Modular polymer platform as a novel approach to head and neck cancer therapy

Head and neck cancer is the sixth most common cancer in the world, with more than 300,000 deaths attributed to the disease annually. Aggressive surgical resection often with adjuvant chemoradiation is the cornerstone of treatment. However, the necessary chemoradiation treatment can result in collateral damage to adjacent vital structures causing a profound impact on quality of life. Here, we present a novel polymer of poly(lactic-co-glycolic) acid and polyvinyl alcohol that can serve as a versatile multidrug delivery platform as well as for detection on cross-sectional imaging while functioning as a fiduciary marker for postoperative radiotherapy and radiotherapeutic dosing. In a mouse xenograft model, the dual-layered polymer composed of calcium carbonate/thymoquinone was used for both polymer localization and narrow-field infusion of a natural therapeutic compound. A similar approach can be applied in the treatment of head and neck cancer patients, where immunotherapy and traditional chemotherapy can be delivered simultaneously with independent release kinetics.


Scientific Reports
| (2022) 12:3592 | https://doi.org/10.1038/s41598-022-07324-y www.nature.com/scientificreports/ patient's daily activities are not restricted during the entire treatment time. An additional benefit of this polymer system includes prophylaxis against tumor recurrence following resection. Viable squamous cell carcinoma cells have previously been recovered from the surgical wound following neck dissection and have been shown to be capable of growing as colonies in vitro; theoretically, these may implant and cause cancer recurrence 10 . The combined use of radiotherapy and chemotherapy has been effective in improving the therapeutic index of radiation therapy for a variety of human cancers 11,12 . Previous studies have reported that the intratumoral administration of chemotherapeutic agents such as cisplatin, in a sustained-release drug delivery system markedly improved antitumor efficacy and reduced systemic toxicity compared with systemically administered drugs 13 . Indeed, even in our previous work, we demonstrated that a poly(e-caprolactone):poly(lactide-cocaprolactone) polymer film loaded with cisplatin can be used for delivering targeted local chemotherapy in a partially resected xenograft animal model of head and neck cancer 14 . Choice of calcium carbonate and TQ. In this study, we used the combinatorial approach of a dual-layer polymer, employing CaCO 3 for the imaging-based facilitation of radiotherapy (RT) and thymoquinone (TQ), a radiosensitizing agent, to enhance RT. Calcium carbonate is an inert material that is easily detected on crosssectional imaging. In addition, it is cheaper and industrially feasible compared with other high-Z metals, such as gold nanoparticles. While operating, surgeons can easily determine in which anatomic areas a tumor was adherent vs. being easily removed. Intraoperatively, the surgeon can place the polymer in the areas of closely adherent tumor, thus guiding the radiation therapist's fields more precisely. Patients can then receive precision treatment, through image guided localization of the polymer implant, minimizing side effects of radiation therapy. Thymoquinone (TQ), a natural product derived from black seed oil, is shown to have anti-cancer properties while being biologically well-tolerated [15][16][17][18] . We established that in a partially resected xenograft animal model of head and neck squamous cell carcinoma the dual-layer CaCO 3 /TQ polymer platform demonstrated a significant reduction in the tumor volume and offered an avenue for the precise localization of areas in the wound bed for adjunct RT.

