CORRECTION NOTICE Magnetoferritin nanoparticles …...CORRECTION NOTICE Magnetoferritin nanoparticles for targeting and visualizing tumour tissues Kelong Fan, Changqian Cao, Yongxin

CORRECTION NOTICE

Magnetoferritin nanoparticles for targeting and

visualizing tumour tissues Kelong Fan, Changqian Cao, Yongxin Pan, Di Lu, Dongling Yang, Jing Feng, Lina Song, Minmin Liang and Xiyun Yan

Nature Nanotechnology 7, 459–464 (2012)

In the Supplementary Information, one of the authors was not mentioned in the author list: Lina Song has now been added. In the section ‘Preparation and characterization of M-HFn particles’ the column used for size-exclusion chromatography was incorrect: it should have been ‘Sepharose 6B’. The synthesis procedure for M-HFn nanoparticles was incorrect: it should have read ‘HFn protein shells were used as a reaction template to synthesize iron oxide nanoparticles according to the method reported by Cao et al.2 with some modification. The solution of 50 ml 100 mM NaCl with HFn (1 mg ml−1) was added to the reaction vessel, synthesized at 65 °C and pH 8.5. Fe(II) (25 mM (NH4)2Fe(SO4)2•6H2O) and stoichiometric equivalents (1:3 H2O2:Fe2+) of freshly prepared H2O2 (8.33 mM) were added. Fe(II) was added in a rate of 100 Fe/(protein min) using a dosing device (800 Dosino) connected with 842 Titrando. After theoretical 5000 Fe/ protein cage were added to the reaction vessel, the reaction was continued for another 5 min. Finally, 200 μl of 300 mM sodium citrate was added to chelate any free iron. The synthesized magnetite-containing HFn (M-HFn) nanoparticles were centrifuged and purified through size exclusion chromatography to remove the aggregated nanoparticles. The concentration of M-HFn nanoparticles was assumed to be the same as that of HFn protein and was determined using a BCA protein assay kit (Pierce). Purified M-HFn nanoparticles were obtained with a yield of about 75%.’ Reference 2 was incorrect and should have read Cao, C. Q. et al. Magnetic characterization of noninteracting, randomly oriented, nanometer-scale ferrimagnetic particles. J. Geophys. Res. 115, B07103 (2010). These errors have been corrected in this file 27 November 2012.

© 2012 Macmillan Publishers Limited. All rights reserved.

SUPPLEMENTARY INFORMATIONDOI: 10.1038/NNANO.2012.209

NATURE NANOTECHNOLOGY | www.nature.com/naturenanotechnology 1

Magnetoferritin nanoparticles for targeting and visualizing tumour

tissues

Kelong Fan1, Changqian Cao2, Yongxin Pan2, Di Lu1, Dongling Yang1, Jing Feng1,

Lina Song1, Minmin Liang1 * & Xiyun Yan1 *

*Corresponding author.

Minmin Liang, PhD. Email : [email protected] Tel: +86 10 6488 8583; Fax: +86 10

6488 8584.

Xiyun Yan, MD. Email: [email protected] Tel: +86 10 6488 8583; Fax: +86 10 6488

8584.


Cell lines, tissues and animal models

Cell lines HT29, K562, U937, HeLa, A375, SKOV3, PC-3, Jurkat, SW1990 and

MDA-MB-231 were obtained from the American Type Culture Collection (ATCC).

SMMC-7721, U251 and MX-1 were from the cell bank of the Committee on Type

Culture Collection of the Chinese Academy of Sciences (CTCC, Shanghai, China).

A375, PC-3, SW1990 and MDA-MB-231 cells were cultured in DMEM medium

containing 10% fetal calf serum. HT29, K562, SMMC-7721, U937, HeLa, SKOV3,

U251, Jurkat and MX-1 cells were cultured in RPMI-1640 medium containing 10%

fetal calf serum. Clinical tumour and normal tissues were obtained from the tissue

bank of the Beijing Tumor Hospital (Beijing) and Aomei Biotechnology Co. (Xi’an,

China). Female BALB/c nude mice were obtained from the Animal Center of the

Chinese Academy of Medical Sciences (Beijing). All animal studies were performed

with the approval of the Chinese Academy of Sciences Institutional Animal Care and

Use Committee. Mice were each injected subcutaneously in one thigh with 0.1 mL of

suspension containing 106 HT29, SKOV3, MX-1 or SMMC-7721 cells. When

tumours reached 0.4-0.6 cm in diameter, they were excised and fixed in 10% buffered

formalin for 24 h before embedding in paraffin. 5 µm paraffin-embedded tumor

xenograft sections were cut and used for subsequent histological staining.

Preparation and characterization of M-HFn nanoparticles.

