Evaluation of expansile nanoparticle tumor localization and efficacy in a cancer stem cell-derived model of pancreatic peritoneal carcinomatosis

Development of a pancreatic peritoneal carcinomatosis in vivo rat model using Panc-1-CSCs

Animal studies were performed in accordance with Boston University School of Medicine (BUSM) Institutional Guidelines. Panc-1-CSCs were isolated as previously described [28]. Briefly, Panc-1 cells in log phase were harvested and subcultured in complete MammoCult^® medium (Stem Cell Technologies, BC, CANADA) in 5% CO₂ in a humidified incubator at 37°C. After 2–3 weeks in culture, Panc-1 cells were harvested and plated in complete MammoCult medium containing 0.5% methylcellulose (Stem Cell Technologies, BC, CANADA) in 100 mm ultra-low attachment plates. Peritoneal tumors were developed from septenary CSCs. The Panc-1-CSC-derived peritoneal tumor model was developed in 4–5-week-old nude^nu/nu female rats (Charles River Laboratories, MA, USA). The intraperitoneal tumor implantation was performed in anesthetized animals using isoflurane maintained at 1–1.5%. Intraperitoneal injection of 2 × 10⁶ CSCs suspended in 1 ml of M2 media was performed in sterile conditions.

eNP synthesis & characterization

eNPs were synthesized according to published protocols using a two-step sonication-induced mini-emulsion and base-catalyzed polymerization procedure [29,30]. Nonswelling control particles were synthesized with 30% wt/wt cross-linker (30X-eNPs; vs 1% cross-linker in the standard eNP formulation), and generic, non-pH responsive poly(lactic-co-glycolic) acid nanoparticles (PLGA-NPs) were synthesized via mini-emulsion precipitation. The encapsulation efficiency of PTX in PTX-eNPs, PTX-30X-eNPs and PTX-PLGA-NPs was determined by HPLC and consistent with published data [15,30]. eNPs were covalently labeled by copolymerization of a rhodamine methacrylate or an iodinated eNP comonomer, as previously described [15,29]. Particle diameter and zeta potential were characterized via dynamic light scattering (DLS; Brookhaven Instruments Corp., NY, USA) and with a ZetaPALS zeta potential analyzer by diluting 10 μl of particles in 3 ml deionized water (∼50 kcps).

eNP swelling was monitored according to published procedures [29] by incubating 50 μl of iodinated eNPs in 2 ml of either 10 mM pH 7.4 phosphate buffer or 10 mM pH 5.0 acetate buffer for 24 h. A 5 μl drop of the diluted suspension was then adsorbed to a carbon-coated grid that had been made hydrophilic by a 30 s exposure to a glow discharge. Excess liquid was removed with filter paper (Whatman #1) and the samples were stained with 0.75% uranyl formate for 30 s. After removing the excess uranyl formate with a filter paper, the grids were examined in a JEOL 1200EX Transmission electron microscope or a TecnaiG2 Spirit BioTWIN and images were recorded with an AMT 2k CCD camera.

PTX release from each formulation was characterized by diluting 120 μl of PTX-eNPs, PTX-30X-eNPs or PTX-PLGA-NPs in 14.8 ml of 10 mM pH 5.0 acetate buffer or 10 mM pH 7.4 phosphate buffer with 0.1% wt/wt sodium dodecyl sulfate (SDS) and placing the total volume in 6″ of 10,000 molecular weight cut-off snakeskin dialysis tubing. Each bag was suspended in a 1000 ml sink of corresponding composition to the diluting buffer (either 10 mM pH 5.0 or pH 7.4 with 0.1% SDS) and 1 ml aliquots removed at 0, 2, 6, 12 and 24 h, and 2, 4, 6, 8 and 12 days. PTX was extracted from the particles via addition of acetonitrile (1:1, vol/vol) followed by filtration with a 0.22 μm filter and quantification via HPLC as above.

eNP cellular uptake & cytotoxicity in vitro

eNP uptake in Panc-1 cells and Panc-1-CSC spheroids was performed following a published protocol [31,32] by incubation of 25 μg/ml Rho-eNPs with cells for various durations (0–24 h). At each time point, cells were washed, fixed with 4% formaldehyde and rhodamine accumulation assess via flow-assisted cell sorting. In vitro cytotoxicity assays were performed as previously reported [30,31]; briefly, cells were incubated with treatments (10 μg/ml–0.1 ng/ml) of PTX-eNPs (1.2 mg PTX/ml; 25 mg polymer/ml), PTX solubilized in 50/50 Cremophor EL/ethanol (CrEL; 1 mg/ml), CrEL without PTX or unloaded-eNPs (25 mg polymer/ml) for 3 days at which point cell viability was assessed via MTS assay (Panc-1) or live-dead staining with propidium iodide (Panc-1-CSCs).

