International Journal of Hyperthermia

ISSN: 0265-6736 (Print) 1464-5157 (Online) Journal homepage: www.tandfonline.com/journals/ihyt20

An optimised spectrophotometric assay for

convenient and accurate quantitation of
intracellular iron from iron oxide nanoparticles

Mohammad Hedayati, Bedri Abubaker-Sharif, Mohamed Khattab, Allen

Razavi, Isa Mohammed, Arsalan Nejad, Michele Wabler, Haoming Zhou, Jana
Mihalic, Cordula Gruettner, Theodore DeWeese & Robert Ivkov

To cite this article: Mohammad Hedayati, Bedri Abubaker-Sharif, Mohamed Khattab, Allen
Razavi, Isa Mohammed, Arsalan Nejad, Michele Wabler, Haoming Zhou, Jana Mihalic, Cordula
Gruettner, Theodore DeWeese & Robert Ivkov (2018) An optimised spectrophotometric assay
for convenient and accurate quantitation of intracellular iron from iron oxide nanoparticles,
International Journal of Hyperthermia, 34:4, 373-381, DOI: 10.1080/02656736.2017.1354403

To link to this article: https://doi.org/10.1080/02656736.2017.1354403

© 2017 The Author(s). Published by Informa View supplementary material

UK Limited, trading as Taylor & Francis
Group

Published online: 31 Jul 2017. Submit your article to this journal

Article views: 6787 View related articles

View Crossmark data Citing articles: 21 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=ihyt20
INTERNATIONAL JOURNAL OF HYPERTHERMIA
2018, VOL. 34, NO. 4, 373–381
https://doi.org/10.1080/02656736.2017.1354403

An optimised spectrophotometric assay for convenient and accurate

quantitation of intracellular iron from iron oxide nanoparticles
Mohammad Hedayatia
, Bedri Abubaker-Sharifa, Mohamed Khattaba, Allen Razavia, Isa Mohammeda,
Arsalan Nejada, Michele Wablera, Haoming Zhoua, Jana Mihalicb, Cordula Gruettnerc, Theodore DeWeesea,d and
Robert Ivkova,d,e,f,g
a
Department of Radiation Oncology and Molecular Radiation Sciences, Johns Hopkins University School of Medicine, Baltimore, MD, USA;
b
Department of Environmental Health Sciences, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA; cMicromod
Partikeltechnologie GmbH, Rostock, Germany; dDepartment of Oncology, Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins
University School of Medicine, Baltimore, MD, USA; eInstitute for NanoBioTechnology, Johns Hopkins University, Baltimore, MD, USA;
f
Department of Materials Science and Engineering, Whiting School of Engineering, Johns Hopkins University, Baltimore, MD, USA;
g
Department of Mechanical Engineering, Whiting School of Engineering, Johns Hopkins University, Baltimore, MD, USA

ABSTRACT ARTICLE HISTORY

We report the development and optimisation of an assay for quantitating iron from iron oxide nano- Received 9 June 2017
particles in biological matrices by using ferene-s, a chromogenic compound. The method is accurate, Revised 7 July 2017
reliable and can be performed with basic equipment common to many laboratories making it conveni- Accepted 8 July 2017
ent and inexpensive. The assay we have developed is suited for quantitation of iron in cell culture
KEYWORDS
studies with iron oxide nanoparticles, which tend to manifest low levels of iron. The assay was vali- Iron oxide nanoparticles;
dated with standard reference materials and with inductively coupled plasma-mass spectrometry (ICP- mass spectrometry; UV/Vis
MS) to accurately measure iron concentrations 1  106 g in about 1  106 cells (1  1012 g Fe per spectrophotometry; iron
cell). The assay requires preparation and use of a working solution to which samples can be directly quantitation assay;
added without further processing. After overnight incubation, the absorbance can be measured with a intracellular iron
standard UV/Vis spectrophotometer to provide iron concentration. Alternatively, for expedited process-
ing, samples can be digested with concentrated nitric acid before addition to the working solution.
Optimization studies demonstrated significant deviations accompany variable digestion times, high-
lighting the importance to ensure complete iron ion liberation from the nanoparticle or sample matrix
to avoid underestimating iron concentration. When performed correctly, this method yields reliable
iron ion concentration measurements to 2  106 M (1  107 g/ml sample).

Introduction iron from biological matrices that provides accuracy and pre-
cision comparable to ICP-MS, the current gold standard. Such
Iron oxide nanoparticles have established themselves as a
an assay would enhance research progress and toxicity
useful tool for biomedical applications ranging from subcellu-
assessments, by enabling quantitative comparisons of cell-
lar manipulation to whole body imaging, and treatment of nanoparticle association that can be related to specific nano-
anaemia and cancer [1–12]. The importance of cell or tissue particle constructs or properties.
iron quantification to assess both efficacy and toxicity is read- ICP-MS is often considered as the gold standard to meas-
ily apparent; however, the most reliable assays, for example ure iron primarily because of high sensitivity – detection lim-
inductively coupled plasma-mass spectrometry (ICP-MS), its to parts per trillion (nanogram per litre) are possible. This
require specialised and costly equipment [13–15]. Existing is a lower detection threshold than achievable with other
less costly and simple chemical assay methods have not spectroscopic techniques such as ICP-atomic emission spec-
been optimised or validated to quantify low levels of iron troscopy (ICP-AES), graphite furnace-atomic absorption spec-
(<1 lg) that are typically recovered from tissue or cell culture troscopy (GF-AAS) or flame atomic absorption spectroscopy
systems mixed with iron oxide nanoparticles [16]. Studies of (AAS) [13,15]. In addition to its high sensitivity, ICP-MS pro-
cancer targeting [17–22], magnetic hyperthermia vides flexibility to detect multiple elements for simultaneous
[17,18,23,24], imaging [8,25], toxicology [26], cell separation interrogation. To be fair, the detection limits possible with
[27] and stem cell tracking [28,29] would benefit significantly ICP-MS for iron quantification require specialised instrument
with development of a more accessible method to quantify configurations having octopole reaction cell/mass filter(s) to