Results
Cells vs TQ and radiation. For determining the optimal TQ dose required to induce LD50, a dose escalation experiment was performed using a panel of one mouse and six human head and neck cell lines: SCCVIISF, FADU, SCC47, A253, Detroit 563, SCC25 and, two squamous cell lung cancer cell lines RH2, A549 respectively. Across all cell lines, a 15 μmol dose of TQ achieved a statistically significant (p < 0.001) LD50 or better as observed via a cell colony formation (clonogenic assay) (Fig. 1A). Therefore, we aimed to achieve a dose of at least 15 μM as the effective dose in the drug delivery polymer layer for the in vitro cell assay. In order to calculate the total dosage, polymers were weighed before implantation (after surgical debulking) and again 30 days later at the termination of the experiment (after removal). The change in the weight reflected the absolute diffusion of TQ from the polymer film. Through this, it was estimated that the total dose delivered over a period of 30 days was 23 mg with an average dose delivery rate of 0.77 mg per day or 4.68 μmol/uM 3 . The mouse cell line was treated with 10 μmol of TQ and irradiated at 2, 4, 6 and 8 Gy to demonstrate that TQ plus radiation had a greater cell killing effect over radiation alone (Fig. 1B). The surviving fraction at 8 Gy was significant with p < 0.05 and with 4.86% (SD 2.7%) for RT only, and 1.2% (SD 0.6%) for RT in combination with 10 μmol of TQ.  Visual examination. At the termination of the experiment (30 days post-polymer implantation) a necropsy was performed for both the treatment polymer group and the control polymer group, and tissues were examined grossly, radiographically and histologically (Figs. 5 and 6). Vehicle control tumors (polymers with no drugs and no RT) had persistence of poorly differentiated carcinomas with completely viable tumors when evaluated microscopically at 40 × with H&E staining. Poorly differentiated cancers appear very abnormal and lack the organizational structures of the tissues they reside in. (Fig. 6). TQ polymer-treated groups demonstrated tumor killing with tumor necrosis along the drug treatment polymer layer (Fig. 6).
Pathological examination. In order to study pathological changes between the treatment groups, histology was performed to evaluate differences at the cellular level. The control group, which received the polymer    www.nature.com/scientificreports/ without TQ or radiotherapy, showed dense nuclei at the polymer placement site, revealing persistent poorly differentiated carcinoma with viable tumor (Fig. 6A). The group treated with the TQ polymer only (Fig. 6B) revealed no tumor regrowth with tumor necrosis and no viable tumor in the treatment area. In contrast, the radiotherapy-only group presented with a narrow band of necrosis along the polymer removal site with still-viable portions of tumor tissue found below (Fig. 6C). The TQ polymer with radiotherapy group (Fig. 6D) presented with no residual tumor present.
FDG-PET/CT assessment. FDG-PET/CT imaging was used to assess tumor metabolic activity and to identify the polymer location relative to the tumor. Figure 5B1-D1 demonstrate FDG-PET/CT imaging with decreased uptake of the PET tracer observed in areas of the tumor directly in contact with the drug-delivering polymer edge. This was in contrast with the control polymer group, which did not receive TQ drug treatment ( Fig. 5B2-D2); where robust cancer persists. TQ-polymer-treated uptake of FDG-PET was half that of the vehicle control with 0.024%ID/g vs 0.055%ID/g, respectively.
Tumor growth dynamics. In order to assess the overall changes in tumor volume, each tumor was measured across their length, width, and height using calipers. In an analysis of the growth rate over the course of 30 days, the radiotherapy plus TQ polymer groups revealed the slowest rate of change (Fig. 7A), and upon terminal histological examination no viable tumor was found. In review with our expert Head and Neck pathology team, this minimal change was likely due to caliper measurement variations due to inflammation and wound healing post treatment. A statistically significant difference was found in the tumor volume when the control polymer group was compared with all other treatment groups (Fig. 7B, P = 0.002). The radiotherapy plus TQ polymer groups demonstrated the smallest tumor volumes overall. RT vs RT + TQ polymer treatment proved significant with p < 0.05.

Discussion
Significance of this study. In this study, we demonstrated that a modular, dual-layer CaCO 3 /TQ polymer provided effective anti-cancer treatment in in vivo cancer xenograft models. A multilayer polymer allowed for the individual loading and adjustment of drugs for cancer treatment. The polymer film physically adheres to the tumor bed contours when placed surgically, thus allowing the drug delivery layer to come into direct contact with the remaining cancer tissues and or cancer bed. The interaction of the polymer film and the tumor surface, along with the slow biodegradation of the polymer promoting localized drug delivery throughout the RT therapy, demonstrated a significant decrease in the tumor growth rate when compared with other groups over the course of the experiment.   15,16 . TQ is a biologically well-tolerated compound 15 , which was further exhibited in our mice, as they showed no local tissue injury at the site of implantation. Our data also demonstrates that the polymer's ability to provide the sustained release of TQ is what drives the therapeutic benefit observed in our study. TQ has been shown to inhibit cancer cell proliferation; arrest cancer cell cycle progression; and, induce pro-apoptotic effects with the targets of Bcl-2, caspases, PPArs, NF-kB, STAT3, MAPK, Akt and ROS 16 . Polymers containing therapeutic agents that can be released slowly over time would continue to provide therapeutic benefit even for microscopic disease (clinical R1 resection) unable to be detected by clinical imaging or visualization. The ability to manipulate the duration of drug release, and the order of drugs released from multiple layers can add great clinical utility in the treatment of patients. The combinational approach of RT and the local release of chemotherapy has been shown to be an effective therapeutic strategy, especially as the local administration of chemotherapeutics does not contribute to systemic side effects 19 . Our platform makes use of this principle, allowing for the synergistic optimization of RT via local radiosensitizing drug delivery to the cancer cells left behind following surgical debulking. The polymer platform allows for a significant reduction in tumor volume and offers an avenue for the precise localization of areas in the wound bed for adjunct RT.
The use of multiple layers composed of various properties (PVA and PLGA) provides the ability to modulate the chemotherapeutic drug delivery rate and the lifetime of the imaging layer independently. The approach is similar to poly (N-isopropylacrylamide) (PNIPAAm) and PLGA for separate release mechanisms 20 . The modular platform allows for multiple layers that deliver controlled and targeted release of radiosensitizers, targeted inhibitors, and immunomodulators. Previous studies demonstrated that magnetic-based core-shell particles with the sustained release of curcumin and the temperature-dependent release of doxorubicin was effective in a melanoma orthotopic mouse model 21 . With the modification of the fabrication process of PVA and PLGA polymers, the controlled release may last as long as seven months 22 .

