Development of a Clinically Viable Strategy for Nanoparticle‐Based Photodynamic Therapy of Colorectal Cancer

Photodynamic therapy (PDT) is an anticancer treatment modality that is poorly adopted into clinical practice for a variety of factors. Though fundamental studies supporting its utility abound, few clinical trials are performed on its use. This may be due to failures in the translation of treatment protocols used in fundamental studies into clinically viable protocols. This study seeks to develop a clinically viable protocol for the treatment of colorectal cancer using nanoparticle‐based PDT by using three complementary animal models. Using the porphysome nanoparticle as a model photosensitizer, several theoretical and practical challenges to the clinical delivery of PDT are addressed including the required drug‐light interval (DLI), the route of administration, and the method of light irradiation delivery. This study finds that nanoparticle‐based PDT can effectively ablate colorectal cancer tumors, that PDT can be effectively delivered after a relatively short DLI, and that safety concerns related to off‐target effects can be mitigated through the peritumoral administration of the photosensitizer and the transanal intraluminal placement of light irradiation. These findings lay the framework for future clinical trials investigating the use of PDT as a component of the multi‐modal approach to the treatment of colorectal cancer.


Introduction
Colorectal cancer is one of the most significant cancers in modern medicine: it is among the most common cancers and is the fourth most common cause of cancer related death, accounting for 9.2% of cancer deaths globally. [1] Current treatment relies on surgical resection of the primary tumor with the potential addition of chemotherapy, or radiotherapy. One of the most significant challenges that exists in the management of colorectal cancer is with respect to rectal tumors arising within a few centimeters of the anal verge. These are traditionally resected using an abdominoperineal resection (APR), which sacrifices the anal sphincter and consigns the patient to a permanent colostomy. This confers a significant compromise in quality of life and predisposes patients to a variety of stoma-related complications for the rest of their lives. [2] There is also evidence that APR is independently associated with a 26% rise in mortality as compared to sphincter-preserving surgery in similar patients. [3] Neoadjuvant therapy in rectal cancer has become increasingly effective in shrinking tumors and enabling patients who would have required an APR to instead receive sphincter-preserving surgery; [4] however, between 23% and 38% of patients with rectal cancer require removal of the anal sphincter at the time of surgery. [3,5] Clinically novel modalities, like photodynamic therapy (PDT) that generates cytotoxic singlet oxygen through interactions between optical light delivery and administration of non-toxic photosensitizers, has emerged as a viable tool for localized treatment of malignant tissues and may be useful adjuncts to existing adjuvant treatment modalities. [6] The use of nanoparticles as photosensitizing agents presents a unique opportunity to simultaneously enhance the delivery of the photosensitizer and avoid systemic adverse effects. [7] Porphysomes (PS) are nanovesicles comprised of porphyrinlipid conjugate subunits that self-assemble in aqueous solution to form a bilayer liposome-like structure. [8] They have demonstrated excellent stability in storage and in vivo compatibility, with no adverse effects observed in mice administered doses of up to 1000 mg kg −1 . [8] PS can exist in two potential forms in biological condition: as an intact vesicle superstructure, and www.advancedsciencenews.com www.advtherap.com as a nanostructure disrupted form in release of porphyrin-lipid monomers. The form of the PS determines its photophysical and photochemical properties. While in its intact form, the PS demonstrates a high degree of self-quenching as a result of dense packing and interaction of porphyrin molecules within the nanoparticle. This high degree of self-quenching means that energy (i.e., photons) absorbed by the intact PS, normally released via fluorescence or production of singlet oxygen, is dissipated as thermal energy. [8] This thermal energy production is the basis of photothermal therapy (PTT); however, relatively high laser fluences (usually ≥500 mW; 1.18 W cm −2 for porphysomes) are required to achieve clinically relevant heating. [8,9] When PS are disrupted into their monomeric form, they become unquenched and "photo-active"-able to participate in photophysical and photochemical reactions, namely fluorescence and singlet oxygen production. This disruption is thought to occur upon cellular uptake of the nanoparticle. Prior research has established that PS can be used as photosensitizers for ablation of various cancer subtypes, largely in subcutaneous xenograft models. [8][9][10][11] Although porphysome-based PDT (PS-PDT) is theoretically possible based on these fundamental studies, the actual application of PS-PDT to colorectal cancer requires further development and optimization before a clinical trial would be feasible. An ideal PDT treatment for colorectal cancer would possess several key features: easy delivery to the target tissue, short drug-light interval (DLI), to facilitate a single day-procedure hospital visit, high tolerability (i.e., minimal off-target effects), and effective ablation of malignant tissue. The goal of the present study was to create a PS-PDT treatment protocol that met these desired features by optimizing drug administration method, DLI for PDT, and light placement approach on three complimentary animal models of colorectal cancer (Scheme 1).

