the United States (SFGC) and the associated RLD Ferrlecit®. We found that diluting the IV iron formulations in 150 mM of saline resulted in significantly higher chelatable labile iron concentrations. This was particularly notable with InFeD, a large molecular weight formulation, that has been shown to release much lower amounts of labile iron when administered to both healthy patients and end-stage renal disease patients on dialysis (Pai et al., 2011). This is likely due to the dilution effect of saline which may compromise the stability of the smaller molecular weight formulations due to reduction in ionic shielding and electrostatic repulsion which maintains formulation stability (Bhattacharjee, 2016). Higher absolute measurements and variability were also observed in labile iron values with Ferrlecit® diluted in saline compared to the generic formulation (Fig. 1). Measurement of dialyzable iron detected inductively coupled plasma mass spectroscopy was shown to be higher when diluted in saline versus water [Jahn et al., 2011]. In contrast, variability in labile iron concentration measurements in the in vitro experiments were markedly lower with use of rat serum as a biologic dilution matrix. The rank order of labile iron release in rat serum followed that which has been shown in previous studies including those conducted in human subjects receiving FDA approved formulations immediately post injection (Ferrlecit > Venofer > InFeD) (Jahn et al., 2007; Pai et al., 2007). The comparative labile iron release profiles in saline versus rat serum suggest that any evaluation of these formulations for application to an IVIVC model should use a validated biorelevant matrix.

Our in vivo studies identified optimal sample collection times for labile iron concentration time profiling and confirmed that labile iron can be measured by an HPLC-DFO assay immediately after injection and up to 6 h. The labile iron concentration time profile was best modeled using a 1-compartment model with no improvement in the fit using multi-compartmental models. It should be noted that the labile iron concentration profile likely largely represents the portion of iron that is directly released from the formulation post-injection and prior to biodistribution, and delayed release of additional labile iron from deposited reservoirs is additionally possible. The labile iron exposure measured by AUC0-4h was similar among the formulations that were consistent with administration of the same dosage. In contrast, the elimination half-life was 2–3 fold longer with the larger molecular weight formulations consistent with clinical data. The pharmacokinetic profiles for the release constant (Kr) clearly were also clustered into two groups: Venofer®, Ferrlecit®, generic sodium ferric gluconate complex (Watson Laboratories, Inc) and InFeD®, Feraheme®, GEH121333. This finding substantiates that molecular weight of the overall complex drives rate of labile iron release and is consistent with the concentration-time profiles in both healthy human subjects and those with various stages of chronic kidney disease (Pai et al., 2007, Pai et al., 2011). Two IV iron formulations (ferric carboxymaltose and iron maltoside) that have larger molecular weights that were not tested in our model have been shown in other studies to have the lowest labile iron release (Neiser et al., 2015). The in vivo Kr values were observed to be similar for Ferrlecit® and SFGC, despite apparent lower in vitro labile iron concentrations for SFGC than in Ferrlecit®, especially in saline. Despite this, a point to point in vitro to in vivo pharmacokinetic parameter correlation could not be identified. Establishing a point to point correlation between in vitro to in vivo pharmacokinetic parameters is complicated by the multifaceted in vivo fates of the IV iron formulations (e.g., rate of clearance from the blood, tissue reservoirs, etc.). However, the early labile release associated with a formulation measurable in vitro should in theory correlate with a measurable in vivo pharmacokinetic parameter. This theoretical framework is supported by our observation of a good correlation between the in vitro rate constant of labile iron release (iKr) to the in vivo maximum labile iron concentration (Cmax). Given the limited number of generic to RLD pairs tested, certainty surrounding the linearity of this relationship requires further study through evaluation of additional formulations. Overall, this study suggests that in vitro labile iron release studies (with or without biorelevant matrix) are more sensitive to formulation differences than in vivo studies. Therefore, in vitro studies to compare potential labile iron release from the formulations may be more discriminative and practical.