CONTACT Robert Ivkov [email protected] Department of Radiation Oncology and Molecular Radiation Sciences, David H. Koch Cancer Research Building,
Rm 442, The Johns Hopkins University School of Medicine, 1550 Orleans St., Baltimore, MD 21231, USA

Present address: Independent consultant, PO Box 41108, Baltimore, MD 21203, USA [email protected]
Supplemental data for this article can be accessed here.
ß 2017 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://creativecommons.org/licenses/by-nc-nd/4.0/),
which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed, or built upon in any
way.
374 M. HEDAYATI ET AL.

discriminate Fe from among the many possible polyatomic dextran matrix [38,39]. The nanoparticles were used as pro-
interferences; and, pristine sample preparation to avoid con- vided by manufacturers.
tamination with ubiquitous environmental Fe [30]. Thus, ICP-
MS and other sensitive mass spectrometry technologies, Nanoparticle loading into cells
although superior in many ways, demand significant financial
Human prostate carcinoma DU145 cells (ATCC, Manassas, VA)
investment to acquire and maintain the instrumentation, and
were maintained and propagated according to the ATCC
to develop and maintain dedicated facility infrastructure.
protocol. Cells were cultured at 37  C and 5% CO2 in T-75 tis-
Colorimetric methods based upon absorption of light by a
sue culture flasks containing RPMI 1640 medium and 10%
sample, that is ultraviolet/visible (UV/Vis) spectrophotometry,
foetal bovine serum (Life Technologies, Carlsbad, CA).
to quantify intracellular iron have long existed [31–33].
Exponentially growing cells were treated by adding iron
Refinement continues to improve their sensitivity and accur-
oxide nanoparticles (50–150 lg/ml Fe) and sterile poly-L-
acy [34–36]. When bound to free iron ions, chelators such as
lysine (PLL, 0.5–2.0 lg/ml; Sigma-Aldrich, St. Louis, MO) to the
the triazines absorb very strongly in the UV/Vis range, with
culture media. After 24 h, the flask was rinsed three times
molar absorptivity of 104 M1 cm1. Many chromogenic
with Dulbecco’s phosphate buffered saline (PBS) to remove
substrates such as bathophenanthroline sulfonate, ferrozine
extracellular nanoparticles. Cells were detached by brief
or ferene have been utilised for iron quantification in human
exposure to 0.25% trypsin-EDTA followed by resuspension in
serum or environmental samples [31–33]. The colorimetric
culture media. After centrifugation and two more washes
assay developed by Riemer et al. [34] incorporates potassium
with PBS the fully suspended cells were counted. Aliquots
permanganate-mediated digestion/oxidation and iron detec- with equal number of cells were transferred to 1.5 ml micro-
tion by ferrozine and is one of the most commonly fuge tubes and frozen to 80  C as cell pellets. To obtain
employed methods for detecting cellular iron. cells with low, medium and high iron content, the concentra-
We report on the development of an optimised triazine- tion of 80 nm BNF-Starch particles was kept constant at
based colorimetric assay with ferene-s, a more sensitive 150 lg/ml while varying the concentrations of PLL (1, 1.5 and
chromogenic compound for iron detection. We introduce a 2 lg/ml respectively).
simplified digestion procedure to ensure accuracy among
varying experimental configurations, and validation of results
Ferene-s assay
with ICP-MS using reference standard materials. The assay pre-
sented is a simple, sensitive, robust and highly adaptable Ferrozine (3-(2-pyridyl)-5,6-diphenyl-1,2,4-triazine-40 ,400 -disul-
method of iron quantitation for a wide variety of cellular appli- fonic acid sodium salt) and ferene-s (3-(2-pyridyl)-5,6-di(2-
cations of iron oxide nanoparticles. It is capable to provide furyl)-1,2,4-triazine-50 ,500 -disulfonic acid disodium salt) were
accurate iron ion concentration measurements to purchased (Sigma-Aldrich, St. Louis, MO). A detailed protocol
2  106 M (1  107 g/ml sample). with lists of materials and equipment for this assay is
included in the Supplementary Materials and Methods.
Materials and methods Briefly, the sample to be measured was directly added to a
1 working solution (5 mM ferene-s, 0.2 M L-ascorbic acid in
Iron standards and nanoparticles 0.4 M ammonium acetate buffer, pH 4.3) in a total volume
All iron standards were prepared from a stock certified stand- of 1 ml. The mixture was incubated at room temperature in
ard reference material (FeCl3, Iron Standard for ICP, the dark for 20 h. The absorbance was measured at 595 nm
1000 ± 2 mg/l Fe in 2% nitric acid, Sigma-Aldrich, St. Louis, in a flat bottom standard 96-well plate (300 ll per well in
MO). Iron oxide nanoparticle formulations used in this study triplicates) using SpectraMax-M5 plate reader (Molecular
generally comprised iron oxide with a polysaccharide (dex- Devices Corporation, Sunnyvale, CA). FeCl3 iron standards
tran or starch) polymer as either coating material (core–shell were always included (0–4 lg/ml). Alternatively, the samples
nanoparticles) or as matrix (iron oxide crystallites in polymer were digested first with concentrated nitric acid (100 ll) for
2 h at 70  C. The digested samples were allowed to cool to
matrix). Specifically, particles used were: 80 and 100 nm plain
room temperature before neutralisation with 10 N NaOH
BNF-Starch and 100 nm plain BNF-Dextran particles (Bionized
(160 ll). The 1 working solution (900 ll) was added to the
NanoFerrite, Micromod Partikeltechnologie GmbH, Rostock,
neutralised sample (100 ll). After 30 min at room temperature
Germany), 20 nm plain nanomagV-D-spio particles and
R

