Cell Culture

Cell lines used in this study included the human mammary adenocarcinoma derived MCF7^GFP-IBD (MCF7^Tet-On (Clontech, Mountain View, CA, USA) stably transfected with doxycycline-inducible GFP fused to the 53BP1 ionizing radiation induced foci binding domain³⁴), TUBO murine mammary carcinoma (derived from BALB-neuT), and B16-F10 murine melanoma (ATCC, Manassas, VA, USA). Cells were cultured in complete growth medium (based on ATCC recommendations for each cell line) supplemented with 1 U/mL penicillin and 1 μg/mL streptomycin, then resuspended in sterile 1× DPBS at a concentration of 1 × 10⁷ cells/mL for injection of 100 μL/mouse. All cell lines used in this study tested negative for mycoplasma.

Mice

Six-week-old C57BL/6, athymic nude, and BALB/c mice were purchased from Harlan Laboratories (Madison, WI, USA). One × 10⁶ in vitro cultured tumor cells (B16F10, TUBO) or 5 × 10⁶ cells (MCF7^GFP-IBD) were injected subcutaneously into the hindlimb of mice to induce tumor growth. In the case of MCF7^GFP-IBD tumors, athymic nude female mice received a surgical implant of an estrogen pellet (1.7 mg 17B-estradiol per pellet, 60 day release, Innovative Research of America, Sarasota, FL, USA) subcutaneously in their backs prior to tumor cell injection to aid in tumor growth. Female mice were used for all experiments.

Doxil (10 mg/kg body weight) was administered via retro-orbital injection in a volume of 100 μL. Tumor size was measured biweekly using calipers with tumor volume calculated using the formula V = lw(h/2). Simvastatin (10 mg/kg, Sigma-Aldrich, St. Louis, MO, USA) was administered via i.p. injection 3 days per week for 1 week prior to and after irradiation. Vascular endothelial growth factor receptor-2 antibody (20 mg/ kg, clone DC101, Bio X Cell, West Lebanon, NH, USA) was administered via i.p. injection once the day before irradiation and once more 2 days after. Imatinib mesylate (100 mg/kg, Selleck Chemicals, Houston, TX, USA) was administered via i.p. injection daily for 1 week prior to and after irradiation. All animal studies were performed in compliance and with the approval of the University of Chicago Institutional Animal Care and Use Committee, ACUP# 70931.

Irradiation

An X-RAD 225Cx small animal irradiator (Precision X-ray Inc., N. Bradford, CT, USA) was used to plan and deliver X-ray ionizing radiation in a two-step process involving computed tomography²⁹ imaging followed by image-guided delivery of a precise treatment dose. The X-ray source was used for both imaging (1.0 mm focal spot) and treatment (5.5 mm focal spot). The subject was immobilized using isoflurane anesthesia and secured on the irradiator stage using surgical tape. Treatment planning began using the X-RAD 225C software and involved two sequential CT images taken through a 2.0 mm Al filter including an initial scout (40 kVp, 0.5 mA, 0.3 mm isotropic voxels), followed by a more detailed full scan (40 kVp, 2.5 mA, 0.1 mm isotropic voxels). DICOM files of the detailed CT scan were exported into a treatment planning application written in MATLAB (The MathWorks Inc., Natick, MA, USA) for selection of the treatment isocenter, planning of an irradiation protocol, and evaluation of the selected treatment fields. In the present experiments, irradiation protocols were designed with fields from two opposing directions, each delivering half the desired total dose. The software also calculated the 3D shift of the animal support stage required to position the chosen target at the radiation isocenter for treatment. Radiotherapy was performed following import of the treatment protocol into the XRAD 225C software, which automated delivery of the planned fields. Treatments were performed at 225 kVp, 13 mA, and 0.3 mm Cu filter, with either 1.5 or 2.0 cm diameter lead collimators providing a dose rate of ~2.5 Gy/min. The irradiator output was calibrated according to the American Association of Physicists in Medicine Task Group 61³⁵ protocol using a Farmer-type chamber. Doses ranging from 2 to 15 Gy were used in this study with total treatment time lasting up to 20 min, including setup.