The inter-formulation differences generated from this study may also be applicable to the evaluation of intra-formulation differences, including batch-to-batch manufacturing variability. Other groups have demonstrated differences in molecular weight, titratable alkalinity kinetics of degradation between lots of the same non-US iron sucrose similar formulation (Lee et al., 2013; Toblli et al., 2012). Toblli et al. characterized the physicochemical characteristics of Venofer® and compared these to several non-US iron sucrose similar products available and in clinical use in Europe and Asia (Toblli et al., 2011, 2012, 2015). Notably, only one product in these comparative analyses complied with United States Pharmacopeia (USP) criteria. Differences in one or more of the following criteria such as pH, titratable alkalinity and turbidity point were observed in all these non-US iron sucrose similar products evaluated. Four of the seven non-US iron sucrose similar products (57%) evaluated in one study (Toblli et al., 2015) had markedly higher molecular weights measured by gel permeation chromatography. Also, animal studies using the 40 mg Fe/kg single IV dose of iron sucrose similar products have shown higher tissue concentrations of pro-inflammatory cytokines, higher intracellular antioxidant enzyme activity, adverse effects on the basic metabolic profiles (elevated liver function tests) and kidney dysfunction (elevated serum creatinine and proteinuria) (Toblli et al., 2011, 2012, 2015). It has been hypothesized that labile iron is principally involved with these observed deleterious effects by generating reactive oxygen species via the Fenton-Haber Weiss reaction. However, labile iron release profiles have not been assessed in any published studies to date comparing Venofer® with non-US iron sucrose and ferric carboxymaltose similar products (Toblli et al., 2011, 2012, 2015). The methods developed in the current study could be applied to evaluate lot-to-lot differences and to test the correlation of labile iron concentrations to biomarkers of oxidative stress.

The findings of our work should be appreciated in the context of several important limitations. We evaluated a single dose of the IV iron formulations and thus the impact of multiple doses cannot be extrapolated from this work. Although the dose selected for these studies was higher than those given clinically, there was no dilution of the formulations required and therefore possible confounding effects of dilution on stability were eliminated. As plasma concentrations of labile iron were measured these measurements may not reflect tissue specific values that may arise from biodistribution differences between formulations. The use of rat serum as a biological matrix to better align with our in vivo testing, to evaluate formulation differences in vitro could be improved further through development of an artificial matrix that mimics human serum in order to be more broadly translatable by other laboratories. Despite these limitations, the methodologies developed through this study serve as a basis towards improving the evaluation of IV iron formulations.

5. Conclusion

This study represents the first data that characterizes labile iron release profiles from complex IV iron formulations both in vitro and in vivo. Good correlations were observed between in vitro Kr and in vivo Cmax values which suggests an in vitro to in vivo correlation model of labile iron release kinetics is possible to develop for potential application to BE. Understanding in vitro and in vivo labile iron release profiles among RLD and generic candidate IV iron formulations provides additional data to improve the safety of these widely used agents.

Conflict of interest/disclosure

Funding for this manuscript was made possible, in part, by the Food and Drug Administration through grant 1U01FD004889-01. Views expressed in this publication do not necessarily reflect the official policies of the Department of Health and Human Services, nor does any mention of trade names, commercial practices, or organization imply endorsement by the United States Government.

Footnotes

Transparency document

Transparency document related to this article can be found online at http://dx.doi.org/10.1016/j.yrtph.2018.05.014.

