Subcutaneous Xenograft Models for Studying PDT in vivo

Abstract

The most facile, reproducible and robust in vivo models for evaluating the anti-cancer efficacy of photodynamic therapy (PDT) are subcutaneous xenograft models of human tumors. The accessibility and practicality of light irradiation protocols for treating subcutaneous xenograft models also increases their value as relatively rapid tools to expedite the testing of novel photosensitizers, respective formulations and treatment regimens for PDT. This chapter will summarize the methods used in the literature to prepare various types of subcutaneous xenograft models of human cancers and syngeneic models to explore the role of PDT in immunooncology. This chapter will also summarize the PDT treatment protocols tested on the subcutaneous models, and the procedures used to evaluate efficacy at the molecular, macromolecular and host organism level.

1. Introduction

For decades, research into photodynamic therapy (PDT) has used mostly subcutaneous xenograft models of human cancer in rodents, such as mice and rats, and has helped the approval of a dozen photosensitizers (PSs) globally with more than 500 active PDT clinical trials.[1-4] Subcutaneous xenograft models residing in an anatomical site that allows for both facile photoirradiation and longitudinal monitoring of tumor volume, perfusion and oxygenation. Thus, these models have understandably been the first option when translocating from in vitro testing to in vivo evaluation of antitumor PDT efficacy.[5, 6] In this chapter, methods of generating and utilizing subcutaneous xenograft models of cancer as tools for evaluating PDT efficacy will be detailed and discussed. These models include human cell line xenografts, syngeneic xenografts, immunocompetent xenografts and patient derived xenografts. Each type of model has been leveraged for a specific purpose and can often be matched to the implications of the resultant findings. Challenges that subcutaneous xenograft models face will also be discussed, along with emerging strategies described in the literature used to address them.

2. Methods

In this section, the methods used for a number of subcutaneous tumor models that have been reported as testing platforms for PDT will be discussed. These include: Implantation of human cell line xenografts (section 2.1), Implantation of syngeneic cell line xenografts (section 2.2) and implantation of patient-derived xenografts (PDX) (section 2.3). Subsequently, we will briefly discuss various PDT treatment strategies (section 2.4). The main advantages of using human cell line xenografts include the flexibility to rapidly evaluate treatment regimens in a broad variety of disease indications from human origin; however, the main disadvantage is the inability to assess the immunological components of therapies. Conversely, the strength of using syngeneic cell line xenografts is their ability to recapitulate secondary responses to PDT regimens mediated by the innate and adaptive immune system. The limited number of established syngeneic cancer cell lines available for testing is a major limitation, in addition to the absence of human molecular targets and cellular sensitivities that are not directly comparable with human disease. PDX tumor models recapitulate the molecular, cellular and structural heterogeneity of patient tumors; however, they suffer from rapid partner cell turnover with the host species and, like human cell line xenografts, can only be tested in immunocompromised host rodent species. The details of how these models are used to evaluate PDT-based regimens will be discussed as follows.

2.1. Implantation of Human Cell Line Xenografts

Owing to the relative simplicity of generating subcutaneous primary cell line tumor models of human cancers, they are the most common type of tumor model to be leveraged for assessing PDT efficacy, in addition to non-PDT treatments also described historically in the literature. This simplicity is also the reason why human cell line xenografts have also been used to measure the PDT efficacy of a broad variety of cancer indications including tumors of the pancreas,[7-9] the brain,[5] [10, 11] and ovaries [12, 13], amongst others. In general, the methods for generating tumors from human cell lines are relatively uniform throughout the literature.