the absorbance was measured at 595 nm in a flat bottom

500 nm nanomagV-D-plain (dextran iron oxide composite par-
R

standard 96-well plate (300 ll per well in triplicates).

ticles, Micromod Partikeltechnologie GmbH, Rostock,
For comparison to the ferrozine assay, the 1 working
Germany) and conjugated or unconjugated magnetic
solution was prepared with 5 mM ferrozine instead of ferene-
microbeads (Anti ErbB2 or Basic MicroBeads, Miltenyi Biotec
s and the absorbance was measured at 560 nm.
Inc., Auburn, CA). BNF particles are core–shell particles with
an iron oxide core (45 nm) comprising multiple iron oxide
Kinetics of iron oxide nanoparticles in working solution
crystallites having parallelepiped shapes with long dimension
of the ferene-s assay
of 15–20 nm [20,23,37,38] and a shell of starch or dextran. In
contrast 20 nm nanomagV-D-spio particles comprise iron
R
Samples were added directly to 1 working solution (total
oxide crystals with diameters of 7–10 nm embedded in volume of 1 ml). The absorbance was measured at 595 nm in
 INTERNATIONAL JOURNAL OF HYPERTHERMIA 375

a flat bottom standard 96-well plate (300 ll per well in tripli- The total amount of iron in an unknown sample could there-
cates) at 30 min intervals for up to 24 h. Absorbance readings fore be determined by dividing the total absorbance units of
for DU145 cells containing medium and high BNF-Starch con- the unknown by this slope.
centrations were adjusted to reflect equivalent number of
cells to that of cells with low Fe content.
Results
Comparison of ferene-s to ferrozine using Fe standards
Inductively coupled plasma-mass spectrometry (ICP-MS)
We compared the most commonly used chromogenic iron-
Frozen cell-nanoparticle samples were thawed at room tem- chelator ferrozine to ferene-s. Ferene-s is structurally similar to
perature, and then prepared for ICP-MS using methods previ- ferrozine (Figure 1) and it has been reported to have a higher
ously described [25,40]. Briefly, thawed samples were molar absorptivity [32,41]. Figure 1 shows UV/Vis spectra of
thermally digested in 50% nitric acid using a two stage ferrozine (A) and ferene-s (B) for FeCl3 standards ranging from
ramp-to-temperature microwave method (MARS5 Xpress 0 to 2 lg in a final volume of 1 ml. Figure 1(C) shows the
microwave, CEM Corporation, Matthews, NC). Digested sam- standard curves for both compounds measured at the peak
ples were diluted for mass spectrometric evaluation and the absorbance for each compound (560 nm for ferrozine and
total iron content of the samples was determined using an 595 nm for ferene-s). The assay developed uses a 96-well plate
Agilent 7500ce ICP-MS (Agilent Technologies, Santa Clara, format to measure optical density based on volumes of 300 ll
CA). An eight-point calibration curve was performed prior to per well. To obtain accurate estimates of the molar absorptiv-
sample analysis. At least three samples were analysed by ICP- ity of the two compounds we measured the absorbance using
MS for each cell preparation. The total iron content per cell a standard 1 ml cuvette (see Supplementary Table S1). The
was calculated, accounting for the number of cells provided calculated mean molar absorptivity ± SD obtained was
as well as dilution factors, as the mean value for these 27 626 ± 523 M1 cm1 for ferrozine and 35 194 ± 831 M1 cm1
samples. for ferene-s. These values are similar to the reported reference
values [41–44]. The molar absorptivity of ferene-s is about 28%
higher than that of ferrozine. Therefore, ferene-s was chosen
Ferrozine assay (Riemer)
as the iron chelator for the assay development.
The assay was performed essentially as reported by Riemer
et al. [34]. Briefly, cell pellets were lysed with 300 ll of Comparison of ferene-s assay to ICP-MS
50 mM NaOH for 2 h at room temperature with constant
mixing using a rotary shaker. An equivalent volume (300 ll) To validate results obtained from the ferene-s assay, we com-
of 10 mM HCl was then added followed by the addition of pared with results obtained from ICP-MS for several commer-
300 ll of freshly prepared solution containing 2.25% cially available iron oxide nanoparticles. The measured iron
KMnO4 and 0.7 M HCl. The samples were then transferred contents of the four different formulations of nanoparticles
to a heat block and heated at 60  C for 2 h. The heated are shown in Table 1. The values obtained by ferene-s assay
samples were allowed to cool to room temperature before are similar to those obtained by ICP-MS. We also compared
addition of 90 ll of freshly prepared iron detection solution the intracellular iron concentration for one of these nanopar-
(6.5 mM ferrozine in 1 M L-ascorbic acid and 2.5 M ammo- ticles (BNF-Starch) that were determined with both methods.
nium acetate). After 30 min, 280 ll aliquots of the samples The mean ± SD of iron in BNF-Starch loaded DU145 cells was
were transferred to a well of a 96-well plate in triplicates (44.3 ± 5.0)  1012 g/cell with ferene-s assay and
and the absorbance at 550 nm was recorded. Iron oxide (45.5 ± 14.6)  1012 g/cell for ICP-MS (Table 1).
nanoparticles and FeCl3, iron standards (0–4 lg) were pre-
pared similarly.
Kinetics of iron oxide nanoparticles in working solution
of the ferene-s assay
Modified protocol to correct for path length for 96-well
The formation of iron ferene-s complex (Fe2þ: ferene-s) in the
plate readers
working solution is rapid for FeCl3 standards (<30 min at
A fixed volume of 300 ll is used for each well, triplicate per room temperature, Supplementary Figure S1). Once formed,
sample, and a standard curve is included in each experiment. the complex is stable (at least 1 week, Supplementary Table
Readings of 96-well plates are dependent on both the con- S6). We note, however, that with other sample matrices the
centration and the volume of the sample, thus each absorb- formation of iron ferene-s complex can be slow depending
ance reading is converted to the total absorbance for the on the formulation of nanoparticles or their intracellular loca-
entire sample (by multiplying with the ratio of the total vol- tion. The kinetics of complex formation obtained for several
ume/volume of sample in the well). The total absorbance different nanoparticles is shown in Figure 2(A). The saturation
therefore depends only on the total amount of iron in the time for the absorbance readings differed among the formu-
sample and not on the concentration. As shown in lations; however, by about 8 h they had all achieved a con-
Supplementary Table S4 and Figure 1(C) the average slope of stant value. The kinetics of complex formation for
total net absorbance for Fe standards is about 1.7 absorb- intracellular BNF-Starch was still slower as shown in Figure
ance units per lg of iron with 4.4% coefficient of variation. 2(B) reaching a constant value at 20 h.
376 M. HEDAYATI ET AL.

