Sustained Supratherapeutic Paclitaxel Delivery Enhances Irreversible Sarcoma Cell Death

Abstract

Risk of locoregional recurrence after sarcoma resection is high, increasing both morbidity and mortality. Intraoperative implantation of paclitaxel (PTX)-eluting polymer films locally delivers sustained, supratherapeutic PTX concentrations to the tumor bed that are not clinically feasible with systemic therapy, thereby reducing recurrence and improving survival in a murine model of recurrent sarcoma. However, the biology underlying increased efficacy of PTX-eluting films are unknown and provide the impetus for this work. In vitro PTX efficacy is time- and dose-dependent with prolonged exposure significantly decreasing PTX IC50 values for human chondrosarcoma (CS-1) cells (153.9 nM at 4 h vs. 14.2 nM at 30 h, p = 0.0001). High dose PTX significantly inhibits proliferation with in vivo PTX-films delivering a dose >130 μM directly to the tumor thereby irreversibly arresting cell cycle and inducing apoptosis in CS-1 as well as patient-derived liposarcoma (LP6) and leiomyosarcoma (LMS20). Supratherapeutic PTX upregulates the expression of p21 in G2/M arrested cells, and irreversibly induces apoptosis followed by cell death, within 4 h of exposure. Microarray analyses corroborate the finding of poor DNA integrity commonly observed as a final step of apoptosis in CS-1 cells and tumor. Unlike low PTX concentrations at the tumor bed during systemic delivery, supratherapeutic concentrations achieved with PTX-eluting films markedly decrease sarcoma lethality in vivo and offer an alternative paradigm to prevent recurrence.

1. Introduction

Sarcomas are a heterogeneous group of malignant mesenchymal tumors for which radical resection is the standard-of-care for localized disease. However, complete resection is often not achieved due to large tumor size, invasion into surrounding structures, or unacceptable postoperative morbidity.(1) Overall, five-year locoregional recurrence (LRR) rates following resection of soft tissue sarcomas range from 10% to as high as 84% depending on tumor location and histologic subtype, and is particularly high with retroperitoneal sarcomas.(2–4) The corresponding five-year overall survival (OS) is approximately 50%.(5)

Given the limitations of radical resection, there is a critical need for novel treatment approaches to prevent LRR of a variety of surgically treated malignancies, including sarcoma.(6,7) We previously demonstrated that local delivery of paclitaxel (PTX) using biocompatible drug-eluting polymer films implanted at time of resection drastically reduces LRR in a murine xenograft model of resected human chondrosarcoma (17% LRR with PTX-films vs. 89% LRR with a systemically-delivered equivalent dose of PTX; p = 0.0002).(8) PTX is released from these films for over 30 days with a 100-fold greater concentration within the local tissues compared to systemic administration of the same 300 μg PTX dose (~1000 nM vs. 10 nM at day 1, respectively). Maximum tolerated dosing studies demonstrate that systemic administration of PTX to achieve the local tissue levels attained with PTX-films is not clinically feasible.(9,10) The ability of drug-eluting polymers to deliver supratherapeutic PTX concentrations locally to the tumor bed represents a new approach to local tumor control while minimizing systemic toxicity.

PTX is a microtubule-stabilizing chemotherapeutic agent that accumulates in cancer cells, resulting in impaired microtubule dynamics, mitotic block, and suspension in a tetraploid state.(11,12) Through microtubule stabilization, PTX halts all mitosis, significantly stunting tumor growth and proliferation. It is known that systemic concentrations of PTX are not sufficient to halt cellular proliferation indefinitely; however, the results of our initial study led us to investigate the extent and means of cell death that occurs within various sarcoma subtypes following exposure to prolonged, supratherapeutic PTX concentrations. The extended survival of animals treated with high-dose PTX films led to our hypothesis that supratherapeutic PTX affords an outcome distinct and more clinically favorable than standard PTX dosing via systemic delivery. Herein, we report time- and dose-dependent efficacy of PTX against various sarcoma lines (CS-1, LP6, LMS20) including patient-derived liposarcoma and leiomyosarcoma subtypes that recapitulate the complex biology and heterogeneity of human sarcomas. We report the cytotoxicity via prolonged cell cycle arrest, mitotic catastrophe, subsequent apoptosis, and ultimately irreversible cell death associated with supratherapeutic PTX levels.

2. Material and Methods

2.1. Cell line and culture conditions

Chondrosarcoma cells (CS-1) derived from a high-grade chondrosarcoma(13) were cultured in RPMI media supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, and 100 μg/mL streptomycin (Invitrogen, Carlsbad, CA). Liposarcoma cells (LP6) derived from a dedifferentiated liposarcoma, and leiomyosarcoma cells (LMS20) derived from a high-grade non-uterine leiomyosarcoma were cultured in DMEM/F12 media supplemented with 10% FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin (Invitrogen, Carlsbad, CA). These various subtypes demonstrate distinct differences in histology, cell growth and proliferation in culture reflecting differences in tumor biology seen in vivo. All cells were cultured at 37 °C and 5% CO₂. Cells were maintained at less than 30 cell passages and checked for consistent morphology by short tandem repeat profiling (ATCC, Manassas, VA). All cell lines were tested for mycoplasma using the MycoAlert® Mycoplasma Detection Kit (Lonza, Basel, Switzerland). Under the Health and Human Services Human Subjects Regulation (45 CFR Part 46) the studies herein completed in this work are not considered Human Subjects Research by virtue of the fact that the human cell lines are from a de-identified source and were received and previously published as such from the source research institute (Dana Farber Cancer Institute). Risks to violation of subject privacy are minimal since no personal health information is recorded. The authors do not have any access to subject identifiers; therefore, the work involving these cell lines is not considered human subject research per the NIH OER decision tool.

