Local and Systemic Antitumor Effects of Photo-activatable Paclitaxel Prodrug on Rat Breast Tumor Models

Abstract

We demonstrated that a large primary and a small untreated distant breast cancer could be controlled by local treatment with our light-activatable paclitaxel (PTX) prodrug. We hypothesized that the treated tumor would be damaged by the combinational effects of photodynamic therapy (PDT) and locally released PTX and that the distant tumor would be suppressed by systemic antitumor effects. Syngeneic rat breast cancer models (single- and two-tumor models) were established on Fischer 344 rats by subcutaneous injection of MAT B III cells. The rats were injected with PTX prodrug (dose: 1 umole kg⁻¹, i.v.), and tumors were treated with illumination using a 690-nm laser (75 or 140 mW cm⁻¹ for 30 min, cylindrical light diffuser, drug-light interval [DLI] 9 h). Larger tumors (~16 mm) were effectively ablated (100%) without recurrence for >90 days. All cured rats rejected rechallenged tumor for up to 12 months. In the two-tumor model, the treatment of the local large tumor (~16 mm) also cured the untreated tumor (4–6 mm) through adaptive immune activation. This is our first demonstration that local treatment with our PTX prodrug produces systemic antitumor effects. Further investigations are warranted to understand mechanisms and optimal conditions to achieve clinically translatable systemic antitumor effects.

INTRODUCTION

Dr. Thomas Dougherty pioneered modern photodynamic therapy (PDT), in which low energy light (visible and near IR), is used as a non- or minimally invasive treatment regime (1-3). PDT is effective for local and regional tumors, with high specificity and without major systemic side effects (4,5). Photofrin (first-generation photosensitizer)-PDT was FDA-approved for esophageal and endobronchial cancers and high-grade dysplasia in Barrett’s Esophagus (5,6). Second-generation photosensitizers were approved in many countries for other types of cancers (7). While PDT showed promising clinical outcomes for chest wall recurrence of breast cancers (8), PDT was not approved for breast cancer treatment. Treating primary breast cancers with PDT using topical illumination has been challenging. It may not be practically feasible to treat such deep-seated large primary breast tumors with topical illumination. However, dramatic advances in optical and imaging technologies make PDT, for example, ultrasound image-guided interstitial PDT, applicable for primary breast cancers (9).

Large tumors are more difficult to ablate with PDT due to the limited tissue penetration of light, while small and thin tumors can be readily ablated. The heterogeneity of tumors and PDT make it even harder to completely kill all of the cells in large tumor masses (10-12). Singlet oxygen (SO) has an extremely short lifetime (~10–320 ns) and limited diffusion distance (<~270 nm compared with ~10–20-um diameter cancer cells) in biological systems (13-15). Thus, SO is generated only during the illumination, and it cannot effectively produce bystander effect. In particular, the temporal and spatial limitations of SO, the major toxic chemical species in PDT, could prevent sustained and thorough cancer cell killing. Cells surviving from PDT effect can lead to recurrence. Although bystander effects in PDT were reported possibly by secondary products (16-18), an active strategy is needed to strengthen bystander effect. It was reasonable to combine PDT and chemotherapy because chemotherapeutic drugs can be complementary with PDT for the enhanced antitumor effect (19-21). Synergistic action of PDT and chemotherapy were reported for overcoming drug resistance (20) and limited light penetration (19), for reducing the tumor burden (22-25) and for enhancing tumor accumulation of therapeutic molecules (26,27). To minimize systemic exposure of anticancer drugs, various drug delivery systems were investigated for colocalized delivery of photosensitizers and anticancer drugs such as polymer carrier (21), polymer or lipid based nanoparticles (28), silica nanocages (29) and upconversion nanoparticles (30). Our laboratory has been working on a prodrug system, which comprised anticancer drug(s) linked with a photosensitizer via SO-cleavable linker for the combination of PDT and site-specific chemotherapy (31-40).

We developed FA-PEG2K-L-Pc-PTX, a paclitaxel (PTX) prodrug, is a folate receptor (FR)-targeting and phthalocyanine (Pc)-linked PTX prodrug (Fig. 1A) (39). Upon illumination with far-red light (690 nm), the prodrug generates SO, which exerts PDT effects, and consequently releases PTX in the illuminated tumor (Fig. 1B) (36). The site-specifically released PTX that can diffuse within the tumor produces sustained bystander effects (Fig. 1C). We previously demonstrated that the FR-mediated cellular uptake of light-activatable prodrugs was more effective than were PDT effects alone (34,36,40).

Figure 1.

(A) Structure of light-activatable PTX prodrug (FA-PEG2K-Pc-L-PTX). (B) Representation of process of singlet oxygen (SO) generation and release of PTX by SO. (C) Graphical representation of interstitial illumination of subcutaneous large tumor using a cylindrical diffuser, combinational antitumor effects and local and systemic tumor controls.