Intravital Imaging

For imaging studies, mice were anesthetized by inhalation of isoflurane gas or i.p. injection of ketamine (90 mg/kg) and xylazine (10 mg/kg). AngioSense 750 (100 μL) (PerkinElmer, Waltham, MA, USA), fluorescein-labeled tomato lectin³⁶ (50 μL in 100 μL of 1× DPBS) (Vector Laboratories, Burlingame, CA, USA), and Doxil (10 mg/kg) (courtesy of the University of Chicago hospital infusion pharmacy) were administered via i.v. tail vein or retro-orbital injection. To examine fluorescence in vivo, the following instruments were utilized:

The Olympus OV100 (Olympus Corporation, Center Valley, PA, USA) is equipped with GFP, RFP, and near-infrared (NIR) fluorescence filter sets enabling multichannel coimaging. For imaging with the NIR blood pool agent AngioSense 750 and NIR iron oxide nanoparticle AngioSPARK 750 (PerkinElmer), these probes were injected via the tail vein (2 nmol/mouse), and fluorescence probe imaging was acquired at 1 min, 30 min, 4 h, 24 h, 3 d, 7 d, and 14 d. Eight-bit planar images were acquired through GFP and the 750 nm channels. To study the whole body distribution of the probes, the lowest magnification lens of 0.14× was used, and the 3–10× zoom level lenses were used to study the blood vessel and microvessel leakage. Tumors were dissected and imaged using the GFP channel for fluorescein-lectin (λ_EX 493 nm, λ_EM 518 nm), RFP channel for Doxil (λ_EX 500 nm, λ_EM 560 nm), and 750 channel for AngioSense (λ_EX 745 nm, λ_EM 800 nm) fluorescence. Imaging analysis was conducted using ImageJ (NIH)³⁷ for brightness and contrast adjustment.

The Xenogen IVIS 200 (Caliper Life Sciences, Hopkinton, MA, USA) was used at 1 h, 4 h, 24 h, 3 d, 7 d, and 14 d after AngioSense injection to noninvasively image and quantitatively measure fluorescent probe permeation and retention in differentially treated tumors. An isoflurane gas anesthesia machine (XG-8 Gas Anesthesia System) was used to anesthetize mice prior to and during imaging. Dissected tumors, as well as other tissues such as liver, spleen, heart, kidney, and lung, were imaged for Doxil loading, AngioSense, and AngioSPARK fluorescence. The radiant efficiency, a relative measure of photon emission from the animal (photons/s/cm²/ sr), was measured in a standardized region of interest (ROI) with variables such as exposure time, binning, and f/stop also standardized. Intestinal Doxil retention was unable to be examined due to autofluorescence from ingested chow.³⁸

To study the fluorescent agents at a higher magnification and resolution, a Leica SP5 Tandem Scanner Spectral 2-Photon confocal microscope (Leica Microsystems, Buffalo Grove, IL, USA) was used for live imaging. This machine provides excitation of six visible laser lines and a tunable NIR pulsed laser (Spectra Physics Mai Tai broadband 710–990 nm). Mice were anesthetized by i.p. injection of ketamine (90 mg/kg) and xylazine (10 mg/kg). Skin flap surgery was used to expose tumor tissue and blood vessels while also avoiding any autofluorescence from the skin. The exposed tumors were placed in a glass-bottom Petri dish filled with 1× PBS, and the dish was placed on the microscope stage for imaging. AngioSense, fluorescein-labeled tomato lectin, Doxil, and SAIVI nanoparticles (100 μL, Invitrogen) were administered via i.v. tail vein or retro-orbital injection. Time course changes in agent fluorescence were imaged with a 20× objective lens plus digital zoom. After imaging, the tumors were dissected for other optical imaging followed by embedding for sectioning.