Recombinant human ferritin shells composed of 100% heavy-chain subunits

were produced in Escherichia coli and purified as described1. Briefly, Escherichia

coli lysate expressing HFn was sonicated on ice and then centrifuged at 10000 g for

30 min. The supernatant was heated at 70°C for 10 min to precipitate most of the

Escherichia coli proteins. After centrifugation, the supernatant was precipitated again

by ammonium sulfate (520 g/L). The precipitate was collected by centrifugation at

22,000×g, and then dissolved in PBS. After dialyzing out the ammonium sulfate,

HFn was purified by size-exclusion chromatography on a Sepharose 6B column. The

final yield of HFn was >100 mgL-1 from the bacterial lysate.

HFn protein shells were used as a reaction template to synthesize iron oxide

nanoparticles according to the method reported by Cao et al2 with some modification.


The solution of 50 mL 100 mM NaCl with HFn (1 mg/ml) was added to the reaction

vessel, synthesized at 65°C and pH 8.5. Fe(II) (25 mM (NH4)2Fe(SO4)2•6H2O) and

stoichiometric equivalents (1:3 H2O2:Fe2+) of freshly prepared H2O2 (8.33 mM) were

added. Fe(II) was added in a rate of 100 Fe/(protein min) using a dosing device (800

Dosino) connected with 842 Titrando. After theoretical 5000 Fe/ protein cage were

added to the reaction vessel, the reaction was continued for another 5 min. Finally,

200 μL of 300 mM sodium citrate was added to chelate any free iron. The synthesized

magnetite-containing HFn (M-HFn) nanoparticles were centrifuged and purified

through size exclusion chromatography to remove the aggregated nanoparticles. The

concentration of M-HFn nanoparticles was assumed to be the same as that of HFn

protein and was determined using a BCA protein assay kit (Pierce). Purified M-HFn

nanoparticles were obtained with a yield of about 75%.

The prepared M-HFn nanoparticles and HFn protein were characterized using

TEM (Tecnai F20, Philips), cryo-TEM (FEI Titan Krios 300kV, FEI, Oregon), HFn

proteins were negatively stained with uranyl acetate for TEM observation while iron

oxide cores encapsulated in HFn proteins were unstained. For cryo-TEM observation,

M-HFn nanoparticle samples were embedded in vitreous ice using an FEI Vitrobot

Mark VI and imaged with an FEI 300-kV Titan Krios cryo-TEM equipped with a

Gatan UltraScan4000 (model 895) 16-megapixel CCD camera. M-HFn nanoparticles

were imaged with an absolute magnification micrograph of 96,000, and the dose for

each micrograph was about 20e-/Å.

A peroxdiase activity test was carried out on M-HFn nanoparticles at room

temperature. M-HFn at 0.5 µM was mixed with 500 mM H2O2 in 0.2 M sodium

acetate buffer (pH 4.5), using 0.2 mg/mL TMB (Sigma) as the substrate. Colour

reactions were recorded 30 min after addition of the substrate. The reaction buffer

used for the DAB (Sigma) substrate was 0.05 M Tris-HCl, pH 7.5.

The prepared M-HFn nanoparticles were further characterized using

size-exclusion chromatography (SEC, Amersham Pharmacia Biotech, Piscataway)

and dynamic light scattering (DLS, DynaPro Titan TC, Wyatt Technology). SEC

analysis were performed on a Superose 12 column installed on a Waters 515 solvent


delivery system (Waters, Milford) equipped with an in-line radioactivity detector and

a Waters UV2487 dual wavelength absorbance detector. DLS analysis was performed

at room temperature. The concentration of M-HFn used was 1.5 mg/ml in PBS buffer.

As shown in Figure S1a, the M-HFn nanoparticles were found to be monodisperse in

solution with a outer diameter of 12~16 nm. In addition, HFn and M-HFn had

identical SEC elution times (Figure S1b,c). These results indicate that the

mineralization process does not significantly perturb the overall protein cage

architecture of HFn and that the iron oxide core is sequestered within the protein

shell.

Figure S1. (a) DLS analysis of M-HFn nanoparticles. M-HFn nanoparticles have

an outer diameter of 12~16 nm (i.e. 6~8 nm in radius). (b) SEC of HFn protein and (c)

M-HFn nanoparticles by in-line UV detection at 280 nm (protein) and 410 nm (iron

oxide core).


Peroxidase activity of M-HFn, natural holoferritin and apoferritin

A comparison of the peroxidase activity of M-HFn, natural holoferritin and

apoferritin was performed on a PVDF membrane. Briefly, 4 µg of M-HFn, natural

holoferritin and apoferritin was respectively mixed with 1 µl of TMB substrate (10

mg/mL), 1 µl of 30% H2O2 and 5 µl of sodium acetate buffer (pH 4.5) and the mixture

was transferred to a PVDF membrane. The image was taken 30 min after transfer.

Mineral cores within ferritin exhibited peroxidase activity, catalyzing the oxidation of

substrate TMB and giving a coloured dot on the membrane. The strong staining

intensity demonstrates the high peroxidase activity of mineral cores. As shown in

Figure S2, M-HFn shows higher peroxidase activity than natural holoferritin due to

their differences in mineral phase composition. The apoferritin control without a

mineral core exhibited no peroxidase activity.

Figure S2. Direct comparison of peroxidase activity between magnetoferritin,

natural holoferritin and apoferritin. Horse spleen holoferritin from Sigma was used

here as natural holoferritin.