Herein, we tested (1) whether a large (~10—16 mm in length) breast tumor could be treated with interstitial illumination with the aid of ultrasound imaging, (2) whether the released PTX could remain in tumor long enough and at a sufficient concentration to kill cancer cells surviving from rapid PDT damage and (3) whether the ablation of the large tumor with PTX prodrug could activate the adaptive immune system via immunogenic cell death (ICD), producing systemic antitumor effects. Relatively large (10 or 16 mm) subcutaneous rat mammary tumors were implanted on Fischer 344 rats. A small portable ultrasound (US) machine was used to guide the fiber insertion into the tumor. DLI was determined based on the pharmacokinetic profiles of the PTX prodrug to achieve both vascular and direct cell killing effects from PDT. A cylindrical light diffuser was used for more effective delivery of light inside the large tumor, overcoming the limited light penetration issue of a topical illumination. Pharmacokinetic profiles of the released PTX were determined in tumor and plasma. US imaging effectively visualized the tumor and the fiber in the tumor. Notably, we observed that not only were the large tumors (~16 mm) effectively ablated by the PTX prodrug treatment, but also that all treated large tumors were cured without recurrence for >90 days. PTX was released and remained at high concentrations (>50 nM, C_max in tumor = 200 nM) within tumors for >48 h, while its systemic concentration remained low (C_max = 34 nM). Most exciting was that the local treatment with PTX prodrug produced systemic antitumor effects, both prophylactic and systemic therapeutic. All rats whose large local tumors were ablated by the PTX prodrug rejected rechallenge with the same cancer cells, demonstrating ICD. Further, the ablation of the local large tumor with PTX prodrug produced a therapeutic systemic antitumor effect that eliminated untreated distant tumors.

MATERIALS AND METHODS

Reagents and materials.

Reagents and analytical-grade solvents were obtained from Sigma-Aldrich, VWR or Fisher Scientific and were used without further purification. Millex-HV sterile filters (0.45 μm, Durapore PVDF membrane) were purchased from Merck Millipore Ltd. FA-PEG2K-Pc-L-PTX was synthesized as reported earlier by our laboratory (39). UV–Vis absorbance was measured on an UV–Vis spectrometer LAMBDA 25 (PerkinElmer) using 10-mm optical path length quartz cuvettes. The fluorescence was recorded on a plate reader (SpectraMax Gemini EM, Molecular Devices) using a 96-well plate, 200 μL in volume, bottom reading, with excitation/emission at 605/680 nm. Alexa Fluor^® 647 anti-rat CD45 antibody (202212, BioLegend) was purchased from BioLegend, CA. Fluorescence images were captured using a Leica confocal microscope SP8 (Leica Microsystems Inc., IL). The bright-field hematoxylin and eosin (H&E) images were taken using a Leica Aperio CS2 slide scanner.

Formulation.

The stock solution of PTX prodrug (4 mM) was prepared in DMSO. The concentration of the prodrug was determined by measuring UV-Vis absorbance at 678 nm with the extinction coefficient 209 967 cm⁻¹M⁻¹. Further 600 μM prodrug was prepared by adding the stock solution into Cremophor EL/Ethanol/PBS solution at 1:1:18 (v/v/v). The formulation was vortexed for 30 s and filtered using a 0.45-μm membrane filter. Concentration of the formulated PTX prodrug was measured by the absorbance at 678 nm. The formulation was prepared fresh and filtered each time for the light diffusion and animal experiments.

Light diffusion in tissue phantoms—frontal distributor vs cylindrical diffuser.

Intralipid solutions (20%) were purchased from Sigma-Aldrich. A frontal light distributor (model FD1), cylindrical light diffuser (model RD10) and isotropic probe (model IP85) from Medlight S.A. (Switzerland) were used. A 690-nm diode laser (MDL-III-690, CNI laser, China) and the power meter (PM100D, Thorlabs) were used. Images were taken using a Canon SX500 IS camera and were analyzed by ImageJ software. Laser power output was set at 100 mW total power in all light diffusion experiments. The total volume of all intralipid solutions was 30 mL. The containers were made by modifying 50-mL Cellstar cell culture flasks (Greiner Bio-one). FA-PEG2K-Pc-L-PTX stock solution (2 μM) was prepared fresh in DMSO before the experiments. Intralipid solutions were prepared by adding 20% intralipid solution to phosphate-buffered saline. FA-PEG2K-Pc-L-PTX (2 μM) in 1% and 10% intralipid solution was prepared by adding 150 μL of FA-PEG2K-Pc-L-PTX (400 μM) formulation solution to the 30 mL solution of 1% and 10% intralipid, respectively. Images of light diffusion in 10% intralipid produced by frontal light distributor (FD) and cylindrical light diffuser (CD) were acquired using the setup depicted in Fig. 2A. A light distributor or diffuser was placed inside the intralipid solution, touching the flask’s frontal wall. The camera was positioned 12 cm from the solution and centered on the light distributor/diffuser using a camera stand. Light power distribution was measured using the setup depicted in Fig. 2C. An isotropic probe was placed touching the middle point of the cylindrical diffuser tip. The distance between the isotropic probe and cylindrical diffuser tip (r) was then gradually increased by 0.1 inch, and the power readings were recorded.