Human cell lines are cultured to near confluence in 2D monolayer, mostly in fetal bovine serum-containing media. In order to implant them to develop into tumors, they are detached from culture, typically using trypsin and concentrated to the required density for subcutaneous injections. As trypsin can rapidly cleave cell surface proteins that are used by the implanted cells to attached to the host species tissue, trypsin-free gentle cell detachment reagents, such as Dispase^®, can be recommended as less perturbative alternatives.[14] The concentrated cell suspensions are then prepared for implantation in PBS, culture media, or combinations of PBS or culture media with Matrigel to enable the stable embedding of the cells under the skin and to prevent leakage. Matrigel has been shown to have no impact on tumor development or on tumor cell metabolism and can be recommended for any tumor implantation protocols to prevent tumor cell loss and to provide consistent tumor morphologies. However, being naturally derived, it can vary in composition which might make it susceptible to batch-to-batch variability.[15] A simpler matrix consisting of Type I Collagen has been shown to be a suitable alternative to Matrigel that allows for effective and reproducible tumor implantation.[16] An example of Matrigel used for implantation is reported in a study by Spring et al. who reported the development of both a subcutaneous and an orthotopic pancreatic ductal adenocarcinoma (PDAC) model using the metastatic AsPC1 cell line to assess the efficacy of a combinatorial PDT-based nanoconstruct. The subcutaneous PDAC model was generated using 6-week-old Swiss nude mice implanted with 1 x 10⁶ AsPC-1 cells per mouse on the hind leg. To aid in the formation of the subcutaneous tumors, cells were implanted in a 50/50 mixture of Matrigel and serum-containing DMEM media, injected in a total volume of 50 μl. Similarly, Mallidi et al. used female Swiss nude mice (6-8 weeks old) and subcutaneously implanted on 3 x 10⁶ U87 human glioblastoma cells suspended in 300 μl in 100% growth factor-reduced Matrigel on the mice backs.[5] Such processes for preparing human cell line xenografts using Matrigel to study PDT are thoroughly reported in the literature for various cancer models including lung cancer,[17] [18] and cervical cancer,[19] amongst others. Thus, for preliminary screening of in vivo PDT efficacy, human cell line-based xenograft tumors can be recommended as the model of choice for simplified and rapid testing. The relatively large number of publications of PDT using human cell line-based xenograft tumors may also serve as a point of reference to help evaluate findings in the context of those already published using the same tumors.

2.2. Implantation of Syngeneic Cell Line Xenografts

Syngeneic tumor models are generated using cancer cell lines implanted into the same species as the host recipients, which are most often mice or rats in studies involving PDT.[20] The major advantage of using syngeneic models is to recapitulate and preserve immunological responses to anticancer therapies, as a result of the immune-compatibility of tumor cells of the same species and the resulting capability to use immunocompetent strains of host animals.[20] However, the degree of relevance of non-human tumors that are designed to model PDT treatment response in human diseases can be questionable. This concern becomes particularly significant when evaluating PDT-based targeted therapies where molecular targets and epitopes require sequence homologies closely matching those of humans to accurately assess outcomes. The majority of the PDT studies in the following methods descriptions for syngeneic tumor implantations are designed to monitor intrinsic host tissue responses to therapy, the most important being anti-tumor immune responses.

Tong et al. used Balb/c mice implanted subcutaneously with the metastatic 4T1 murine breast adenocarcinoma cells to evaluate the efficacy of benzoporphyrin derivative monoacid ring A (BPD-MA) in combination with Adriamycin.[21] Cell suspensions (200 μl in what is assumed to be culture media) containing 8 x 10⁶ 4T1 cells were implanted subcutaneously into the right flank of Balb/c mice until tumors reached 100-150 mm³ in volume, as determined by caliper measurements of tumor dimensions. In this study, no specific motivation to use a syngeneic model was mentioned, although the methods described here can be used in future studies involving the analysis of host responses, molecular or immunological, that are otherwise unaccounted for when using human cell line or tissue model xenografts.

In another study using a syngeneic model of squamous cell carcinoma, the efficacy of PDT using Foscan was evaluated in combination with a ceramide analogue. [22] In this study, the authors required an intact immune system that is provided by a syngeneic model to evaluate the multifaceted in vivo response to anti-tumor PDT. Given that ceramide signaling is known to modulate the immune system, a syngeneic model was particularly important for evaluating this specific combination regimen.[23] Female syngeneic C3H/HeN mice were implanted with 1 x 10⁶ SCCVII squamous carcinoma cells subcutaneously in the lower back and were allowed to grow for 7-10 days in order to reach 6-8 mm in diameter. The study found that combining PDT with the ceramide analogue improved survival in the syngeneic model and that intrinsic ceramides served as predictive biomarkers for tumor responses to PDT.