Use of CaCO 3 for polymer localization.
One novel feature of the study design was the use of CaCO 3 for visualization on cross-sectional imaging. In prior studies, inert metals or radiopaque hydrogel (e.g., TraceIT Tissue Marker) have been used as fiducial markers for image-guided radiotherapy, external beam radiation and brachytherapy localization 23,24 . The benefit of using CaCO 3 is the use of a natural compound that is absorbed by the body over time 25 . This is the first study of its kind to demonstrate the utility of embedded CaCO 3 in a polymer as a potential fiduciary marker for image-guided radiotherapy. While operating, surgeons can easily determine in which anatomic areas a tumor was adherent vs. being easily removed. Intraoperatively, the surgeon can place the polymer in the areas of closely adherent tumor, thus guiding the radiation therapists' fields more precisely. The flexibility of the polymer sheet allows for it to deform with the tissues it resides. The polymers' flexibility is a unique feature compared to traditional markers used for localization. The shape of the polymer sheet within the patient can then guide conformal radiation therapy minimizing secondary side effects of radiotherapy.

Conclusion
The management of head and neck cancer patients poses a considerable challenge to the surgeon and the radiation oncologist. Herein, we report a novel dual-layer polymer film that is capable of delivering antitumor therapeutic agents with significant tumor kill. This polymer wrap was designed to be applied intraoperatively to the surgical bed after removing or debulking the tumor, thus allowing for enhanced post-operative radiation treatment, and also functioning as a platform for the delivery of therapeutic agents and immunomodulators. Besides www.nature.com/scientificreports/ its clinically relevant features, the modular nature of this polymer platform provides an elegant approach to future investigations, allowing us to seek out specific molecular targets. We have the capacity to build polymers that specifically target each patient's tumor as we can profile the tumors and then target them selectively on the polymer platform. As more combinations of the polymer platform are developed, direct polymer therapy may play an important role in the armamentarium against oral and head and neck carcinoma, as well as many other cancers and human diseases.