Figure 2.

(A) Setup for visualizing light distribution in intralipid tissue phantom using cylindrical light distributor and frontal light diffuser. (B) Light diffusion of cylindrical light diffuser (i) and frontal diffuser (iii) of light fibers in 10% intralipid PBS solution (tissue phantom) at 100 mW power (ii and iv) pictures of those in the air (*from vendor website). (C) Setup for measuring light power distribution at different distances perpendicular to the center of a cylindrical light distributor. (D) Relative photon density distribution produced in 0–10% intralipid (IL), with or without 2 μM PTX prodrug (PD).

Cell line, animals and tumor system.

The 13762 MAT B III (called MAT B III herein) mammary gland adenocarcinoma cell line was acquired from ATCC (cat # CRL-1666) and used to create the syngeneic subcutaneous breast tumor models. MAT B III cells were grown in McCoy’s medium (Corning, VA) supplemented with 10% fetal calf serum (FBS) and 1% penicillin–streptomycin antibiotics (Corning, VA). Four-to-six-week-old Fischer 344 female rats (120–140 g) obtained from Charles River Laboratories, Inc. were housed in microisolator cages in a temperature-controlled room with a 12 h light/dark cycle. All animal experiments were approved by the IACUC of OUHSC. For the single-tumor model, a tumor (~10 or ~16) mm was established on the shoulder of a rat by subcutaneous injection of 1 million cells in 200 μL phosphate-buffered saline (PBS). In the two-tumor model, a large tumor (~16 mm) on the shoulder and a small (~4-6 mm) tumor on the lower rear flank were established; this group was considered a “Large/Small” tumor model (Fig. 7B). In another group, two separate small tumors (~4–6 mm each) were established on the right flank; this group was considered a “Small/Small” tumor model (Fig. 7C). Cancer cells (1 million cells) were subcutaneously injected on the nontreated rear flank for rechallenge experiments. Tumor growth was monitored using digital calipers. The longest axis of the tumor (l) and the axis perpendicular to l (w) were used to calculate tumor volume (lw²/2).

Figure 7.

Prophylactic and therapeutic systemic antitumor effects of PTX prodrug. (A) Rechallenge study schedule of MAT B III cell injection using a single-tumor model. (B and C) Therapeutic systemic antitumor effect study schedule with two-tumor models (large/small and small/small tumor models). (D) Tumor growth curves of G6 with rechallenge compared with control tumors. Two-tumor growth curves of small/small (E) and large/small tumor (Fi) models: T1 (treated tumor) and T2 (nontreated distant tumor on flank region). Fii. Enlarged graph for T2 growth.

Local antitumor efficacy using the single-tumor model.

Subcutaneous breast tumors were grown as described above. Two different tumor sizes (~10 or ~16 mm) were used for evaluating the local antitumor effects of the PTX prodrug with interstitial illumination. The prodrug dose was 1 μmole kg⁻¹ based on our previous studies in mouse tumor models (34) and delivered intravenously via tail vein injection. Light (690-nm diode laser) was delivered interstitially at 75 mW cm⁻¹ (or 140 mW cm⁻¹) for 30 min through a cylindrical light diffuser (41). Tumors were illuminated 9 h postprodrug administration (DLI = 9 h). The position of the interstitial cylindrical light diffuser inside the tumor was guided by US imaging (GE Healthcare, LOGIQ e BT12, L8-18i-RS linear array transducer). During the illumination, the rats were anesthetized with 2.5% isoflurane in oxygen gas. The rats were divided into 6 groups (n = 4): Group 1: control (no treatment); Group 2: drug control (prodrug only); Group 3: light control (illumination only, 140 mW cm⁻¹); Group 4: ~10-mm tumor treated at 140 mW cm⁻¹; Group 5: ~10-mm tumor treated at 75 mW cm⁻¹; and Group 6: ~16-mm tumor treated at 75 mW cm⁻¹. Tumor volume was measured every day starting from the illumination date (day 0). After the illumination, the body weight of each rat was monitored daily.

Ultrasound imaging with power Doppler mode.

The power Doppler mode of the ultrasound machine was used to visualize any detectable active blood flow within and around the tumor, before and after illumination. A rat with a ~10-mm tumor was i.v.-injected with 1 μmole kg⁻¹ prodrug. After 9 h, the tumor was illuminated at 140 mW cm⁻¹ for 30 min using a cylindrical light diffuser. Power Doppler mode scanning was performed before, at the end of 30 min illumination, and three and eight days after illumination on the tumor using an L8-18i scanning probe, which was connected to the LOGIQ™ system.

