Rational evaluation of the therapeutic effect and dosimetry of auger electrons for radionuclide therapy in a cell culture model

Article abstract of DOI:10.1007/s12149-017-1225-9

Objective: Radionuclide therapy with low-energy auger electron emitters may provide high antitumor efficacy while keeping the toxicity to normal organs low. Here we evaluated the usefulness of an auger electron emitter and compared it with that of a beta emitter for tumor treatment in in vitro models and conducted a dosimetry simulation using radioiodine-labeled metaiodobenzylguanidine (MIBG) as a model compound. Methods: We evaluated the cellular uptake of ¹²⁵I-MIBG and the therapeutic effects of ¹²⁵I- and ¹³¹I-MIBG in 2D and 3D PC-12 cell culture models. We used a Monte Carlo simulation code (PHITS) to calculate the absorbed radiation dose of ¹²⁵I or ¹³¹I in computer simulation models for 2D and 3D cell cultures. In the dosimetry calculation for the 3D model, several distribution patterns of radionuclide were applied. Results: A higher cumulative dose was observed in the 3D model due to the prolonged retention of MIBG compared to the 2D model. However, ¹²⁵I-MIBG showed a greater therapeutic effect in the 2D model compared to the 3D model (respective EC₅₀ values in the 2D and 3D models: 86.9 and 303.9?MBq/cell), whereas ¹³¹I-MIBG showed the opposite result (respective EC₅₀ values in the 2D and 3D models: 49.4 and 30.2?MBq/cell). The therapeutic effect of ¹²⁵I-MIBG was lower than that of ¹³¹I-MIBG in both models, but the radionuclide-derived difference was smaller in the 2D model. The dosimetry simulation with PHITS revealed the influence of the radiation quality, the crossfire effect, radionuclide distribution, and tumor shape on the absorbed dose. Application of the heterogeneous distribution series dramatically changed the radiation dose distribution of ¹²⁵I-MIBG, and mitigated the difference between the estimated and measured therapeutic effects of ¹²⁵I-MIBG. Conclusions: The therapeutic effect of ¹²⁵I-MIBG was comparable to that of ¹³¹I-MIBG in the 2D model, but the efficacy was inferior to that of ¹³¹I-MIBG in the 3D model, since the crossfire effect is negligible and the homogeneous distribution of radionuclides was insufficient. Thus, auger electrons would be suitable for treating small-sized tumors. The design of radiopharmaceuticals with auger electron emitters requires particularly careful consideration of achieving a homogeneous distribution of the compound in the tumor.

Full text of DOI:10.1007/s12149-017-1225-9

Annals of Nuclear Medicine
https://doi.org/10.1007/s12149-017-1225-9
ORIGINAL ARTICLE
Rational evaluation of the therapeutic effect and dosimetry of auger
electrons for radionuclide therapy in a cell culture model
1
2
3
4
2
Ayaka Shinohara · Hirofumi Hanaoka · Tetsuya Sakashita · Tatsuhiko Sato · Aiko Yamaguchi ·
3
5,6
Noriko S. Ishioka · Yoshito Tsushima
Received: 20 October 2017 / Accepted: 7 December 2017
The Japanese Society of Nuclear Medicine 2017
©
Abstract
Objective Radionuclide therapy with low-energy auger electron emitters may provide high antitumor efficacy while keeping
the toxicity to normal organs low. Here we evaluated the usefulness of an auger electron emitter and compared it with that
of a beta emitter for tumor treatment in in vitro models and conducted a dosimetry simulation using radioiodine-labeled
metaiodobenzylguanidine (MIBG) as a model compound.
1
25
125
131
Methods We evaluated the cellular uptake of I-MIBG and the therapeutic effects of I- and I-MIBG in 2D and 3D
1
25
PC-12 cell culture models. We used a Monte Carlo simulation code (PHITS) to calculate the absorbed radiation dose of
I
1
31
or I in computer simulation models for 2D and 3D cell cultures. In the dosimetry calculation for the 3D model, several
distribution patterns of radionuclide were applied.
Results A higher cumulative dose was observed in the 3D model due to the prolonged retention of MIBG compared to the
1
25
2
D model. However, I-MIBG showed a greater therapeutic effect in the 2D model compared to the 3D model (respective
1
31
EC values in the 2D and 3D models: 86.9 and 303.9 MBq/cell), whereas I-MIBG showed the opposite result (respective
5
0
1
25
EC values in the 2D and 3D models: 49.4 and 30.2 MBq/cell). The therapeutic effect of I-MIBG was lower than that of
5
0
1
31
I-MIBG in both models, but the radionuclide-derived difference was smaller in the 2D model. The dosimetry simulation
with PHITS revealed the influence of the radiation quality, the crossfire effect, radionuclide distribution, and tumor shape on
the absorbed dose. Application of the heterogeneous distribution series dramatically changed the radiation dose distribution
1
25
125
of I-MIBG, and mitigated the difference between the estimated and measured therapeutic effects of I-MIBG.
1
25
131
Conclusions The therapeutic effect of I-MIBG was comparable to that of I-MIBG in the 2D model, but the efficacy
1
31
was inferior to that of I-MIBG in the 3D model, since the crossfire effect is negligible and the homogeneous distribution
of radionuclides was insufficient. Thus, auger electrons would be suitable for treating small-sized tumors. The design of
radiopharmaceuticals with auger electron emitters requires particularly careful consideration of achieving a homogeneous
distribution of the compound in the tumor.