Saturation binding assay

The binding affinity of HFn to TfR1 was measured using a saturation binding

assay. A total of 5×105 SMMC-7721 cells were incubated with increasing amounts

of FITC-conjugated HFn at concentrations ranging from 0 to 400 nM for 1 h at 4°C.

The total binding of FITC-conjugated HFn was calculated after cells were washed

(Figure S3a, black line). Nonspecific binding was determined as the binding of

FITC-conjugated HFn to non-TfR1 sites. In order to measure nonspecific binding, the


same reaction mixtures were prepared with the addition of an excess of (40 µM)

unconjugated HFn. Nonspecific binding of FITC-conjugated HFn was obtained after

washing with cold PBS (Figure S3a, red line), and the specific binding of

FITC-conjugated HFn was calculated by subtracting nonspecific binding from total

binding (Figure S3b). The dissociation constant (Kd) was calculated using GraphPad

Prism 4.0. As shown in Figure S3, the saturation binding curve clearly shows that the

binding of HFn to SMMC-7721 cancer cells is saturable and that the binding can be

significantly inhibited by adding an excess dose of unconjugated HFn. This confirms

again that HFn binding is specific. Scatchard analysis demonstrated that HFn binds to

SMMC-7721 cells with a high affinity (Kd of 50 nM).

Figure S3. Saturation binding curve for FITC-HFn binding to TfR1 on

SMMC-7721 cells. (a) Total (black line) and non-specific binding (red line). Excess

unconjugated HFn was used to determine nonspecific binding. Each data point

represents the average value from triplicate wells. (b) Specific binding was obtained

after subtraction of non-specific binding from total binding. Binding of FITC-HFn to

SMMC-7721 cells is saturable. Binding is inhibited by excess unconjugated HFn. The

Kd of HFn was 50 nM.

Reactivity of HFn to human cancer cells

The binding activity of HFn protein shells to cancer cells was confirmed using

human hepatocellular carcinoma, colon carcinoma, breast adenocarcinoma, melanoma,

erythroleukemia, cervical carcinoma, ovarian carcinoma, prostate carcinoma,


glioblastoma, histiocytic lymphoma and T-cell leukemia cell lines by flow cytometric

analysis. As illustrated in Figure S4, HFn strongly bound to these cancer cell lines,

demonstrating the ability of HFn to universally recognize cancer cells.

Figure S4. Flow cytometric analysis of HFn reactivity with human cancer cells.

The ten cancer cell lines examined were glioblastoma (U251), ovarian carcinoma

(SKOV-3), histiocytic lymphoma (U937), pancreatic cancer (SW1990), cervical

carcinoma (HeLa), prostate carcinoma (PC-3), breast adenocarcinoma

(MDA-MB-231), melanoma (A375), erythroleukemia (K562), and T-cell leukemia

(Jurkat).

Antibody and transferrin blocking studies

Antibody blocking study was performed on SMMC-7721 cells. Briefly, 0.3 µM

of FITC-HFn was added to the wells in the presence or absence of a 10-fold molar

excess of anti-TfR1 mAbs. After incubation for 1 h on ice, the cells were washed

three times in cold PBS and then collected. Cell-bound fluorescence was measured by

flow cytometry. In addition, the incubated cells were also examined under a confocal

laser scanning microscope. Results are shown in Figure S5. Anti-TfR1 mAb

completely blocks the binding of HFn to SMMC-7721 cancer cells as determined by

both flow cytometry and confocal studies, further confirming that TfR1 is the binding

receptor of HFn and mediates its specific binding to cancer cells.


Figure S5. (a) Fluorescence staining of SMMC-7721 human liver cancer cells

using FITC-conjugated HFn in the presence (right) or absence (left) of a 10-fold

molar excess of anti-TfR1 mAbs. (Scale bar = 50 µm). (b) Flow cytometry analysis of

the binding of FITC-conjugated HFn to SMMC-7721 cancer cells in the presence or

absence of a 10-fold molar excess of anti-TfR1 mAbs. (n = 3, bars represent means ±

SD)

Competitive binding by transferrin was tested using SMMC-7721 cells. Briefly,

5X105 cells were incubated for 1 h on ice with 0.3 µM FITC-HFn and transferrin at

concentrations ranging from 0 to 40 µM. After washing three times in cold PBS, the

fluorescence intensity of cells was measured by flow cytometry. Figure S6 shows that

transferrin competes with HFn for binding to TfR1. However, transferrin at a 100-fold

molar excess only inhibits HFn binding by 50%.


0 10 20 30 40100

120

140

160

180

200

220

Fluo

resc

ence

Inte

nsity

Transferrin(µM)

Figure S6. Transferrin competes with HFn for binding to TfR1. Binding of

FITC-HFn to SMMC-7721 cells is only partially inhibited by transferrin.