Pharmacokinetic study of PTX prodrug.

Rats bearing ~ 10-mm s.c. tumors were injected with PTX prodrug (2 μmole kg⁻¹) via the tail vein. Blood samples were drawn from the animals at 0.08, 1, 4, 9, 24 and 48 h postdosing and were processed to collect plasma. The rats were euthanized by isoflurane overdose at specified time intervals after PTX prodrug injection (0.08, 0.5, 12, 24 and 48 h) to collect tissue of major organs, including liver, spleen, kidney, lung and intestine, muscle, skin and tumor. The collected tissues were rinsed with PBS and blotted dry. Excised tissues (~100 mg) were homogenized, centrifuged, extracted in acetonitrile (200 μL), and further reconstituted in DMSO (200 μL). Prodrug fluorescence was measured using a plate reader (Ex/Em: 605/680 nm). The concentration of the prodrug in collected serum and tissue extract was determined relative to the standard curve and expressed in μM units (with tissue density = 1 g mL⁻¹).

Quantification of released PTX in tumor and plasma.

We monitored the released PTX in tumor and plasma in rats with ~16-mm tumors, before and after illumination. The prodrug (1 μmole kg⁻¹) was administered by tail vein injection. Control blood samples were drawn at 9 h postinjection (before illumination). Blood samples were also collected at 0.5, 6, 12, 24 and 48 h postillumination (690 nm laser at 75 mW cm⁻¹ for 30 min), each time point in triplicate. Tumor biopsies were taken at the same time points and weighed, then snap-frozen until analysis. Concentrations of PTX in both rat plasma separated from blood samples and tumor tissue homogenates were determined by ultra-high performance liquid chromatography with tandem mass spectrometric detection (UHPLC-MS/MS). All samples were analyzed using a previously validated assay (39), with a calibration range of 1–1,000 ng mL⁻¹ in plasma and 5–10 000 pg mg⁻¹ in 100 mg mL⁻¹ tissue homogenate. Briefly, 10× volume of MTBE (methyl-tert-butyl ether) was added to plasma or tissue homogenate in a liquid-liquid extraction. The mixture was vortexed and centrifuged, and the supernatant was transferred to a collection plate to be dried down under nitrogen. PTX was reconstituted with (60/40/0.1, v/v/v) water/methanol/formic acid. The plate was then vortexed and centrifuged for 5 min at 2000 rpm, 4°C. Finally, the reconstituted solution was injected onto a Waters Symmetry Shield^® RP18 column, 2.1 × 50 mm, 3.5 μm before mass spectrometric detection in the positive ion mode.

Physiologically based pharmacokinetic (PBPK) modeling to predict the released PTX in plasma and tumor.

We previously developed a PBPK model to describe the kinetics of the PTX prodrug and the released PTX in plasma and tumor in mice and to compare with the kinetics of i.v.-injected PTX itself (39). The detailed model scheme and build-up process were presented in our earlier study (40). In the present study, to predict the concentration–time profiles of the released PTX in plasma and tumor, we applied the same PBPK model with rat physiological parameters (42), including organ volumes, blood flows, and fraction of vascular and interstitial space in each tissue. Additionally, PTX clearance was scaled up to rats using the allometric equation: CLrat = 0.55·BW^0.86 (43). Partition coefficients (k_p) and uptake rate (k_uptake) for the prodrug and PTX were assumed to be same between rats and mice.

Histology.

Tumors (10 and 16 mm) were collected from rats without illumination and washed with PBS. After wet blotting, the tumor tissue was embedded in an OCT mold and frozen in liquid nitrogen. Tumor sections (5–10 μm) were obtained using a cryo-microtome and sections were fixed in 95% ethanol for 15 s. The sections were stained with H&E, dehydrated through an alcohol gradient, and then covered with coverslips using mounting medium. Images of the stained tissue sections were digitalized using a bright-field whole slide scanner at 4× magnification (Aperio CS2, Leica Biosystems, IL). For immunohistochemical (IHC) staining, tissue sections with 5-to-10-μm thickness were obtained from the cryomold using a cryo-microtome. Tissue sections on glass slides were dried overnight at room temperature and fixed by immersing the slides in precooled (−20°C) acetone. Tissue sections were incubated in 0.3% H₂O₂ and rinsed with PBS. Fetal bovine serum (10%) in PBS was applied over tissue sections as a blocking agent. After washing, the sections were incubated with the Alexa Fluor^® 647 anti-rat CD45 antibody (202212, BioLegend) for 1 h at room temperature. Then, the tissue sections were washed with PBS and counter stained with DAPI. Finally, the fluorescent-labeled tissue images were captured under confocal microscopy. The Alexa Fluor^® 647-tagged antibody was excited at 594 nm and emission was collected at 665 nm. The DAPI fluorescence signal was collected by Ext/Em 358/461 nm.