Keywords Radionuclide therapy · Auger electron emitter · 125/131I-Metaiodobenzylguanidine (125/131I-MIBG) · 3D cell
culture model · Computer simulation
Introduction
Based on the concept of delivering radionuclides selectively
to tumor sites while sparing normal tissues from radiation
toxicity, radionuclide therapy has been an attractive cancer
treatment for decades. The clinical success of radionuclide
therapy with β-emitters for certain types of cancer spurred
the demand for the expanded application of radionuclide
therapy. However, the administration of a sufficiently tumori-
cidal dose is sometimes restricted due to the maximum toler-
ant dose limitation. This is partially because of the relatively
Electronic supplementary material The online version of this
article (https://doi.org/10.1007/s12149-017-1225-9) contains
supplementary material, which is available to authorized users.
*
Hirofumi Hanaoka
hanaokah@gunma-u.ac.jp
Extended author information available on the last page of the article
Vol.:(0
1
123456789)
3

Annals of Nuclear Medicine
1
31
low linear energy transfer (LET) (<0.2 keV/μm) and long
path length (1–10 mm in tissue) of β-rays. In particular,
when the tumor size is smaller than the path length of β-rays,
a substantial amount of the radiation energy escapes from
the tumor site and causes normal tissue toxicity [1]. These
limitations have led to the consideration of α-emitters as the
radionuclides of choice for radionuclide therapy. The higher
LET (−80 keV/μm) and shorter path length (50–100 μm) of
α-particles compared to β-rays brought about improved ther-
apeutic effects for micrometastases and disseminated tumors
analogs, labeled with either β-emitting I or auger electron-
1
25
emitting I, and in light of MIBG’s well-defined intracel-
lular localization [16]. We evaluated the therapeutic effects
1
25
131
of I-MIBG and I-MIBG in two-dimensional (2D) and
three-dimensional (3D) cell culture models. The 3D cell cul-
ture model, also known as a tumor spheroid, well simulates
in vivo tumor conditions better than 2D cell culture model,
and the 3D cell culture model can thus accommodate the
crossfire effect to some extent. We estimated the absorbed
radiation dose of each condition using the Monte Carlo code
for particle transport simulation (i.e., Particle and Heavy
Ion Transport code System, PHITS) which can analyze the
transport of almost all radiation particles in 3D matter [17].
Based on the results obtained, we discuss the usefulness
and suitable applications of an auger electron emitter for
tumor treatment. A new evaluation method to predict the
therapeutic effect of radionuclide therapy is also proposed.
[
2–4]. However, the widespread application of α-emitters
has not been achieved because of these emitters’ limited
production and availability.
Another potential candidate is low-energy auger elec-
trons. Auger electrons have a tenfold higher LET than β-rays
(
−26 keV/μm), and their simultaneous action can cause a
dense ionization within several cubic nanometers around the
point of decay. The dense reactive chemical species pro-
duced by auger electrons can cause biological effects similar
to those of α-particles [1, 5–7]. In addition, the commercial
Materials and methods
1
23 125
availability of auger electron emitters, such as I, I, and
1
11
In, gives auger electrons an additional advantage over
α-emitters.
DNA-incorporated, auger electron emitter-labeled deoxy-
General
1
25
Na I (carrier-free, specific activity approx. 650 GBq/mg
as iodine) was purchased from American Radiolabeled
uridine derivatives have indeed demonstrated very high radi-
otoxicity in vitro [8, 9] and they showed several successful
clinical results in patients with cancer (including otherwise
incurable metastatic pancreatic cancer in the central nerv-
ous system) [10]. Despite these encouraging results, further
foundational evaluations are needed for wide acceptance of
the use of auger electron emitters for radionuclide therapy
in clinical settings.
1
31
Chemicals (St. Louis, MO), and Na I (specific activ-
ity > 185 GBq/mg as iodine) was purchased from Perki-
nElmer (Waltham, MA). The reversed-phase high-perfor-
mance liquid chromatography (RP-HPLC) analysis was
performed with a C-18 column (Capcell Pak C18 MGII,
4.6 × 150 mm, Shiseido, Tokyo) at the flow rate of 1 ml/
min eluted with a linear gradient of water containing 0.1%
trifluoroacetic acid (TFA) and acetonitrile containing 0.1%
TFA from 90:10 to 10:90 in 30 min. All other chemicals
Previous comparisons of the therapeutic effect of auger
electron emitters and β-emitters in in vivo xenograft mod-
els have reported conflicting results [1, 11]. Because the
therapeutic effect of each radionuclide depends on various
factors, especially the tumor size and the intra-tumor distri-
bution of radiopharmaceuticals [12], it would be impracti-
cal to attempt to comprehensively evaluate the therapeutic
effects of each radiopharmaceutical in every possible condi-
tion in in vivo models. Therefore, the influence of the radia-
tion quality and dose on the biological radiotoxicity should
be investigated using a more simplified and generalized
method. In vitro cell culture models allow us to evaluate the
therapeutic effect of each radionuclide under uniform condi-
tions [2, 4, 9, 13]. Moreover, with the use of the obtained
dose–response relationship of each radionuclide, computer
simulations enable the estimation of the tumor-absorbed
dose [14, 15], and thus the therapeutic effect of radionuclide
therapy in individual tumors in various conditions.
1
used were of the highest purity available. H NMR spec-
tra were obtained on an ECA-600 (600 MHz) spectrometer
(JEOL, Tokyo), and chemical shifts were identified using
NMR predictor software (ChemAxon, Budapest, Hungary).