PS-PDT is Effective in the Ablation of Subcutaneous Colorectal Cancer via Intravenous Injection
Porphysomes were synthesized according to previously reported protocols. [8] Briefly, porphyrin-lipid conjugates are first created through the acylation of pyropheophorbide with 1-palmitoyl-2-hydroxy-sn-glycero-2-phosphocholine. These porphyrin-lipids are then mixed with cholesterol and PEG2000-DSPE (Avanti Polar Lipids) in chloroform and dried under a gentle stream of nitrogen gas to form a thin film. The film was then hydrated in phosphate-buffered saline, subjected to several freeze-thaw cycles, and extruded through a 100 nm polycarbonate membrane at 65°C to give PS with size ≈100 ± 5 nm, polydispersity index of <0.15 determined by dynamic light scattering (Malvern Instruments, Malvern U.K.), and concentration quantified by UV/vis spectrophotometry (Varian Inc. Palo Alto, CA). [8] All animal experiments were approved by the University Health Network Animal Care Committee. Experiments adhered with all relevant institutional, provincial, and federal requirements. Athymic nude mice bearing subcutaneous HT-29 human colorectal cancer xenografts and BALB/c mice bearing subcutaneous CT26 syngeneic murine colorectal cancer tumors were treated with intravenously administered porphysome-based photodynamic therapy (IV-PS-PDT): animals received intravenously administration Scheme 1. Optimizing PS-PDT treatment protocol for colorectal cancer treatment in pre-clinical animal models.
of PS and then the tumors received light irradiation of 135 J cm −2 by a 671 nm laser at a DLI of 24 h. At both animal models, the IV-PS-PDT treated groups demonstrated significant tumor ablation and improved survival compared with the untreated control groups (Figure 1). This verified that IV-PS-PDT was theoretically effective in treating both human colorectal cancer tumors in an immunocompromised host, and syngeneic colorectal cancer tumors in an immunocompetent host using previously established IV-PS-PDT treatment parameters.

Reducing Drug-Light Interval for IV-PS-PDT to Facilitate Single-Day Treatment
To explore the feasibility of reducing DLI such that a PDT treatment could be completed in a single day, BALB/c mice bearing subcutaneous CT26 murine colorectal tumors were treated with IV-PS-PDT at a PS dose of 7.5 mg kg −1 , light fluence of 135 J cm −2 and varied DLI of 1, 3, 6, or 24 h. These DLIs were selected to represent the range of feasible intervals between the administration of a photosensitizer to a human patient and the time to delivery of light, with intervals beyond 6 h considered too great to be completed in a same-day setting. All treatment protocols successfully reduced tumor volume compared with the untreated control group, with the greatest treatment effect seen at a DLI of 3 or 1 h.
Tukey's multiple comparisons tests detected significant differences between all groups except for between the 3 h DLI group and the 1 h DLI group (Figure 2A, p < 0.0001). In addition, Figure 1. A) Tumor volume change of HT-29 tumor bearing nude mice after treatment with IV-PS-PDT at three doses (5, 7.5, or 10 mg kg −1 , n = 5 each) and untreated control animals (n = 4); these treatments were significantly different from the untreated animals, p = 0.0636, p = 0.0210, p < 0.0001, respectively. B) Corresponding overall survival of HT-29 tumor bearing nude mice after treatment with IV-PS-PDT at various PS doses (n = 5 each) and untreated control animals (n = 4); curves are not significantly different p = 0.668. C) Tumor volume change of CT26 tumor bearing BALB/c mice after treatment with IV-PS-PDT at 7.5 mg kg −1 (n = 5) and untreated control animals (n = 5); *p < 0.05, ***p < 0.0005. D) Overall survival of CT26 tumor bearing BALB/c mice after treatment with IV-PS-PDT (n = 5) and untreated control animals (n = 5); curves are significantly different, p = 0.0023.
animals treated at all DLIs demonstrated a survival advantage as compared to the untreated control group (p < 0.0001, Figure 2B). Comparing the two best performing DLI groups (i.e., 3 h DLI and 1 h DLI) to the control group reveals that these two treatment regimens produce an identical survival benefit, with a 95.8% reduction in death (p = 0.002) for each comparison. One animal from the 3 h DLI group achieved a complete tumor response, defined as the absence of tumor upon dissection at the end of the 30 days follow-up period; no other animals from any group achieved a complete tumor ablation. This demonstrates that not only a single-day treatment protocol is feasible, but it may be superior to a treatment protocol that requires a prolonged DLI.