(A) (B)

(C) 8
y = 1.71x
7 R2 = 0.99
6 y = 1.34x
Total Absorbance

R2 = 0.99
5
4
3
ferene-s
2
ferrozine
1
0
0 1 2 3 4 5
Total Fe (µg)
Figure 1. Comparison of the spectra of the iron standards (0–2 lg Feþ3) in working solution. Ferene-s (A), ferrozine (B), and the corresponding standard curves
(0.1–4 lg Feþ3) measured at the peak absorbance for each compound (560 nm for ferrozine and 595 nm for ferene-s), (C).

Table 1. Iron concentrations obtained from various iron oxide nanoparticles. varying times at 70  C. As shown in Figure 3, a 1-h digestion
Nanoparticles ICP-MSa (lg Fe) Ferene-sa (lg Fe) was sufficient to yield constant values of measured iron,
V
R
nanomag -D-spio 4.7 ± 0.3 5.0 ± 0.3 implying that liberation of the intracellular iron with longer
BNF-Starch 5.7 ± 0.6 5.8 ± 0.4 digestion times of up to and over 16 h did not yield additional
Miltenyi anti-ErbB2 microbeads 4.9 ± 0.4 4.8 ± 0.2 significant differences of measured iron concentrations.
Miltenyi basic microbeads 5.1 ± 0.9 4.6 ± 0.3
Intracellular BNF-Starch 45.5 ± 14.6 (n ¼ 7) 44.3 ± 5.0 (n ¼ 20)
a
Data are reported as mean ± SD, and n ¼ 2 unless otherwise noted.
Comparison to Riemer’s method
We next compared results obtained from the ferene-s assay
Digestion with acid
of nanoparticle iron quantification to results obtained using
Intracellular iron can be released more rapidly by digestion the method presented by Riemer et al. [34], one of the most
with concentrated nitric acid. Figure 3 shows the results common methods used for these measurements. Measured
obtained from DU145 cells loaded with medium BNF-Starch iron content obtained using Riemer’s method is significantly
concentration after digestion with concentrated nitric acid for lower than the iron concentrations measured by the ferene-s
 INTERNATIONAL JOURNAL OF HYPERTHERMIA 377

(A) 2.5 (B) 5

4.5

2 4

3.5

Absorbance (@595 nm)

Absorbance (@595 nm)

1.5 3

2.5

1 2

Miltenyi anti Erb2 microbeads 1.5

Miltenyi basic microbeads
0.5 BNF-Dextran plain 100 nm 1
BNF-Starch plain 100 nm
BNF-Starch plain 80 nm Du145 cells with high BNF-Starch
0.5
nanomag® -D plain 500 nm Du145 cells with medium BNF-Starch
Nanomag-D-spio plain 20 nm Du145 cells with low BNF-Starch
0
0
0 1 2 3 4 5 6 7 8 0 5 10 15 20

Time (hrs.) Time (hrs.)

Figure 2. Time course of absorbance at 595 nm for ferene-s assay in working solution. Iron oxide nanoparticles with various formulations (A) and intracellular BNF-
Starch iron oxide nanoparticles (B), DU145 cells with low, medium and high Fe content). Absorbance readings are adjusted to reflect equivalent number of cells.