Electrospray ionization mass spectroscopy (ESI-MS) data
were obtained with a model LCMS-2020 mass spectrometer
(Shimadzu, Kyoto, Japan).
Preparation of 125I‑ and 131I‑MIBG
The stannyl precursor of radiolabeled MIBG (N,N′-
bis(tert-butyloxycarbonyl)-3-(trimethylstannyl) benzyl-
guanidine) was synthesized according to the procedure
of Vaidyanathan et al. [18] with some modifications.
N,N′-bis(tert-butyloxycarbonyl)-3-iodobenzylguanidine
(143.6 mg, 0.30 mmol) in 10 ml of anhydrous toluene was
added with (Ph P) Pd (34.7 mg, 0.03 mmol) and hexam-
In this study, we selected radioiodine-labeled metaiodo-
benzylguanidine (MIBG) as a model compound due to the
high availability of its chemically and biologically similar
3
4
ethylditin (412.8 mg, 1.26 mmol) and refluxed under a N
2
atmosphere for 2 h. The reaction mixture was cooled to
1
3

Annals of Nuclear Medicine
room temperature, filtered over a Celite pad, and washed
with ethyl acetate. The filtrate was concentrated and puri-
fied by column chromatography using a hexane–ethyl
In vitro cellular therapeutic studies
For the evaluation of the added radioactivity-dependent
1
25
131
acetate gradient to give 37.1 mg (24.2%) of the stannyl
cytotoxic effect of I-MIBG and I-MIBG in vitro, we
performed cellular therapeutic studies in the 2D and 3D
cell culture models. PC-12 cells were prepared as described
1
precursor as a colorless oil: H NMR (CDCl ): 0.27 (s,
3
9
H), 1.35 (s, 9H), 1.49 (s, 9H), 5.16 (s, 2H), 7.20–7.22
1
25
(
d, 2H), 7.35–36 (d, 1H), 7.42 (s, 1H), 9.31–48 (d, 2H).
above, and then 100 µl of medium-containing I-MIBG
+
131
ESI-MS Calc’d for C H N O Sn (M + H) : m/z 514.2,
or
I-MIBG (0.10, 0.30, 1.0, 3.0, 10, or 30 MBq/ml)
2
1
36
3
4
found: 514.2.
was added to each well of 96-well plates, and the plates
were incubated for 3 days at 37 °C. After the plates were
washed with PBS, the cell viability was measured with a
Cell Counting Kit-8 (Dojindo, Kumamoto, Japan) accord-
ing to the manufacturer’s protocol. The results of this assay
are presented as the % cell viability relative to non-treated
controls. The total effective radioactivity within a single cell
(MBq/cell) was calculated using the cumulative % uptake of
radioactivity and analyzed by sigmoidal added radioactivity-
response fitting using SigmaPlot ver. 11.
For the radioiodination, 100 µl of the stannyl precur-
sor (1 mg/ml) dissolved in methanol containing 1% acetic
1
25/131
125
acid was mixed with 2–10 µl of Na
I solution ( I:
1
31
1
11.0 MBq, I: 124.2 MBq) and 5 µl of N-chlorosuc-
cinimide (10 mg/ml) in methanol in a small vial, and the
reaction mixture was incubated at room temperature for
1
0 min. Then, 100 µl of 6 M aqueous HCl was added, and
the reaction mixture was incubated at room temperature for
6
h. After the pH of the reaction mixture was neutralized,
purification was performed by RP-HPLC, and the solvent
was removed in vacuo. The radiochemical purity of MIBGs
was determined by RP-HPLC.
Absorbed radiation dose estimation by simulation
with PHITS
We estimated the absorbed radiation dose of each model
and each radioiodine-labeled MIBG with the use of PHITS,
a three-dimensional Monte Carlo particle transport simula-
tion code that simulates various radiation behaviors in sub-
stances using nuclear reaction models as well as nuclear and
atomic data libraries. PHITS has been used in fields such as
engineering, medicine, and science [17, 19]. The accuracy
of PHITS for the dosimetry of beta-emitting isotopes was
well validated [20].
In vitro cellular uptake studies
1
25
Cellular uptake studies of I-MIBG were performed in
both a 2D cell culture model and a 3D cell culture model.
For the 3D cell culture model, Nunclon Sphera microplates
(
Thermo Fisher Scientific, Waltham, MA), which support
the consistent formation of tumor spheroids, were used. The
rat pheochromocytoma cell line PC-12 (American Type
Culture Collection ATCC, Manassas, VA) were seeded
Briefly, to estimate the absorbed radiation dose per analy-
sis unit (Gy/(Bq/mL)) by PHITS, we defined the simula-
tion areas for 2D and 3D simulation models. The simulation
models were designed with Fiji, an open source image-pro-
cessing distribution of ImageJ (http://imagej.net/Fiji). From
the experimentally determined diameter of the PC-12 cells
(20 μm), each cell was assumed as a cube 20 μm on a side
5
into 2D and 3D 96-well plates (1.0 × 10 cells/well), then
pre-incubated overnight. The disk-shaped tumor spheroids
approx. 2 mm in diameter were formed in the 3D plates.