Prophylactic antitumor effect using a rechallenge experiment.

The cured rats from Group 6 (~16-mm tumor treated at 75 mW cm⁻¹) in the single-tumor study were rechallenged with 1 × 10⁶ MAT B III cells at 1, 3, 6 and 12 months post-treatment (Fig. 7A). As same-age rat controls, the cells were injected into same-age rats to confirm the tumorigenicity. The rats were monitored for 90 days after 12 months of rechallenge.

Therapeutic systemic antitumor effect using two-tumor models. Two-tumor models were used to test the therapeutic systemic antitumor effects of our prodrug treatment (Fig. 7B,C).

Large/small tumor model.

A primary tumor was established on the shoulder of the rat by subcutaneous injection of 1 × 10⁶ MAT B III cells. The distant second tumor was induced on the right rear flank (1 × 10⁶ MAT B III) of the same rat on the 10^th day. The rats underwent prodrug treatment (prodrug: 1 μmole kg⁻¹) and illumination (75 mW cm⁻¹ for 30 min) as described above when the size of the primary tumor was about 16 mm and the second tumor was about 4–6 mm. The treated rats were observed for 90 days.

Small/small tumor model.

The shoulder tumor was illuminated using a frontal light diffuser at 75 mW cm⁻² for 30 min.

Statistical analysis.

Statistical analysis was performed using Prism 6 (GraphPad Software, CA). Survival curves were compared and analyzed using the log-rank test. P < 0.05 or 0.01 was considered statistically significant.

RESULTS AND DISCUSSIONS

Light diffusion in tissue phantoms

Dosimetry in interstitial PDT was investigated by a number of groups, which is recently reviewed (44). We wanted to have better sense of light diffusion used in our treatment. We visualized light distribution from two types of fibers (cylindrical light distributor and frontal diffuser) using 10% intralipid solution (optical tissue phantom) (45,46) (Fig. 2A). The cylindrical diffuser made a much larger photon distribution volume than did the frontal diffuser at the same laser power (100 mW, Fig. 2B). Thus, it is logical to choose the cylindrical light diffuser for interstitial illumination treating a large tumor (44,47). To examine the impact of concentration of intralipid (light scattering particle) and prodrug (photon absorber) on photon distribution, we also measured the photon density using an isotropic diffuser (Fig. 2C) at various vertical distances from the fiber center (r). As the concentration of intralipid increased from 0 to 10%, the photon density near the fiber increased (2–81 AU at the contact) due to the higher scattering by intralipid particles (Fig. 2D). With the addition of PTX prodrug (2 μM) to the solution, the photon density decreased (16→9 and 81 → 35 AU at the contact for 1 and 10% IL solution, respectively) due to the absorption by the prodrug.

Determination of DLI based on pharmacokinetics of PTX prodrug

Tumor damage by illumination with PTX prodrug can occur through four combinational mechanisms: three PDT effects (direct cell kill, vascular damage and antitumor immune activation) and site-specific PTX chemotherapeutic effects (Fig. 1C). Illumination time point, aka DLI (drug-light interval), is one of the determinant parameters in PDT, determining tumor damage mechanisms (vascular vs direct cell kill) and overall antitumor efficacy (40,48,49). To achieve maximum tumor damage by both vascular damage and direct cell killing, we determined the pharmacokinetic profiles of the PTX prodrug. Following the i.v. injection of 2 μmole kg⁻¹, the prodrug plasma concentration–time profile exhibited a multi-exponential decline with a long terminal half-life (t_1/2 = 49 h, vs ~5.6 h for Taxol [Cr EL/EtOH-formulated PTX]) (50) and a small volume of distribution (Vss = 56 mL kg⁻¹, vs 4939 mL kg⁻¹ for Taxol) (Fig. 3). The longer retention of the prodrug might be caused by its larger molecular size, which could reduce its renal clearance. Receptor-mediated uptake could also trap the prodrug in cells, mostly in lysosomes (39), for a long time. Lysosome trapping of PTX prodrug was reported in our previous study (39). The 2K PEG spacer increased the solubility of prodrug, which could restrict the distribution of the prodrug to tissues, compared with PTX. Uptake was high in highly perfused tissues, such as the liver, kidney, spleen and lung, with slow clearance rates. Uptake in skin, muscle and intestine was much lower. In particular, uptake to muscle was much lower than was uptake to the other tissues. At 9 h postadministration, the prodrug showed relatively high concentrations in both plasma (10 μM) and tumor (2.4 μM), similar to our mouse PK data (40). Thus, a DLI of 9 h was chosen to achieve maximum tumor damage via both vascular and cellular PDT effects (40). At 9 h, tumor/skin and tumor/muscle concentration ratios (16 and 17) were quite high for more selective tumor damage (Fig. 3).

Figure 3.