Methods
Cell culture. Head and neck cancer cell line panels were ordered from ATCC and mouse derived squamous cell carcinoma, SCCVIISF, was spontaneously captured from a mouse tumor and grown in culture. All cell lines were cultured in the following conditions: 37 °C, 5% CO 2 incubator with Dulbecco's Modified Eagle's Medium (DMEM), 10% fetal bovine serum (FBS), and 1% penicillin/streptomycin. The media was changed twice a week until the cells were confluent. Confluent cells were then passaged or harvested for use.
Multilayer films were prepared layer-by-layer, and the fabrication of each layer depended on various factors of the material, including its rigidity, toughness, adhesion, degradability, porosity, hydrophilicity, and loading capacity, among others. One fabrication technique involved a mechanical spinner followed by heat and evaporation of the solvent. First, 10 g of the PVA powder were added to 100 mL of deionized water and then centrifuged for 1 h at 95 °C to obtain a 10% (w/v) PVA solution. Then, 0.4 g of calcium carbonate (CaCO 3 ) microparticles were dispersed in 10 mL of a PVA solution after bath sonication for 5 min and the mixture was vortexed for 2 min (Fig. 1). The obtained homogeneous solution was dropped onto a plastic petri dish and placed in a shaker so that the solution could be spread evenly. Subsequently, the petri dish was transferred to an oven at 70 °C for 2 h to ensure that the water was completely evaporated. Finally, the first layer (CaCO3/PVA film) for CT imaging was obtained. Furthermore, 0.5 g of the PLGA powders were added in 10 mL of DCM and then centrifuged for 30 min to obtain 5% (w/v) PLGA solution. Then, 20 mg of the TQ powders were dispersed in 2 mL of a PLGA solution after bath sonication for 5 min and the mixture was vortexed for 2 min. The resulting mixture was dropped on the first layer in the petri dish and then evaporated with ambient air under proper ventilation for 30 min. The multilayer polymer film was imaged using scanning electron microscopy (Fig. 2). Tensile stress was measured for the polymer as shown in Fig. 3. Finally, multilayer films with the first layer (CaCO 3 /PVA film) for CT imaging and the second layer (TQ/PLGA film) for chemotherapy were obtained (Fig. 4).  26 guidelines and protocols were approved and followed. For testing the polymer platform, 20 8-week-old C3H/HeJ mice were injected with 400,000 cells of a mouse derived squamous cell carcinoma line, SCCVIISF, in the right rear flank. Tumor growth was assessed with calipers three times per week following polymer implantation for 18-31 days to evaluate the antitumor efficacy of the different treatments. The control mice required euthanasia as determined by the Animal Research Committee of University of California at Los Angeles and in accordance with the American Veterinary Medical Association (AVMA) guidelines for the euthanasia of Animals (2020) 27 due to the tumor burden. The lengths, widths, and heights (in mm) of the tumors were measured and the tumor volume (cm 3 ) was calculated according to the formula: Tumor volume = π/6 × length × width × height. When tumors reached an average size of 0.5-1 cm, all animals underwent surgery to debulk their tumors by 50%. This was done to approximate the surgical situation when a patient's tumor is unresectable, and some tumor is left behind prior to polymer therapy. Animals were then randomly assigned to the various treatment groups. The treatment groups included: (1) Inert polymer with no drug; (2) TQ polymer; (3) Inert polymer + 3 × 4 Gy RT; (4) TQ polymer + 3 × 4 Gy RT. No systemic TQ treatment was given. Each tumor bed was covered with 2.25 cm 2 of the polymer cut in the shape of the remaining tumor. The polymer was draped over the tumor edges and sutured in place.
Radiation therapy. Mice in the RT group were anaesthetized with ketamine/xylazine and had eye ointment applied to their eyes for days 1, 2, and 3 post-surgery for RT treatment. The mice were positioned under ½ inch of lead shielding leaving only the tumor exposed. An X-ray dose was delivered at 0.4299 Gy/min for 9.3 min until 4 Gy of the total dose was received. Mice received 3 × 4 Gy RT, which is the scaled comparable dose given to head and neck cancer patients.
MicroPET imaging. In vivo small animal imaging was conducted at the Crump Institute's Preclinical Imaging Technology Center. The mice were injected via the lateral tail vein with a radiotracer (70 μCi for 18F-FDG and 200-400 μCi for 11C-L-glutamine), and they then underwent 60 min of uptake under 2% isoflurane anesthesia, followed by microPET (G8 PET/CT, PerkinElmer) and microCT (CrumpCAT, Arion Hadjioannou laboratory) imaging. The quantification of 18F-FDG uptake was done using AMIDE software by drawing a region of interest over the tumor and the entire body, as well as by calculating each of the maximum uptake values (SUVmax) as the percent of the injected dose per gram (%ID/g).

Scientific Reports
| (2022) 12:3592 | https://doi.org/10.1038/s41598-022-07324-y www.nature.com/scientificreports/ Tissue collection. After the animals were sacrificed, a gross necropsy examination was conducted of the surrounding tissues. No discernable differences at the implant site could be observed between the groups. Tumor specimens were collected for sectioning and hemotoxin and eosin staining were done through the UCLA Translational Core Pathology Lab. Histopathological examination was performed with the assistance of a senior pathologist at UCLA Medical Center.
Estimation of TQ dose rate delivery. Imbedded polymers were weighed before implantation and upon resection. The imaging and drug treatment layer were peeled apart and weighed separately, as the imaging layer was formulated to last three months where the drug treatment layer was formulated to last 1-1.5 months. Subtracting the difference in weights from before and after one month of implantation, we found that the CaCO 3 imaging layer lost 5% of its weight and the drug treatment layer lost 60% of its weight, equating to an estimated delivery of 0.77 mg per day or 4.68 uM/uM 3 .

Statistics.
The statistical significance was set at P = 0.05. The dose-escalation experiment for the optimal TQ dose required to induce LD50 across eight cell lines was generated via a fitted, non-linear regression curve and determined via a one-sample t-test. The tumor volume was compared between the treatment groups with a one-way analysis of variance model. In addition, the statistical significance of differences in the tumor volume when comparing the control group with all other treatment groups was determined via an unpaired two-sample t-test. A linear quadratic survival curve depicting TQ-treated SCCVIISFs versus the control was analyzed via GraphPad Prism.