The AAPS Journal, Vol. 19, No. 5, September 2017 ( # 2017)
`DOI: 10.1208/s12248-017-0126-0
`
`Review Article
`Theme: Nanotechnology in Complex Drug Products: Learning from the Past, Preparing for the Future
`Guest Editors: Katherine Tyner, Sau (Larry) Lee, and Marc Wolfgang
`
`Physicochemical Characterization of Iron Carbohydrate Colloid Drug Products
`
`Peng Zou,1,2 Katherine Tyner,1 Andre Raw,1 and Sau Lee1
`
`Received 7 April 2017; accepted 13 July 2017; published online 31 July 2017
`
`Iron carbohydrate colloid drug products are intravenously administered to
`Abstract.
`patients with chronic kidney disease for the treatment of
`iron deficiency anemia.
`Physicochemical characterization of iron colloids is critical to establish pharmaceutical
`equivalence between an innovator iron colloid product and generic version. The purpose of
`this review is to summarize literature-reported techniques for physicochemical characteriza-
`tion of iron carbohydrate colloid drug products. The mechanisms, reported testing results,
`and common technical pitfalls for individual characterization test are discussed. A better
`understanding of the physicochemical characterization techniques will facilitate generic iron
`carbohydrate colloid product development, accelerate products to market, and ensure iron
`carbohydrate colloid product quality.
`KEY WORDS: ferumoxytol; iron colloid; iron dextran; iron sucrose; sodium ferric gluconate.
`
`1 Center for Drug Evaluation and Research, US Food and Drug
`Administration, 10903 New Hampshire Ave, Silver Spring, Mary-
`land 20993, USA.
`2 To whom correspondence should be addressed. (e-mail:
`peng.zou@fda.hhs.gov)
`
`Abbreviations: AAS Atomic absorption spectroscopy, AFM Atomic
`force microscopy, AUC Analytical ultracentrifugation, BDI
`Bleomycin-detectable iron, DLS Dynamic light scattering, DPP
`Differential pulse polarography, DSC Differential scanning calorim-
`etry, EDX Energy-dispersive X-ray, EMR Electron magnetic
`resonance, EPR Electron paramagnetic resonance, ESA Electroki-
`netic sonic amplitude, ESR Electron spin resonance, EXAFS
`Extended X-ray absorption fine structure, FT-IR Fourier transform
`infrared spectroscopy, GPC Gel permeation chromatography,
`HPLC High-performance liquid chromatography, ICP-MS Induc-
`tively coupled plasma mass spectrometry, MDA Malondialdehyde,
`Maap Apparent molecular weight, Mn Number-average molecular
`weight, Mw Weight-average molecular weight, MPS Mononuclear
`phagocyte system, NEXAFS Near-edge X-ray absorption fine
`structure, NMR Nuclear magnetic resonance, NPP Normal pulse
`polarography, NTA Nitrilotriacetate, NTBI Nontransferrin-bound
`iron, OGD Office of Generic Drugs, SLS Static light scattering,
`SQUID Superconducting quantum interference device, STEM
`Scanning transmission electron microscope, TBA Thiobarbituric
`acid, TBI Transferrin-bound iron, TEM Transmission electron
`microscopy, TEM/NBED Transmission electron microscopy/nano
`beam electron diffraction, TEM/SAED Transmission electron
`microscopy/selected area electron diffraction, TGA Thermal gravi-
`metric analysis, USP United States Pharmacopeia, UV/Vis Ultravi-
`olet-visible spectroscopy, VSM Vibrating sample magnetometer,
`XAS X-ray absorption spectroscopy, XANES X-ray absorption
`near-edge structure, XRD X-ray diffraction
`
`INTRODUCTION
`
`Iron colloid drug products are intravenously adminis-
`tered as iron replacement therapies for the treatment of iron
`deficiency anemia in patients with chronic kidney disease
`receiving hemodialysis and supplemental epoetin therapy (1).
`After intravenous administration,
`iron colloids enter the
`bloodstream and are processed by the mononuclear phago-
`cyte system (MPS) (2). Once internalized by macrophages in
`the liver, spleen, and bone marrow, iron colloids are delivered
`to lysosomes and iron ions are released from the colloids and
`become part of the intracellular iron pool which is available
`for uses in biological processes. If
`iron is not needed
`immediately, the cell stores it in the form of ferritin or
`hemosiderin. If iron is needed elsewhere in the body, MPS
`will release iron from the cell and the extracellular protein
`transferrin will bind iron ions to form transferrin-bound iron
`(TBI) and deliver iron to where they are needed. If the serum
`transferrin is saturated with iron, the released iron weakly
`bind to serum components such as albumin to form
`nontransferrin-bound iron (NTBI) which causes oxidative
`stress and toxicity when taken up by the liver and heart (3).