2.2. Cytotoxicity Assay

A standard MTS assay was utilized for in vitro measurement of cell viability in response to PTX. CS-1 cells were plated at 3,000 cells per well, and upon reaching 70–80% confluence, exposed to various concentrations of PTX solubilized in 1:1 Cremophor EL/ethanol (PTX C/E) for either a 4 or 30 h exposure. Colorimetric readings were measured on a Spectramax M5 plate reader (Molecular Devices, Sunnyvale, CA), and viability calculated as a percentage of absorbance for non-treated cells at each time point.

2.3. Confocal Microscopy

200,000 CS-1 cells were seeded onto a 6-well plate with 1.5 thickness glass coverslips inside. Once adherent 24 h later, cells were treated with 10 nM PTX C/E for 30 h. After treatment, cells were washed in cold PBS to remove trace PTX and media components, and fixed in 2% paraformaldehyde in PBS for 20 minutes. Coverslips were mounted onto glass microscope slides with Prolong Gold Antifade (ThermoFisher Scientific, Waltham, MA) containing DAPI (4′,6-diamidino-2-phenylindole) to stain nuclei.

2.4. Mass Spectrometry: Intratumoral drug levels as a function of distance from PTX films

Tissues were prepared for mass spectrometry (MS) analysis by cryosectioning resected tumors at 5 μm thickness using a HM550 Cryostat (ThermoFisher Scientific, Waltham, MA). MS analysis was performed using liquid extraction surface analysis (LESA) (TriVersa NanoMate, Advion, Ithaca NY) coupled to an ion trap mass spectrometer (amaZon speed, Bruker, Billerica MA). Analyzed samples were also stained with hematoxylin and eosin and imaged using a Z1 Observer (Carl Zeiss Microscopy LLC, Peabody MA).

2.5. Histology and Immunohistochemistry

CS-1 tumors were harvested and fixed in 2% paraformaldehyde in PBS. Tissue mounting and staining was performed at the Beth Israel Deaconess Medical Center (BIDMC) histology core (Boston, MA). Briefly, tissues were embedded in paraffin blocks to show both the tumor-film interface and the opposite side of the tumor. Tissues were sectioned at a thickness of 5 μm with a microtome before deparaffinization and staining.

Slides were stained with anti-Ki67 antibody (Cell Signaling Technology, Danvers, MA) at a dilution of 1:50 overnight at 4 °C. The slides were washed and treated with goat, anti-rabbit HRP-polymer secondary antibody (Abcam, Cambridge, MA) for 2 hours. Slides were washed and developed using a peroxidase substrate DAB kit (Vector Laboratories, Burlingame, CA) and were counterstained with hematoxylin before mounting.

2.6. Flowcytometric Analysis of Cell Cycle

For cell cycle analysis, PTX treated CS-1, LP6, and LMS20 monolayers were fixed with ice-cold 70% EtOH and then incubated 30 minutes with FxCycle PI/RNase staining solution (Life Technologies, Woburn, MA) prior to flow cytometric analysis. Annexin V staining was performed using the Dead Cell Apoptosis Kit using FITC-labelled annexin V and propidium iodide (ThermoFisher, Waltham, MA). For p21 analysis, cells were fixed in 4% paraformaldehyde at 37° for 10 minutes and permeabilized by slowly adding 90% methanol. Following a 1% BSA wash, cells were incubated with rabbit anti-p21 monoclonal antibody (Cell Signaling Technology, Danvers, MA) for 1 h following the manufacturer’s protocol and counterstained with DAPI prior to analyzing on a BD LSRFortessa flow cytometer (BD Biosciences, San Jose, CA) in FACS buffer. At least 10,000 cells per sample per replicate were counted.

2.7. Organotypic Ex Vivo Tumor Culture and Analysis

Untreated CS-1 tumor slices were cut at 250 μm thickness using a VT1000S vibratome as previously described(14) (Leica Biosystems, Buffalo Grove, IL) with an amplitude of 0.28 mm at a speed of 0.22 mm/s. Slices were placed on organotypic culture inserts (EMD Millipore, Darmstadt, Germany).

2.8. Preparation of PTX-Films

PTX-loaded poly(glycerol monostearate-co-ε-caprolactone) (PGC-C18) films were synthesized and loaded with PTX following our previously published protocol.(8) Briefly, PGC-C18 films were synthesized with and without 300 μg PTX (3 mg PGC-C18 & 300 μg PTX at 10% wt/vol (polymer/solvent) in dichloromethane) followed by solvent casting onto a biodegradable collagen-based scaffold (1.0 × 1.0 cm; Peri-Strips Dry with Veritas Collagen Matrix, Synovis, St. Paul, MN) with a depth of 40 μm and total width of 500 μm.