Immunohistochemistry

Tumors were dissected from mice immediately following sacrifice and were either fixed in 10% formalin for paraffin embedding or embedded in OCT and stored at −80 °C. Tissue processing, embedding, sectioning, and staining were performed by the Human Tissue Resource Center at the University of Chicago. Antibodies for CD31 (Santa Cruz Biotechnology, Inc., Dallas, TX, USA, sc-1506, 1:200), α-SMA (DAKO, Carpinteria, CA, USA, M0851, 1:100), cleaved Caspase 3 (Cell Signaling Technology, Danvers, MA, USA, #9661, 1:200), Collagen Type IV (EMD Millipore, Billerica, MA, USA, AB8201, 1:100), and Collagen Type I (Abcam, Cambridge, MA, USA, ab34710, 1:1600) were used for IHC of tissue sections. These antibodies are routinely used throughout the literature as markers for the cell types and tissues specified herein, though some off-target signal detecting other cell types may be observed. After deparaffinization and serial rehydration, tissue sections were treated with citrate buffer pH 6.0 (S1699, DAKO) in a microwave oven. Primary antibodies were applied on tissue sections for 1 h at room temperature in a humidity chamber. Following a TBS wash, tissue sections were incubated with biotinylated anti-rabbit IgG (1:200, BA-1000, Vector laboratories) for 30 min at room temperature. Antigen–antibody binding was detected using the VECTASTAIN Elite ABC kit (Vector Laboratories, PK-6100) and diaminobenzidine (DAKO, K3468) system. Tissue sections were briefly immersed in hematoxylin for counterstaining and were covered with cover glasses.

For tissue immunofluorescence, cryosections were blocked with 10% goat serum and 0.1% Triton X-100 in PBS and incubated with 1:100 primary antibody overnight at 4 °C, washed, incubated with 1:400 fluorescent secondary antibodies (Invitrogen, Carlsbad CA, USA) for 30 min, washed, and then mounted in 4′,6-diamidino-2-phenyl-indole mounting medium (Vector Laboratories).

Images were captured using the Zeiss Axioskop, Zeiss Axiovert 200, and Leica SP-2 confocal microscopes. Relative quantification of staining was performed using the Color Deconvolution plugin for ImageJ to separate diaminobenzidine staining from hematoxylin. Three to five images per tumor were used for quantification. The brightness, contrast, color saturation, and color temperature of the representative images shown in Figures 2 and ​and33 and Supplementary Figures 1, 2, and 3 were adjusted in Microsoft Powerpoint in order to normalize the background coloring and brightness.

Open in a separate window

Figure 2

Radiation alters the tumor microenvironment to enhance delivery and retention. (A) Dose response of radiation-enhanced delivery examined by confocal microscopy of MCF7^GFP-IBD tumors injected with the endothelial stain tomato lectin and blood pool agent AngioSense demonstrates a threshold between 2 and 5 Gy at 3 days after irradiation. Scale bar = 100 μm. (B) Hematoxylin-stained (purple) tissue sections of MCF7^GFP-IBD xenograft tumors excised at 17 days after irradiation display dose-dependent changes in immunohistochemical staining (brown) for CD31 (endothelium), α-SMA (pericytes), Collagen IV (basement membrane), and Collagen I (extracellular matrix). Scale bar = 200 μm. (C) Quantification of immunohistochemical staining by percent of microscopic field. *p ≤ 0.05 relative to 0 Gy control, n = 3. (D) Transmission electron microscopy of thin sections of MCF7^GFP-IBD xenograft tumor excised 3 days after 5 Gy demonstrates disorganization of interstitial collagen (*). EC = endothelial cell, BL = basal lamina, L = vessel lumen. Scale bar = 2 μm.