1
25
Medium-containing I-MIBG (5 kBq, 100 µl) was added
to each well, and the cells were incubated for 0, 10, 30 min,
1
, 3, 6 h, 1, 2, and 3 days at 37 °C. Cells were washed once
3
with phosphate-buffered saline (PBS) and lysed with 0.2 M
NaOH. The radioactivity of each cell lysate was measured
by a well-type gamma counter (ARC-7001; Hitachi Aloka
Medical, Tokyo). In addition, the protein concentration of
each well was measured using the Modified Lowry Protein
Assay Kit (Thermo Fisher Scientific) according to the manu-
facturer’s protocol.
(volume 8000 μm ) (Fig. 1a). Using this cube as a standard
voxel, the matrix size was determined as 200×200×75 (4
mm×4 mm×1.5 mm). Based on the size of the tumor 3D
spheroid (2 mm diameter and 0.5 mm height, Fig. 1b), the
2D tumor area was placed at the center of the bottom of the
matrix within a radius of 1 mm (a disk comprised 7,860 cell
cubes). Since the height of the tumor spheroid was approx.
0.5 mm, the 3D tumor area was defined as a cylinder com-
prised 25 layers of the disk (196,500 cell cubes) (Fig. 1a).
To take into account the radiation exposure from the radio-
nuclide remaining in the medium, we defined the outside
area of each tumor model in the matrix as the culture area.
We next determined the absorbed radiation dose per anal-
1
25
The uptake of I-MIBG was calculated as the percent-
age of the added activity/mg protein, and the time activity
curve was plotted. The cumulative % uptake of radioactiv-
ity within the observation period (3 days, % dose/mg pro-
tein×day) was calculated by applying a lognormal peak fit-
ting curve model using SigmaPlot ver. 11 (Systat Software,
San Jose, CA). The same calculations were performed for
1
25
131
ysis unit (Gy/(Bq/mL)) in each tumor area for I or
I
1
31
I-MIBG, taking decay correction into consideration.
using PHITS, with the assumption that each radionuclide is
1
3

Annals of Nuclear Medicine
Fig. 1 The cylindrical tumor area and the culture area for the PHITS simulation (a) and the spheroid of PC-12 cells (b)
distributed either homogeneously or heterogeneously. For
the heterogeneous distribution model, it is assumed that
MIBG is penetrated from the surface of the spheroid and is
not absorbed from the bottom. We created an exponential
radioactivity decreasing map with the following equation:
N = N × exp(− ln 2/D × d) (Suppl. Fig. S1). Here, N is
determine the absorbed dose–response relationship, the
obtained radiation doses in the 2D simulation model were
plotted against the surviving fraction (% cell viability) from
the therapeutic studies, and a fitting curve [SF=exp(−ln 2/
LD ×D), where SF is the surviving fraction, LD is the
5
0
50
median lethal dose (Gy), and D is the absorbed dose (Gy)]
obtained using ORIGIN (MicroCal Software, Northampton,
MA). The surviving fractions in the 3D model were then
estimated for each radioactivity using the absorbed radia-
tion dose obtained from the 3D simulation models and the
formula as described above. We compared these values with
the experimental therapeutic results. For the heterogeneous
distribution model, we calculated the % cell viability for
each cube, and the averaged values of all cubes were deter-
mined as the survival fraction.
0
1/2
the relative radioactivity in each voxel at the depth d, N
is the radioactivity at the outer spheroid surface (=1), D
0
1
/2
is the half-value depth (μm), and d is the depth from the
surface (μm). The dose distribution was calculated for D
1
/2
=
10, 30, 50, and 250 μm. The intensity of radioactivity in
each cell was normalized according to the predetermined
total intensity (196,500). The total intensity was defined by
setting the intensity of radioactivity in each cell to 1 in the
homogeneous distribution model.
The absorbed radiation dose of tumor from the radionu-
clide in the tumor and medium was determined individually,
and then combined to calculate the total absorbed radiation
dose. The concentration of intratumoral radioactivity was
calculated based on the results from the abovementioned
cellular uptake studies (cumulated radioactivity/cellular vol-
ume, Bq/mL). Finally, the absorbed radiation dose (Gy) of
each condition was calculated by multiplying these values
and the absorbed radiation dose per analysis unit (Gy/(Bq/
mL)) in PHITS.
Statistical analysis
Data are expressed as mean± SD. The results were analyzed
using the unpaired t test. Differences were considered sig-
nificant when p values were <0.05.
Results
The micro radiation dose from point radiation source of
In vitro cellular uptake studies
1
25
I placed at the center of sphere with a radius of 50 μm was
1
25
also estimated using PHITS. From the obtained estimates,
the absorbed radiation dose distribution was plotted against
the distance.
For the initial few hours, the cellular uptake of I-MIBG
was increased in a time-dependent manner and reached
a peak at 6 h in both the 2D and 3D cell culture mod-
els (Fig. 2). In the 2D cell culture model, the cellular
radioactivity was sharply decreased by 24 h to < 50% of
the peak, and it became stable after that. In contrast, the
cellular radioactivity in the 3D cell culture model was
gradually decreased up to 72 h, and 50% of the radioactiv-
ity was finally released. Accordingly, the cumulative %
Survival estimation of the 3D model by computer
simulation
To evaluate the applicability of PHITS for the prediction
of the therapeutic effect in the more complicated biologi-
cal situation, we performed survival estimation studies. To
1
25
uptake value of the radioactivity of I-MIBG and that of
1
3

Annals of Nuclear Medicine
In vitro cellular therapeutic study
In our cellular therapeutic studies, we observed radioactiv-
ity-dependent cytotoxic effects in each tracer and cell culture
model combination, but the effective radioactivity is dis-
tinct in each setting. At an added radioactivity>1.0 MBq/
1
25
ml, I-MIBG reduced the cell viability in the 2D cell cul-
ture significantly more than in the 3D cell culture (p<0.05;
1
31
Fig. 3a). I-MIBG showed an inverse effect: the reduction
rate of cell viability was significantly greater in the 3D cell
culture than that in the 2D cell culture at an added radioac-
tivity>3.0 MBq/ml (p<0.05; Fig. 3b). We calculated the
total effective radioactivity within the cells by multiply-
ing each added radioactivity and the cumulative % uptake,
and then plotted these values against the therapeutic effect
1
25
Fig. 2 Time course of the cellular uptake of I-MIBG in PC-12 cells
in the 2D and 3D culture models. All results are mean± SD (n≥5).