2.9. In Vivo Xenograft Models for Assessment of Cytotoxicity and Measurement of Pre-treatment Tumor Lethality

All in vivo work presented in this study were approved by the Massachusetts General Hospital Institutional Animal Care and Use Committee. To assess differences in intratumoral PTX delivery in vivo as a function of delivery method, implanted tumors were harvested from Nu/J mice on day 10 following in vivo treatment with 300 μg PTX given systemically or via PTX-films. PTX drug levels within plasma and tumor were assessed via HPLC using normal tissues in non-treated animals as negative drug controls (BASi).

To assess differences in CS-1 lethality after in vitro PTX exposure, CS-1 cells were first treated with 10 nM or 1000 nM PTX for a duration of 4 or 30 hrs. Two million PTX-exposed CS-1 cells were subsequently injected subcutaneously in the dorsum of Nu/J female mice (Jackson Laboratory, Bar Harbor, ME). Mice were euthanized when tumors reached 2 cm, animals appeared moribund, or if alive at the termination of the study at 60 days.

2.10. Statistical Analysis

Continuous variables were analyzed using 2-tailed t-test or 1-way ANOVA, where indicated. Bonferroni post-hoc analysis was used for ANOVA testing. For survival analysis, the Kaplan-Meier method was used. GraphPad Prism 5.0 (La Jolla, CA) was used for the analyses. Post-hoc analysis of survival was carried out via Cox Proportional Hazards Model in SAS (Cary, NC). A two-sided p-value of < 0.05 was considered statistically significant.

2.11. Microarray Processing and Statistical Analysis

The Affymetrix Human Gene 1.0 ST Array, which encompasses approximately 67,000 transcripts, was used. For in vitro microarray analysis, 150,000 CS-1 cells were plated into a 6-well plate and treated with either 10 nM or 1000 nM PTX for 16–18 hours before extracting RNA. For in vivo samples, tumors treated with IP PTX and PTX Films were harvested to extract RNA. RNA integrity was assayed via BioAnalyzer before proceeding with microarray. Preparation of RNA, hybridization, and scanning of microarrays were performed according to the manufacturer’s protocol (Affymetrix).

Background subtraction and transcript normalization were completed with the robust multi-array average algorithm via the GeneSpring GX software. Fold change values for genes were calculated as the ratio of the signal values of the PTX Film or 1000 nM PTX groups compared to the IP PTX or 10 nM PTX groups, respectively. Gene expression ratios of ≥2 or ≤−2 were considered differentially expressed using the Benjamini-Hochberg procedure to determine statistical significance.

2.12. Data Availability

Raw data for the microarray work were generated at the Dana Farber Cancer Institute Microarray Core. Derived data supporting the findings of this study are available from the corresponding author upon request.

3. Results

3.1. Efficacy of PTX is Dose and Duration Dependent

To evaluate the cytotoxic potency of PTX against CS-1 tumor cells, PTX was titrated at various concentrations for either a 4 h or 30 h duration of exposure. Given that the clearance of PTX following systemic administration occurs rapidly, within 4 h post-injection(15), systemic delivery was approximated in vitro by a short exposure time of 4 h. This was compared to a 30 h prolonged PTX exposure to represent the continuous prolonged release kinetics (> 1 month) evident with PTX-films.(8) PTX efficacy against CS-1 chondrosarcoma cells is time-dependent as evidenced by a 10-fold decrease in IC50 with prolonged PTX exposure (Figure 1A; IC50 = 153.9 nM at 4 h vs. 14.2 nM at 30 h, p = 0.0001).

Figure 1. Paclitaxel Efficacy is Time- and Dose-Dependent.

A) The PTX IC50 for the CS-1 cell line is time-dependent and decreases by an order of magnitude from 153.9 nM to 14.7 nM with 4 hour vs 30 hour treatment. N=5 B) Exposure to 10 nM PTX induces condensation and irregular packing of DNA in CS-1 tumor cells revealed by DAPI staining and confocal laser microscopy. C) PTX inhibits cellular proliferation ratio only at high concentration (red) compared to low concentration or no treatment (blue and black, respectively). D) 1000 nM PTX similarly induces irregular DNA condensation in CS-1 tumor cells.

To assess the importance of PTX dosing in relation to improved efficacy of PTX-films, CS-1 cells were treated with 10 nM and 1000 nM PTX to represent PTX concentrations from systemic and local delivery present in the tumor resection bed, respectively.(8) Low dose PTX (10 nM), the tissue concentration measured following systemic administration, was sufficient to condense CS-1 tumor nuclei (Figure 1B), but not inhibit cellular proliferation. CS-1 cell proliferation was expressed as a ratio of total number of viable CS-1 cells measured initially following a 4 h in vitro PTX exposure compared to the number of viable CS-1 cells present at 72 hours. After 72 h, a proliferation ratio of 5.18 and 5.06 was observed for untreated CS-1, and 10 nM PTX-treated CS-1, respectively (Figure 1C; p = 0.95). Only with treatment of high dose 1000 nM PTX, was inhibition of cell proliferation persistent with a decreased ratio of 0.99 (p = 0.034) in addition to DNA condensation in the nuclei (Figure 1D).