Fluorescence and mineral core-peroxidase staining analysis

We assessed the co-localization of fluorescence staining and mineral

core-peroxidase staining by incubating HT-29 cancer xenograft tumours with

FITC-conjugated M-HFn. Briefly, two sequential sections from an HT-29 cancer

xenograft were first stained with FITC-labeled M-HFn. One section was examined

after M-HFn-peroxidase staining by light microscopy while the other was examined

by fluorescence microscopy. By comparing these sequential sections we can see that

the fluorescence staining co-localized with mineral-peroxidase staining in tumour

cells (Fig. S7).


Figure S7. Fluorescence staining co-localized with mineral core-peroxidase

staining in tumour cells on the sequential sections. Two sequential sections from an

HT-29 cancer xenograft were stained both with FITC-labeled M-HFn. One section

was examined after M-HFn-peroxidase staining by light microscopy while the other

was examined by fluorescence microscopy. (Scale bar = 50 µm )

HFn staining of human tissue arrays

To study the potential correlation of HFn binding with the grade and growth

pattern of tumours, tissue arrays including hepatocellular carcinoma, lung squamous

cell carcinoma, cervical squamous cell carcinoma, prostate adenocarcinoma, ovarian

serous papillary carcinoma, and colonic adenocarcinoma (about 20 cases/type), and

their corresponding normal and lesion tissues (about 5 cases/type) were incubated

with FITC-labeled HFn (1 µM) at 4°C overnight. The stained tissues were imaged

under a confocal laser scanning microscope (Olympus). As shown in Figure S8-13,

HFn based fluorescence staining positively correlated with differentiation, grades and

growth patterns of hepatocellular carcinoma, lung squamous cell carcinoma, cervical

squamous cell carcinoma, prostate adenocarcinoma, ovarian serous papillary

carcinoma and colonic adenocarcinoma. The corresponding normal tissues, necrotic

tumours, chronic inflammatory tissues and hyperplastic tissues showed negative

staining.


Figure S8. HFn staining of a lung tissue array. Tissues were incubated with

FITC-labeled HFn. (a) Grade 1 lung squamous cell carcinoma showing weak staining

intensity with 10-20% positive cells. (b) Grade 2 lung squamous cell carcinoma

showing medium staining intensity with 20-50% positive cells. (c) Grade 3 lung

squamous cell carcinoma showing strong staining intensity with >50% positive cells.

(d) Necrotic carcinoma tissue showing negative staining. (e) Normal lung showing

negative staining. (f) Normal lung with congestion showing slight staining. The

tumours were graded according to Gleason’s grading system. Scale bar = 100 µm


Figure S9. HFn staining of a liver tissue array. Tissues were incubated with

FITC-labeled HFn. (a) Grade 1 hepatocellular carcinoma tissue showing medium

staining intensity with 50-80% positive cells. (b) Grade 2 hepatocellular carcinoma

showing strong staining intensity with >80% positive cells, (c) Grade 3 hepatocellular

carcinoma tissue showing strong staining intensity with >80% positive cells. (d)

Normal liver showing negative staining. (e) Liver tissue of hepatitis showing negative

staining. (f) Cirrhotic tissues showing negative staining. The tumours were graded

according to Gleason’s grading system. Scale bar = 100 µm


Figure S10 HFn staining of a cervical tissue array. Tissues were incubated with

FITC-labeled HFn. (a) Grade 1 cervical squamous cell carcinoma showing weak

staining intensity with 20-50% positive cells. (b) Grade 2 cervical squamous cell

carcinoma showing medium staining intensity with 20-50% positive cells. (c) Grade 3

cervical squamous cell carcinoma showing strong staining intensity with >50%

positive cells. (d) Cancer adjacent normal cervical tissue showing negative staining. (e)

Cervical chronic inflammatory tissue showing negative staining. (f) Normal cervical

tissue showing negative staining. The tumours were graded according to Gleason’s

grading system. Scale bar = 100 µm


Figure S11 HFn staining of a prostate tissue array. Tissues were incubated with

FITC-labeled HFn. (a) Grade 1 prostate adenocarcinoma showing weak staining

intensity with 10-20% positive cells. (b) Grade 2 prostate adenocarcinoma showing

medium staining intensity with 20%-50% positive cells. (c) Grade 3 prostate

adenocarcinoma showing medium staining intensity with 50-80% positive cells. (d)

Grade 4 prostate adenocarcinoma showing strong staining intensity with >80%

positive cells. (e) Prostatic hyperplasia showing negative staining. (f) Prostate smooth

muscle showing negative staining. (g) Cancer adjacent normal prostate tissue showing

negative staining. (h) Grade 1 prostatic intraepithelial neoplasia showing slight

staining only in prostate duct epithelium. The tumours were graded according to

Gleason’s grading system. Scale bar = 100 µm


Figure S12. HFn staining of an ovary tissue array. Tissues were incubated with

FITC-labeled HFn. (a) Grade 1 ovarian serous papillary carcinoma showing medium

staining intensity with 50-80% positive cells. (b) Grade 2 ovarian serous papillary

carcinoma showing strong staining intensity with >80% positive cells. (c) Grade 3

ovary serous papillary carcinoma showing strong staining intensity with >80%

positive cells. (d) Cancer adjacent normal ovary showing negative staining. The

tumours were graded according to Gleason’s grading system. Scale bar = 100 µm

Figure S13. HFn staining of a colon tissue array. Tissues were incubated with

FITC-labeled HFn. (a) Grade 1 colonic adenocarcinoma showing weak staining

intensity with 10-20% positive cells. (b) Grade 2 colonic adenocarcinoma showing

strong staining intensity with 20-50% positive cells. (c) Grade 3 colonic

adenocarcinoma showing strong staining intensity with >50% positive cells. (d)