Investigation of PS-PDT via Peritumoral Administration
Photosensitizers for internal disease are traditionally delivered via an intravenous route of administration, which relies predominantly on the selectivity of the light administration to avoid offtarget effects. Previous attempts to treat colorectal cancer with IV-administered photosensitizers have resulted in a complication rate of at least 23% with the majority of these complications attributable to off-target effects resulting directly from intravenous administration of the photosensitizer. [6a] Peritumoral (PT) administration of photosensitizers can eliminate such effects and make a PDT treatment protocol more feasible. Colorectal cancer is particularly well-suited to peritumoral photodynamic therapy (PT-PDT) because such tumors are readily and routinely accessible endoscopically, and the surrounding areas are often tattooed for easy future localization. Optical fibers could also be easily delivered through many endoscopes/colonoscopes for the application of light irradiation for PDT.
To investigate peritumorally administered porphysome-based photodynamic therapy (PT-PS-PDT), BALB/c mice bearing syngeneic subcutaneous colorectal cancer tumors were treated with peritumorally administered PS (114 μg in 50 μL PBS) and received 135 J cm −2 of light irradiation at DLIs of 24, 3, or 1 h and compared with untreated controls. Tumors' growths were significantly different between groups (p < 0.0001), with the 3 h DLI group achieving the greatest degree of tumor ablation ( Figure 3A) but no significant differences in survival between the three treatment groups ( Figure 3B, p = 0.517). This establishes  that PT-administrated photosensitizer can produce an effective treatment response while providing a means of reducing the complication rate in future PDT treatment protocols.

PS-PDT Treatment of Orthotopic Human Colorectal Cancer in Rats
To assess the feasibility of light delivery and PDT to a clinically relevant (e.g., low rectal) tumor, an orthotopic human colorectal cancer (HT-29) model was produced in nude rats. These models were developed to grow human colorectal tumors in the rectal tissue of rats and to reproduce the technical challenges and real-world complications of delivering PDT to such tumors in humans. Initially, rats were treated with IV-PS-PDT (5 mg kg −1 , 67.5 J cm −1 , 24 h DLI) with intraluminal rectal light delivery via a bare fiber optic ( Figure 4A).
These animals initially demonstrated a significant, but tolerable response to therapy. Up until post-treatment day 10, tumor volume trended down and was significantly different from untreated control animals beginning at day 6 post-treatment (p < 0.05, Figure 4B). On post-treatment day 10, treated animals began to show signs of rectal perforation, namely large fecaliths and masses in the peri-rectal region ( Figure 4C). In addition, several of the animals demonstrated evidence of rectovaginal fistula when rectal irrigation was found to be draining through the vagina. Upon sacrifice and necropsy, three of the four animals were found to have developed rectovaginal fistulae; one of these was also found to have developed a posterior rectal perforation. Small tumors were found in the three animals that suffered perforation; upon dissection of the animal that had not suffered perforation, no tumor was found, indicating a complete response to treatment.
To limit the off-target effects of the PDT treatment, PT-PS-PDT was also performed by peritumoral administration of PS (218 μg in 100 μL) prior to intraluminal irradiation with 135 J cm −1 of light, and a DLI of 3 h. In addition, a hemispherically opaque sheath was used to shield non-tumor bearing rectal tissue from irradiation to further reduce off-target effects ( Figure 5A). This PT-PS-PDT treatment protocol produced an effective response ( Figure 5B). Treated animals demonstrated no significant adverse events; notably, no animals developed rectal perforation or rectovaginal fistula. The response to treatment appeared less significant, with a brief period of edema in the first 3 days, followed by signs of healing beginning on post-treatment day 6 ( Figure 5C).