`Therefore, the MPS uptake, in vivo stability, and iron release
`rate of
`iron colloids can dramatically impact
`their
`biodistribution, efficacy, and toxicity.
`Iron colloids are composed of an iron oxyhydroxide or
`iron oxide core and a complex carbohydrate coat with an
`average particle size ranging from 8 to 24 nm (4). The
`carbohydrate shell stabilizes and protects the iron core from
`hydrolysis, precipitation, and polymerization until the deliv-
`ery of iron colloids to the MPS. The carbohydrate shell also
`reduces the direct release of iron into the bloodstream before
`
`1359
`
`1550-7416/17/0500-1359/0 # 2017 American Association of Pharmaceutical Scientists
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1092 - Page 1
`
`

`

`1360
`
`NDA
`
`021135
`017807
`010787
`040024
`017441
`
`020955
`
`022180
`203565
`
`Table I. Parental Iron Colloid Product NDAs Approved by the US FDA
`
`Zou et al.
`
`Trade name
`
`Venofer®
`Proferdex®
`Imferon®
`Dexferrum
`INFeD®
`(Cosmofer® outside
`of the USA)
`Ferrlecit®
`
`Feraheme®
`Injectafer®
`(Ferinject® outside of
`North America)
`
`Generic name
`
`Iron sucrose
`Iron dextran
`Iron dextran
`Iron dextran
`Iron dextran
`
`Current sponsor
`
`Approved date
`
`Luitpold
`New River
`Fisons
`Luitpold
`Watson Labs
`
`November 06, 2000
`March 26, 1981
`April 25, 1957
`February 23, 1996
`April 29, 1974
`
`Sodium ferric gluconate
`complex in sucrose
`Ferumoxytol
`Ferric carboxymaltose
`
`Sanofi Aventis
`
`February 18, 1999
`
`AMAG Pharmas Inc.
`Luitpold in the USA and
`Vifor Pharma outside
`of North America
`
`June 30, 2009
`July 25, 2013
`
`NDA new drug application
`
`the iron colloids are delivered to the MPS and, hence,
`prevents the formation of NTBI and oxidative toxicity. As a
`result, the iron colloid products coated with various carbohy-
`in vivo
`drate shells exhibit differences in MPS uptake,
`stability, iron release profiles, and the content of labile and
`low molecular weight iron species (4).
`The intravenously administered iron colloid drug prod-
`ucts approved by the US FDA for the treatment of iron
`deficiency anemia are summarized in Table I. Differences
`exist among the formulations regarding the size of carbohy-
`drate group and iron colloids, surface property, stability of the
`iron-carbohydrate complex, and the rate at which iron is
`released from iron colloids. Among these approved iron
`colloid products,
`iron sucrose injection and iron dextran
`injection have individual United States Pharmacopeia (USP)
`monographs.
`The FDA has published product-specific bioequivalence
`draft guidances for the generic version of
`iron sucrose
`injection (5), ferumoxytol
`injection (6) and sodium ferric
`gluconate in sucrose injection (7), ferric carboxymaltose
`injection (8), and iron dextran injection (9). In the draft
`guidances, the applicants are recommended to conduct a
`comparative clinical bioequivalence study between the inno-
`vator product and generic version using total iron in serum
`and transferrin-bound iron in serum as pharmacokinetic
`endpoints. In addition, physicochemical characterization
`studies are recommended to demonstrate the same physico-
`chemical properties of iron colloids between the innovator
`product and generic version. The physicochemical properties
`which may affect the safety and efficacy of iron colloid drug
`products are summarized in Table II and divided into three
`categories:
`the whole nanoparticle properties,
`iron core
`properties, and carbohydrate shell properties. These physico-
`chemical properties are likely correlated with the pharmaco-
`kinetics and tissue distribution of iron colloids and may also
`impact in vivo stability and iron release kinetics.
`The European Medicines Agency (EMA) expresses
`different opinions on the data required to demonstrate
`bioequivalence of intravenous iron colloid products (10). In
`addition to comparative physicochemical characterization and
`plasma pharmacokinetic study in humans, EMA recommends
`
`nonclinical studies to compare tissue distribution and toxicity
`of generic and innovator iron colloid drugs in animal models.