Normal colon showing negative staining. The tumours were graded according to

Gleason’s grading system. Scale bar = 100 µm

M-HFn-based peroxidase-like reaction mechanism

To understand the mechanism of the M-HFn-based peroxidase-like reaction, we

detected the formation of •OH during the reaction by electron spin resonance

(ESR),using 5, 5-dimethyl-1-pyrroline-N-oxide (DMPO) as the spin-trapping reagent,

a method well-known to effectively determine •OH. Briefly, 100 µL reaction mixtures


were taken after 10 min reactions at room temperature and mixed with 20 µl of 200

mM DMPO to form a DMPO-OH adduct. The amount of hydroxyl radical was

determined from ESR signals using a Bruker model ESP 300E spectrometer. ESR

measurements were carried out at room temperature under the following conditions:

microwave power 15.89 mW, modulation amplitude 3.081 G, scan range of 100.00 G,

modulation frequency 100.00 kHz, and center field 3485.00 G. As shown in Figure

S14 (a, b), •OH was produced during the peroxidase-like reaction in the presence of

both M-HFn nanoparticles and H2O2. With the addition of the •OH scavenger, ethanol,

the formed •OH disappeared (Figure S14 c) and the peroxidase activity of M-HFn

nanoparticles decreased to 20% of the original activity (Figure S14 d,e), indicating

that the •OH formed during the peroxidase-like reaction is responsible for the

catalytic oxidation of peroxidase substrate to give the colored precipitate at the site of

its target.

Based on these results, we propose the following mechanism for the

M-HFn-based peroxidase-like reaction. With the addition of H2O2 and peroxidase

substrate into the M-HFn reaction solution, diffusing H2O2 enters the ferritin cavity

through its hydrophilic channels and interacts with the iron oxide core of M-HFn to

generate •OH on the surface of the iron core. The generated •OH then oxidizes the

peroxidase substrate (e.g., DAB) diffused nearby to form an insoluble colored

precipitate at the site of M-HFn, which is targeted to cancer cells. The colored

precipitates are only formed at the site of M-HFn since •OH radicals are highly

reactive and short-lived, and can only oxidize nearby substrates to give colored

precipitates. The clear boundary between tumour and normal tissues on M-HFn

stained tissue slides (Figure S15) also proves that the colored precipitates are

generated right at the site of M-HFn-targeted cancer cells, and the oxidized colored

precipitates do not diffuse away from their targets.


Figure S14 ESR spectra of DMPO-OH adducts in the M-HFn-based peroxidase-like

reaction. (a) H2O2/40 mM DMPO; (b) H2O2/0.5 µM M-HFn nanoparticles/40 mM

DMPO; (c) H2O2 /M-HFn nanoparticles/40 mM DMPO/•OH scavenger ethanol; The

reaction was initiated by adding 500 mM H2O2. All the mixtures were in 200 mM

acetate buffer (pH = 4.5). (d) Peroxidase activity of M-HFn nanoparticles with (right

tube) or without the •OH scavenger ethanol. M-HFn catalyzed the oxidation of

peroxidase substrate TMB in the presence of H2O2 to give a colored product. (e)

Absorbance at 652 nm of the TMB reaction solution catalytically oxidized by M-HFn

with or without the •OH scavenger ethanol.


Figure S15 The M-HFn-stained tumour area is clearly demarcated from normal

tissues. A HT-29 colon cancer xenograft was stained by M-HFn and visualized by

DAB color development. The clear boundary between tumour and normal tissues on

M-HFn stained tissue slides demonstrates that the colored precipitates are generated

right at the site of M-HFn-targeted cancer cells, and that the oxidized colored

precipitates do not diffuse away from their targets. (Scale bar = 50 µm)

Comparison of immunohistochemical and M-HFn approaches

Sequential liver tissue sections containing tumours were stained respectively by

M-HFn nanoparticles and anti-TfR1 Abs. Paraffin-embedded tissues of two

hepatocellular carcinoma cases identified by a pathologist were deparaffinized in

xylene and then hydrated progressively in an ethanol gradient. After quenching

endogenous peroxidase activity, the tissue sections were blocked with goat serum, and

then incubated with M-HFn or polyclonal rabbit anti-TfR1 antibody, respectively. The

stained sections were analyzed under a microscope and the results were assessed by

two independent pathologists. M-HFn-stained tissues showed positive staining of

tumour cells and negative staining of normal liver cells (Figure S16 a, c). Anti-TfR1

antibody-based immunohistochemical staining could not distinguish between tumours

and normal liver tissues. (Figure S16 b, d).