Open in a separate window

Figure 3

Independence of radiation-enhanced delivery and endothelial integrity. (A) Immunohistochemistry of MCF7^GFP-IBD xenograft tumors excised 17 days after irradiation suggests simvastatin protects endothelial staining (brown, CD31), preventing depletion after treatment with 5 Gy. No significant effects were seen on α-SMA, Collagen IV, or Collagen I (brown). Purple = hematoxylin, nuclei. Scale bar = 200 μm. (B) Simvastatin protects endothelium against destruction after 15 Gy. (C) Relative quantification of IHC staining in MCF7^GFP-IBD tumor sections. % area denotes the area of an image stained calculated using an ImageJ macro (details in methods). *p ≤ 0.05 relative to no treatment control, n = 3. (D) In vivo imaging revealed that, despite endothelial integrity (green, tomato lectin), AngioSense still permeated from the vasculature into the MCF7^GFP-IBD tumor interstitium. Scale bar = 100 μm. (E) MCF7^GFP-IBD tumor retention of AngioSense, measured using IVIS fluorescence quantification, does not appear limited by the endothelial protection afforded by simvastatin treatment. AngioSense was administered 3 days after IR. *p ≤ 0.05 relative to no treatment control; ^#p ≤ 0.05 relative to statin only, at the same time point, n = 5.

Transmission Electron Microscopy

Dissected tumors were fixed overnight in a 2% glutaraldehyde, 4% paraformaldehyde in 0.1 M sodium cacodylate buffer then washed with 0.1 M sodium cacodylate buffer 3× for 5 min each. Buffer was then removed, and the tumor samples were incubated in 1% osmium tetroxide in 0.1 M sodium cacodylate buffer for 60 min, followed by two 5 min washes with sodium cacodylate buffer, and one 5 min wash with pH 5.1 maleate buffer. Tumor samples were then stained in 1% uranyl acetate in maleate buffer for 60 min, followed by three 5 min washes in maleate buffer and serial dehydration. Polymerization with Spurr resin was performed in an embedding oven at 60 °C for 1 to 2 days after the tumors were infiltrated by 2:1 propylene oxide/Spurr resin 2× for 30 min, then 1:1 propylene oxide/Spurr resin 2× for 30 min, and finally 100% Spurr resin 6× for 60 min. The blocks were trimmed and cut to 90 nm ultrathin sections using a Reichert-Jung Ultracut E and mounted on 200 mesh thin bar copper grids. The specimens then were stained with uranyl acetate and lead citrate. Each specimen was examined under 300 kV in a FEI Tecnai F30 transmission electron microscope (FEI, Hillsboro, OR, USA) with a Gatan CCD digital camera (Gatan, Pleasanton, CA, USA).

Radiation Effects on the Microenvironment Influencing Tumor Delivery

Since the advent of fractionated radiotherapy, efforts have focused on optimizing dose and delivery to obtain maximal cytotoxic effects on tumor cells while sparing normal tissue. Nonetheless, studies have long implicated damage to tumor stroma and vasculature as critical determinants of success for therapeutic radiation. Endothelial cells serve a key role in the microvascular permeability barrier and are considered critical targets of radiation.^22,41 Using fluorescein-conjugated tomato lectin to label endothelial cells, in vivo confocal imaging revealed bright, continuous staining facing the lumen of microvessels in unirradiated MCF7^GFP-IBD xenograft tumors. However, by 3 days after 5 or 15 Gy, only weak and patchy staining of endothelium was observed (Figure 2A) concomitant with increased extravasation of AngioSense into the tumor parenchyma. Immunohistochemistry (IHC) using endothelial cell-specific CD31 antibody confirmed these results, demonstrating a significant decrease in immunoreactivity (p ≤ 0.01) after 5 or 15 Gy (Figure 2B,C). The damage to endothelial cells was not associated with increased perivascular apoptosis (Figure 2C, Supplementary Figure 1). Pericytes are contractile cells that form a sheath around endothelial cells, controlling capillary diameter and perfusion while contributing to the regulation of permeability.^42,43 Like endothelial gaps in tumor vasculature, incomplete coverage by pericytes may contribute to tumor vascular permeability and the passive EPR effect.⁴⁴ IHC using alpha smooth muscle actin (α-SMA) as a pericyte marker revealed a significant loss of α-SMA staining (p ≤ 0.05) 17 days after treatment with 2 to 15 Gy in MCF7^GFP-IBD xenograft tumors (Figure 2B,C).