60

50

40
pg Fe/cell

30

20

10

Figure 3. Digestion with concentrated nitric acid. Intracellular BNF-Starch iron oxide nanoparticles digested at various times at 70  C prior to addition to the work-
ing solution.

method for all formulations comprising core–shell particles ferene-s or ICP-MS (Figure 4(B)). Between them, ICP-MS and
(Figure 4(A)). Interestingly, dextran matrix formulations – ferene-s yielded consistent results.
Miltenyi beads and nanomagV-D-spio particles – yielded simi-
R

lar measured iron values between the two methods.

When comparing iron concentration measured from cells Modified protocol to correct for path length for 96-well
that internalised BNF-Starch nanoparticles, we find again the plate readers
intracellular iron measurements obtained using Riemer’s Results of comparing iron concentration measurements
method are significantly lower than those obtained using obtained from 96-well plate reader are shown in
378 M. HEDAYATI ET AL.

(A) 18 ferene-s method (B) 140 Riemer's method

16 ferene-s method
Riemer's method 120
ICP-MS

Estimated mass of Fe (pg/cell)

Estimated Fe (mg/ml) 14
12 100

10 80
8
60
6
4 40
2
20
0
0

Figure 4. Comparison with Riemer’s method. Iron oxide nanoparticles with various formulations (A) and intracellular BNF-Starch iron oxide nanoparticles (B).
Identical samples for each formulation of nanoparticles or DU145 cells loaded with BNF-Starch were prepared in duplicates.

Supplementary Table S4 and Figure 1(C). The average slope among the most widely used and are commercially available.
of total net absorbance for Fe standards is about 1.7 absorb- They are generally optimised for use in either clinical or
ance units per lg of iron with 4.4% coefficient of variation. environmental applications [26,33,36,41–45]. A common fea-
The total amount of iron in an unknown sample could be ture of nearly all spectrophotometer-based methods of iron
determined by dividing the total absorbance units of the mass quantitation is an initial iron ion liberation (“digestion”)
unknown by this slope. The slope of Fe standard curves is process that separates individual iron ions from the biological
highly reproducible and reliable when performed in this or environmental sample matrices. A reducing agent is then
manner. Supplementary Table S7 shows the results from five used to convert the ferric (Fe3þ) iron in the sample to fer-
experiments performed on different days with the mean ± SD rous, Fe2þ, iron. The measured concentration of iron in the
of 1.71 ± 0.06 (3.3% CV). sample will be inaccurate and typically underestimated if any
iron in the sample remains bound to the matrix material, or
remains in the trivalent state (typically the result of vigorous
Discussion
digestion of sample with corrosive agents).
Applications of iron oxide nanoparticles include many exam- Spectrometric techniques, such as ICP-MS or ICP-AES, are
ples in biology and medicine. Characterization of their effects typically sensitive and require no reduction of the iron ions
on biological processes, including toxicity, in cell culture after their liberation. On the other hand, sample preparation
model systems requires accurate determination of nanopar- and instrument configuration can negate the high sensitivity
ticle, that is iron ion concentration to evaluate concentration- by failing to resolve differences with sample components or
dependent effects. For imaging and hyperthermia, accurate other contaminants. Furthermore, equipment costs and infra-
iron quantification in the cell or biological sample material is structure demands make acquisition and maintenance of
required to correctly interpret the translational implications ICP-MS equipment prohibitive thus limiting access for many
of dose–response effects. Many methods are available to research laboratories.
characterise iron concentrations, however the most reliable With the simplified method reported here, which requires
and accurate methods require expensive equipment and reagents and a UV/Vis spectrophotometer, we describe the
elaborate sample preparation protocols. Thus, accurate and preparation of a working solution to which samples can be
reliable iron concentration measurements are inaccessible to directly added; and, after thorough mixing the solution is
researchers who lack the necessary infrastructure and equip- incubated at room temperature overnight. The working solu-
ment. In this report we described the development of a sim- tion comprises a 0.2 M acetate buffer at pH 4.3 and 0.2 M
ple, rapid, low-cost and reliable method to quantify iron ion ascorbic acid acting as the reducing agent. Given adequate
concentration in samples containing iron oxide nanoparticles. time this acidic solution is sufficient to liberate iron ions from
We have optimised the assay for use with various formula- the nanoparticles and the sample matrix. Complete liberation
tions of iron oxide nanoparticles as well as their intracellular of iron ions depends upon the nanoparticle formulation,
concentration. Furthermore, the assay is validated by com- nanoparticle coating(s), and other sample details. It is worth
parison with standard reference materials and ICP-MS. noting that selection of nanoparticles for the present study
Colorimetric methods utilising chromogenic compounds was limited to a subset of commercially available nanopar-
such as ferrozine or ferene to assay iron concentrations are ticle constructs comprising iron oxide and dextran or starch
 INTERNATIONAL JOURNAL OF HYPERTHERMIA 379