Discussion
While PDT as a modality has been extensively studied for decades, its deployment in the treatment of human disease has been limited largely to palliative and second line (or later) Adv. Therap. 2023, 6, 2200342 Figure 4. A) Intra-rectal light delivery set-up using a bare fiber placed adjacent to the tumor intraluminally in rat model. B) Tumor growth curves after orthotopic HT-29 rat model received IV-PS-PDT (5 mg kg −1 of PS and light dose of 67.5 J cm −1 ) (n = 4) at a DLI of 24 h, compared with untreated control animals (n = 4). Values expressed as mean ± standard error of the mean. C) Representative photo series demonstrating the tumor regional response to IV-PS-PDT at i) immediately prior to treatment, day 0; ii) post-treatment day 1; iii) post-treatment day 3; iv) post-treatment day 6; v) post-treatment day 8; vi) post-treatment day 10; vii) post-treatment day 10 following sacrifice and fecal disimpaction; and viii) post-treatment day 10 following sacrifice and dissection, with a blunt probe demonstrating a rectovaginal fistula.
indications. The management of many cancers has recently moved toward a multi-modal approach, including rectal cancer. While historically treated primarily with surgical resection, rectal cancer is frequently treated with neoadjuvant (pre-operative) chemoradiotherapy as part of a multi-modal approach. PDT as a modality lacks the clinical evidence, provider experience, and materials and equipment to supplant any existing clinical treatment regimen for cancer; however, it could find clinical utility as part of a multi-modal approach, in conjunction with-for instance-chemotherapy, radiotherapy, and surgery for rectal cancer. The present study sought to develop a nanoparticlebased, clinically relevant approach to PDT that facilitates future translational work by answering critical questions about the safety, efficacy, and feasibility of PDT for rectal cancer.
This study is the first to establish the effectiveness of PS-PDT for colorectal cancer in two independent models-one being a human colorectal cancer that can establish effectiveness upon human colorectal cancer cells, and the other being a murine colorectal cancer that can establish effectiveness in a more relevant organismal context given the more intact immune system and active thymus gland of BALB/c mice.
While PS and PS-related nanoparticles (e.g., folate-PS) have been investigated as PDT photosensitizers for several malignancies including bronchogenic cancer, mesothelioma, and breast  4). Values expressed as mean ± standard error of the mean. C) Representative photo series demonstrating the regional response to PT-PS-PDT at i) immediately prior to treatment with tumor location outlined in black, day 0; ii) post-treatment day 1; iii) post-treatment day 3; iv) post-treatment day 6; v) post-treatment day 8; and vi) post-treatment day 10. cancer, this is the first study to confirm effectiveness in colorectal cancer. This is especially relevant considering recent studies highlighting the varied mechanisms of PDT resistance extant in different cancer cells lines, particularly colorectal cancer lines, that could disrupt the oxidative stress response that leads to apoptosis following PDT. [12] The existence of such resistance mechanisms means that the generalizability of PDT efficacy cannot be assumed across all cancer disease sites, or even between populations of colorectal cancer. Abrahamse et al. found that mechanisms of PDT resistance varied significantly between two human colorectal cancer cell lines in vitro, suggesting that (as seen in chemotherapy) some tumors will respond to PDT to a greater extent than others, depending on the tumor biology. [12b] Beyond the differences in cellular biology between previously studied cancers and colorectal cancer is the fact that this study is the first to investigate PS-PDT in a gastrointestinal malignancy of any kind. This is relevant from a practical and translational perspective given that luminal gastrointestinal malignancies would likely be among the most feasibly treated with PDT, given their relative ease of access through an endoscopic approach.
This study also established that PS nanoparticles can be effective photosensitizers at early time-points, with a peak in effectiveness at a DLI of 3 h. While this optimal DLI is shorter than expected with a nanoparticle-based photosensitizer driven by the enhanced permeability and retention effect, it is within the range of expectations for a cellular-targeted photosensitizer (hours to days). Conversely, this DLI is longer than would be expected for a vascular photosensitizer, most of which use DLIs within minutes of administration. [13] In addition, the fact that the 3 h DLI was more effective than the 1 h DLI suggests that the effects observed here cannot be solely explained by a vascular effect.
It has been reported that Verteporfin distributes into tumor tissue 3 h after administration by binding with low-density lipoproteins (LDLs) leading to cellular uptake via LDL receptor-mediated endocytosis. [14] It can then induce cellular-targeted photosensitivity at these time-points in addition to vascular-targeted photosensitizing effects due to the existence of remaining photosensitizer molecules in the vasculature. [13] The similarities here suggest that PS may be similarly acting through a combined cellularand vascular-targeting mechanism. This would explain why effective PDT is possible at both very short (1, 3 h) and very long (24, 48 h) time-points.