Figure S16 A side-by-side comparison of standard antibody-based

immunohistochemical and M-HFn-based approaches in two hepatocellular

carcinoma cases identified by pathologists. Sequential liver tissue sections

containing tumours were stained respectively by M-HFn nanoparticles and anti-TfR1

antibody. (a, c) M-HFn-stained tissues showed positive staining of tumour cells and

negative staining of normal liver cells. (b, d) Anti-TfR1 antibody-based

immunohistochemical staining could not distinguish tumours from normal liver

tissues. Scale bar = 100 µm

We further did a literature search to establish the range of sensitivities and

specificities of different antibodies-based detections to make valid comparsions with

our M-HFn-based approach. The search results are shown as following in

Supplemental Table 1. Of the 56 different antibodies widely used in the detection of

the 15 main cancer biomarkers for 7 different types of cancer reported in the literature,

only the best ones (AMACR antibody P504S: sensitivity of 80%~95%; specificity:


79%~100%; and PSA antibody ER-PR8: sensitivity of 82%~94.4%, specificity of

100%) had a similar tumour detection sensitivity and specificity with our M-HFn.

However, they can only detect one type of cancer. When M-HFn nanoparticles were

used to screen 474 clinical specimens from patients with nine types of cancer, it had a

total sensitivity of 98% and a total specificity of 95% (Table 1 in manuscript),

indicating that M-HFn are an excellent reagent for cancer screening.

Table S1 Sensitivity and specificity of various antibody-based

immunohistochemistry and immunofluorescence detection methods. The

antibodies listed below represent most of the antibody population reported in the

literature or available commercially for cancer biomarker detection.

Tumor Biomarker Antibody (Source) Sensitivity Specificity References

Hepatocellular carcinoma

CEA 11-7 (DAKO) 46.88~79% 52.44~97% [3],[4],[5] AFP α-AFP mAb (DAKO) 17~61.5% 97.00% [5],[6] TfR OKT9, 5E9, RBC4, B3/25, Tu15 97.06% 86.67% [7]

Lung carcinoma

TTF-1 SPT24 (Novocastra), 8G7G3/1 (DAKO） 65~87% 92%~100% [8],[9]

p53 DO-7 (Novocastra), PAb240 (Oncogene Science)

46~79.2% 83.7~100% [10],[11],[12]

CK-7 M7018 (DAKO) 75.00% 77% [13]

Colonic adenocarcinoma

CEA 11-7 (DAKO) 50~59.4% 53.23% [3],[14]

p53 PAb240 (Oncogene Science), α-p53 mAb (DAKO), DO-7 (Novocastra)

49~80% 91.63% [15],[16],[17]

Ki-67 α-Ki-67 mAb (CHANGDAO), α-Ki-67 mAb (DAKO)

21.6%, 87.5%

100%, 86.67%

[18],[19]

Cervical squamous cell carcinoma

p53 BP53-12 (BioGenex), DO-7 (Novocastra Laboratories), α-p53 mAb (DAKO)

17.1~85.7% 87~100% [20],[21],[22], [23],[24]

Ras α-Ras mAb RP35, α-Ras mAb Y13 259 38.7~100% 50~90% [25],[26]

C-erbB-2 α-C-erbB-2 pAb, NCL-CBll (Novocastra laboratories)

12.1~38.7% 83~100% [27],[28],[29],[30]

Prostate adenocarcinoma

EPCA α-EPCA pAb 84~94% 85% [31],[32] AMACR P504S (Corixa) 80~95% 79~100% [33],[34],[35]

PSA ER-PR8 (DAKO), A0562 (DAKO)

82~94.4%, 100%

100%, 68%

[36],[37],[38]

Ovarian serous papillary carcinoma

ER 6F-11 (Novocastra Laboratories) 63~86.36% 95~97.7% [39],[40],[41] WT1 6F-H2（DAKO） 82~86% 95~97% [39],[41]

p53 DO-7 (Novocastra Laboratories), PAb1801 (Cambridge Research Biochemicals)

55~73.7% 61.54~100% [39],[42],[43], [44],[45]

Breast Her-2 28 different antibodies including TAB250 6~82% 92~100% [46], [47]


carcinoma (Berlex biosciences), 2H11 (Genetech), and 3E8 (Genetech)


References

1. Rucker, P. et al. Recombinant ferritin: Modulation of subunit stoichiometry in

bacterial expression systems. Protein Eng. 10:967–973 (1997).

2. Cao, C. Q. et al. Magnetic characterization of noninteracting, randomly oriented,

nanometer-scale ferrimagnetic particles. J. Geophys. Res. 115, B07103 (2010).

3. Porcell, A.I., De Young, B.R., Proca, D.M. & Frankel, W.L.

Immunohistochemical analysis of hepatocellular and adenocarcinoma in the liver:

MOC31 compares favorably with other putative markers. Mod Pathol. 13,

773-778 (2000).