coatings or matrices for development of methodology and dependent study matrix to identify the optimal digestion and
validation. Such constructs have demonstrated widespread sample preparation procedure(s). On the other hand, the fer-
utility in biomedical research and clinical applications. ene-s procedure described above demonstrates consistent
All possible and available nanoparticles having similar com- iron ion concentration measurements as validated by com-
position; however, were not tested but methods described parison with ICP-MS for several different iron oxide nanopar-
can be used. Iron oxide nanoparticles comprising inorganic ticles and cell-internalised BNF-Starch nanoparticles (see
coatings or matrices, for example SiO2, Au, Zn, etc., were not Table 1).
tested in the present study. Such particle constructs may To develop our assay, the chromogenic compound ferene-s
require modified digestion reagents and time. In such cases; was chosen because its higher molar absorptivity compared
however, the methods described herein for measurement of to ferrozine (Figure 1, Supplementary Table S1). The molar
liberated iron ion concentrations following appropriate diges- absorptivity was calculated from the Beer–Lambert law:
tion may still be used. A ¼ Elc; where A is the absorbance, E is the molar absorptiv-
For the described samples containing cell matrices, the ity, l is the path length and c is the concentration of the
time needed to achieve complete iron ion liberation also sample. When using a standard UV/Vis cuvette (usually with
depends upon location of nanoparticles outside or within a fixed path length of 1 cm) the light passes through the
intracellular compartments or organelles of the cell. sample horizontally and is therefore independent of the sam-
Absorbance readings measured from the Miltenyi microbeads ple volume. The ferene-s assay protocol reported here uses a
became constant within 3 h after addition to the working 96-well plate format for absorbance readings to facilitate
solution. This was approximately half the time needed for high throughput processing. It is important to note however
BNF-Starch particles (Figure 2(A)). We may consider that such that microplate readers use a vertical light path and the dis-
differences of digestion time may be due to the aggregated tance for light to travel through the sample depends on the
(BNF) vs. dispersed iron oxide crystallite (Miltenyi beads) con- sample volume. To compare the absorbance readings
figurations of the nanoparticle constructs. Further investiga- between standard 1-ml UV/Vis cuvettes and 96-well plates
tion, however is needed to determine the cause of the we performed four independent experiments with iron stand-
difference in digestion times. When associated with cellular ards ranging in concentration from 0 to 4.0 lg Fe/ml. In each
material, however the BNF-Starch particles required >12 h to experiment the same sample was used for both readings.
achieve constant values (Figure 2(B)). FeCl3 used as standard
The results are shown in Supplementary Tables S3 and S4.
reached saturation in fewer than 30 min (Supplementary
Absorbance readings were about 20% lower when measured
Figure S1) suggesting that reduction of Fe3þ to Fe2þ by
with 96-well plates. This is expected because the approxi-
ascorbic acid is not the limiting factor. Instead, liberation of
mate height of the 300 ll solution in each well is 0.88 cm.
iron ions from within the nanoparticle or sample matrix
When adjusted to a height of 1 cm the readings remained
appears to vary significantly. For faster processing, the sam-
about 8–9% lower than when measured with 1-ml cuvettes.
ples can be digested with concentrated nitric acid prior to
The plate reader used for our absorbance measurements
addition to the working solution. As shown in Figure 3, a 1-h
incorporates a path-check option in its software which auto-
digestion with acid at 70  C was sufficient for complete
matically adjusts the absorbance readings to match the read-
release of iron from intracellular BNF-Starch nanoparticles.
ings from a 1-ml cuvette. When using this option the
When using nitric acid to effect sample digestions, it is
absorbance readings from the 96-well plates were similar to
important to neutralise the digested sample solution with
those obtained using a 1-ml cuvette (compare
base, for example NaOH, in order to maintain pH ¼ 3–6
Supplementary Tables S3 and S5).
before adding to the working solution [32].
If available, the use of this option in the UV/Vis spectro-
The method described by Riemer et al. [34] utilises a
potassium permanganate and HCl mixture to effect sample photometer enables more consistent results. This option
digestion (2 h incubation at 60  C). Comparison of results however may be unavailable with some models of plate
obtained by the ferene-s and Riemer methods in Figure 4(A) readers therefore a modified method is needed to obtain
and (B) demonstrates that the latter method may significantly consistency among experimental values. We described a pro-
underestimate the iron concentration depending upon the cedure to account for variations of path length that employs
nanoparticle construct under study. Thus, caution should be use of the slope of absorbance calibration curve obtained
exercised when interpreting results obtained by Riemer’s from the reference standard solutions, to yield reproducible
method for iron oxide nanoparticles. When compared against and reliable measurements.
ICP-MS (see Figure 4(B)), the consistency between the ferene-
s and ICP-MS values implies that the lower iron concentration Conclusion
obtained using Riemer’s method may be due to incomplete
iron ion liberation. With this method, longer digestion times We present a detailed protocol for iron concentration deter-
or digestion by a concentrated acid may be necessary to mination using a ferene-s assay. The results were obtained
avoid underestimating iron content in the sample. Further, from several iron oxide nanoparticle constructs used for bio-
absent a validation with an orthogonal and sensitive method- medical applications. We also present results of iron concen-
ology, that is ICP-MS, it is recommended to evaluate concen- tration measurements obtained from cell culture comparing
tration measurements using Riemer’s method following an three methods to demonstrate that, depending upon
analysis of results obtained from a time- and concentration- method chosen, sample digestion may yield unreliable
380 M. HEDAYATI ET AL.