In addition to mice, rats have also been used to generate syngeneic tumors for the investigation of PDT based treatment regimens. Owing to the larger size of rats, more complex and thorough PDT procedures are possible, such as interstitial illumination and real-time dosimetry. An example of such a study was published by the group of Ronald Moore, who generated subcutaneous R3327-H prostate tumors and AY-27 bladder tumors in syngeneic Fischer CDF344 male rats.[24] R3327-H tumors and AY-27 tumors were initially passaged subcutaneously in donor rats that are a cross between Fischer and Copenhagen breeds. Before the tumors grew to a maximum of 5000 mm³ in the donor rats, they were excised and cut into 3 mm chunks in sterile Hank’s balanced salts solution. The R3327-H tumors and AY-27 tumors were then implanted subcutaneously into the flanks of the recipient Fischer CDF344 male rats and were allowed to grow for 16-20 weeks or 10-14 days, respectively. Following ALA-PDT, tumor dimensions were monitored using calipers and the volumes were calculated using the ellipsoid formula.

Syngeneic models have also been used to explore the immunological role of PDT. A seminal report by Korbelik and Dougherty demonstrated that splenocytes from PDT-sensitized or X-ray sensitized immunocompetent mice bearing syngeneic subcutaneous tumor models were able to enhance the antitumor efficacy of PDT when injected into immunodeficient mice with syngeneic subcutaneous tumor models of the same tumor type. [25] Syngeneic EMT-6 mammary sarcoma cells and Meth-A fibrosarcoma cells (1 x 10⁶ cells) were implanted subcutaneously in the lower dorsum of 7-9-week-old Balb/cJ mice (immunocompetent) and Balb/cJ-scid.TO (severe combined immunodeficient) female mice were also implanted in the same manner. Splenocytes were isolated from spleens by mild mechanical separation of the tissue, selective lysis of erythrocytes using ammonium chloride and filtration through 50 mm polyester membranes. Two days prior to subcutaneous tumor implantation of the recipient Balb/cJ-scid.TO immunodeficient mice, the freshly prepared splenocytes were injected into the tail veins of the recipient mice at a density of 1-2 x10⁷ cells/mouse. The donor mice were then subject to Photofrin-PDT as described before. The authors found that adoptive transfer of splenocytes from X-ray sensitized tumor bearing donor mice had a significant protective effect on the regrowth of the subcutaneous tumors following PDT. Even more so, the adoptive transfer of splenocytes from PDT sensitized EMT-6 tumor bearing donor mice resulted in a complete cure of EMT-6 tumor bearing recipient mice. This effect was supported by the finding that adoptive transfer of splenocytes from PDT sensitized Meth-A tumor bearing donor mice had no inhibitory effect on the growth rate on the EMT-6 tumor bearing recipient mice. The strength of this protocol lies in its ability to accurately probe which leukocytes are responsible or play a major contributing role in the immunological destruction of residual tumor cells following PDT.

Syngeneic murine xenograft models bearing two contralateral subcutaneous tumors have proven to be a powerful tool to explore the impact of PDT on a phenomenon known more widely in oncology as the abscopal effect. The abscopal effect occurs when distant tumors that represent metastatic disease are eradicated by the adaptive immune system following the localized focal treatment of the primary neoplasm.[26] Examples of such focal treatments include radiotherapy [27], thermal ablation [28] and PDT[29]. By confining light irradiation to one of two contralateral tumors, the immunological effects of PDT can be monitored relatively simply with a degree of control that is not observed with protocols requiring the adoptive transfer of immune cells. An important study by Mroz et al. demonstrated that the contralateral subcutaneous re-challenge of mice with colon adenocarcinomas that had been initially treated with PDT 90 days prior was rejected in more than 95% of mice tested.[30] Importantly, contralateral tumor re-challenge was only rejected following PDT of mice implanted with CT26.CL25 tumors expressing the immunogenic tumor antigen β-galactosidase. Wild type CT26 (CT26WT) colon adenocarcinoma cells or CT26.CL25 cells transfected with the lacZ gene encoding the model tumor associated antigen β-galactosidase, were implanted into immunocompetent Balb/c mice subcutaneously on the right thigh at a density of 3.5 x 10⁵ cells per mouse. Vernier calipers were used to monitor the tumor dimensions until the diameters reached 5-7mm, approximately nine days following implantation. 90 days following PDT with BPD, surviving mice were re-challenged with 3.5 x 10⁵ CT26TW or CT26.CL25 cells subcutaneously implanted in the contralateral left thigh and the tumor volumes were monitored for another 60 days.