`The tissue distribution study includes the assessment of iron
`distribution in at least three biological compartments: plasma
`and red blood cells, reticulo-endothelial system (RES) (liver
`and spleen), and pharmacological/toxicological target tissues
`(bone marrow, kidney, liver, lung, or heart) in animal models.
`In case that minor differences are observed in physicochem-
`ical characterization, nonclinical or human pharmacokinetic
`studies, a therapeutic equivalence study is recommended to
`address their impact on efficacy and safety.
`Although the critical physicochemical properties have
`been recommended in the FDA product-specific draft guid-
`ances, the specific techniques for iron colloid characterization
`have not been systemically reviewed. The availability of
`reliable characterization techniques is critical for demonstrat-
`ing the same physicochemical properties of two iron colloid
`products. To fill the knowledge gap on iron colloid charac-
`terization, we have summarized the physicochemical tests that
`have been reported in the literature to characterize iron
`colloid drug products (Table II). This review provides an
`overview of currently available iron colloid characterization
`techniques and a summary of recent advances in physico-
`chemical characterization of iron colloids, which may facili-
`tate the development of iron colloid drug products and assure
`drug product quality.
`
`CHARACTERIZATION OF PHYSICOCHEMICAL
`PROPERTIES OF THE WHOLE PARTICLES
`
`After an iron colloid enters systemic circulation via
`intravenous infusion,
`the particle is tagged by plasma
`proteins for recognition (opsonization) and processed by
`the MPS system. After
`internalized by marcrophages
`residing in the liver and spleen,
`iron colloids accumulate
`in lysosomes and release free irons. The tissue distribution,
`iron release kinetics, efficacy, and safety of iron colloids are
`affected by a number of physicochemical properties of the
`whole particles including chemical composition of
`iron
`colloids, molecular weight distribution, particle size and
`distribution, content of
`low molecular weight
`iron and
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1092 - Page 2
`
`

`

`Physicochemical Characterization of Iron Colloids
`
`1361
`
`Table II. Physicochemical Characterization of Iron Complex Drugs
`
`Attributes
`
`Reasons to recommend
`the attributes
`
`Tests
`
`Whole particle
`
`Equivalence in stoichiometric
`ratios of iron, free and
`bound carbohydrate, and
`other relevant components
`Molecular weight distribution
`(Mw, Mn, and Mw/Mn)
`Low molecular weight iron species
`
`Labile iron
`
`Particle size distribution
`
`Iron core
`
`Iron core size and morphology
`
`Crystallinity
`
`Iron environment (valence state
`of iron, the spin state of ferric,
`and the type of coordination state
`of iron atoms and details
`of ligand binding)
`Fe3+ to Fe2+ reduction potential
`and Fe(II) content
`Magnetic properties
`
`Carbohydrate shell Carbohydrate composition and
`carbohydrate-iron
`core interactions
`
`Surface charge
`
`Characterization of
`polysaccharides
`
`Q1/Q2
`
`USP requirement
`
`Formation of NTBI
`and potential oxidative toxicity
`Formation of NTBI and
`potential oxidative toxicity
`Related to opsonization, MPS
`uptake, PK and tissue distribution
`Related to iron release rate, MPS
`uptake, PK and tissue distribution
`Related to iron release
`and in vivo stability
`Related to iron release
`and in vivo stability
`
`Iron assay (atomic absorption),
`carbohydrate assay (HPLC), or
`elemental analysis
`(AAS, ICP-MS, or EDX)
`AUC or GPC
`
`Dialysis or ultrafiltration
`
`Bleomycin assays or iron
`chelation assays
`DLS or AFM
`
`TEM (diameter), AFM, Mössbauer
`spectroscopy, and XRD
`Mössbauer spectroscopy, Raman, XRD,
`TEM/SAED or TEM/NBED, XANES
`Mössbauer spectroscopy, EPR,
`Raman, and UV/Vis
`
`Related to in vivo stability and to
`detect impurities
`To measure overall sameness