3.2. PTX-Films Deliver Supratherapeutic Drug Levels to Tumor to Inhibit Recurrence

Our previous studies demonstrated that the local tissue concentration of PTX in a non-tumor bearing animal on day 1 following PTX-film administration was 797 ± 412 ng/g and remained high through 30 days.(8) Given that solid tumors have been shown to further concentrate PTX over time(16), PTX levels were hypothesized to be greater within the tumor itself as a result of PTX-film exposure, further improving anti-tumor efficacy. Therefore, to determine tumor PTX concentration, tumor-bearing mice underwent partial resection leaving residual tumor (Figure 2A) and were treated with either systemic 300 μg PTX via intraperitoneal (IP) injection or 300 μg PTX via PTX-film implantation before the remaining tumors were harvested 10 days later for determination of intratumoral drug concentration. HPLC-MS showed significantly higher average PTX concentrations within the tumor treated with PTX-films (3190 ± 638 ng/g) as compared to negligible PTX remaining following systemic IP PTX therapy (2.1 ± 0.72 ng/g, p = 0.037, Figure 2B), despite an equivalent total dose administered. Of note, despite the high intratumoral concentration of PTX in the PTX-film group, there was minimal (<1.0 ng/g) PTX detected within the plasma at day 10 in either IP PTX or PTX-film treated groups with no statistical difference between the groups. To further characterize the PTX delivery gradient within the tumor, LESA-mass spectrometry (MS) was performed on a sliced section of tumor beneath the PTX-film. LESA-MS analysis revealed the intratumoral PTX concentration at 1 mm below the film was extremely high at 134.8 μM and decreased to less than 0.250 μM approximately 5 mm into the tumor (Figure 2C).

Figure 2. Recurrence depends on paclitaxel concentration to the tumor bed post-resection.

A) CS-1 tumors are implanted, grown, and resected prior to treatment with either intraperitoneal PTX or PTX-loaded films. Locoregional recurrence occurs despite IP PTX but not with PTX-films. B) PTX-loaded films afford an average intratumoral PTX concentration 3 orders of magnitude greater than systemic administration of the mean tolerated dose 24 hours after treatment. There is no difference in PTX plasma levels between IP administration or film consistent with the minimal side effects evident with PTX- film despite the high concentration inside the tumor. C) PTX films are capable of delivering high dose chemotherapeutic directly to the film-tumor interface. PTX concentration decreases as a function of distance from the film within the tumor. * = below limit of detection (250 nM)

H&E staining of CS-1 tumor treated with either film or IP PTX was performed to assess cell death in residual tumor post-resection (Figure 3). Control tumor samples that were untreated after resection showed minimal regions of cell death. Tumor samples treated with IP PTX similarly showed limited cell death within the tumor tissue (Figure 3A and B). However, tumors that were treated with PTX-films showed significant cell death in a spatially-dependent manner with the greatest death occurring at the tumor-film interface (Figure 3C, dotted line) as indicated by cells with missing or distorted nuclei. Ki67 staining revealed that residual tumor treated with PTX-film was not actively dividing, unlike untreated tumor (Supplemental Figure S1).

Figure 3. Histological Analysis of CS-1 PDXs Reveal Significant Cell Death Near PTX Films.

CS-1 PDX tumors were resected and either treated with 300 μg PTX, PTX films, or untreated. Tumors or residual tissue were harvested 10 days after resection and treatment. A and B) The untreated control and IP PTX-treated PDXs show bulk tumor with no indication of cell death. C) Significant cell death within the tissue is observed as a function of distance from the film (dotted line, top). The mouse skin is indicated by the red arrow (top) and cell death with the arrows. In the middle 40X panel, arrows are indicating misshapen nuclei, a hallmark of cell death within the tumor.

3.3. High-Dose PTX Prevents Tumor Growth via a p21-Associated, Irreversible G2/M Cell Cycle Arrest

To investigate the extent of cell death from PTX-films, the effect of PTX dose and treatment length on CS-1, LP6, and LMS20 cells in vitro was investigated via cell cycle analysis. Flow cytometric analysis of the cell cycle showed that a short (4 h) PTX exposure at 1000 nM was sufficient to induce mitotic arrest, as shown by the shift from G1 to G2/M phases (CS-1: 66.2% G2/M, LP6: 58.7% G2/M, LMS20: 53.6% G2/M; Figure 4A) 30 h after treatment. In all three cell lines, the ratio of G1:G2/M phases shifts by 30 h to less than 1 in the 1000 nM PTX treatment groups (CS-1: 0.201, LP6: 0.394, LMS20: 0.597, respectively) and is significantly smaller than the 10 nM PTX or control groups where the G1:G2/M ratio is greater than 1. Seventy-two hours after high dose treatment, nearly all LP6 and LMS20 cells were suspended in G2/M (mitotic arrest), and the G1:G2/M ratio for both cell lines substantially decreases for the 1000 nM PTX treatment group, but remains greater than 1 in the 10 nM and control groups. Interestingly for CS-1, which exhibits a faster proliferation time, the G2/M-arrested cells are largely gone by 72 h, being replaced by a significant proportion of tumor cells demonstrating DNA fragmentation and death in the form of < 2N ploidy (39% Sub-G1, Figure 4A). Only a small population of tumor cells remain in G1 at 72 h (16% after 1000 nM PTX vs 67% after 10 nM PTX). Compared to 4 h PTX exposure, this G2/M shift was further exacerbated by 30 h of continuous treatment with 1000 nM PTX as prolonged, high dose PTX prevented this escape and suspended 61% of CS-1 tumor cells in G2/M arrest with virtually no G1 population left (Figure 4B). We investigated p21 expression, a regulator of cellular responses to microtubule damage, in these G2/M arrested tumor cells noted after 30 h high dose PTX exposure and found significant upregulation in p21 expression by an order of magnitude compared to cells in G2/M arrest following low dose or no PTX treatment (Figure 4B). In contrast, the 10 nM PTX-treated group did not exhibit mitotic arrest after either 4h or 30 h of continuous PTX treatment and did not significantly alter the cell cycle or p21 expression compared to the untreated control.