A recent study by Lu et al. leveraged contralateral subcutaneous syngeneic xenograft models to investigate the immunomodulatory effects of PDT using a nanoconstruct containing a chlorin-based PS, in synergy with an immune checkpoint inhibitor of indoleamine 2,3-dioxygenase (IDO).[31] PDT of the primary tumor in combination with IDO inhibition lead to tumor antigen presentation and T cell proliferation, that lead to the destruction of the contralateral tumor. Contralateral tumor models of MC38 and CT26 colon adenocarcinomas were generated in C57BL/6 mice and Balb/c mice, respectively, by implanting 2 x 10⁵ cells into the left flank and 1 x 10⁶ cells into the right flank of each mouse.

Due to the simplicity of generating syngeneic tumor xenografts that can develop in the presence of a functional immune system, syngeneic models are recommended as the model of choice for assessing the immunological aspects of PDT treatment response. Weaknesses associated with the subcutaneous physiological location of these syngeneic tumor xenografts are outweighed by the unique ability to monitor contralateral untreated tumors that helps predict the capacity of PDT treatment to control metastatic and recurrent disease in patients.

2.3. Implanting Patient-Derived Xenografts (PDX)

The complex and heterogeneous patient-derived xenograft (PDX) models of cancer are a powerful testing platform due to their increased physiological relevance and stronger predictive power for human tumor response to anticancer treatment regimens.[32] In addition, the fact that they are passaged through multiple generations of immunodeficient mice avoids the non-physiological selective pressures that tumor tissues experience under ex vivo and in vitro tissue culture. PDX tumor models have also been leveraged to test the efficacy of PDT regimens when implanted subcutaneously in mice. The strain of mice used for PDX implantation are most commonly immunocompromised, such as the athymic Swiss nude strain, to prevent immunogenicity and rejection of the xenograft. [32] However, the innate immune system present in the Swiss nude mice can prove to be problematic and an immunodeficient SCID strain is preferred, which lacks T cells and B cells.[33] SCID mice can oftentimes produce T cells and B cells as they age, therefore Non-Obese Diabetic (NOD) – SCID mice are sometimes preferred due their absence of T cells and B cells and dysfunctional natural killer cells.[34, 35] The absence of fully functional immune systems in Swiss nude mice, SCID and NOD-SCID mice can limit evaluation of PDT to only direct anti-PDX tumor effects, without any impact of the innate and adaptive immune systems.

Nwogu et al. investigated the efficacy of a newly reported photosensitizer, termed PS1, in a murine PDX model of non-small cell lung cancer (NSCLC) and compared it to PDT using the approved PS Photofrin.[36] The subcutaneous PDX model of was developed using tumor tissue that was surgically resected from 85 NSCLC patients from 2000-2010, that were histologically verified as primary tumors disease. To add to the heterogeneity of the model, tumor tissue of differing cellular origin and staging were used. The resected patient tumor tissue was cut to 2x2 mm dimension and were immediately passaged in SCID mice. For the PDT experiments, second passage PDX tumor tissue from the first generation SCID mice were cut to the same 2x2 mm dimensions and implanted subcutaneously onto the abdominal wall of four groups of SCID mice. When the tumor diameters reached 4-6mm, PDT using Photofrin was performed.

A recent study by Lin et al. explored the “trimodal” therapeutic effects of a nanoconstruct integrating PDT using a pyropheophorbide A derivative, photothermal therapy and doxorubicin chemotherapy in PDX models for bladder cancer obtained from Jackson Laboratories, namely BL269, BL440, BL645, and BL293.[37] The PDX models were passaged subcutaneously on the flank in NOD SCID gamma (NSG) mice and implanted orthotopically on the wall of the bladder, or subcutaneously on the flank. The nanoconstruct, termed nanoporphyrin, was administered intravesically into the bladder 2 days following orthotopic implantation for subsequent PDT.