The fitting curves for the 2D cell culture model (solid line) and 3D
cell culture model (dashed line) are shown
(
Fig. 3c, d).
1
25
The opposite therapeutic effect of
I-MIBG and
1
31
3
131
I-MIBG was higher in the 3D culture model (2.58 × 10
I-MIBG was also observed even after effective radioac-
3
and 2.30 × 10 % dose/mg protein × day, respectively)
tivity correction was applied. Based on the sigmoidal fit-
ting curve, the EC (half-maximal effective concentration)
3
than those in the 2D cell culture model (1.59 × 10 and
5
0
3
125
1
.43 × 10 % dose/mg protein × day, respectively).
values of I-MIBG in the 2D and 3D cell culture models
were 86.9 and 303.9 MBq/cell, respectively (Fig. 3c), and
1
31
those of I-MIBG were 49.4 and 30.2 MBq/cell, respec-
tively (Fig. 3d).
1
25
Fig. 3 Cell viability of PC-12 cells treated with
I-MIBG (a)
for 3 days in the 2D and 3D culture models against the total effective
radioactivity within the cells. All results are mean± SD (n≥5). The
fitting curves for the 2D cell culture model (solid line) and 3D cell
culture model (dashed line) are shown
131
or I-MIBG (b) in the 2D and 3D culture models. All results are
mean± SD (n≥5). *p<0.05 (2D vs. 3D cell culture model). Cell via-
1
25
131
bility of PC-12 cells treated with I-MIBG (c) and I-MIBG (d)
1
3

Annals of Nuclear Medicine
Absorbed radiation dose estimation by simulation
with PHITS
The representative absorbed dose distribution image from
1
25 131
the intratumoral I/ I-MIBG in the homogeneous and
heterogeneous 3D distribution model was shown in Suppl.
Fig. S2 and S3, respectively. In the homogeneous distri-
bution model, the calculated averaged absorbed dose per
1
25
analysis unit of I in the 3D simulation model was compa-
rable to that in the 2D simulation model [Suppl. Table S1,
−
12
− 12
1
.96 × 10
and 1.89 × 10
Gy/(Bq/ml), respectively],
1
31
whereas the average absorbed dose of I in the 3D simu-
lation model was much higher than that in the 2D simula-
−
12
−12
tion model [Suppl. Table S1, 23.6×10
and 12.4×10
Fig. 4 Surviving fraction of therapeutic studies with the absorbed
Gy/(Bq/ml), respectively]. In the 3D models, the averaged
absorbed dose was also estimated layer by layer. In case of
dose in the 2D model. The solid line is the unified fitting curve for
1
25
131
I-MIBG and I-MIBG
125
I, there was little difference in the averaged absorbed dose
−
12
between each layer (Suppl. Table S2), 1.88–1.98 × 10
Gy/(Bq/ml). In contrast, the averaged absorbed dose of
those values in the half-value depth of the 50-µm model
were well-matched with the measured results. In the case
1
31
I showed a gentle gradient from the surface to the center
−
11
−11
131
(
Suppl. Table S2, 1.22×10 and 1.81×10 Gy/(Bq/ml),
of I-MIBG, the heterogeneous distribution models also
for the top and middle layers, respectively).
mitigate the difference in the survival rate from the measured
data, but the separation was not completely eliminated even
in the half-value depth of the 10-µm model.
Reflecting the high total effective radioactivity of both
MIBGs in the 3D cell culture model compared to that in
the 2D cell culture model, the estimated absorbed radiation
1
25
131
dose of both I-MIBG and I-MIBG was higher in the
3
D model than those in the 2D model (4.28 vs. 2.49 Gy for
Discussion
1
25
131
I-MIBG and 33.45 vs. 5.04 Gy for I-MIBG, respec-
1
25
tively), on the assumption of 1.0 MBq/ml administration.
The representative absorbed radiation dose distribution
image and distance radiation dose plot from point source of
In the therapeutic studies of the 2D model,
I-MIBG
1
31
showed therapeutic effect comparable to that of I-MIBG,
125
even though I-MIBG does not reside in the nucleus. Auger
electrons can cause cytotoxicity even when the tracers are
located in the cytoplasm [21]. The detailed mechanisms of
the cytotoxicity remain unknown. Probable factors contrib-
1
25
the I were shown in Suppl. Fig. S4. Absorbed dose from
1
25
the point source I is mainly concentrated within close
proximity to the source (less than 0.5 μm in radius), but to
a lesser extent, the dose reached up to approximately 10 μm
in radius.