Figure 4. Sustained Supratherapeutic PTX induces irreversible and p21-associated cell death.

Cells were treated with 10 or 1000 nM PTX for 4 or 30 hours, then analyzed immediately, 30 h, or 72 h after treatment. A) Immediately after a 4 h treatment, there is no difference in cell cycle between treatment and untreated controls. After 30 hours, majority of sarcoma tumor cells treated with high-dose paclitaxel are arrested in the G2/M phases. After 72 hours, CS-1 cells are dead as evident by the increase in <2N ploidy population. LP6 and LMS20 cells remain G2/M arrested at this time. The low-dose PTX groups do not differ from the untreated control. N=3 B) p21 expression was measured for cells arrested in G2/M (bracket) and is significantly upregulated for CS-1 treated for 30 h with high dose PTX and assessed 30 h after the start of treatment. C) (Scheme) Resected tumor is cut into cross sections with a thickness of 250 μm for ex vivo analysis inclusive to tumor stroma. p21 expression is also increased following high dose PTX in organotypic tumor cultures in the nuclei (black arrows) and cytosol (red arrow heads), compared to low dose PTX or untreated CS-1 cross sections.

These findings were also corroborated in an ex vivo organotypic experiment to assess the p21 expression histologically in the tumor microenvironment. Briefly, CS-1 tumor cross sections were treated with either 1000, 10, or 0 nM PTX or via PTX-film in a transwell for 4 h and were stained for p21 expression at 30 h (Figure 4C, scheme). Cross sections treated with 10 nM PTX exhibited a similar low-expression pattern of nuclear p21 compared to the untreated control. Conversely, sections exposed to 1000 nM PTX or a PTX-film implant showed significant upregulation of p21 in both the nuclei and cytoplasm (Figure 4C).

3.4. Supratherapeutic PTX Induces Apoptosis in an Irreversible and Dose-Dependent Manner

Given the possibility of cell death escape, we assessed the irreversibility of high-dose PTX-mediated cell death in an in vivo implantation experiment. CS-1 tumor cells were exposed in vitro to 10 nM or 1000 nM PTX for 4 h or 30 h, and screened for viability prior to subcutaneous injection into Nu/J mice in a murine xenograft model (Figure 5, scheme). CS-1 cells that were treated with 10 nM PTX for either 4 or 30 h did engraft but trended towards a slower albeit not significantly different growth rate than the untreated control, leading to smaller tumor volumes 28 days post-injection (Figure 5A). However, CS-1 cells that were treated with 1000 nM PTX showed significantly less tumor growth with nearly half of the mice in both the 4 and 30 h treatment groups failing to grow tumors at all (Figure 5A; p = 0.0041 and 0.0055 for 4 and 30 h treatment, respectively vs. controls).

Figure 5. High-Dose Paclitaxel Induces Apoptosis and DNA Damage.

A and B) CS-1 cells that were treated in the same manner were injected into Nu/J mice to establish CS-1 xenografts and assess lethality of the treated CS-1 tumor cells (Scheme). A) CS-1 pre-treated with PTX show a dose-dependence, but not time-dependence on tumor volume. CS-1 pre-treated with high-dose PTX form significantly smaller tumors if any at all (4 h: p = 0.0041, 30 h: p = 0.0055 compared to control). B) CS-1 tumor cells pre-treated with low-dose PTX remain viable in vivo resulting in similar survival of the xenograft host as with untreated CS-1 tumor implants. In contrast, CS-1 treated with high-dose PTX before implantation exhibit decreased lethality and host survival is superior at the 60-day designated time point. log rank p = 0.0003. C and D: N = 6–11. C) Propidium iodide and annexin V staining shows high dose PTX treatment induces increased apoptosis and cell death in LP6, LMS20, and CS-1 after 30 hours. Low dose PTX treatment does not show statistical difference between untreated cells for all three lines. N=3.

