Cancer Cell, Volume 16

Supplemental Data

Tissue-Penetrating Delivery of Compounds

and Nanoparticles into Tumors

Kazuki N. Sugahara, Tambet Teesalu, Priya Prakash Karmali, Venkata Ramana Kotamraju, Lilach

Agemy, Olivier M. Girard, Douglas Hanahan, Robert F. Mattrey, and Erkki Ruoslahti

Figure S1. Tumor models used in the phage display selections

(A) PPC1 tibia xenografts: Upper panel, X-ray picture taken with Image Station In Vivo FX

of the area shown in the lower panel. Arrows point to the tumors in the tibia.

(B) GFP-PC-3 tumors disseminated from an injection of one million tumor cells into the

left ventricle of the heart. The mouse was imaged under UV light using the Illumatool

Bright Light System LT-9900. Tumors that grew in the bones (e.g., jaws, radius, femur)

were used for the phage selection.

Figure S2. Homing of synthetic iRGD peptide to orthotopic xenografts and

spontaneous mouse tumors

Approximately 200 µg of FAM-iRGD was intravenously administered to mice bearing

tumors. After 2 hrs, organs were collected and viewed under UV or white light.

Arrowheads point to the tumors.

(A) Xenografts were brain and tibia xenografts of a human prostate cancer PPC1, and

orthotopic xenografts of 22Rv1, a human pancreatic carcinoma MIA PaCa-2, and a

human breast cancer BT474. The bladder in the 22Rv1 model is strongly fluorescent due

to excreted peptide in the urine.

(B) De novo mouse tumors were pancreatic islet tumors of RIP-Tag2 mice and cervical

tumors of K14-HPV16 mice.

Figure S3. iRGD does not accumulate in normal organs

Confocal images of normal organs from mice bearing 22Rv1 orthotopic tumors injected

with FAM-iRGD are shown. Representative images from 3 tumors are shown. Green,

FAM-iRGD; red, CD31; blue, DAPI. Scale bars = 100 µm. (Compare with Figure 2B).

Figure S4. iRGD phage extravasates and penetrates into tumor tissue and cells

Confocal images of PDAC tumors from transgenic mice injected with iRGD, RGD-4C, or

CG7C phage. Arrows point to blood vessels positive for phage, and asterisks denote the

tumor ducts. Green, T7 phage; red, CD31; blue, DAPI. The upper right panel shows a

magnified view of the dotted square in the upper left panel. The inset in the RGD-4C

phage panel shows a blood vessel targeted by this phage. Note that iRGD phage has

spread widely through the tumor parenchyma, while RGD-4C phage targets the blood

vessels but stays in close association with the vasculature. CG7C phage shows no tumor

homing. Representative images from three independently studied tumors for each phage

are shown. Scale bars = 50 µm.

Figure S5. Whole body imaging of PDAC mice injected with FAM-iRGD micelles or

FAM control micelles labeled with Cy7

Images were taken 3 hrs after the injection of the micelles. Only the shaved areas of the

skin delineated by the blue dotted lines are shown. Green, 800 nm (Cy7); red, 700 nm

(background fluorescence). Representative images from three independently studied

mice for each micelle are shown. The saw tooth appearance of the fluorescent organs is

caused by the mouse’s breathing.

Figure S6. Integrin-mediated binding of iRGD phage to PPC1 cells

(A) Integrin expression in PPC1 cells analyzed by flow cytometry. The profiles represent

the values of unstained cells (light gray), and cells incubated with mouse IgG (dark gray)

or appropriate integrin antibodies (unshaded with arrows) as primary antibodies.

(B) Inhibition of iRGD phage binding to PPC1 cells by antibodies against integrins or

control mouse IgG. Statistical analysis was performed with Student’s t-test. n = 3; error

bars, s.e.m.; triple asterisk, p < 0.001.

Figure S7. Integrin expression in M21 cells

Integrin expression in M21 cells and its variant were analyzed by flow cytometry. The

profiles represent the values of unstained cells (light gray), and cells incubated with

mouse IgG (dark gray) or appropriate integrin antibodies (unshaded with arrows) as

primary antibodies.

Figure S8. Peptide fragments recovered from PPC1 cells treated with FAM-iRGD

peptides

PPC1 cells were incubated with FAM-iRGD peptide (FAM at the N-terminus) at 37˚C for

90 min in the presence of a proteasome inhibitor MG132. Peptide fragments were

recovered with an anti-FITC affinity column and analyzed by mass spectrometry. Note the

presence of FAM-CRGDK [M + H] (m/z: 1,049) and FAM-CRGDK [M + Na] (m/z: 1,072),

and the absence of the full-length FAM-iRGD (m/z: 1,419). No major peptide fragments

were recovered from mouse IgG column used as an isotype control for the anti-FITC

affinity column, or when a lysate of PPC1 cells that had not been exposed to FAM-iRGD

was fractionated on anti-FITC (not shown). An iRGD peptide labelled with FAM at the

C-terminus yielded no GPDC-FAM (mw: 988; an expected fragment of iRGD after

CRGDK is cleaved off) from the cells (not shown). This suggests that only the

neuropilin-1-binding N-terminal fragment (CRGDK) internalizes, which could happen if

the iRGD peptide is proteolytically cleaved at the K-G bond and the disulfide bond is

reduced before internalization. Omitting MG132 yielded only peptides smaller than

FAM-CRGDK (not shown), suggesting that intracellular FAM-CRGDK is degraded in

proteasomes.