In vivo localization or Rho-eNPs to PPC tumors

Animal studies were performed in accordance with BUSM Institutional Guidelines. Xenograft PPC nude rats were randomly assigned to different study groups. At designated time points after tumor initiation (2 weeks, 5 weeks and end-stage), Rho-eNPs were injected intraperitoneally following sterile procedure under isoflurane anesthesia and aseptic preparation of the injection site. Animals received injections of 100 μL, 300 μL, or 1 ml of Rho-eNPs (25 mg polymer/ml; 0.2% wt/wt rhodamine/polymer) in the dose-response study and 1 ml in the time-course study. Fluorescent imaging was performed 24 h after Rho-eNP injection unless otherwise specified.

Simulating intraoperative visualization, fluorescent imaging of Rho-eNP localization was performed after midline abdominal incision and retraction under isoflurane anesthesia in a sterile manner using a Woods lamp (long-wave UV), and digital photography was used to document fluorescence. Ex vivo fluorescence imaging was also performed using the Woods lamp and digital photography or an Olympus stereo-fluorescence microscope for analysis of cut-surfaces of tumors and normal organs. In the presence of ascites fluid, the fluid was collected and placed into sterile conical tubes for fluorescence imaging with a Woods lamp, and documentation by digital photography. Thereafter, rats were euthanized via removal of vital organs under deep isoflurane anesthesia.

PTX-eNP efficacy against PPC

Thirty-five 4–5-week-old nude^nu/nu female rats were injected intraperitoneally with 2 × 10⁶ Panc1-CSCs/rat using the same septenary Panc-1-CSC preparation. Treatments (1×/week × 4 weeks) commenced 2 weeks post cell injections. PTX (10 mg/kg; n = 9), PTX-eNPs (10 mg/kg; n = 9), unloaded eNPs (equivalent polymer dose; n = 9) or saline (n = 8) were administered to PPC tumor-bearing rats. All injections were performed under isoflurane anesthesia and in sterile manner. Rats were monitored daily for health concerns; body weights were obtained weekly; observing monitors were blinded to the treatment group. At the end stage, rats were assessed and euthanized with photo-documentation of ascites, peritoneal tumors and abdominal organ states, and tumor and organ collection for further analysis.

Development of PPC in vivo rat model using Panc-1-CSCs

To develop a xenograft tumor model of PPC, we isolated and employed a heterogeneous pool of Panc-1-derived CSCs isolated for anoikis resistance as defined by growth, survival, and spheroid formation in methylcellulose suspension culture through eight passages (Figure 1A). Confocal microscopy of immunostained Panc-1-CSC spheroids revealed the pairwise coexpression of CD133 and CXCR4 – markers of increased metastatic potential [33,34] – and the dual-endothelin1/VEGF signal peptide receptor, DEspR (Figure 1B).

Following in vitro validation of the Panc-1-CSC line, we developed a metastatic PPC xenograft model in nude rats via an intraperitoneal injection of 2 × 10⁶ Panc-1-CSCs, which resulted in rapid tumor establishment throughout the peritoneum. Both large (>1 cm) and small (<0.5 cm) tumors were observed, with larger tumors exhibiting hard nodular features consistent with stromal desmoplasia (Figure 1C). Notably, some rats developed bowel obstruction (Figure 1D), jaundice due to obstruction of the hepatic portal and cachexia (data not shown), as well as ascites (Figure 3 & 4). Histological evaluation of tumors stained with Masson trichrome revealed that even micro tumors (∼1 mm) exhibited fibrotic deposition (Figure 1E & F).

The percent tumor take-rate (i.e., the number of animals that develop tumor divided by the total number of animals xenografted) for animals xenografted with 2 × 10⁶ Panc-1 cells was 0%. In contrast, xenografting with an equivalent number of Panc-1-CSC cells yielded a tumor take-rate of 66% (Figure 1G). By pretreating animals with cyclophosphamide, the Panc-1-CSC take-rate was improved to >95% and this strategy was used to produce sufficient animals for the ensuing imaging and efficacy studies. In-life, animals displayed comorbidities including ascites, both serous and hemorrhagic, cachexia, with associated gut hypoplasia and decreased activity, gut obstruction, bowel dysfunction and jaundice. At end stage/necropsy, animals displayed retroperitoneal metastases and presented with significant variation in PPC tumor size (Figure 1H).