References

  1. Bailie GR, Larkina M, Goodkin DA, Li Y, Pisoni RL, Bieber B, Mason N, Tong L, Locatelli F, Marshall MR, Inaba M, Robinson BM, 2015. Data from the Dialysis Outcomes and Practice Patterns Study validate an association between high intravenous iron doses and mortality. Kidney Int. 87 (1), 162–168. [DOI] [PubMed] [Google Scholar]
  2. Balakrishnan VS, Rao M, Kausz AT, Brenner L, Pereira BJ, Frigo TB, Lewis JM, 2009. Physicochemical properties of ferumoxytol, a new intravenous iron preparation. Eur. J. Clin. Invest 39, 489–496 2009. [DOI] [PubMed] [Google Scholar]
  3. Barrow M, Taylor A, García Carrión J, Mandal P, Park BK, Poptani H, Murray P, Rosseinsky MJ, Adams DJ, 2016. Co-precipitation of DEAE-dextran coated SPIONs: how synthesis conditions affect particle properties, stem cell labelling and MR contrast. Contrast Media Mol. Imaging 11 (5), 362–370. [DOI] [PubMed] [Google Scholar]
  4. Bhattacharjee S, 2016. DLS and zeta potential - what they are and what they are not? J. Contr. Release 235, 337–351. [DOI] [PubMed] [Google Scholar]
  5. Brissot P, Ropert M, Le Lan C, Loréal O, 2012. Non-transferrin bound iron: a key role in iron overload and iron toxicity. Biochim. Biophys. Acta 1820 (3), 403–410. [DOI] [PubMed] [Google Scholar]
  6. Charytan DM, Pai AB, Chan CT, Coyne DW, Hung AM, Kovesdy C, et al. 2015. Considerations and challenges in defining optimal iron utilization in hemodialysis. J. Am. Soc. Nephrol 26 (6), 1238–1247. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Danielson BG, 2004. Structure, chemistry, and pharmacokinetics of intravenous iron agents. J. Am. Soc. Nephrol 15 (Suppl. 2), S93–S98. [DOI] [PubMed] [Google Scholar]
  8. Duncan R, Gaspar R, 2011. Nanomedicine(s) under the microscope. Mol. Pharm 8 (6), 2101–2141. [DOI] [PubMed] [Google Scholar]
  9. Elford P, Bouchard J, Jaillet L, Pearson N, Rogue A, Sabadie C, Forster R, 2013. Biodistribution and predictive hepatic gene expression of intravenous iron sucrose. J. Pharmacol. Toxicol. Meth 68 (3), 374–383. [DOI] [PubMed] [Google Scholar]
  10. Jahn MR, Mrestani Y, Langguth P, Neubert RHH, 2007. CE characterization of potential toxic labile iron in colloidal parenteral iron formulations using off-capillary and on-capillary complexation with EDTA. Electrophorsis 28, 2424–2429. [DOI] [PubMed] [Google Scholar]
  11. Jahn MR, Andreasen HS, Futterer S, Nawroth T, Schünemann V, Kolb U, Hofmeister W, Muñoz M, Bock K, Meldal M, Langguth P, 2011. A comparative study of the physicochemical properties of iron isomaltoside 1000 (Monofer®), a new intravenous iron preparation and its clinical implications. Eur. J. Pharm. Biopharm 78, 480–491. [DOI] [PubMed] [Google Scholar]
  12. Lee ES, Park BR, Kim JS, Choi GY, Lee JJ, Lee IS, 2013. Comparison of adverse event profile of intravenous iron sucrose and iron sucrose similar in postpartum and gynecologic operative patients. Curr. Med. Res. Opin 29 (2), 141–147. [DOI] [PubMed] [Google Scholar]
  13. Martin-Malo A, Merino A, Carracedo J, Alvarez-Lara MA, Ojeda R, Soriano S, Crespo R, Ramirez R, Aljama P, 2012. Effects of intravenous iron on mononuclear cells during the haemodialysis session. Nephrol. Dial. Transplant 27 (6), 2465–2471. [DOI] [PubMed] [Google Scholar]
  14. Molday RS, MacKenzie D, 1982. Immunospecific ferromagnetic iron-dextran reagents for the labeling and magnetic separation of cells. J. Immunol. Meth 52 (3), 353–367. [DOI] [PubMed] [Google Scholar]
  15. Molony JT, Monda KL, Li S, Beaubrun AC, Gilbertson DT, Bradbury BD, Collins AJ, 2016. Effects of epoetin alfa titration practices, implemented after changes to product labeling, on hemoglobin levels, transfusion use, and hospitalization rates. Am. J. Kidney Dis 68 (2), 266–276. [DOI] [PubMed] [Google Scholar]