Although PDX models boast of a genomic, cellular and structural fidelity to the patient tumor, the inconsistency and dynamism of cell type proportions can result in inconsistencies in treatment response between in vivo passages. It has been shown that PDX tumor stromal cells, such as fibroblasts, are rapidly replaced by host species fibroblast, and thus the advantage of true tumor heterogeneity is dampened by dynamic heterogeneity in tumor development as a function of time when transplanted into mice. [38]

2.4. PDT Treatment Regimens

2.4.1. Photosensitizer administration routes

For the majority of PDT studies using subcutaneous xenograft tumor models, the treatment protocol follows a similar pattern, whereby the intravenous administration of a PS or a sensitizer formulated into a solubilizing vehicle is ensued by discrete PS-light intervals. PSs and their respective formulations have been administered intravenously through tail vein injections [5, 7, 21, 24, 25], intraperitoneally [22] and intratumorally [31]. Intravenous tail vein administration is the most common route used in pre-clinical in vivo studies using subcutaneous xenograft models for PDT, whereby the pharmacokinetics of common PSs is well studied and well understood. Implications on the effects of certain PS-light intervals are also the most well understood for intravenous administration, although this will be discussed in greater detail in the following section. Intraperitoneal administration is less common, and the pharmacokinetics of tumor delivery are not as well understood. Thus, future studies using intraperitoneal administration of photosensitizing agents would likely require a separate set of biodistribution experiments, either longitudinally using fluorescence imaging of the PS molecules or terminally using tissue extraction protocols and subsequent analysis. Intratumoral administration bears little relevance to the clinical application of the PDT regimens being tested. However, it serves a very specific purpose when the goal is to confine the sensitizer and its concomitant combination agents to one point of focus to evaluate secondary biological effects in distant sites such as contralateral tumors.

2.4.2. Selection of PS-light interval

PDT, in addition to other activatable therapies with a narrow radius of damage, provides a unique level of precision for differentially targeting the tumor endothelium or parenchyma. The duration of PS-light intervals is dictated by the option of inducing vascular PDT (short PS-light interval, ca. 15 minutes), direct tumor cell phototoxicity (long PS-light interval, ca. 3 hours or more), or a combination of both (intermediate PS-light interval; ca. 90 minutes.).[39] These intervals of 15 mins, 90 mins, and 3 hours or more is somewhat limited to amphiphilic PS molecules administered as free molecules or weakly associated with inert vehicles. PS-light intervals are different for photosensitizing entities with differing pharmacokinetics and pharmacodynamics, such as strongly hydrophilic sensitizers with ambiguous intracellular uptake rates and stable nanocarriers with longer circulation half-lives. For such photosensitizing entities, separate experiments are required to individually determine the time-dependent tumor compartment localization. Such studies are time consuming and technically challenging, hence more commonly with new photosensitizing entities, investigators either establish the time of maximal tumor uptake as the time for irradiation, or select arbitrary time-points, such as 24 hours.

To exemplify the impact of the drug-light interval on therapeutic efficacy, PDT has been performed in subcutaneous U87 glioblastoma xenografts using 0.5 mg/kg of BPD formulated in a liposome that mimics the clinical formulation, Visudyne.[5] The BPD formulation was injected intravenously via a tail vein injection, with a comparison of either a 1-hour or 3-hour PS-light interval. The subcutaneous U87 tumors were irradiated with an Intense 690 nm diode laser light at an irradiance of 100 mW/cm² delivering a total fluence of 100 J/cm2. The authors found that the 3-hour PS-light interval was less tumoricidal than the 1-hour PS-light interval for vascular PDT, and they found that tumoricidal efficacy correlated with tumor oxygenation.

This unique spatiotemporal selectivity of tumor phototoxicity by tuning the PS-light interval is also only specific to parenterally administered exogenous photosensitizing entities, and thus alternative PDT treatments, such as ALA-based therapies are further complicated by a rate limiting endogenous metabolic conversion into the active PS protoporphyrin IX.

2.4.3. Light application

The second determinant of PDT dosimetry is the light application to activate the sensitizers and control the localization and extent of RMS generation. PDT light dosimetry is most often described in the literature two parameters: power density, also known as irradiance, expressed as W/cm² and fluence expressed as J/cm². The relationship between energy, power and time is described in Equation 1. Equation 1 is used to determine the irradiation time and irradiance needed to achieve certain fluences of light and the light application protocols described in the literature for PDT regimens in subcutaneous xenograft models will be discussed in this section. Although, fluence is the principle metric of PDT dosimetry, it is becoming increasingly evident that irradiance must be regulated to optimize outcomes, with numerous reports claiming that the same fluence administered to tumors is more efficacious when delivered using lower irradiances.[40, 41]

$$Energy(J)=Power(W)\times Time(s)$$ Equation 1:

In the remainder of this section, examples of established PDT protocols for a variety of photosensitizers and photosensitization strategies are provided.