The associated survival of each animal following implantation of PTX-treated CS-1 also varied with PTX dose and treatment time (Figure 5B; n = 7 or 8, p = 0.0003). Although exposure to 10 nM did delay tumor development compared to untreated controls, 100% of mice that received CS-1 pre-treated with only 10 nM developed tumors and succumbed by 60 days suggesting that tumors can escape cell death despite PTX exposure, even if exposure is prolonged (Figure 5B, blue traces). In contrast, significantly fewer tumors developed in mice inoculated with CS-1 pre-treated with 1000 nM PTX for 4 or 30 h (58% and 62% respectively, Figure 5A and B, red traces) and thus survival was prolonged in these high-dose PTX cohorts.

To determine whether the lower in vivo viability of tumor cells following high-dose PTX exposure was occurring via apoptosis as the result of the prolonged cell cycle arrest noted after 4 h PTX treatment, cell-surface annexin V and propidium iodide stains were used in flow cytometric analyses in CS-1, LP6, and LMS20 cell lines. Even within 30 h of high-dose PTX exposure, a significant proportion of cells were noted to be apoptotic or already dead in all three sarcoma cell lines. These rates of apoptosis and death were significantly higher than the low-dose PTX-treated or untreated groups respectively (CS-1: 39% vs 26% or 15%; LP6: 28.0% vs 12.3% or 7.93%; and LMS20: 31.2% vs 20.9% or 14.9%; Figure 5C). Minimal apoptosis was observed in the untreated or 10 nM PTX groups with no significant difference between these groups.

These findings were further explored via microarray analysis of the 1000 nM and 10 nM PTX treated CS-1 cells in vitro. Transcripts were screened for at least two-fold change in expression, with p < 0.05 between the high dose and the low dose treatment groups. No genes were significantly upregulated within the high dose treatment (Figure 6A); however, transcripts of histone isoforms were significantly downregulated (p < 0.05; >2-fold expression change) in the high-dose PTX treatment group. Specifically, high dose PTX inhibited transcription of all histones, H1, H2A, H2B, H3, and H4 (Figure 6B), suggesting that DNA metabolism or processing is significantly hindered via exposure to supratherapeutic PTX. This is further supported by the significant downregulation of MIR100 and CCNE2, both of which are implicated in cell cycle maintenance. Additionally, a stark downregulation of pro-hypertrophic genes CTGF(17) and ANKRD1(18), in addition to the FAM111B gene that encodes a protein of unknown function (Supplemental Figure S2) was observed.

Figure 6. Microarray Analysis Reveal Significant Differential Expression Between High and Low Dose PTX Treated CS-1.

A and B) CS-1 monolayers treated with low 10 nM and high 1000 nM PTX were subjected to microarray analysis. A) Volcano plot of the in vitro transcriptional analysis that reveals a select few genes downregulated with high-dose PTX. B) A stark downregulation of histone mRNAs validated the finding high dose PTX induced DNA damage and apoptosis. C and D) Volcano plot analysis shows a substantial number of transcripts that are differentially expressed (>2-fold) between in vivo sample. C) Red points represent individual transcripts that are significantly downregulated while blue points represent transcripts that are significantly upregulated. Black points are either not differentially expressed, or differentially expressed in a statistically insignificant manner. D) A significant downregulation of histone mRNA expression is observed with PTX-Film treated CS-1 tumors compared to that of equivalently-dosed systemic PTX. B and D) p-values are listed above each column pair.

The microarray analysis was extended to in vivo PTX-film-treated CS-1 tumors, revealing that histones H2A, H2B, and H4 were significantly down-regulated in tumors in vivo (p=0.011, 0.019 & 0.0076, respectively), just as they were in the in vitro high dose PTX analysis (Figure 6C and D). Additionally, a downregulation in minichromosome maintenance proteins involved in DNA processing and replication was observed, supporting the observed cell cycle arrest (Supplemental Figure S3).

4. Discussion

In this work, we show using patient-derived soft-tissue sarcomas, which reflect relevant tumor biology and are clinically relevant, that high and sustained doses of paclitaxel, delivered by PTX-eluting films, decreases tumor cell proliferation, and leads to cell cycle arrest and cell death. With systemic PTX delivery, it is impossible to achieve such anti-tumor effects due to rapid dilution and clearance of PTX and lethal systemic toxicity. These differential findings are first demonstrated in vitro. Treatment duration plays an important role in the cytotoxicity of PTX against CS-1 tumor cells as shown by the order-of-magnitude decrease in IC50 following 30 h versus 4 h duration of PTX exposure (Figure 1A). The hindrance of cellular proliferation with multi-day high-dose PTX treatment further supports this dose-dependency (Figure 1C). These results indicate that an ideal clinical regimen would consist of a concentrated and prolonged exposure of the tumor to supratherapeutic PTX levels.