results. The ferene-s method developed was validated against sciences: evidence from analysis of selected USGS reference
ICP-MS demonstrating reproducible measurements of Fe samples. Chem Geol 83:133–48.
[14] Elsaesser A, Taylor A, de Yan
es, et al. (2010). Quantification of
concentrations to 2  106 M (1  107 g/ml sample).
nanoparticle uptake by cells using microscopical and analytical
The reliability and convenience of the ferene-s protocol dem- techniques. Nanomedicine 5:1447–57.
onstrated can facilitate rapid and accurate iron oxide nano- [15] Patil US, Adireddy S, Jaiswal A, et al. (2015). In vitro/in vivo tox-
particle concentration measurements for a range of icity evaluation and quantification of iron oxide nanoparticles. Int
biological and toxicological experiments utilising cell culture J Mol Sci 16:24417–50.
[16] Rimkus G, Bremer-Streck S, Gr€ uttner C, et al. (2011). Can we
models.
accurately quantify nanoparticle associated proteins when con-
structing high-affinity MRI molecular imaging probes? Contrast
Media Mol Imaging 6:119–25.
Disclosure statement [17] DeNardo SJ, DeNardo GL, Miers LA, et al. (2005). Development of
Dr. Cordula Gruettner is an employee of micromod Partikeltechnologie, tumor targeting bioprobes (111In-Chimeric L6 monoclonal anti-
GmbH, manufacturer of BNF and nanomagV-D-spio particles used in the body nanoparticles) for alternating magnetic field cancer therapy.
R

studies. Dr. Robert Ivkov is an inventor on several issued and pending Clin Cancer Res 11:7087s–92s.
patents. All patents are assigned to Johns Hopkins University or Aduro [18] DeNardo SJ, DeNardo GL, Natarajan A, et al. (2007). Thermal dos-
Biotech, Inc. All other authors declare no competing interests. imetry predictive of efficacy of 111In-ChL6 nanoparticle AMF-
induced thermoablative therapy for human breast cancer in mice.
J Nucl Med 48:437–444.
Funding [19] Natarajan A, Gruettner C, Ivkov R, et al. (2008). NanoFerrite par-
ticle based radioimmunonanoparticles: binding affinity and in
This work was supported by an award from the Prostate Cancer vivo pharmacokinetics. Bioconjugate Chem 19:1211–18.
Foundation/Safeway-STAR. ICP-MS work was supported in part by the [20] Gr€uttner C, M€uller K, Teller J, et al. (2007). Synthesis and antibody
Maryland Cigarette Restitution Fund Program at Johns Hopkins conjugation of magnetic nanoparticles with improved specific
Bloomberg School of Public Health and the NIEHS Center P30 ES00319. power absorption rates for alternating magnetic field cancer ther-
apy. J Magn Mag Mat 311:181–6.
[21] Baiu DC, Artz NS, McElreath MR, et al. (2015). High specificity tar-
References geting and detection of human neuroblastoma using multifunc-
tional anti-GD2 iron-oxide nanoparticles. Nanomedicine (Lond)
[1] Gupta AK, Gupta M. (2005). Synthesis and surface engineering of
10:2973–88.
iron oxide nanoparticles for biomedical applications. Biomaterials
[22] Behnam Azad B, Banerjee SR, Pullambhatha M, et al. (2015).
26:3995–4021.
Evaluation of a PSMA-targeted BNF nanoparticle construct.
[2] Johannsen M, Gneveckow U, Eckelt L, et al. (2005). Clinical hyper-
Nanoscale 7:4432–42.
thermia of prostate cancer using magnetic nanoparticles: presen-
[23] Dennis CL, Jackson AJ, Borchers JA, et al. (2009). Nearly complete
tation of a new interstitial technique. Int J Hyperther 21:637–47.
regression of tumors via collective behavior of magnetic nanopar-
[3] Lee H-Y, Li Z, Chen K, et al. (2008). PET/MRI dual-modality tumor
imaging using arginine-glycine-aspartic (RGD)–conjugated radiola- ticles in hyperthermia. Nanotechnology 20:395103.
beled iron oxide nanoparticles. J Nucl Med 49:1371–9. [24] Laurent S, Dutz S, H€afeli UO, Mahmoudi M. (2011). Magnetic fluid
[4] Huang H, Delikanli S, Zen H, et al. (2010). Remote control of ion hyperthermia: focus on superparamagnetic iron oxide nanopar-
channels and neurons through magnetic-field heating of nano- ticles. Adv Colloid Interface Sci 166:8–23.
particles. Nat Nanotechnol 5:602–6. [25] Wabler M, Zhu W, Hedayati M, et al. (2014). Magnetic resonance
[5] Maier-Hauff K, Ulrich F, Nestler D, et al. (2011). Efficacy and safety imaging contrast of iron oxide nanoparticles developed for hyper-
of intratumoral thermotherapy using magnetic iron-oxide nano- thermia is dominated by iron content. Int J Hyperthermia
particles combined with external beam radiotherapy on patients 30:192–200.
with recurrent glioblastoma multiforme. J Neurooncol [26] Hussain SM, Hess KL, Gearhart JM, et al. (2005). In vitro toxicity
103:317–24. of nanoparticles in BRL 3A rat liver cells. Toxicol In Vitro 19:975–83.
[6] McCormack PL. (2012). Ferumoxytol: in iron deficiency anaemia in [27] Xu H, Aguilar ZP, Yang L, et al. (2011). Antibody conjugated mag-
adults with chronic kidney disease. Drugs 72:2013–22. netic iron oxide nanoparticles for cancer cell separation in fresh
[7] Kirschbaum K, Sonner JK, Zeller MW, et al. (2016). In vivo nano- whole blood. Biomaterials 32:9758–65.
particle imaging of innate immune cells can serve as a marker of [28] Kasten A, Gr€ uttner C, K€uhn J-P, et al. (2014). Comparative in vitro
disease severity in a model of multiple sclerosis. Proc Nat Acad study on magnetic iron oxide nanoparticles for MRI tracking of
Sci 113:13227–32. adipose tissue-derived progenitor cells. PLoS One 9:e108055.
[8] Liu Y, Yang Y, Zhang C. (2013). A concise review of magnetic res- [29] Kilian T, Fidler F, Kasten A, et al. (2016). Stem cell labeling with
onance molecular imaging of tumor angiogenesis by targeting iron oxide nanoparticles: impact of 3D culture on cell labeling
integrin avb3 with magnetic probes. Int J Nanomed 8:1083–93. maintenance. Nanomedicine (Lond.) 11:1957–70.
[9] Xu C, Sun S. (2013). New forms of superparamagnetic nanopar- [30] Pro€frock D, Prange A. (2012). Inductively coupled plasma-mass
ticles for biomedical applications. Adv Drug Deliv Rev 65:732–43. spectrometry (ICP-MS) for quantitative analysis in environmental
[10] Bullivant JP, Zhao S, Willenberg BJ, et al. (2013). Materials charac- and life sciences: a review of challenges, solutions, and trends.
terization of Feraheme/Ferumoxytol and preliminary evaluation of Appl Spectrosc 66:843–68.
its potential for magnetic fluid hyperthermia. Int J Mol Sci [31] Stookey LL. (1970). Ferrozine—a new spectrophotometric reagent
14:17501–10. for iron. Anal Chem 42:779–81.
[11] Marchal S, El Hor A, Millard M, et al. (2015). Anticancer drug deliv- [32] Hennessy DJ, Reid GR, Smith FE, Thompson SL. (1984). Ferene – a
ery: an update on clinically applied nanotherapeutics. Drugs new spectrophotometric reagent for iron. Can J Chem 62:721–4.
75:1601–11. [33] Fish WW. (1988). Rapid colorimetric micromethod for the quanti-
[12] Bashir MR, Bhatti L, Marin D, Nelson RC. (2015). Emerging applica- tation of complexed iron in biological samples. Methods Enzymol
tions for ferumoxytol as a contrast agent in MRI. J Magn Reson 158:357–64.
Imaging 41:884–98. [34] Riemer J, Hoepken HH, Czerwinska H, et al. (2004). Colorimetric
[13] Jenner GA, Longerich HP, Jackson SE, Fryer BJ. (1990). ICP-MS—a ferrozine-based assay for the quantitation of iron in cultured cells.
powerful tool for high-precision trace-element analysis in earth Anal Biochem 331:370–5.
 INTERNATIONAL JOURNAL OF HYPERTHERMIA 381