`in iron core and detect impurities
`Related to opsonization,
`MPS uptake, PK,
`tissue distribution, iron release
`rate and in vivo stability
`Related to opsonization,
`MPS uptake, PK, and
`tissue distribution
`Related to in vivo stability
`
`Polarography, cerimetric titration
`
`VSM or SQUID
`
`FT-IR, thermal analysis
`(TGA or DSC), changes in
`particle size under dilution,
`or polarography
`Electrophoretic mobility
`(e.g., zeta potential),
`potentiometric titration
`NMR, size exclusion
`chromatography, copper assay
`
`HPLC high-performance liquid chromatography, AAS atomic absorption spectroscopy, ICP-MS inductively coupled plasma mass
`spectrometry, EDX energy-dispersive X-ray, Mn number-average molecular weight, Mw weight-average molecular weight, USP United States
`Pharmacopeia, AUC analytical ultracentrifugation, GPC gel permeation chromatography, NTBI nontransferrin-bound iron, MPS mononuclear
`phagocyte system, DLS dynamic light scattering, AFM atomic force microscopy, TEM transmission electron microscopy, XRD X-ray
`diffraction, TEM/SAED transmission electron microscopy/selected area electron diffraction, TEM/NBED transmission electron microscopy/
`nano beam electron diffraction, XANES X-ray absorption near-edge structure, EPR electron paramagnetic resonance, VSM vibrating sample
`magnetometer, SQUID superconducting quantum interference device, FT-IR Fourier transform infrared spectroscopy, TGA thermal
`gravimetric analysis, DSC differential scanning calorimetry, NMR nuclear magnetic resonance
`
`labile iron, and stability of iron colloids. The characteriza-
`tion of
`these physicochemical properties of
`the whole
`particles is discussed in this section.
`
`Stoichiometric Ratios of Iron, Free and Bound Carbohydrate,
`and Other Excipients
`
`The iron in iron colloids is measured by ICP-MS, atomic
`absorption spectroscopy, or atomic emission spectroscopy,
`while the unbound carbohydrates are measured by chromato-
`graphic methods. The presence of unbound and loosely
`bound carbohydrates,
`low molecular weight Fe complex,
`excipients, and impurities may mask the real elemental
`composition in iron colloids. It is desired to completely
`remove these interfering species by dialysis or other
`
`approaches prior to elemental analysis. The amounts of
`carbon and hydrogen in dialyzed iron colloids can be
`measured by CHN analysis, energy dispersive X-ray elemen-
`tal analysis, atomic absorption, or other elemental analysis
`techniques.
`Elemental analysis of dialyzed iron core obtained from
`Venofer® and Ferrlecit® has been reported by Kudasheva
`et al. (11). From the molar ratio of iron to carbon (1.45:1) for
`Venofer® determined by elemental analysis, they concluded
`that one sucrose molecule of 12 carbon atoms is present for
`17 atoms of iron. For iron-gluconate complex, the elemental
`analysis gives a molar ratio of iron to carbon as 1.30:1,
`indicating that one gluconate of 6 carbon atoms is present for
`7.8 atoms of Fe. Considering the molecular formula of
`Ferrlecit® [NaFe2O3(C6H11O7)(C12H22O11)5]n = 200,
`the
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1092 - Page 3
`
`

`

`1362
`
`Zou et al.
`
`measured gluconate to iron ratio (1:7.8) revealed that only a
`small fraction of total gluconate binds to the iron core.
`Accurate elemental analysis depends on a complete removal
`of unbound and loosely bound carbohydrates, low molecular
`weight Fe complex, excipients, and impurities. In another
`reported study (12), the percentages (wt%) of C, H, N, and
`Fe in dialyzed Venofer® were determined as 36.31 ± 0.33,
`5.15 ± 1.87, 0.08 ± 0.12, and 5.7 ± 0.3, respectively, indicating
`that one sucrose molecule of 12 carbon atoms is present for
`2.5 atoms of iron. The high percentage of carbon indicated
`that the sample dialysis was likely incomplete, illustrating
`some of
`the limitations of elemental analysis for the
`characterization of iron colloids.