4. Chu, P.G.G., Ishizawa, S., Wu, E. & Weiss, L.M. Hepatocyte antigen as a marker

of hepatocellular carcinoma - An immunohistochemical comparison to

carcinoembryonic antigen, CD10, and alpha-fetoprotein. Am. J. Surg. Pathol. 26,

978-988 (2002).

5. Minervini, M.I. et al. Utilization of hepatocyte-specific antibody in the

immunocytochemical evaluation of liver tumors. Mod Pathol. 10, 686-692

(1997).

6. Ali S. Sawan. The Diagnostic Value of Immunohistochemistry in the Diagnosis of

Primary and Secondary Hepatic Carcinomas. JKAU: Med. Sci. 16, 37-48(2009).

7. Sciot, R. et al. Transferrin receptor expression in human hepatocellular carcinoma:

an immunohistochemical study of 34 cases. Histopathology 12, 53-63 (1988).

8. Comperat, E. et al. Variable sensitivity and specificity of TTF-1 antibodies in

lung metastatic adenocarcinoma of colorectal origin. Mod Pathol. 18, 1371-1376

(2005).

9. Wu, M. et al. p63 and TTF-1 immunostaining. A useful marker panel for

distinguishing small cell carcinoma of lung from poorly differentiated squamous

cell carcinoma of lung. Am. J. Clin. Pathol. 119, 696-702 (2003).

10. Papadakis, E. et al. Ras and p53 expression in non-small-cell lung-cancer patients

- p53 over-expression correlates with a poor prognosis. Int. J. Oncol. 1, 403-413

(1992).

11. Iggo, R., Gatter, K., Bartek, J., Lane, D. & Harris, A.L. Increased expression of


mutant forms of p53 oncogene in primary lung cancer. Lancet 335, 675-679

(1990).

12. Levesque, M.A., Tadross, L., Diamandis, E.P. & D'Costa, M. Comparison of

immunofluorometry and immunohistochemistry for the detection of p53 protein

in lung cancer specimens. Am. J. Clin. Pathol. 107, 308-316 (1997).

13. Su, Y.C., Hsu, Y.C. & Chai, C.Y. Role of TTF-1, CK20, and CK7

immunohistochemistry for diagnosis of primary and secondary lung

adenocarcinoma. Kaohsiung. J. Med. Sci. 22, 14-19 (2006).

14. Ok S.B., Tae S.L., Sung D.P., CEA Expressions in Colorectal tumor. J. Korean

Soc. Coloprotocol. 20, 39-44 (2004).

15. Huh, J.W., Lee, J.H. & Kim, H.R. Expression of p16, p53, and Ki-67 in colorectal

adenocarcinoma: a study of 356 surgically resected cases.

Hepatogastroenterology 57, 734-740 (2010).

16. Kanavaros, P. et al. Expression of p53, p21/waf, bcl-2, bax, Rb and Ki67 proteins

in colorectal adenocarcinomas. Med Oncol. 16, 23-30 (1999).

17. Petrisor, O., Giusca, S.E., Sajin, M., Dobrescu, G. & Caruntu, I.D. Ki-67, p53 and

bcl-2 analysis in colonic versus rectal adenocarcinoma. Rom. J. Morphol. Embryol.

49, 163-171 (2008).

18. Lin, M.X., Wen, Z.F., Feng, Z.Y. & He, D. Expression and significance of Bmi-1

and Ki67 in colorectal carcinoma tissues. Ai Zheng 27, 1321-1326 (2008).

19. Ma, Y.L. et al. Immunohistochemical analysis revealed CD34 and Ki67 protein

expression as significant prognostic factors in colorectal cancer. Med Oncol. 27,

304-309 (2010).

20. Haensgen, G. et al. Tumor hypoxia, p53, and prognosis in cervical cancers. Int. J.

Radiat. Oncol. Biol. Phys. 50, 865-872 (2001).

21. Akasofu, M. & Oda, Y. Immunohistochemical Detection of P53 in Cervical

Epithelial Lesions with or without Infection of Human Papillomavirus Type-16

and Type-18. Virchows. Arch. 425, 593-602 (1995).

22. Hunt, C.R., Hale, R.J., Buckley, C.H. & Hunt, J. p53 expression in carcinoma of

the cervix. J. Clin. Pathol. 49, 971-974 (1996).


23. Mittal, K.R., Lin, O., Chan, W., Goswami, S. & Demopoulos, R.I. Cervical

squamous dysplasias and carcinomas with immunodetectable p53 frequently

contain HPV. Gynecol. Oncol. 58, 289-294 (1995).

24. Kainz, C. et al. Mutant p53 in patients with invasive cervical cancer stages IB to

IIB. Gynecol. Oncol. 57, 212-214 (1995).

25. Agnantis, N.J. et al. Immunohistochemical study of ras oncogene expression in

endometrial and cervical human lesions. Eur. J. Gynaecol. Oncol. 9, 360-365

(1988).

26. Sagae, S. et al. ras oncogene expression and prognosis of invasive squamous cell

carcinomas of the uterine cervix. Cancer 63, 1577-1582 (1989).