[35] Rad AM, Janic B, Iskander ASM, et al. (2007). Measurement of [40] Hedayati M, Thomas O, Abubaker-Sharif B, et al. (2013). The effect of
quantity of iron in magnetically labeled cells: comparison among cell cluster size on intracellular nanoparticle-mediated hyperthermia:
different UV/VIS spectrometric methods. Biotechniques 43:627–35. is it possible to treat microscopic tumors? Nanomedicine 8:29–41.
[36] Dadashzadeh ER, Hobson M, Bryant LH, et al. (2013). Rapid spec- [41] Derman DP, Green A, Bothwell TH, et al. (1989). A systematic
trophotometric technique for quantifying iron in cells labeled evaluation of bathophenanthroline, ferrozine and ferene in an
with superparamagnetic iron oxide nanoparticles: potential trans- ICSH-based method for the measurement of serum iron. Ann Clin
lation to the clinic. Contrast Media Mol Imaging 8:50–6. Biochem 26:144–7.
[37] Bordelon DE, Cornejo C, Gruttner C, et al. (2011). Magnetic nano- [42] Carter P. (1971). Spectrophotometric determination of serum iron
particle heating efficiency reveals magneto-structural differences at the submicrogram level with a new reagent (ferrozine). Anal
when characterized with wide ranging and high amplitude alter- Biochem 40:450–8.
nating magnetic fields. J Appl Phys 109:124904. [43] Gibbs CR. (1976). Characterization and application of ferrozine
[38] Dennis CL, Krycka KL, Borchers JA, et al. (2015). Internal magnetic iron reagent as a ferrous iron indicator. Anal Chem 48:1197–201.
structure of nanoparticles dominates time-dependent relaxation [44] Artiss JD, Vinogradov S, Zak B. (1981). Spectrophotometric study
processes in a magnetic field. Adv Funct Mater 25:4300–11. of several sensitive reagents for serum iron. Clin Biochem
[39] Gr€ utt W, et al. (1997). Preparation and char-
uttner C, Teller J, Sch€ 14:311–15.
acterization of magnetic nanospheres for in vivo application. In: [45] Pieroni L, Khalil L, Charlotte F, et al. (2001). Comparison of batho-
Hafeli UO, Sch€utt W, Teller J, Zborowski M, eds. Scientific and clin- phenanthroline sulfonate and ferene as chromogens in colorimet-
ical application of magnetic carriers. New York: Plenum Press, ric measurement of low hepatic iron concentration. Clin Chem
53–68. 47:2059–61.