eNP synthesis & characterization

eNPs (Figure 2A), 30X-eNPs and PLGA-NPs were synthesized according to published protocols with a mean diameter/polydispersity index (PDI) of 150 nm/0.19, 145 nm/0.21 and 210 nm/0.15, respectively, as determined by dynamic light scattering (Figure 2D) [15,29,30]. All particles were negatively charged with zeta potentials greater than -40 mV (Figure 2D). The mean encapsulation efficiency of PTX in PTX-eNPs, PTX-30X-eNPs and PTX-PLGA-NPs loaded with 5% PTX (wt drug/wt polymer) was 86, 92 and 78%, respectively, as determined by HPLC. Transmission electron micrographs of eNPs incubated under pH 5 conditions for 24 h revealed an increase in particle diameter and a decrease in particle density (i.e., less contrast/signal) as compared with particles incubated at pH 7.4. Twenty-four hours was chosen as the incubation time for particle swelling studies based upon previously conducted assays [16,29,30] demonstrating observable changes in particle size and shape within this timeframe. Last, PTX release under sink conditions was pH- and time-dependent for PTX-eNPs, which demonstrated rapid, triggered release at pH 5.0 and slow, eventual leakage at pH 7.4 (Figure 2C). PTX release from heavily cross-linked PTX-30X-eNPs was not pH-dependent and mirrored the slow rate of leakage from PTX-eNPs at pH 7.4. PTX-PLGA-NPs released >80% of their payload within 48 h regardless of pH.

eNP cellular uptake & cytotoxicity in vitro

The time-course accumulation of Rho-eNPs in differentiated Panc-1 tumor cells and tumor-initiating Panc-1-CSCs was evaluated by flow cytometry. The time course analysis revealed increasing fluorescence from 1–24 hrs concordant with increasing particle uptake given the same number of fluorescent Panc-1 tumor cells (Figure 3A1, left). Within 4 h of coincubation at 37°C, over 80% of Panc-1 cells demonstrated Rho-eNPs uptake (Figure 3A1, right). This cellular uptake of Rho-eNPs by Panc-1 tumor cells was almost completely inhibited at 4°C (<10% uptake after 24 h coincubation; Figure 3A2) suggesting active, energy-dependent cellular uptake.

In Panc-1-CSCs, which are grown in spheroid clusters, 30% of cells exhibit fluorescence at time 0 (i.e., Rho-eNPs pipetted onto cells and immediately removed followed by washing; Figure 3B1, right), but with lower fluorescence-intensity compared with the 4 and 24 h time points (Figure 3B1, left). Similarly to the increased fluorescence at 4 and 24 h, confocal microscopy revealed a combination of particle internalization and adhesion to the cell surface (Figure 3B2).

PTX-eNPs demonstrated dose-dependent cytotoxicity similar to that of free PTX against both Panc-1 cells grown in monolayer (IC₅₀ = 5.6 ng/ml, IC₅₀ = 2.0 ng/ml, respectively) and Panc-1-CSC spheroids grown in suspension (IC₅₀ = 13.8 ng/ml, IC₅₀ = 2.1 ng/ml, respectively) (Figure 3C1 & C2). Interestingly, the Cremophor EL/ethanol excipient used to formulate free PTX was found to be significantly cytotoxic at the highest concentrations tested while unloaded-eNPs (vehicle control for PTX-eNPs) exhibited only minimal cytotoxicity at concentrations required to deliver equivalent amounts of PTX.

Time course analysis of Rho-eNP localization to PPC tumors

We evaluated the in vivo tumor localization of Rho-eNPs using the Panc-1-CSC-derived human xenograft PPC nude rat model. First, we examined the time-course of distribution of rhodamine fluorescence following injection of a 1 ml dose of Rho-eNPs (25 mg polymer/ml). A hand-held Woods Lamp (long wave UV) enabled straightforward and facile visualization of rhodamine fluorescence within the peritoneal cavity. As shown in Figure 4, time course analysis of identical intraperitoneal doses of Rho-eNPs revealed increasing fluorescence in tumor areas from 1–24 h. Specifically, at the 1 h time point, minimal Rho-eNP accumulation was observed in the tumor with significant fluorescent signal still detected in the ascites and peritoneal fluid (Figure 4A). Increasing Rho-eNP fluorescence was observed within the tumors up to 4 h and then appeared to stabilize out to 24 h after injection. Rho-eNP fluorescence was observed to increase in proportion to the number of tumors with: greater omentum>gastro-splenic ligament>small intestinal mesentery>retroperitoneal tumors. Notably, in all rats, there was no detectable fluorescence in normal abdominal or pelvic organs (Figure 4), nor in the lungs, heart and brain (data not shown).