Regardless of the underlying mechanism, this study establishes that not only is nanoparticle-based PDT possible at early DLIs, but it may also be superior to longer DLIs. If this proves to be possible in human populations, patients may be able to undergo PDT under a "same day" visit, dramatically reducing the cost of the therapy (to both the patient and healthcare system), the burden of the therapy upon the patient, and improving the feasibility of PDT as a modality. In contrast, longer DLIs would necessitate either admission for an overnight stay, or a second hospital visit the following day.
This study also confirmed that the off-target effects of nanoparticle-based PDT can be mitigated through PT administration of the nanoparticle. This is particularly relevant for colorectal cancer as a luminal disease that is easily accessed endoscopically for the administration of both the photosensitizer and the light irradiation. [15] PT administration also enables a degree of tumor-specificity that cannot be achieved even by the EPR ef-fect or other nanoparticle-based delivery methods, given that a very high dose can be administered directly to the target tissue while avoiding systemic or regional distribution to a large extent. Thus, the investigation of PT administration is particularly clinically relevant in the management of colorectal cancer. This study found that PT administration results in statistically and clinically significant tumor ablation and control, and a survival benefit in treated versus untreated animals. This data is consistent with prior studies of PT and intra-tumoral photosensitizer injection for PDT that demonstrate the principle that this route of administration produces effective PDT. [16] The orthotopic rat model of human colorectal cancer used here was developed specifically to replicate the clinically challenging scenario of low rectal cancer that may threaten the anal sphincters. Using rats for this model enabled light delivery to be done endoluminally through transanal placement of the optical fiber. Though this model has drawbacks, it is the most relevant model in existence for studying this disease and treatment.
This study demonstrated that IV-PS-PDT could not be safely delivered; the rate of rectal perforation in the experimental group was 75%. While posterior rectal perforation might be expected given the posterior location of the rectal tumors, anterior rectal perforation (i.e., into the vagina, forming a fistula) is not. This may be explained by the accumulation of PS in the vaginal or rectal tissue to the point that these healthy tissues are photosensitized and tissue damage results from the administration of PDT. This hypothesis is supported by the circumferential edema, erythema, and sloughing seen when the rectal mucosa was examined at necropsy at short intervals after treatment. This in turn may be because the doses of PS or light used are excessive but may indicate that the biodistribution of the PS varies significantly between species. It may also suggest that the minimum therapeutic dose of PDT is very close to the maximum tolerable dose for healthy tissues; however, this would require specific experimental investigation to verify.
Ultimately, these rectal tumors were successfully treated with a protocol involving PT administration of the photosensitizer and use of an optical sheath that confined the light exposure to only areas of disease. PT-PS-PDT treated animals experienced more modest inflammatory responses and correspondingly modest decreases in tumor volume as compared with the IV-PS-PDT animals; however, it resulted in no off-target effects. Given the safety and tolerability of this PDT protocol, even a modest effect could be useful in the context of a multi-modality approach to neoadjuvant therapy including chemoradiotherapy, especially given that such a highly localized PDT treatment (unlike chemotherapy and radiotherapy) can be repeated as needed to achieve the desired effect (e.g., clearing the anal sphincter to avoid APR). A similar approach to that described here-using PT administration and directed irradiation-could be used in future clinical trials of colorectal PDT to limit both the local and distant off-target effects observed in prior studies of colorectal PDT.

Conclusion
This study sought to develop a clinically relevant and feasible approach to the administration of nanoparticle-based PDT to colorectal cancer. Using three complementary in vivo models, we established that nanoparticle-based PDT is effective in colorectal www.advancedsciencenews.com www.advtherap.com cancer, that it can be delivered with a short DLI (facilitating clinical deployment), and that previously documented off-target effects associated with colorectal PDT can be mitigated through PT-administration of the photosensitizer. We also established in a clinically relevant model, that rectal tumors can be feasibly and effectively treated with PT-administered photosensitizer and subsequent intraluminally directed light irradiation in a manner similar to what could be expected upon clinical deployment. These findings can be used to design future clinical trials and may support the incorporation of PDT as a modality alongside chemotherapy, radiotherapy, and surgery in the treatment of colorectal cancer.