To administer supratherapeutic concentrations of PTX to the tumor bed, we prepared PTX-loaded films for implantation in a murine recurrence model of human chondrosarcoma (Figure 2A). As shown by LESA-MS analysis, PTX penetrates and is distributed through the entirety of the residual tumor with a concentration of approximately 250 nM on the side opposite the film. Perhaps more important is the approximately 500-fold higher PTX concentration (>130 μM) located immediately under the PTX-film, i.e., the location where residual tumor cells are located and where the risk of recurrence after surgery is highest, 10 days after placement of the PTX-film (Figure 2C). For comparison, an 18 mg/kg intraperitoneal bolus of PTX in mice results in peak plasma concentration of 13 ± 3.1 μg/mL at 2 h, with a half live of 0.7 h where as the films result in an initial plasma concentration of nearly 30 ng/ml on the first day but are less than 1 ng/ml by day 10–15.^8,(19) Other strategies, such as entrapping PTX in silica nanoparticles(20) or liposomes(21), aim to prolong half-life and overall exposure (AUC_0−∞) of PTX, but fail to release PTX directly to the tumor at concentrations as high as the films described in this work without systemic toxicities. In the clinic, 275 mg/m² Taxol administered as a 6 h intravenous infusion reaches a peak plasma concentration of 8 μM immediately post-infusion in humans—far below the intratumoral PTX concentration (250 nM to 130 μM) achieved via PTX-films at 10 days.(22) In addition, to achieving PTX concentrations considerably higher than the 50 nM reported to be critical for PTX efficacy(23), the prolonged duration of exposure to such high PTX concentrations is also with PTX-loaded films, further increasing efficacy in vivo. Furthermore, the films allow for the localized administration of high dose PTX exclusively to the tumor bed with minimal detectable plasma concentrations after 10 days (Figure 2B) and even 30 days (338.4 ng/g)(8) compared to the brief half-life of 0.83 h observed after intravenous infusion.(22) In retrospect, an examination of tumor response to 10 nM vs 1000 nM PTX underestimates the local PTX exposure achieved within the tumor via films, given that the current study demonstrates an even greater average intratumoral concentration using this local delivery platform.

Next, we evaluated the efficacy and impact on cell cycle kinetics of supratherapeutic PTX via flow cytometric cell cycle analyses. PTX is a known potent anti-mitotic agent via stabilization of cellular tubulin(24,25), however a short 4 h treatment of PTX at the 10 nM dose achieved systemically does not elicit a cytotoxic effect to the CS-1, LP6, and LMS20 cultures even up to 72 hours after treatment (Figure 4A). Conversely, a 4 h treatment of high-dose PTX arrests the majority of CS-1, LP6, and LMS20 cells in G2/M within 30 hours of treatment, as evidenced by a decreased G1:G2/M ratio significantly lower than 1. Interestingly, LP6 and LMS20 cells remain G2/M-arrested even after 72 h of washout, while CS-1 cells show extensive cell death after 72 hours. This difference in cell cycle may be explained by differences in tumor biology among the different sarcoma subtypes, slow growth of the LP6 and LMS20 cell lines in 2D culture, or greater chemoresistance. Differences in the kinetics of cell death in response to high-dose PTX will be the focus of future work as this may yield insight into unique aspects of molecular susceptibility or mechanisms of cell death in the setting of supratherapeutic drug delivery. P21 staining of PTX-treated CS-1 cells and ex vivo CS-1 tumor cross sectional slices shows a dose-dependent increase in p21 expression, consistent with previous literature.(26–28) Canonically, p21 serves as a regulator to microtubule damage by inducing G1 arrest, protecting cells from PTX toxicity during mitosis(26). Interestingly, with high-dose 1000 nM paclitaxel, cells arrest in G2/M despite exhibiting higher p21 expression in both the nucleus and cytoplasm (Figure 4). Cytoplasmic p21 promotes cell survival in an Akt-dependent manner and is correlated with worse prognosis in some cancers.(26,29–32) One explanation for the observed increase in p21 expression and cytoplasmic translocation noted with high dose PTX exposure in CS-1 is perhaps the attempt to ameliorate PTX toxicity as cytoplasmic p21 has been previously shown to mediate cisplatin and 5-fluorouracil sensitivity in testicular, ovarian, and colorectal cancer.(33–35) Similarly, taxane (docetaxel or PTX) sensitivity is reported to be proportional to the expression of microRNA miR-100 (MIR100).(36,37) Interestingly, our PTX-film treated CS-1 tumor cells exhibit suppressed MIR100 expression and thus should be resistant to PTX (Figure S2), however, the high dose delivered from the PTX-film results in mitotic catastrophe and appears to allow chemotherapeutic resistance mechanisms to be overcome. This suggests a more complex regulation of cell cycle and fate that is irreversibly decided within 4 h of high dose PTX treatment.

To further investigate this phenomenon of escaping tumor cell cycle arrest to recover viability after PTX exposure, we treated CS-1 cells with PTX prior to tumor implantation. CS-1 cells pre-treated with low dose PTX (10 nM) resulted in smaller tumors with a significantly slower growth rate and lower success rate than the untreated control in a dose-, but not time-dependent manner (Figure 5A). However, CS-1 treated with 1000 nM showed significantly inhibited tumor growth, with absence of tumor formation altogether in nearly half of the mice. In addition, the inability of CS-1 cells treated with high-dose PTX to escape cell arrest and recover led to significantly longer survival (>60% for 2+ months) in the host animals compared to both low-dose PTX-treated and untreated groups, which had similar median survival of ~30 days (Figure 5B).