Dose-response analysis

Next, we evaluated dose-response effects on localization via intraperitoneal administration of 100 μl, 300 μl and 1 ml doses of Rho-eNPs (all 25 mg polymer/ml) with fluorescence-imaging 24 h post injection at euthanasia. As shown in Figure 5, we detected increasing fluorescence within the tumors with increasing Rho-eNP dose. As with the previous time-course study, Rho-eNP fluorescence was localized to tumor-rich areas including the: greater omentum, lesser omentum and small intestinal mesentery. No fluorescence was observed in normal, abdominal and retroperitoneal organs. Significant particles accumulation was observed in the microtumors.

Rho-eNP localization in early- & late-stage PPC

To determine whether tumor stage/size was a determinant of Rho-eNP localization, we administered a 1 ml injection of Rho-eNPs to animals either 2 (early peritoneal dissemination) or 5 weeks (late-stage disease) postxenografting, and imaged the peritoneum 24 h postinjection. Significant accumulation of Rho-eNPs was observed in earlier-stage tumors with bright signal in many, although not all, microtumors (Figure 6A). A similar trend was noted at the 5-week tumor stage (Figure 6B). No empirical difference was observed between early- and end-stage accumulation of Rho-eNPs in microtumors. No fluorescence was detected in normal abdominal and retroperitoneal organs or within organs of the reticuloendothelial system (e.g., kidney; Figure 6C, top row). Interestingly, when several large ‘macro tumors’ (∼1 cm diameter) showing fluorescence accumulation were harvested and bisected, Rho-eNPs were found to have penetrated throughout the entire tumor and not just in the superficial layers (Figure 6C, middle row). Furthermore, fluorescence microscopy analysis confirmed tumor-cell localization and sparing of normal mesothelium and tumor microvasculature (Figure 6C, bottom right panel.) Histological evaluation of tissues revealed that Rho-eNPs accumulated both intracellularly and in the extracellular matrix, thus indicating the potential for intracellular delivery of a drug payload.

PTX-eNP efficacy against PPC

Having observed Panc1 and Panc-1-CSC uptake of eNPs and cytotoxicity from PTX-eNPs in vitro, as well as eNP tumor localization in vivo, we next determined the efficacy of PTX-eNPs in vivo. Beginning 2 weeks postxenografting, rats with established PPC tumors were treated with four weekly 1 ml doses of PTX-eNPs (experimental; 10 mg/kg PTX), PTX (clinical formulation control; 10 mg/kg), unloaded-eNPs (vehicle control) or saline (tumor growth control). Both PTX treatments improved survival compared with the saline and unloaded-eNP control groups (median survival 26 and 29 days, respectively; p < 0.05; Figure 7A). However, there was no significant difference in overall survival between the PTX-eNP and PTX groups. The study concluded 50 days postxenografting, at which point gross analysis of tumor burden in all remaining animals (average tumor burden scores [0 none, +1 mild, +3 moderate, +5 marked] in three regions: greater omentum, mesentery and retroperitoneum) revealed a trend toward less tumor burden in PTX-eNP-treated animals compared with PTX and unloaded-eNP-treated animals. However, no statistically significant differences in tumor burden were observed (Figure 7B).

Interestingly, while overall survival was not statistically significantly different between PTX-eNP- and PTX-treated rats, their respective in-life activity and degree of cachexia marked by muscle mass and gut size at end stage were observed to be markedly different. PTX-eNP-treated rats were observed to be more active in-life and demonstrated minimal cachexia as confirmed at necropsy; however, the majority of animals displayed significant ascites. The PTX-treated rats were markedly cachectic and exhibited decreased activity and gut hypoplasia at necropsy, while a minority displayed ascites.

The empirical observations of cachexia-associated gut hypoplasia were borne out in histological analysis of gut diameters and wall thickness. To achieve comparable gut segments, analysis was limited to the duodenum within 1 cm of the gastroduodenal junction. Histological analysis of diameters showed smaller duodenal diameters in the PTX-treated group compared with age-matched nontreated controls (Figure 8A & B). However, PTX-eNP-treated rats exhibited slightly larger diameters compared with control nontreated nude rats suggesting dilation of the duodenum or less postmortem contraction upon fixation. The latter is likely since analysis of wall thickness reveals that nontreated control duodenal wall thickness is twofold greater than in PTX-eNP-treated animals (p < 0.01) (Figure 8B). Notably, duodenal wall thickness in PTX-treated animals is threefold less than controlled nontreated nude rat duodenum.