Experimental Section
Cell Culture and Preparation: HT-29 human colorectal cancer cells (American Type Culture Collection, Manassas, VA, USA) were cultured in McCoy's 5A (MilliporeSigma, Burlington, MA, USA) media supplemented with fetal bovine serum (MilliporeSigma, Burlington, MA, USA) at a final concentration of 10% in sterile T-175 cell culture flasks (Corning Inc., Corning, NT, USA). Cells were maintained at an ambient temperature of 37°C, and an air mix of 95% air, 5% CO 2 . When cells reached confluence, they were subcultured when cell population expansion was required, or they were prepared for injection into animals to prepare tumor models. All cell handling was performed in a certified Biosafety Cabinet (Labconco Corporation, Kansas City, MO, USA). Cells were maintained and used between passage numbers 3 and 10.
CT26.WT murine colorectal cancer cells (American Type Culture Collection, Manassas, VA, USA) were cultured in RMPI-1640 (MilliporeSigma, Burlington, MA, USA) media supplemented with fetal bovine serum (Mil-liporeSigma, Burlington, MA, USA) at a final concentration of 10% in sterile T-175 cell culture flasks (Corning Inc., Corning, NT, USA). Cells were maintained at an ambient temperature of 37°C, and an air mix of 95% air, 5% CO 2 . When cells reached confluence, they were subcultured when cell population expansion was required, or they were prepared for injection into animals to prepare tumor models. All cell handling was performed in a certified Biosafety Cabinet (Labconco Corporation, Kansas City, MO, USA). Cells were maintained and used between passage numbers 3 and 10.
To subculture all cells, they were first incubated with 0.25% Trypsin-EDTA (Wisent Bioproducts, St. Bruno, QC, Canada) at 37°C for 5 min. They were then resuspended in complete medium, at a 1:5 ratio of Trypsin-EDTA to medium. The resulting mixture was then centrifuged at 1000 rpm for 5 min. Prior to centrifugation, 1 mL of this mixture was aliquoted into a 1.5 mL Eppendorf tube for cell counting. Cells were stained with 0.4% Trypan blue (Gibco-ThermoFisher Scientific, Waltham, MA, USA) and counted via light microscopy with stained cells aliquoted onto a hemacytometer (Hausser Scientific, Horsham, PA, USA). The cell pellet was then resuspended to achieve the desired cell concentration based upon the total cell count. Resuspended cells were plated into new T-175 cell culture flasks.
To prepare cells for injection into animal models, cells underwent incubation with Trypsin-EDTA, centrifugation, and counting, as in the subculturing protocol. The cell pellet was resuspended in a mixture of 1:1 phosphate buffered saline and Matrigel (Corning Inc., Corning, NY, USA) to achieve a cell concentration of 5 × 10 6 cells (for murine models with HT-29 injections) or 2 × 10 6 cells (for murine models with CT26 injections) per injection volume of 100 μL and kept on ice. 100 μL of this mixture was aspirated into 0.5 mL 30G insulin syringes (Becton, Dickinson and Company, Franklin Lakes, NJ, USA) and kept on ice until injection. To prepare cells for injection into orthotopic rectal rat models, cells underwent incubation with Trypsin-EDTA, centrifugation, and counting, as described. The cell pellet was resuspended in a mixture of 1:1 phosphate buffered saline and Matrigel (Corning Inc., Corning, NY, USA) to achieve a cell concentration of 1 × 10 7 cells per injection volume of 50 μL and kept on ice. 50 μL of this mixture was aspirated into 0.3 mL 30G insulin syringes (Becton, Dickinson and Company, Franklin Lakes, NJ, USA) and kept on ice until injection.
In Vivo Human Colorectal Cancer Subcutaneous Xenograft Model: All animal studies were conducted in compliance with institutional guidelines (University Health Network, Toronto, ON, Canada, Animal Use Protocols #4299 and #6278). Female Hsd:athymic nude Foxn1 nu mice (Envigo, Indianapolis, IN, USA) aged 6-8 weeks were obtained and maintained in the Animal Resource Centre at the Princess Margaret Cancer Research Tower (University Health Network, Toronto, ON, Canada). All mouse handling was performed in a certified Biosafety Cabinet. After a minimum of 1 week of acclimation time following delivery, animal anesthesia was induced and maintained using 2.5% inhaled isoflurane, and mice were laid in the left lateral decubitus position in the Biosafety Cabinet. The skin over the right shoulder blade was cleaned with 70% isopropanol and was tented using forceps. Previously prepared HT-29 cell/Matrigel suspensions (5 × 10 6 cells in 100 μL) were injected in the subcutaneous space over the right shoulder blade to form a bleb. Animals were then recovered from anesthesia and returned to standard maintenance conditions. Animals were monitored with body weight and tumor measurements taken three times per week. All animal studies were conducted in compliance with institutional guidelines (University Health Network, Toronto, ON, Canada).