Thus, we investigated whether G2/M cell cycle arrest leads to DNA fragmentation and/or cell death via apoptosis.(38,39) Annexin V and propidium iodide staining in flow cytometry reveals that even a 4 h exposure to high-dose PTX induced an apoptotic-like phenotype in a dose-dependent manner evident within 30 h after treatment (Figure 5C). PTX treatment, especially in the 1000 nM dose cohort, led to significant decreases in viability and increases in apoptosis within 30 hours of PTX exposure in CS-1, LP6, and LMS20 sarcoma cells (Figure 5C). In the high-dose group, the apoptotic CS-1 cells noted within 30 h went on to die by 72 h, as noted by the cell cycle analysis (Figure 4). In contrast, the apoptotic and dead cell populations within the low-dose group were not statistically different from the untreated control, indicating that CS-1 are able to overcome chemotherapeutic effects at low concentrations resulting in tumor escape and subsequent tumor growth. LP6 and LMS20 cells do not exhibit the same cell death 72 h after high dose PTX treatment (Figure 4), but instead remain arrested in the G2/M phase. A possible explanation for this observation is the slower proliferation rate of these sarcoma cell lines compared to the CS-1 cell line, and the mechanism of action of PTX, which affects cells that are undergoing mitosis. This difference in G2/M kinetics will be further explored in future studies for mechanistic susceptibility of the various sarcoma subtypes.

In addition to performing cellular proliferation analysis and cell cycle analyses, we also conducted in vitro and in vivo microarray analysis of the CS-1 at low and high dose to identify transcriptional changes that highlight the induction of mitotic arrest and subsequent apoptosis. Significant downregulation of histones occurs with the 1000 nM PTX treatment group (Figure 6B and D). This finding is consistent with our hypothesis that the degradation of histone mRNAs is a hallmark of cellular stress when DNA replication is inhibited by supratherapeutic doses of PTX.(40,41) The cytotoxic effects of PTX induce genotoxic stress and subsequent DNA damage in various cell types including CS-1 with DNA damage linked to inhibition of cyclin E-Cdk2 and cell cycle arrest upon activation of the G1 checkpoint.(42–44) To this point, paclitaxel treatment downregulates the gene transcript encoding for cyclin E (CCNE) in osteosarcoma resulting in cell cycle arrest.(45) Similarly, high dose PTX exposure downregulated CCNE in our chondrosarcoma line used in this study (Supplemental Figure S2).

Soft-tissue sarcomas are classically PTX-resistant in vivo and in vitro, as evidenced by local recurrence and death in a surgically resected patient-derived xenograft (PDX) model of chondrosarcoma, when treated with PTX delivered systemically via IP injection. PTX efficacy is only significant at high intratumoral concentrations unachievable via systemic administration. There is therefore significant room for improvement in strategies to increase PTX delivery efficiency. It is evident from the in vitro analyses, that high concentrations of PTX are able to completely inhibit the growth of CS-1 tumor cells and afford significant cell death. The trend of large shifts from G1 to G2/M arrest are present in all three sarcoma subtypes tested, indicating marked susceptibility to high dose PTX. The difference in kinetics from G2/M arrest to cell death does differ between the fast growing CS-1 vs the slower, more challenging LP6 and LMS20 cell lines, but the high degree of apoptosis demonstrated after 30 hrs in all 3 lines suggests that susceptibility to high dose PTX may be conserved among various sarcoma subtypes (Figures 4 and 5). This suggests that PTX films would therefore yield greater rates of cell death even in tumors which clinically behave as PTX-resistant. The postsurgical implantation of a PTX-loaded polymer film, carrying an equivalent PTX dose, prevents recurrence in this same CS-1 PDX model.(8,46–48) Similarly, histological analysis of exposed residual tumors demonstrates no evidence of drug resistance to our supratherapeutic PTX films with significant cell death being seen within 10 days of film implantation (Figure 3 and Figure 5B), whereas tumor treated with IP PTX display limited cell death and clinically resistant growth, thus highlighting the dose dependency of tumor recurrence and resistance.

With the increasing development of novel drug delivery platforms, researchers are revisiting current standards of efficacy achievable via traditional chemotherapeutics, and are aiming to improve drug delivery in conjunction with decreasing systemic toxicity.(49) Generally, tumor phenotype, location, and surrounding tissue significantly impact the efficacy of chemotherapy. Hypoxic, acidic, and nutrient-deprived tumor microenvironments, as well as higher interstitial fluid pressure within tumors compromise intratumoral delivery and efficacy via systemic PTX administration.(50,51) PTX-films change the paradigm for local drug delivery and achievable intratumoral drug levels, and thus may alter the “tumor resistance” behavior for many currently available chemotherapeutics. Supratherapeutic PTX levels, achieved with our PTX-films, are not clinically feasible via systemic administration due to significant mortality and morbidity. Our current findings suggest that chemotherapeutics traditionally thought to be ineffective may warrant re-evaluation in the setting of an enhanced drug delivery platform with an understanding that supratherapeutic doses of chemotherapeutic agents may improve the efficacy classically attributed to those agents. From a drug development perspective, the implications of a shift in drug delivery strategies will inform and change the way in which drug candidates are evaluated and selected based on in vitro and in vivo performance. From a clinical perspective, these findings open an opportunity for augmenting efficacy of already established chemotherapeutics, and provide a novel methodology for administration of new pharmacologic agents moving forward.