Pharmacokinetic profiles of PTX prodrug in plasma, tumors and various tissues following an i.v. administration (2 μmole kg⁻¹) of PTX prodrug in rats with a subcutaneous tumor. Bottom right panel: Ratios of tumor/skin and tumor/muscle prodrug concentration at 9, 24 and 48 h postadministration.

US imaging-guided fiber insertion

Location of the fiber within the tumor was easily monitored by ultrasound imaging (Fig. 4A,B). Using current fiber engineering technology, a thin optical fiber can be readily fabricated to various active lengths and thicknesses adaptable for interstitial illumination into the breast (51). Ultrasound imaging has been used in the clinic to guide therapy or collect samples of breast cancers (52-54). Thus, the clinical adaption of US image-guided light therapy to primary breast cancers is very feasible.

Figure 4.

(A) Tumor with an inserted cylindrical light diffuser at 140 mW cm⁻¹. (B) Ultrasound image of the tumor with an inserted fiber.

Local antitumor effect with a single-tumor model

Six groups (three control groups and three treatment groups) were used to evaluate the antitumor effects of the interstitial illumination with the i.v.-administered PTX prodrug, with varying light intensity and tumor size. The control groups were as follows: G1, no treatment; G2, prodrug (1 μmole kg⁻¹) only; and G3, illumination only (140 mW cm⁻¹). The treatment groups were as follows: G4, 10-mm tumor at 140 mW cm⁻¹; G5, 10-mm tumor at 75 mW cm⁻¹; and G6, 16-mm tumor at 75 mW cm⁻¹. All illumination lasted 30 min. There was no noticeable difference in tumor growth rates between the control group (G1) and prodrug-only or illumination-only groups (G2 and G3) (Fig. S1). For the medium-sized tumors (10 mm), lower light intensity treatment seemed to produce better antitumor effects than did higher intensity treatment (G5> G4 [median survival: 14 vs 8.5 days], Fig. 5A). However, the difference was not significant (P = 0.179). At the same intensity treatment (75 mW cm⁻¹), the large tumors (16 mm) responded much better than the medium-sized tumors (G6 » G5, Fig. 5A). To our surprise, all four large tumors were ablated, and no recurrence was observed for 90 days.

Figure 5.

(A) Tumor growth curves of individual rats: (i) G4 (10-mm tumor, 140 mW cm⁻¹ × 30 min), (ii) G5 (10-mm tumor, 75 mW cm⁻¹ × 30 min) and (iii) G6 (16-mm tumor, 75 mW cm⁻¹ × 30 min) vs G1 (control, no treatment). Dose of PTX prodrug = 1 μmole kg⁻¹ for G4, G5 and G6. G2 (prodrug only, 1 μmole kg⁻¹) and G3 (light only, 140 mW cm⁻¹ × 30 min). The curves for G2 and G3 are shown in Fig. S1. (iv) Kaplan–Meier survival curve represented by considering tumor size reached a maximum of 5000 mm³ according to humane endpoints (G4 vs G5, P > 0.05, ns; G1, G2, G3, G4, G5 vs G6, ***P < 0.001). (B) Representative images of H&E staining and CD45 staining of 10-mm tumor (i and iii) and 16-mm tumor (ii and iv). Necrotic area of the large tumor was washed away during the staining (blue arrow area). The red arrows indicate a tightly packed healthy tumor region. The yellow arrows on CD45 staining (iii and iv) show the presence of tumor-infiltrated lymphocytes (TILs, red; DAPI, blue). (C) Representative photographs of treated tumors on rats in G1, G4, G5 and G6 at various time points after illumination. Photographs of all groups are shown in Fig. S4.

These antitumor results were counter-intuitive. The medium-sized tumors were expected to be damaged more. However, the large tumors were controlled much more effectively at the same light intensity (75 mW cm⁻¹, G6 » G5). To explore this unexpected result, we investigated histological differences of medium- and large-sized tumors at the time of treatment (no prodrug and no illumination). The large tumor showed extensive necrotic areas within the tumor, and a part of the necrotic area was washed away during the slide sample preparation (Fig. 5Bi,iii). The tumor cell density was comparatively lower in the large tumor than in the medium-sized tumor, which had densely packed healthy tumor cells (Fig. S2A). Furthermore, tumor-infiltrated lymphocytes (TILs) were more extensive throughout the large tumor than the medium tumor (Fig. 5Bii,iii). TILs resided mostly along the tumor rim of the medium-sized tumor (Fig. 5Biii).

It was interesting to see that such light treatment efficacy depended on tumor conditions other than simply on tumor size. We speculate two main possibilities for this counter-intuitive result based on optical properties in tissues where scattering and absorption are main factors (55-57). First, photons might be trapped closer to the fiber due to the higher light scattering from the higher cell density in the medium tumor (similar to 1% IL vs 10% IL, Fig. 2D). That circumstance could limit the light propagation to tumor edges from the fiber, thus limiting the direct damage. Actively growing cells outside the medium tumor might overcome the partial damage by the treatment. Second, TILs in the larger tumor might enhance the antitumor tumor immune response in the large tumor. The large tumors disappeared slowly over the course of 2–3 weeks post-treatment (Fig. 5C). Such sustained tumor damage would not be made by two rapid PDT damage mechanisms (direct cell kill and vascular damage), due to the temporal limitations of singlet oxygen. Such a slow effect could be produced by immune responses, and in part by released and trapped PTX.