1
25
uted to the cytotoxicity of I-MIBG would be the direct
effects such as membrane stress and mitochondrial dysfunc-
tion, and to a certain degree, the bystander effect [16]. In
addition, micro dose distribution analysis revealed that auger
Survival estimation of 3D model by computer
simulation
125
electrons from I-MIBG, which is stored in norepinephrine
storage granules in the cytoplasm, could occasionally hit
1
25
The absorbed dose in the 2D simulation model was plot-
ted against the surviving fraction, and fitted by exponential
curve (Fig. 4). Note that the surviving fraction is assumed
to be radionuclide independent and depends solely on the
absorbed dose. When the survival rate for the 3D model was
calculated under the assumption of the homogeneous distri-
bution of MIBG in the 3D simulation model, the therapeutic
the nucleus. The absorbed dose from I is distributed up
to 10 μm in radius to a lesser extent, though the substan-
tial amount of dose is distributed within a radius of 0.5 μm.
1
25
These results provide the rationale for the use of I-MIBG
to evaluate the usefulness and application of auger electrons.
Since the crossfire effect by auger electrons is negligible,
1
25
it can be assumed that the therapeutic effect of I-MIBG
is solely dependent on the cumulative radioactivity. The
cumulative % uptake of radioactivity in the 3D model was
higher than that in the 2D model due to the difference in the
release pattern of MIBG. The release pattern of the cellular
radioactivity in the 2D model was steep, whereas that in the
3D model was gradual. This difference would be attributable
1
25
131
effect of both I-MIBG and I-MIBG was overestimated
(
Fig. 5). When the heterogeneous distribution models were
applied, the separation of the estimated survival fraction
and the experimental results becomes smaller in both the
1
25
131
125
I-MIBG and I-MIBG models. In case of I-MIBG,
the estimated survival rate becomes markedly high, and
1
3

Annals of Nuclear Medicine
Fig. 5 Estimation of the cell viability of PC-12 cells in the 3D model
the assumption of a homogeneous distribution of MIBG. The other
markers are the estimated survival rate data under the assumption of
a heterogeneous distribution of MIBG with various half-value depths
(μm)
125
131
treated with
I-MIBG (a) or
I-MIBG (b). Gray bars are the
experimental results in the 3D cell culture model using the same data
as in Fig. 3. White circles are the estimated survival rate data under
1
25
to the density of the cells; the higher cellular density in the
of I-MIBG dramatically changed the distribution of the
radiation dose (Suppl. Fig. S3), and the discrepancy between
the measured and estimated values became smaller. These
results strongly suggested the existence of distribution het-
erogeneity of MIBG in the spheroids, thus reinforcing the
requirement for radiopharmaceuticals used with auger elec-
tron emitters to be distributed homogeneously in the tumor
for radionuclide therapy.
3
D model relative to the 2D model makes it easier for the
surrounding cells to reuptake MIBG more frequently. How-
1
25
ever, I-MIBG showed a greater therapeutic effect in the
D model than that in the 3D model even after the effective
2
radioactivity was corrected using the cumulative % uptake.
Since the compounds’ penetration in the 3D spheroid is lim-
ited [22, 23], the level of radiation exposure was too weak
to cause a cytocidal effect in the inner part of the spheroid
1
31
However, in the case of I-MIBG, the heterogeneous
distribution series still overestimated the therapeutic effect,
probably due to the suboptimal curve fitting. In general, the
surviving fraction of cells irradiated by external radiation
dose D can be expressed by an exponential fitting curve,
1
25
where the concentration of I-MIBG was low. In contrast,
despite the heterogeneous distribution in the 3D spheroid,
131
I-MIBG showed higher therapeutic effect in the 3D model
than that in the 2D model due to the crossfire effect [11].
These results indicate that the design of radiopharmaceuti-
cals with auger electron emitters requires particularly care-
ful consideration to achieve the homogeneous distribution
of the compound in the tumor, compared to β-emitters or
α-emitters.
2
exp(− αD − βD ) with positive α and β coefficients. The
application of the fitting model to our data results in an
overestimation of the surviving fraction, especially at higher
radioactivity. Thus, the establishment of the optimal for-
mula for estimating the survival fraction of the radionuclide
therapy needs further investigation.
Assuming a homogeneous distribution of MIBG, the
PHITS simulation well reflected the difference in the radia-
tion quality and the crossfire effect between β-ray and auger
electrons (Suppl. Fig. S2). In both the 2D and 3D models,
Considering the clinical situation, a consensus regarding
the absorbed dose–response relationship for radionuclide
therapy has not yet been reached, even with the accumu-
lated clinical evidence of external radiation therapy. This
is presumably not only because of the insufficiency of the
number of case series of radionuclide therapy but also due
to the immaturity of the technique for accurately assessing
the radiopharmaceutical-derived absorbed radiation doses.
Simulation with PHITS enables the estimation of radiation
dose distributions of any given tumor for which biodistribu-
tion data of the respective radiopharmaceutical are available,
and this will eventually contribute to the establishment of
robust absorbed dose–response relationships for radionu-
clide therapy. This approach may also permit the patient-
based selection of the suitable radionuclide and radiophar-
maceuticals for individual tumors.
1
25
I-MIBG showed a comparable average absorbed dose. In
131
contrast, I-MIBG showed a much higher average absorbed
dose in the 3D model compared to that in the 2D model,
1
31
and the average absorbed doses of I-MIBG increased
gradually from the surface layer to the middle layers. How-
ever, there was a discrepancy between the measured and
estimated values when the therapeutic effect of MIBG in
the 3D model was estimated based on the homogeneous
distribution model. We therefore also calculated the simu-
lation models under the assumption of the heterogeneous
distribution. Using PHITS, we can set a micro-distribution
of the radiation source and calculate an absorbed dose at
an arbitrary point. The heterogeneous distribution series
1
3

Annals of Nuclear Medicine
The advantage of β-emitters over auger electron emitters
increases with an increase in the size of the tumor. In our
the radiation dose for any given tumor for which the distri-
bution of the respective radiopharmaceutical is available.