`
`Molecular Weight Distribution
`
`The average molecular weight and molecular weight
`distribution are indicators of the quality and stability of iron
`colloids and can be used to detect variations in product
`quality and potential degradation or aggregation of iron
`colloids. Usually,
`the weight-average molecular weight
`(Mw), the number-average molecular weight (Mn), and the
`polydispersity index (Mw/Mn) are measured to reflect the
`molecular weight distribution of the iron colloids. In the
`USP monograph of iron sucrose injection, the acceptance
`criteria of molecular weight are as follows: Mw = 34,000–
`60,000 Da, Mn ≥ 24,000 Da, and Mw/Mn ≤ 1.7. The
`apparent molecular weight of sodium ferric gluconate in
`sucrose complex (Maap 289,000–440,000 Da) is provided in
`the label of Ferrlecit®. In addition to Mw, Mn, and Mw/Mn,
`the Maap of sodium ferric gluconate in sucrose complex
`needs to be measured (13,14). It is inappropriate to directly
`compare the molecular weight of iron sucrose and sodium
`ferric gluconate in sucrose complex because of the differ-
`ences in analytical methods (chromatographic conditions,
`columns, and calibration standards). The common analytical
`methods for determining molecular weight distribution of
`iron colloids include size exclusion chromatography (SEC)
`or gel permeation chromatography (GPC), analytical ultra-
`centrifugation (AUC), and static light scattering (SLS). For
`these analytical methods,
`the selection of appropriate
`molecular weight calibration standards and calibration
`curve plotting method are crucial for accurate measurement
`of iron colloid molecular weight.
`
`Particle Size and Size Distribution
`
`Among the techniques used to determine the particle
`size distribution of carbohydrate-coated iron colloids, dy-
`namic light scattering (DLS) is most widely used. The
`particle size of iron colloids measured by DLS is summa-
`rized in Table III (4). DLS measurement is quick and
`sample preparation is very simple. However, the results are
`affected by the experimental conditions. For comparative
`particle size measurements, samples should be measured
`under the same experimental conditions. The DLS method
`should be appropriately validated for robustness and
`accuracy. The existence of free sucrose and/or gluconic acid
`may probably affect the particle size measured by DLS (11).
`
`is desired to measure the size and size
`it
`Therefore,
`distribution of samples by DLS before and after sample
`dialysis. It was reported that the removal of free carbohy-
`drates did not cause aggregation or degradation of iron
`colloids (11,14,18).
`
`Low Molecular Weight Iron
`
`Low molecular weight iron (or Bfree iron^) is the iron
`impurities which can be physically separated from iron
`colloids by dialysis or ultrafiltration (4,13,15,19). One of
`the concerns for iron colloid drug products is the potential
`toxicity of low molecular weight iron in the formulations
`(20). According to the USP monograph for iron sucrose
`injection (21), no additional peaks in the polarogram can
`be used to demonstrate the absence of
`low molecular
`weight iron complex. However, the quantitation of
`low
`molecular weight
`iron species by polarography is not
`sensitive (22). Low molecular weight
`iron can also be
`measured by dialysis or ultrafiltration. In the label of
`sample dialysis against a 12–14-kDa cutoff
`Ferrlecit®,
`membrane over a period of up to 270 min is recom-
`mended for the measurement of
`low molecular weight
`iron and the acceptable limit is 1% of total iron. Previous
`studies (Table IV) have revealed that the amount of low
`m o l e c u l a r w e i g h t
`i r o n
`r a n k s
`a s
`f o l l o w s :
`Ferrlecit® > Venofer® ≈ INFeD® > Feraheme®,
`Injectafer®, and Monofer® (4,13).
`The amount of low molecular weight iron determined
`by ultrafiltration depends on filter cutoff, pH of the buffer,
`and centrifugation speed. The amount of low molecular iron
`determined by dialysis depends on the components in
`dialysis buffer, the buffer pH, the volume ratio of iron
`colloid suspension and dialysis buffer, and the cutoff of
`dialysis membrane. For example, a higher percentage of
`dialyzable iron was detected when 0.9% sodium chloride
`solution at pH 7.5 was used as the dialysis buffer compared
`with pH 7.5 water (4). Therefore, a comparative quantita-
`tion of low molecular weight iron complex in two iron
`colloid samples should be conducted under the same
`experimental conditions and in the presence of control
`samples.