In Vivo Subcutaneous Syngeneic Colorectal Cancer Model: Female BALB/c (AnNHsd) mice (Envigo, Indianapolis, IN, USA) aged 6-8 weeks were obtained and maintained in the Animal Resource Centre at the Princess Margaret Cancer Research Tower (University Health Network, Toronto, ON, Canada). All mouse handling was performed in a certified Biosafety Cabinet. After a minimum of 1 week of acclimation time following delivery, animal anesthesia was induced and maintained using 2.5% inhaled isoflurane, and mice were laid in the left lateral decubitus position in the Biosafety Cabinet. The fur overlaying the right shoulder blade was clipped short using electric clippers, then depilatory cream was applied to the region and left in situ for 45-60 s. The cream was then removed with gauze. Application of depilatory cream was repeated until all visible fur was removed from the region. Once all fur was removed, the skin was cleaned with 70% isopropanol and was tented using forceps. Previously prepared CT-26 cell/Matrigel suspensions (2 × 10 6 cells in 100 μL) were injected in the subcutaneous space over the right shoulder blade to form a bleb. Animals were then recovered from anesthesia and returned to standard maintenance conditions. Animals were monitored with body weight and tumor measurements taken three times per week.
In Vivo Orthotopic Human Rectal Cancer Model: Female RNU nude rats (Envigo, Indianapolis, IN, USA) aged 6-8 weeks were obtained and maintained in the Animal Resource Centre at the Princess Margaret Cancer Research Tower (University Health Network, Toronto, ON, Canada). After a minimum of 1 week of acclimation time following delivery, animal anesthesia was induced and maintained using 2.5% inhaled isoflurane, and rats were laid in a supine position. The anus was dilated using a blunt probe and any stool in the rectal vault was removed using forceps and/or using a flush of sterile saline directed through a soft catheter. An adjustable speculum was placed in the anal canal while an assistant retracted the anterior rectal wall anteriorly using forceps and the posterior rectal mucosa was gently grasped with forceps and prolapsed through the anus. Previously prepared HT-29 cell/Matrigel suspension (1 × 10 7 cells in 50 μL) was injected submucosally into the posterior wall of the rectum. Animals were then recovered from anesthesia and returned to standard maintenance conditions. Animals were monitored with body weight and tumor measurements taken three times per week. All animal studies were conducted in compliance with institutional guidelines (University Health Network, Toronto, ON, Canada).
Porphysome and Light Administration: Porphysomes were prepared according to previously published protocols. [8] Following randomization, porphysome was administered intravenously by bolus tail vein injection, or peritumorally by guiding the tip of the needle into the tumor and injecting the photosensitizer in a fan shaped distribution with a back-andforth motion to ensure homogeneous distribution of the drug throughout the tumor. Appropriate delivery was visually confirmed by assessing for green-black subcutaneous discoloration throughout the tumor. PS was administered at various doses per experimental design using weightbased dosing; each dose was diluted with phosphate-buffered saline to a total volume of 200 μL per dose for murine intravenous administration, or to a total volume of 3 mL per dose for rat intravenous administration. For peritumoral administration in mice, stock solutions were diluted with phosphate-buffered saline to a total volume of 50 μL per dose such that each dose contained 112.5 nmol of porphysome. For peritumoral administration in rats, stock solutions were diluted with phosphate-buffered saline to a total volume of 100 μL per dose such that each dose contained 225 nmol of porphysome.
A diode laser with a wavelength of 671 nm was used for irradiation (Laserglow Technologies, Toronto, ON, Canada). For murine models, under general anesthesia (2-3% inhaled isoflurane) and with minimal ambient lighting, the treatment light was delivered via an optical fiber as a 10 mm diameter circular spot centered on the tumor. The energy density was measured prior to each treatment using a calibrated power meter (Thorlabs Inc., Newton, NJ, USA). A total light dose of 135 J cm −2 was delivered at 100 mW cm −2 over 1350 s (22.5 min). Animals were protected from scattered light using an opaque mask with a 10 mm diameter hole.