Figure S9. Penetration of iRGD and CRGDK phage into PPC1 and M21 cells

Cells were treated with iRGD or CRGDK phage for 1 hr at 37˚C followed by an acid buffer

wash to remove the phage that bound to the cell surface. Note that CRGDK phage

penetrated more efficiently than iRGD phage into PPC1 cells that have high expression of

neuropillin-1, whereas it is reversed in M21 cells. Statistical analysis was performed with

Student’s t-test. n = 3; error bars, s.e.m.; triple asterisk, p < 0.001.

Figure S10. Confocal microscopy images of PPC1 cells incubated with phage

expressing iRGD and other RGD peptides

The cells were stained for phage (green), neuropilin-1 (red), and nuclei (blue). The right

panels of the upper two panels are high magnification view of the dotted area in the left.

Note that iRGD phage penetrated PPC1 cells and colocalized with neuropilin-1

(arrowheads), whereas an iRGD phage variant CRGDGGPDC, CRGDC, and RGD-4C

failed to do so (arrows). Representative images from three independently studied

experiment for each phage are shown. Scale bar = 20 µm.

Figure S11. Expression of av integrins and neuropilin-1 in 22Rv1 tumor cells.

(A) αv integrin and neuropilin-1 expression in cultured 22Rv1 cells analyzed by flow

cytometry. The profiles represent the values of cells incubated with rabbit IgG (gray), an

αv-integrin antibody (left, black), or a neuropilin-1 antibody (right, black).

(B) Confocal images of orthotopic 22Rv1 tumors from mice injected with 200 μg of

FAM-iRGD peptide (green). Serial cryosections were stained for av integrins (red, upper

panel) or neuropilin-1 (red lower panel), and CD31 (magenta). Representative images

from three tumors are shown. Scale bars = 100 μm.

Figure S12. Cytotoxicity and tumor homing of abraxane conjugates

(A) In vitro tumor cell treatment with various abraxane conjugates. Cultured 22Rv1 cells

were treated for 30 min at room temperature with non-targeted abraxane (ABX), or

abraxane conjugated with iRGD (iRGD-ABX) or CRGDC (CRGDC-ABX). The cells were

incubated for 48 hrs and cell death was quantified by MTT assays (n = 3). Statistical

analyses were performed with Student’s t-test. Error bars, s.e.m.; single asterisk, p <

0.05; double asterisk, p < 0.01; triple asterisk, p < 0.001.

(B) In vivo tumor homing of abraxane conjugates. Confocal microscopy images of 22Rv1

ortjotopic tumors from mice injected with the indicated abraxane conjugates at a

paclitaxel equivalent of 3 mg/kg. The particles were allowed to circulate for 3 hrs.

Representative images from three tumors for each conjugate are shown. Green,

abraxane; red, CD31; blue, DAPI. Scale bars = 100 μm.

Figure S13. Subcutaneous 22Rv1 tumor treatment with iRGD-coated abraxane

(A) Abraxane quantification in a subcutaneous 22Rv1 xenograft model. Abraxane was

intravenously injected into tumor mice 3 hours earlier and captured from tumor extracts

with a taxol antibody, followed by detection with a human albumin antibody. n = 3 for each

group.

(B) Mice bearing subcutaneous 22Rv1 xenografts were intravenously injected with

peptide-coated abraxanes every other day at 3 mg paclitaxel/kg/injection. The treatment

was continued for 12 days. The number of mice per group was 8.

Statistical analysis was performed with Student’s t-test in (A) and ANOVA in (B). Error

bars, s.e.m.; n.s., not significant; single asterisk, p < 0.05; triple asterisk, p < 0.001.

Figure S14. Integrin and neuropilin expression on cultured BT474. The expression

of αv integrins and neuropilin-1 in cultured BT474 tumor cells analyzed by flow cytometry.

The profiles are from cells incubated with rabbit IgG (gray), an αv-integrin antibody (left,

black), or a neuropilin-1 antibody (right, black).

Figure S15. Cytotoxicity of abraxane conjugates to cultured BT474 cells

In vitro tumor cell treatment with various abraxane conjugates. Cultured BT474 cells were

treated for 30 min at room temperature with non-targeted abraxane (ABX), or abraxane

coated with iRGD (iRGD-ABX) or CRGDC (CRGDC-ABX). The cells were incubated for

48 hrs and cell death was quantified by MTT assays (n = 3). Note that about 300 fold more

abraxane is required to provide similar cell toxicity in the BT474 cells than in 22Rv1 cells

(see Figure 12A). Statistical analyses were performed with Student’s t-test. Error bars,

s.e.m.; single asterisk, p < 0.05; double asterisk, p < 0.01.