`
`Labile Iron
`
`Different from dialyzable low molecule weight iron or free
`iron, labile iron is iron impurities which physically bind to iron
`carbohydrate matrix and are not dialyzable in saline but readily
`released in blood circulation (4,15). In the presence of 4.5%
`bovine serum albumin in dialysis buffer (Table IV), however,
`the amount of dialyzable iron dramatically increased (13),
`indicating that labile iron can be dialyzed in the presence of
`plasma protein or iron chelation agents (4). If serum transferrin
`is saturated, the released labile iron in the bloodstream forms
`NTBI, which potentiates oxidative stress and inflammation, then
`resulting in direct cellular damage and possibly increasing the
`risk of atherosclerotic disease (24). The content of labile iron
`can be measured under physiological conditions using several
`reported assays, which are divided into two categories:
`bleomycin assays and chelation-based assays (25). As shown in
`Table IV, the content of labile iron roughly follows the sequence:
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1092 - Page 4
`
`

`

`Physicochemical Characterization of Iron Colloids
`
`1363
`
`Table III. The Mean Size of Iron Core and Whole Particle of Iron Colloids
`
`Iron colloid products
`
`Mean diameter of iron core (nm)
`
`Mean diameter of whole particle (nm)
`
`Ferrlecit
`Venofer
`Feraheme
`INFeD
`Injectafer
`Monofer
`Refs.
`
`TEM
`
`4.1 ± 1.7
`5.0 ± 0.8
`6.2 ± 1.4
`5.6 ± 1.2
`11.7 ± 4.4*
`6.3 ± 1.2
`(4)
`
`XRD
`
`Mössbauer
`
`3.4
`3.3
`6.4
`4.4
`4.3
`4.2
`(4)
`
`N.A.
`2.5
`5–10
`N.A.
`3.3
`2.5
`(15,16)
`
`AFM
`
`3 ± 1
`4 ± 2
`N.A.
`N.A.
`N.A.
`N.A.
`(11,17)
`
`DLS
`
`8.6
`8.3
`23.6
`12.2
`23.1
`9.1
`(4)
`
`* The mean diameter of an agglomeration of several cores. Single cores are not definable. TEM transmission electron microscopy, XRD X-ray
`diffraction, AFM atomic force microscopy, DLS dynamic light scattering, N.A. not available
`
`Ferrlecit®≈Venofer®>INFeD®>Dexferrum®>Feraheme®≈Inj-
`ectafer® ≈ Monofer®. The content of labile iron measured by
`bleomycin assay is lower than that determined by EDTA
`chelation assay or transferrin binding assay (Table IV). This
`might probably be explained that the bleomycin assay specifi-
`cally detects redox-active iron only while the chelation-based
`assay detects the total amount of labile iron and free iron (25).
`Both bleomycin assay and chelation assay usually cannot
`distinguish free iron and labile iron, and the amount of free
`iron is included in the measured labile iron (15). The values of
`labile iron assays are dependent on binding affinity of selected
`chelators and the concentration of chelators. Therefore, large
`cross-lab variability in labile iron measurement is usually
`observed. It is important to compare the labile iron of two iron
`colloid products using the same assay and under the same
`experimental conditions.
`
`Bleomycin Assay
`
`The conventional bleomycin assay was first reported for
`the selective measurement of radical-promoting,
`loosely
`bound iron in biological fluids (26), and the detected iron is
`generally named bleomycin-detectable iron (BDI).
`Bleomycin is an anticancer drug which induces DNA
`degradation in the presence of ferrous ion. In a conventional
`bleomycin assay, bleomycin is incubated with DNA, ascorbic
`acid which converts ferric ions into ferrous ions, and the
`biological sample. The formation of bleomycin-Fe(II) com-
`plex causes degradation of DNA. DNA damage is then
`quantified by the formation of malondialdehyde (MDA) from
`the deoxyribose moiety of the DNA using the thiobarbituric
`acid (TBA) test. The formation of (TBA)2-MDA adduct
`causes UV absorption at 532 nm, and therefore,
`the
`
`Table IV. Contents of Low Molecular Weight Iron and Labile Iron in Iron Colloid Drug Products
`
`Iron colloid drugs
`
`Low molecular
`weight iron (free iron)
`
`Free iron + labile iron
`
`Dialysis in pH
`7.5 NaCl
`solution (4)
`
`Ultra-
`fi l t r a t i o n
`(13)
`
`Dialysis in 4.5%
`albumin (13)
`
`Bleomycin
`assay (13)
`
`Transferrin
`binding (100–125 mg
`dose) (19)
`
`Transferrin
`binding (200 mg
`dose) (4)
`
`EDTA
`c h e l a t i o n
`(23)
`
`Iron dextran
`(Dexferrum®
`or DexIron®)
`Low MW
`iron dextran
`(INFeD®)
`Sodium ferric
`gluconate
`(Ferrlecit®)
`Iron sucrose
`(Venofer®)
`Ferumoxytol
`(Feraheme®)
`Iron
`carboxymaltose
`(Injectafer®)
`Iron isomaltoside
`(Monofer®)
`
`–
`
`–
`
`–
`
`–
`
`2.5%
`
`0.207%
`
`0.30%
`
`0.70%
`
`0.19%
`
`3.4%
`
`1.338%
`
`2.36%
`
`4.80%
`
`1.40%
`
`5.8%
`
`0.067%
`
`<0.002%
`
`<0.002%
`
`<0.002%
`
`0.038%
`
`0.001%
`
`2.80%
`
`0.55%
`
`–
`
`–
`
`–
`
`–
`
`0.69%
`
`0.07%
`
`–
`
`–
`
`4.5%
`
`–
`
`–
`
`–
`
`–
`
`2.1%
`
`3.2%
`
`3.5%
`
`0.8%
`
`0.5%
`
`0.9%
`
`–
`
`0.8%
`
`3.6%
`
`3.8%
`
`–
`
`–
`
`–
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1092 - Page 5
`
`

`

`1364
`
`Zou et al.