  16. Neiser S, Rentsch D, Dippon U, Kappler A, Weidler PG, Göttlicher J, Steininger R, Wilhelm M, Braitsch M, Funk F, Philipp E, Burckhardt S, 2015. Aug. Physico-chemical properties of the new generation IV iron preparations ferumoxytol, iron isomaltoside 1000 and ferric carboxymaltose. Biometals 28 (4), 615–635. [DOI] [PubMed] [Google Scholar]
  17. Pai AB, Boyd AV, McQuade CR, Harford A, Norenberg JP, Zager PG, 2007. Comparison of oxidative stress markers after intravenous administration of iron dextran, sodium ferric gluconate, and sucrose in patients undergoing hemodialysis. Pharmacotherapy 27 (3), 343–350. [DOI] [PubMed] [Google Scholar]
  18. Pai AB, McQuade C, Olp J, Hicks P, Conner T, 2011. Non-transferrin bound iron (NTBI), cytokine activation and intracellular reactive oxygen species generation in hemodialysis (HD) patients receiving iron dextran (ID) or iron sucrose (IS). Biometals 24, 603–613. [DOI] [PubMed] [Google Scholar]
  19. Pai AB, Meyer DE, Bales BC, Cotero VE, Pai MP, Zheng N, Jiang W, 2017. Performance of redox active and chelatable iron assays to determine labile iron release from intravenous iron formulations. Clin Transl Sci. 10 (3), 194–200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Shi Q, Pisani LJ, Lee YK, Messing S, Ansari C, Bhaumik S, Lowery L, Lee BD, Meyer DE, Daldrup-Link HE, 2013. Evaluation of the novel USPIO GEH121333 for MR imaging of cancer immune responses. Contrast Media Mol. Imaging 8, 281–288. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Slotki I, Cabantchik ZI, 2015. The labile side of iron supplementation in CKD. J. Am. Soc. Nephrol 26 (11), 2612–2619. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Stein J, Dignass A, Chow KU, 2012. Clinical case reports raise doubts about the therapeutic equivalence of an iron sucrose similar preparation compared with iron sucrose originator. Curr. Med. Res. Opin 28 (2), 241–243. [DOI] [PubMed] [Google Scholar]
  23. Tesoro A, Novakovic J, Thiessen JJ, Spino M, 2005. Validated HPLC assay for iron determination in biological matrices based on ferrioxamine formation. J Chromatogr Analyt Technol Biomed Life Sci 823, 77–83. [DOI] [PubMed] [Google Scholar]
  24. Tinkle S, McNeil SE, Mühlebach S, Bawa R, Borchard G, Barenholz YC, Tamarkin L, Desai N, 2014. Nanomedicines: addressing the scientific and regulatory gap. Ann. N. Y. Acad. Sci 1313, 35–56. [DOI] [PubMed] [Google Scholar]
  25. Toblli JE, Cao G, Oliveri L, Angerosa M, 2011. Assessment of the extent of oxidative stress induced by intravenous ferumoxytol, ferric carboxymaltose, iron sucrose and iron dextran in a nonclinical model. Arzneimittelforschung 61 (7), 399–410. [DOI] [PubMed] [Google Scholar]
  26. Toblli JE, Cao G, Oliveri L, Angerosa M, 2012. Comparison of oxidative stress and inflammation induced by different intravenous iron sucrose similar preparations in a rat model. Inflamm. Allergy - Drug Targets 11 (1), 66–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Toblli JE, Cao G, Angerosa M, 2015. The induction of oxidative/nitrosative stress, inflammation, and apoptosis by a ferric carboxymaltose copy compared to iron sucrose in a non-clinical model. J. Clin. Diagn. Res 9 (12) FF08–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Yang Y, Shah RB, Faustino PJ, Raw A, Yu LX, Khan MA, 2010. Thermodynamic stability assessment of a colloidal iron drug product: sodium ferric gluconate. J Pharm Sci 99, 142–153 2010. [DOI] [PubMed] [Google Scholar]

RESOURCES