For rat models, under general anesthesia (2-3% inhaled isoflurane), the anus of each animal was dilated with a blunt dissecting probe and any fecal matter was removed from the rectal vault with forceps or saline irrigation. Light was delivered via transanal placement of an intraluminal fiber optic with a 1 cm long cylindrical diffusing tip. For early pilot trials, this fiber optic was placed directly into the rectal vault, adjacent to the tumor for treatment; in later studies, the fiber optic was placed into a custommade fiber optic sheath which was placed adjacent to the tumor. The laser source was a 671 nm diode laser (Laserglow Technologies, Toronto, ON, Canada). The energy density was measured prior to each treatment using a calibrated power meter (Thorlabs Inc., Newton, NJ, USA) and a spherical integrating sensor (Thorlabs Inc., Newton, NJ, USA). When the fiber sheath was used, the fiber optic was placed into the sheath and the entire sheath was introduced into the spherical integrating sensor for calibration of energy density. Light doses were administered at 67.5 or 135 J cm −1 (of fiber optic diffuser-in this study, fiber optic diffuser was 1 cm) at a power of 50 or 100 mW cm −1 over 1350 s (22.5 min).
Study Metrics: For mice, following treatment, animals were followed for 30 days, and the tumor size was measured using digital vernier calipers three times per week. Photographs of tumors were taken, and the animals were weighed on each occasion. Humane sacrifice by anesthetic and CO 2 overdose was carried out after 30 days or if tumors reached 500 mm 3 in volume, whichever came first. Animals were considered to have reached a "mortality" endpoint if they required sacrifice because their tumor volume was >500 mm 3 . For rats, following treatment, animals were followed for 30 days, and the tumor size was measured using digital vernier calipers three times per week. Photographs of tumors were taken, and the animals were weighed on each occasion. Humane sacrifice by anesthetic and CO 2 overdose was carried out after 30 days or if animals suffered from therapy related complications that affected their systemic well-being. All therapy related complications were collected as qualitative metrics.
PDT Study Design: To investigate the effectiveness of PS-PDT in an HT-29 human colorectal cancer murine model, animals were enrolled when HT-29 tumors measured 5 mm in maximum diameter. They were then randomized to one of three treatment groups or a control group. PS was administered at one of three drug doses-5, 7.5, or 10 mg of pyropheophorbide per kg of animal body weight (n = 5 per group)-to tumor bearing mice, followed by light exposure (135 J cm −2 , 100 mW cm −2 , 22.5 min) to the tumor after a DLI of 24 h. A fourth, no treatment control group (n = 4) received no treatment or intervention. Based on the results of prior studies demonstrating the equivalence of the no treatment control group with drug only and light only control groups, and given the ethical consequences of including all possible negative control groups for all studies, the only negative control group used was the no treatment control group.
To investigate the effectiveness of PS-PDT in a CT26 murine colorectal cancer model across multiple DLIs, animals were enrolled when CT26 tumors measured 3.5 mm in maximum diameter. They were then randomized to one of four treatment groups, or a fifth, untreated control group (n = 5). PS was administered at the optimal dose as identified in the HT-29 trial, and PDT was delivered after (a DLI of) either 1, 3, 6, or 24 h (n = 5 per group). For all animals, the duration between enrollment and treatment was kept constant (i.e., 24 h) to ensure that tumor sizes at the time of treatment were comparable.
For rat studies, animals were enrolled into experimental groups between 12 and 15 days following tumor cell injection. For the intravenous administration protocol, rats were block-randomized (by tumor volume) to receive either treatment with IV-PS-PDT (5 mg kg −1 , 67.5 J cm −1 , n = 4) or no treatment (n = 4). For the peritumoral administration protocol, rats received treatment with PT-PS-PDT (225 nmol, 135 J cm −1 , n = 5).
Statistics Analysis: Complete response was defined as the absence of detectable tumor upon dissection at the end of the 30 days follow-up period (i.e., cure). Differences in relative tumor size between three or more groups were analyzed using one-way repeated measures ANOVA with Tukey adjustment for multiple comparisons when multiple groups were compared and using Dunnet's correction when groups were compared to a control group. Differences in tumor size between two groups were compared using multiple unpaired t-tests with Welch correction. Log-rank (Mantel-Cox) tests were conducted for comparison of survival curves and cure rates; hazard ratios were calculated using Mantel-Haenszel tests. Statistical analysis was performed using GraphPad Prism software; p < 0.05 were considered significant.