The other counter-intuitive result was the slightly better antitumor effect with lower light intensity than with higher light intensity (G5 ≥ G4). Some reports showed that low irradiance PDT was more effective than higher irradiance PDT (58-60). Although there was no significant difference in survival (P> 0.05), light intensity and dose-dependent studies are warranted. The survival rate of G6 was far greater (100% tumor-free) than that of all other groups (P < 0.01) (Fig. 5Aiv). No significant bodyweight loss was observed in any treated animal groups (Fig. S3).

The main advantage of our prodrug over a combination of PDT and systemic chemotherapy is elimination of systemic side effect by the systemically administered anticancer for the local antitumor effect. In addition, PDT and site-specifically released PTX worked cooperatively at the multiple levels based on the complementary properties of PDT and PTX, which was discussed in detail in our recent paper (40). PTX combined PDT ablate the tumor effectively with minimized toxic effects. The differential mechanistic actions of PDT and PXT combined together and given better antitumor effect (61,62).

Pharmacokinetic profiles of released PTX in tumor and plasma: model-predicted and experimental data

PTX that is site-specifically released in tumor (~30% yield) by illumination can have a cell killing effect that occurs in addition to the three main anticancer mechanisms of PDT: direct cell kill, vascular damage and antitumor immune responses (Fig. 1C). Our in vitro PKPD model suggested that the retention of the released PTX within the tumor is one of key parameters for the PTX-mediated anticancer effect of the light-activatable PTX prodrug (38). Since the released PTX in tumor could be cleared mainly via blood circulation, the blood flow rate in the tumor could affect the clearance of PTX from tumor (Q_T). Vascular damage by PDT is known to be rapid, even during the illumination in vascular PDT conditions, and can also result in temporal stasis of blood flow in PDT-treated tumors (63-65). Using the power Doppler mode of US imaging, we noted that active blood flow was temporally halted in the tumor treated with PTX prodrug + illumination (Fig. 6B). We hypothesized that the rapid vascular damage by PDT could lower the clearance of PTX from the tumor. This study revealed the release of PTX in tumor after illumination indicating safety of our prodrug system.

Figure 6.

(A and B) Concentration–time profiles of the released PTX in tumor and plasma upon illumination (140 mW cm⁻¹) for 30 min at 9 h after i v. dosing 1 μmole kg⁻¹ PTX prodrug in rats with 10-mm tumor. The open circle symbols represent the observed data. The lines represent the model predictions using our developed PBPK model with rat physiology parameters and PTX clearance allometrically scaled to rats. For comparison purpose, we also included the concentration–time profiles of PTX in plasma and tumor expected following an i.v. bolus dose of 5 mg (5.9 μmole) kg⁻¹ PTX in rats. The open square symbols indicate the observed PTX data from an i.v. bolus dose of PTX, which was obtained from the literature (79). (C) Semi-quantitative US images with power Doppler mode at various time points before and after illumination (140 mW cm⁻¹ × 30 min, 9 h after 1 μmole kg⁻¹ PTX prodrug) with 10-mm tumor.

We used a PBPK model to predict the kinetics of released PTX in tumor and plasma. The measured concentrations (symbols) are presented in Fig. 6A. PTX prodrug treatment yielded higher and sustained PTX in tumor, and minimal PTX in plasma as compared with PTX administered by i.v. bolus (5 mg kg⁻¹ = 5.9 μmole kg⁻¹, a safe dose for antitumor study in rats) (66,67). The plasma AUC of released PTX was 1/8.6 of that of Taxol (691 vs 5977 h nM for 1 μmole kg⁻¹ 2K-PTX vs 5 mg kg⁻¹ [5.9 μmole] Taxol). The released PTX showed a longer retention at a higher concentration in tumor than did the i.v.-administered PTX. The concentration of released PTX remained above 50 nM in tumors more than 48 h after illumination. We postulate that this longer retention was caused by vascular damage from PDT effects, which slowed the tumor blood flow and thus the clearance rate of released PTX from the tumor (Fig. 6C).