1
25
present in vitro studies, the therapeutic effect of I-MIBG
1
31
was lower than that of I-MIBG in both the 2D and 3D
models, though the radionuclide-derived difference was
smaller in the 2D model. In the simulation study by Wolf-
Compliance with ethical standards
Conflict of interest The authors state that there is no conflict of interest
1
25
to declare.
gang et al. [24], the therapeutic effects of I-MIBG and
1
31
I-MIBG on tumor spheroids were equal for tumors up
to approx. 100 μm in diameter. The therapeutic effect of an
auger electron emitter can be enhanced using radiopharma-
ceuticals that accumulate in the nucleus [5–7]. From these
factors, we can deduce that auger electrons are suitable for
treating tumors with a small size such as micro-metastases
and disseminated small tumors rather than the tumors of
a certain size. In addition, our PHITS simulation clearly
References
1
.
Behr TM, Béhé M, Löhr M, Sgouros G, Angerstein C, Wehrmann
E, et al. Therapeutic advantages of Auger electron-over beta-emit-
ting radiometals or radioiodine when conjugated to internalizing
antibodies. Eur J Nucl Med. 2000;27:753–65.
2
.
Andersson H, Elgqvist J, Horvath G, Hultborn R, Jacobsson L,
Jensen H, et al. Astatine-211-labeled antibodies for treatment of
disseminated ovarian cancer: an overview of results in an ovarian
tumor model. Clin Cancer Res. 2003;9:3914–21.
1
25
demonstrated that the radiation exposure from I was
restricted specifically in the tumor lesion, whereas that from
1
31
3. Teiluf K, Seidl C, Blechert B, Gaertner FC, Gilbertz KP, Fernan-
dez V, et al. α-Radioimmunotherapy with ²¹³Bi-anti-CD38 immu-
noconjugates is effective in a mouse model of human multiple
myeloma. Oncotarget. 2015;6:4692–703.
I extended widely to the surrounding area. These results
indicating the expectedly very low side effect of auger elec-
tron may enable higher radiation dose administration to the
patients. Radionuclide therapy with auger electrons has the
potential to provide a better therapeutic effect than that with
β-emitters in the case of small tumors.
4. Hagemann UB, Wickstroem K, Wang E, Shea AO, Sponheim K,
Karlsson J, et al. In vitro and in vivo efficacy of a novel CD33-
targeted thorium-227 conjugate for the treatment of acute myeloid
leukemia. Mol Cancer Ther. 2016;15:2422–31.
There were several limitations in this study. Although
the difference in the therapeutic effect between auger elec-
trons and β-rays was clearly demonstrated in our spheroid
model, the size of the spheroid might affect the distribution
patterns of radiopharmaceuticals and the degree of cross-
fire effect. To simplify the simulation models, the shapes of
the cells and the spheroid were deformed, but even under
these assumptions, the estimated results agreed well with
the experimental results. We will be able to obtain more
precise results using more accurate models. We did not per-
form in vitro experiments multiple times in the exact same
condition, due to the usage limitation of radionuclide in our
facility. But similar tendencies were observed in the prelimi-
nary experiments. Besides, all in vitro therapeutic studies
were performed at the same time and in the same condition.
Taken together, these factors support the reproducibility of
our results and validate the usefulness of auger electrons.
5
6
7
.
.
.
Kassis AI. The amazing world of auger electrons. Int J Radiat
Biol. 2004;80:789–803.
Kassis AI. Molecular and cellular radiobiological effects of Auger
emitting radionuclides. Radiat Prot Dosimetry. 2011;143:241–7.
Sastry KS. Biological effects of the Auger emitter iodine-125: a
review. Report No. 1 of AAPM Nuclear Medicine Task Group No.
6. Med Phys. 1992;19:1361–70.
8
.
Kassis AI, Sastry KS, Adelstein SJ. Kinetics of uptake, retention,
1
25
and radiotoxicity of IUdR in mammalian cells: Implications
of localized energy deposition by Auger processes. Radiat Res.
1
987;109:78–89.
9. Kassis AI, Adelstein SJ, Haydock C, Sastry KS, McElvany
KD, Welch MJ. Lethality of Auger electrons from the decay
of bromine-77 in the DNA of mammalian cells. Radiat Res.
1
982;90:362–73.
1
0. Rebischung C, Hoffmann D, Stefani L, Desruet MD, Wang K,
Adelstein SJ, et al. First human treatment of resistant neoplastic
1
25
meningitis by intrathecal administration of MTX plus IUdR.
Int J Radiat Biol. 2008;84:1123–9.
1
1. Rutgers M, Buitenhuis CK, van der Valk MA, Hoefnagel CA,
131
125
Voûte PA, Smets LA. [ I] and [ I] metaiodobenzylguanidine
therapy in macroscopic and microscopic tumors: a comparative
study in SK-N-SH human neuroblastoma and PC12 rat pheochro-
mocytoma xenografts. Int J Cancer. 2000;90:312–25.