`
`concentrations of BDI in the biological sample can be
`quantified using a standard curve of the UV absorption at
`532 nm (27). However, the formation of (TBA)2-MDA
`adduct requires relatively harsh conditions. For example, the
`mixture is incubated at 80°C for 20 min to allow the
`formation of (TBA)2-MDA adduct. Side reactions occur at
`high temperature, which may result in the generation of
`misleading data (27).
`To overcome the limitations of conventional bleomycin
`assay, Burkitt et al. modified the assay by using ethidium-
`binding to measure DNA damage caused by bleomycin-
`Fe(II) complex (27). Ethidium bromide is a dye for DNA
`staining. The intercalation of ethidium bromide into DNA
`enhances fluorescence, while the DNA damage caused by
`bleomycin-Fe(II) complex can decrease the fluorescence
`intensity. A nonlinear calibration curve is established between
`BDI concentration and fluorescence intensity. The ethidium-
`binding-based bleomycin assay was found particularly re-
`sponsive to damage induced by iron at very low concentration
`end of the standard curve, but saturation occurred at higher
`concentrations. The range of
`iron concentration in the
`standard curve can be adjusted by changing the concentration
`of DNA in the incubation.
`
`Iron Chelation Assays
`
`labile iron is first
`For iron chelation-based assays,
`mobilized with a chelating agent and then the chelation
`complex is separated (optional) and measured by different
`methods, including HPLC, colorimetric measurement, fluo-
`rescence measurement, and atomic absorption spectropho-
`tometry (AAS). The reported chelation agents for labile iron
`assay include Ferrozine® (4), EDTA (23), nitrilotriacetate
`(NTA) (25), oxalate (28), chromazurol B (29), bathophenan-
`throline (30), polyphenols in tea extract (15), and others.
`The most widely used chelation agent for labile iron
`determination is Ferrozine®, which can detect not only low
`molecular weight iron but also labile iron and transferrin-
`bound iron in serum (4,19). Briefly, the iron colloid sample
`is incubated in human plasma at room temperature, and
`then the mixture is added with detergent, thiourea, and
`citric acid followed by the addition of sodium ascorbate and
`Ferrozine. The detergent serves to clarify plasma sample,
`acidic buffer lowers pH to <2 to free Fe(III) from
`transferrin, and ascorbic acid reduces Fe(III) to Fe(II)
`which forms complex with Ferrozine. The UV absorption
`of the complex is measured at 562 nm. One concern for this
`procedure is that the harsh conditions such as low pH may
`cause degradation of iron colloids, resulting in overestima-
`tion of labile iron. To avoid the harsh conditions, Van Wyck
`et al. (19) removed iron colloids from the plasma incubation
`mixture using an alumina column before the Ferrozine test.
`Only the protein-bound labile iron and protein-bound free
`iron were measured.
`A capillary electrophoresis method coupled with EDTA
`chelation was used to determine labile iron in iron colloid
`samples (23). Different from Ferrozine which formed Fe(II)-
`chelate, EDTA formed Fe(III)-EDTA chelate and therefore
`reduction with ascorbic acid was not required. The Fe(III)-
`EDTA chelate was separated on a capillary electrophoresis
`and detected at 246 nm.