Prophylactic systemic antitumor effect determined with tumor rechallenge

All Group 6 rats, whose tumors were cured, were rechallenged with the same cancer cells (MAT B III, 1 million) on their rear flanks. The study model with experimental schedule is shown in Fig. 7A. All rats blocked the formation of tumor at 1, 3, 6 and 12 months (Fig. 7D). Three-month-old and 12-month-old control rats did not block tumor formation. Thus, the blocking of tumor growth was not due to their ages. Two cured rats were not included in the 12-month rechallenge due to their deaths for unknown reasons during the 10^th and 11^th months post-treatment. Our demonstrated prophylactic systemic antitumor effects suggest that prodrug treatment activated a systemic immune response against tumor via immunogenic cell death (ICD) (68). PDT and photothermal therapies have been reported to activate the adaptive immune system through ICD via releasing tumor cell debris (69). T-cell-mediated tumor antigen-specific immunity has been reported to clear rechallenged tumor in PDT-treated mouse models (70-72). ICD by PDT was tested for a cancer vaccine to activate tumor-specific T cells (73-76). However, this is the first demonstration of such ICD and prophylactic systemic antitumor effects of the PTX prodrug. Further studies are underway to investigate the mechanisms of activation of immune systems and differences in ICD by PTX prodrug vs other methods such as PDT, chemotherapy and photothermal effects.

Therapeutic systemic antitumor effect evaluated with two-tumor models

To assess the therapeutic systemic antitumor effect of localized PTX prodrug treatment on the distant untreated tumor, we used two-tumor models (Fig. 7B,C). Initially, we used a small/small tumor model (Fig. 7C) to determine the systemic antitumor effect on a small distant tumor (T2) following PTX prodrug treatment administered to one small tumor (T1). As shown in Fig. 7E, the growth of treated tumors was delayed, but there was no significant antitumor effect on untreated tumors. The growth of untreated flank tumor was normal, and all tumors reached the endpoint in 15 days after cell inoculation, as observed in the normal control group G1 (Fig. 5A). We assumed that ICD with the small tumor could not generate sufficient tumor antigens to produce a measurable systemic therapeutic effect. Thus, we used a large/small tumor model to increase the dose of ICD. A large tumor (16 mm) and a small tumor (4–6 mm) were established on the shoulder and rear right flank (Fig. 7B). Only the large tumor was interstitially illuminated after 9 h of intravenous injection of PTX prodrug (1 μmole kg⁻¹). Growth of both tumors was monitored to determine the therapeutic efficiency and local and systemic antitumor effects. The primary large tumors shrank slowly over the duration of 2 weeks at a rate similar to that of the single-tumor model (Fig. 7F vs G6, Fig. 5A). Surprisingly, the distant untreated tumor was eradicated completely in one week of prodrug treatment (Fig. 7Fii). No recurrence was observed for 90 days in any rat. All rats in this group also blocked the tumor formation of rechallenged cancer cells (Fig. S5), demonstrating activation of the adaptive immune system via ICD. The results of the two-tumor models showed that local PTX prodrug treatment can activate the adaptive immune system to sufficiently produce therapeutic and prophylactic antitumor effects, but its strength may depend on the treatment dose (represented by tumor size). It is also possible that high TILs in the large and necrotic tumors (Fig. 5B) might play a role in ICD-initiated systemic antitumor effects (77,78). To successfully translate our results to the clinic, in-depth mechanistic and systemic understandings of the decisive factors for both local and systemic antitumor effects are needed. In the future, we will investigate the effects of controllable variables, for example, dose of light and prodrug, size of tumors, and combination with other therapeutics.

CONCLUSION

US image-guided interstitial illumination was an effective treatment for the large breast tumors (16 mm) in rats. A common US machine, which is available in most clinics, was sufficient to locate the fibers inside the tumor. About 30% of the prodrug released PTX in tumor, and clearance of PTX from tumor was slowed and then halted, presumably due to the rapid vascular damage from PDT effects. All 16-mm tumors were cured, without recurrence for 90 days, by PTX prodrug treatment. The prodrug treatment to the large tumor produced both prophylactic and therapeutic antitumor effects on all treated rats through ICD of the local prodrug treatment. Our proof-of-concept study clearly demonstrates the feasibility of US-guided interstitial treatment of a large tumor, such as primary breast cancers, and that local treatment can induce systemic antitumor effects to kill metastatic cancers. For clinical translation of our exciting preliminary results, further studies are warranted to understand key determinant factors to yield maximal systemic antitumor effects from this local prodrug therapy.

Supplementary Material

SI

Figure S1. Tumor growth curves in G2 and G3.

Figure S2. Large scale images of Fig. 5B.

Figure S3. Body weight change of G1–G6.

Figure S4. Photos of the treated tumor areas.

Figure S5. Two tumor growth curves of Large/Small tumor model.

Abbreviations:

CD

cylindrical light diffuser

DLI

drug-light interval

FA-PEG2K-L-Pc-PTX (PTX prodrug)

folate receptor-targeting phthalocyanine-linked paclitaxel

FD

frontal light distributor

ICD

immunogenic cell death

MAT B III

mammary gland adenocarcinoma cell line

PD

prodrug

PDT

photodynamic therapy

SO

Singlet oxygen

US

ultrasound