1
25
In conclusion, I-MIBG showed a therapeutic effect in
both 2D and 3D cell culture models, but its therapeutic effi-
1
31
cacy was lower than that of I-MIBG, especially in the 3D
model. Thus, radiation therapy with auger electrons would
be suitable for small-sized tumors such as small metastases.
In addition, the design of radiopharmaceuticals with auger
electron emitters requires the very careful consideration of
how to achieve a homogeneous distribution of the compound
in the tumor. In the present study, the simulation with PHITS
1
1
2. Brans B, Bodei L, Giammarile F, Linden O, Luster M, Oyen WJ,
et al. Clinical radionuclide therapy dosimetry: the quest for the
“Holy Gray”. Eur J Nucl Med Mol Imaging. 2007;34:772–86.
3. Govindan SV, Goldenberg DM, Elsamra SE, Griffiths GL, Ong
GL, Brechbiel MW, et al. Radionuclides linked to a CD74 anti-
body as therapeutic agents for B-cell lymphoma: comparison of
Auger electron emitters with beta-particle emitters. J Nucl Med.
2000;41:2089–97.
125
14. Bousis C, Emfietzoglou D, Hadjidoukas P, Nikjoo H. Monte Carlo
single-cell dosimetry of auger-electron emitting radionuclides.
Phys Med Biol. 2010;55:2555–72.
estimated the absorbed radiation dose of I-MIBG, and the
obtained dose–response relationship provided an accurate
1
25
prediction of the therapeutic effect of I-MIBG in the 3D
spheroid. This approach may thus enable the evaluation of
1
5. Bousis C, Emfietzoglou D, Nikjoo H. Monte Carlo sin-
gle-cell dosimetry of I-131, I-125 and I-123 for targeted
1
3

Annals of Nuclear Medicine
radioimmunotherapy of B-cell lymphoma. Int J Radiat Biol.
kernels using particle and heavy ion transport code system:
PHITS. App Radiat Isot. 2017;128:199–203.
2
012;88:908–15.
1
6. Boyd M, Ross SC, Dorrens J, Fullerton NE, Tan KW, Zalut-
sky MR, et al. Radiation-induced biologic bystander effect
elicited in vitro by targeted radiopharmaceuticals labeled with
α-, β-, and auger electron-emitting radionuclides. J Nucl Med.
21. Pouget JP, Santoro L, Raymond L, Chouin N, Bardiès M, Bascoul-
Mollevi C, et al. Cell membrane is a more sensitive target than
cytoplasm to dense ionization produced by auger electrons. Radiat
Res. 2008;170:192–200.
2
006;47:1007–15.
22. Sagnella SM, Duong H, MacMillan A, Boyer C, Whan R, McCa-
rroll JA, et al. Dextran-based doxorubicin nanocarriers with
improved tumor penetration. Biomacromol. 2014;15:262–75.
23. Groh CM, Hubbard ME, Jones PF, Loadman PM, Periasamy N,
Sleeman BD, et al. Mathematical and computational models of
drug transport in tumours. J R Soc Interface. 2014;11:20131173.
24. Weber W, Weber J, Senekowitsch-Schmidtke R. Therapeutic effect
1
1
1
2
7. Sato T, Niita K, Matsuda N, Hashimoto S, Iwamoto Y, Noda S,
et al. Particle and heavy ion transport code system PHITS, version
2
.52. J Nucl Sci Technol. 2013;50:9,913–23.
8. Vaidyanathan G, Affleck DJ, Alston KL, Zalutsky MR. A tin pre-
211
cursor for the synthesis of no-carrier-added [*I]MIBG and [ At]
MABG. J Label Compd Radiopharm. 2007;50:177–82.
1
31
125
9. Sato T, Niita K, Matsuda N, Hashimoto S, Iwamoto Y, Noda S,
et al. Overview of the PHITS code and its application to medical
physics. Prog Nucl Sci Technol. 2014;4:879–82.
of m-[ I]- and m-[ I]iodobenzylguanidine on neuroblastoma
multicellular tumor spheroids of different sizes. Cancer Res.
1996;56:5428–34.
0. Shiiba T, Kuga N, Kuroiwa Y, Sato T. Evaluation of the accuracy
of mono-energetic electron and beta-emitting isotope dose-point
Affiliations
1
2
3
4
2
Ayaka Shinohara · Hirofumi Hanaoka · Tetsuya Sakashita · Tatsuhiko Sato · Aiko Yamaguchi ·
3
5,6
Noriko S. Ishioka · Yoshito Tsushima
1
4
5
Department of Heavy Ion Beam Medical Physics
and Biology, Gunma University Graduate School
of Medicine, 3-39-22 Showa, Maebashi 371-8511, Japan
Nuclear Science and Engineering Center, Japan Atomic
Energy Agency, 2-4 Shirakata, Tokai 319-1195, Japan
Department of Diagnostic Radiology and Nuclear Medicine,
Gunma University Graduate School of Medicine, 3-39-22
Showa, Maebashi 371-8511, Japan
2
3
Department of Bioimaging Information Analysis, Gunma
University Graduate School of Medicine, 3-39-22 Showa,
Maebashi 371-8511, Japan
6
Research Program for Diagnostic and Molecular Imaging,
Division of Integrated Oncology Research, Gunma
University Initiative for Advanced Research (GIAR), Gunma
University Graduate School of Medicine, 3-39-22 Showa,
Maebashi 371-8511, Japan
Quantum Beam Science Research Directorate, National
Institutes for Quantum and Radiological Science
and Technology, 1233 Watanuki-machi, Takasaki 370-1292,
Japan
1
3