To detect radiation effects on the extracellular matrix surrounding the tumor vasculature, we performed IHC for collagen types I and IV, major components of the stroma and basal lamina, respectively, that may contribute to the barrier limiting macromolecular delivery. A significant increase in collagen I staining was observed following 15 Gy IR, but no change was detected in the lower doses studied (Figure 2B,C). Irradiation did not significantly alter anti-collagen IV immunoreactivity either but the characteristic perivascular organization appeared disrupted (Figure 2B,C). Thin section transmission electron microscopy (TEM) at 3 days after a single 5 Gy dose demonstrated disordered alignment and packing of perivascular collagen fibers, consistent with the collagen IV IHC results (Figure 2D).

Radiation-Enhanced Tumor Delivery Independent of Loss of Endothelial Integrity

Since irradiation dose-dependent effects were observed on endothelial cells in the tumor stroma, we sought to examine whether endothelial damage might be necessary or sufficient to observe an enhanced delivery effect. Beyond their effects on cholesterol, HMG-CoA reductase inhibitors (statins) can suppress injury-induced endothelial permeability (e.g., Jacobsen et al.⁴⁵) including radiation-induced damage.^46,47 Statin effects on the microvasculature have been ascribed to reduced geranyl geranylation of Rho and/or other small GTPases. Treating mice every other day with simvastatin for a week prior to and 1 week following irradiation prevented endothelial disruption, even following 15 Gy irradiation (Figure 3A–C). Despite the rescue of CD31 immunoreactivity in mice treated with simvastatin and 5 Gy, radiation still induced enhanced macromolecular delivery (Figure 3C–E).

As a complementary test for whether damaging the endothelium might be sufficient to recapitulate the effects of radiation on macromolecular delivery, we transiently stripped tumor microvasculature of endothelium by treating tumor-bearing animals with a high dose of an antibody targeting the vascular endothelial growth factor receptor-2 (VEGFR-2).^48,49 When examined at 24 h after administration of anti-VEGFR-2, tumor microvasculature displayed significantly decreased CD31 (endothelium), α-SMA (pericyte), and collagen IV (basal lamina) immunoreactivity. However, the endothelial damage did not affect either passive EPR-dependent or radiation-enhanced delivery of AngioSense (Supplementary Figure 2A–C).

An alternative target of radiation affecting delivery might be pericytes. Imatinib mesylate, an inhibitor of platelet-derived growth factor receptor β (PDGFRβ), can deplete pericytes and disrupt tumor vascular integrity.⁵⁰ Inhibition of PDGFRβ-signaling by imatinib led to significant decreases in not only tumor pericyte staining but also endothelial and basal lamina immunoreactivity (Supplementary Figure 3A,B). Imatinib alone appeared to increase passive delivery above that due to the passive EPR effect. In combination with radiation, imatinib did not augment delivery above that displayed by radiation alone (Supplementary Figure 3C). These data are consistent with pericyte integrity and/or basal lamina as radiation targets.

Taken together, the lack of an increase in radiation-enhanced delivery following additional endothelial damage, plus the failure of simvastatin to suppress radiation-enhanced delivery despite protecting the endothelium from IR-associated damage, suggest that damage to pericytes, basement membrane, and/or extracellular matrix might have a primary role.

This work was supported by the NCI via R01s CA164492 and CA199663 and by the University of Chicago Ludwig Center for Metastasis Research. We acknowledge Dr. Helena Mauceri for assistance with animal studies, Drs. Chin-Tu Chen and Patrick La Riviere and the staff of the Integrated Small Animal Imaging Research Resource (iSAIRR) for assistance with animal imaging, Dr. Vytas Bindokas for assistance with microscopy, Yimei Chen for performing electron microscopy, the staff of the Human Tissue Resource Center for tissue processing, embedding, and immunohistochemistry staining, and the Infusion Center pharmacy for providing PEGylated liposomal doxorubicin.