27. Kristensen, G.B., Holm, R., Abeler, V.M. & Trope, C.G. Evaluation of the

prognostic significance of cathepsin D, epidermal growth factor receptor, and

c-erbB-2 in early cervical squamous cell carcinoma. An immunohistochemical

study. Cancer 78, 433-440 (1996).

28. Hale, R.J., Buckley, C.H., Fox, H. & Williams, J. Prognostic value of c-erbB-2

expression in uterine cervical carcinoma. J. Clin. Pathol. 45, 594-596 (1992).

29. Busmanis, I. Biomarkers in carcinoma of the cervix: emphasis on tissue-related

factors and their potential prognostic factors. Ann. Acad. Med. Singapore. 27,

671-675 (1998).

30. Oka, K., Nakano, T. & Arai, T. C-Erbb-2 Oncoprotein Expression Is Associated

with Poor-Prognosis in Squamous-Cell Carcinoma of the Cervix. Cancer 73,

668-671 (1994).

31. Uetsuki, H. et al. Expression of a novel biomarker, EPCA, in adenocarcinomas

and precancerous lesions in the prostate. J. Urol. 174, 514-518 (2005).

32. Dhir, R. et al. Early identification of individuals with prostate cancer in negative

biopsies. J. Urol. 171, 1419-1423 (2004).

33. Jiang, Z., Li, C.Z., Fischer, A., Dresser, K. & Woda, B.A. Using an AMACR

(P504S)/34 beta E12/p63 cocktail for the detection of small focal prostate

carcinoma in needle biopsy specimens. American Journal of Clinical Pathology.

123, 231-236 (2005).


34. Beach, R. et al. P504S immunohistochemical detection in 405 prostatic specimens

including 376 18-gauge needle biopsies. Am. J. Surg. Pathol. 26, 1588-1596

(2002).

35. Jiang, Z. et al. P504S/alpha-methylacyl-CoA racemase: a useful marker for

diagnosis of small foci of prostatic carcinoma on needle biopsy. Am. J. Surg.

Pathol. 26, 1169-1174 (2002).

36. Leite, K.R. et al. Cdx2, cytokeratin 20, thyroid transcription factor 1, and

prostate-specific antigen expression in unusual subtypes of prostate cancer. Ann.

Diagn. Pathol. 12, 260-266 (2008).

37. Parwani, A.V. et al. Immunohistochemical staining of precursor forms of

prostate-specific antigen (proPSA) in metastatic prostate cancer. Am. J. Surg.

Pathol. 30, 1231-1236 (2006).

38. Varma, M., Morgan, M., Jasani, B., Tamboli, P. & Amin, M.B. Polyclonal

anti-PSA is more sensitive but less specific than monoclonal anti-PSA -

Implications for diagnostic prostatic pathology. American Journal of Clinical

Pathology. 118, 202-207 (2002).

39. Garcia-Velasco, A. et al. Prognostic value of hormonal receptors, p53, ki67 and

HER2/neu expression in epithelial ovarian carcinoma. Clin. Transl. Oncol. 10,

367-371 (2008).

40. Rao, B.R. & Slotman, B.J. Endocrine role in ovarian cancer. Endocr-Relat. Cancer.

3, 309-326 (1996).

41. Lee, B.H., Hecht, J.L., Pinkus, J.L. & Pinkus, G.S. WT1, estrogen receptor, and

progesterone receptor as markers for breast or ovarian primary sites in metastatic

adenocarcinoma to body fluids. American Journal of Clinical Pathology. 117,

745-750 (2002).

42. Kupryjanczyk, J. et al. p53 gene mutations and protein accumulation in human

ovarian cancer. Proc. Natl. Acad. Sci. USA. 90, 4961-4965 (1993).

43. Lobna Ayadi. et al. Correlation Between Immunohistochemical Biomarkers

Expression and Prognosis of Ovarian Carcinomas in Tunisian Patients. World

Journal of Oncology. 1, 118-128 (2010).


44. Rohlke, P. et al. p53 is a persistent and predictive marker in advanced ovarian

carcinomas: multivariate analysis including comparison with Ki67

immunoreactivity. J. Cancer Res. Clin. 123, 496-501 (1997).

45. Righetti, S.C. et al. A comparative study of p53 gene mutations, protein

accumulation, and response to cisplatin-based chemotherapy in advanced ovarian

carcinoma. Cancer Res. 56, 689-693 (1996).

46. Gancberg, D. et al. Sensitivity of HER-2/neu antibodies in archival tissue samples

of invasive breast carcinomas. Correlation with oncogene amplification in 160

cases. Am. J. Clin. Pathol. 113, 675-682 (2000).

47. Press, M.F. et al. Sensitivity of HER-2/neu antibodies in archival tissue Samples:

potential source of error in immunohistochemical studies of oncogene Expression.

Cancer. Res. 54, 2772-2777 (1994).


Documents

CORRECTION NOTICE Magnetoferritin nanoparticles …...CORRECTION NOTICE Magnetoferritin nanoparticles for targeting and visualizing tumour tissues Kelong Fan, Changqian Cao, Yongxin