`
`For the assay of labile iron in iron sucrose and sodium
`ferric gluconate complex in sucrose, acidic conditions can
`cause hydrolysis of sucrose to form glucose and fructose. Due
`to their free ketone group, glucose and fructose can reduce
`Fe(III) in the assay to Fe(II). Therefore, when a Fe(III)
`chelate is used to measure labile iron of iron sucrose and
`sodium ferric gluconate complex in sucrose samples under
`acidic conditions, the autoreduction by reducing sugars may
`result in underestimation of labile iron.
`
`IRON CORE CHARACTERIZATION
`
`The efficacy and safety of iron colloids are affected by
`not only the properties of the whole particle but also the iron
`core properties. For example, macrophage uptake and tissue
`distribution of iron colloids may be impacted by the iron core
`size and morphology (31). The in vivo stability and iron
`release rate are correlated with iron core size, crystallite
`structure, and iron core environment. For iron colloids with
`magnetic property, comparative magnetic characterization
`can detect the potential impurities and subtle differences in
`iron core structure and environment (32).
`
`Iron Core Size and Morphology
`
`Iron carbohydrate complexes are composed of an iron
`core surrounded by associated carbohydrate ligands. Al-
`though DLS can be used to determine the hydrodynamic size
`of iron colloids, it is unable to measure the size of the iron
`core. The diameter of the iron core can be measured by
`transmission electron microscopy (TEM), X-ray diffraction
`(XRD), Mössbauer spectroscopy, and other appropriate
`techniques. TEM and atomic force microscopy (AFM) are
`frequently used to investigate the particle morphology (e.g.,
`shape and agglomeration status of the particles).
`
`Transmission Electron Microscopy
`
`TEM is frequently used to measure the diameter and
`morphology of the iron core of iron colloid products. The
`carbohydrate shell usually cannot be detected by TEM
`without appropriate staining due to the lower electron density
`of the carbohydrate shell. Jahn et al. reported the mean
`diameter of Venofer®, Ferrlecit®, Feraheme®, INFeD®,
`Injectafer®, and Monofer® iron core as 5.0 ± 0.8, 4.1 ± 1.7,
`6.2 ± 1.4, 5.6 ± 1.2, 11.7 ± 44, and 6.3 ± 1.2 nm, respectively
`(Table III) (4). It is worthy to note that the mean diameter of
`the Injectafer® particle is the mean diameter of an agglom-
`eration of several
`iron cores. Individual
`iron cores of
`Injectafer are not able to be resolved under TEM. The
`average iron core size of Feraheme® measured by TEM was
`reported by Balakrishnan et al. as 6.4 ± 0.4 nm (13). Although
`it was difficult to accurately determine the iron core sizes of
`Venofer®, Ferrlecit®, and INFeD® due to an undefined iron
`core margin, Balakrishnan et al. ranked the average core size
`as follows: Feraheme® > INFeD® > Venofer® ≥ Ferrlecit®
`(13). Inconsistent with Jahn’s results, Kudasheva et al. (11)
`reported the iron core size distribution of dialyzed Venofer
`and Ferrlecit ranging from 1.0 to 6.5 nm (mean 3 ± 2 nm) and
`0.9 to 3.5 nm (mean 2 ± 1 nm), respectively. Bullivant et al.
`reported the mean diameter of Feraheme® core as 3.25 nm
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1092 - Page 6
`
`

`

`Physicochemical Characterization of Iron Colloids
`
`1365
`
`Fig. 1. Cryo-TEM micrographs of a Ferrlecit®, b Venofer®, c INFeD®, and d Feraheme® (34)
`
`(33). These inconsistencies are likely caused by the differ-
`ences in sample concentrations, sample preparation, instru-
`ment variability, instrumental parameters, and data analysis.
`Morphology analysis is often performed concurrently
`with the iron core size determination. A recent study
`conducted at an FDA lab showed that cryo-TEM where the
`samples were prepared in liquid nitrogen could generate a
`higher resolution image (Fig. 1) and a more accurate
`measurement of iron core size compared with room temper-
`ature TEM because the sample dehydration at room temper-
`ature may potentially cause aggregation (34). The cryo-TEM
`data showed that the iron cores of Venofer®, Ferrlecit®,
`INFeD®, and Feraheme® share a spherical morphology. As
`shown in Fig. 1d, Feraheme® exhibits clusters of multiple
`particles rather than individual particles. The clusters were
`observed in the cryo-TEM images of both undiluted
`Feraheme® sample and 100-fold diluted sample. Consistently,
`clusters of multiple particles were also observed in TEM and
`STEM images of Feraheme® a