In a previous study by our group, a liposomal nanoconstruct was prepared containing the photosensitizer BPD in the membrane and a polymeric PEG-PLGA nanoparticle encapsulated in the core, which entrapped the anti-metastatic and anti-angiogenic small molecule inhibitor, XL184.[7] The combination nanoconstruct was referred to as photoactivable multi-inhibitor nanoliposome (PMIL). Treatment in this subcutaneous AsPC-1 model was initiated following intravenous administration of 0.25 mg/kg BPD equivalent and 0.1 mg/kg XL184 equivalent contained in the PMIL construct. The subcutaneous tumors were irradiated 60 minutes following intravenous administration using a 690 nm diode laser delivering a light fluence of 75 J/cm² at an irradiance of 100 mW/cm². The study concluded that only with the combinatorial nanoconstruct, PDT treatment reduced tumor volumes down to 10% of those in untreated control mice and less than 40% in those mice treated with the monotherapies.

Similarly, in a subcutaneous 4T1 breast tumor model, 5 mg/kg Adriamycin was administered intravenously through tail vein injections four days after the tumors reached their required volume of 100-150 mm³. The following day, 1 mg/kg of BPD was administered intravenously, and the tumors were irradiated with 120 J/cm² of 690 nm laser light 24 h later. [21] It was found that the PDT-Adriamycin combination therapy was just under two-fold more potent at controlling tumor growth in vivo, as compared to PDT alone, and prolonged mice survival to the same degree.

In a separate study using SCCVII subcutaneous xenograft model for squamous cell carcinoma, intraperitoneal injection of Foscan formulations was also explored. [22] Foscan was formulated in a 2/3/5, v/v combination of ethanol/polyethyleneglycol400/water and was administered intraperitoneally at a dose of 0.1 mg/kg. Following a 24-hour duration, the subcutaneous SCCVII tumors were irradiated with 650 nm light at a fluence of 50 J/cm² and an irradiance of 80-90 mW/cm². The light was generated using a Sciencetech FBQTH high throughput illuminator. For the mice receiving the combination treatment with a ceramide analogue, C6-pyridinium ceramide dissolved in water was injected intraperitoneally at a dose of 80 mg/kg either 24 hours prior to PDT or immediately after. Up to 90 days following PDT, mouse survival was only ca. 35%, whereas in combination with the ceramide analogue, survival was ca. 80%.

Subcutaneous R3327-H prostate tumors and AY-27 bladder tumors in Fischer CDF344 rats have also been treated with PDT after the tumors reached 1000 mm³ in volume. [24] Intravenous administration of 500 mg/kg of ALA dissolved in PBS by tail vein injection in the rats was performed four hours prior to interstitial photoirradiation of the subcutaneous tumors. The tumors were irradiated with 630 nm light delivered by a tunable Coherent CR-599 argon-pumped dye laser through a beam equally split into 8 quartz fibers with terminal diffusing tips that were placed interstitially into the tumors, equally spaced 7 mm apart in a hexagonal, icosahedral pattern. The wavelength and power output of light set to 80 mW for each individual fiber was measured using a LaserTherapeutics power meter and the eighth fiber was kept attached to the power meter for real time monitoring of light dosimetry during the irradiation protocol. Four hours after injection of ALA, the tumors were irradiated with light doses ranging from 1000 – 3000 J. The authors showed that with a 3000 J treatment of PDT, all subcutaneous R3327-H prostate tumors and AY-27 bladder tumors were cured and only mild reductions in tumor perfusion were recorded, as is consistent with tumor cell targeted PDT regimens like ALA-PDT.

In an immune competent mouse model, PDT was performed on the Balb/cJ mice six days following implantation of EMT-6 and Meth-A tumors in groups of 8-10 mice per arm. [25] Briefly, 10 mg/kg of Photofrin was administered intravenously and the tumors were irradiated 24 h later using 630 nm light (A5000 unit with a 1-kW xenon bulb, Photon Technology International, Inc.) at an irradiance of 120-130 mW/cm² and a fluence of either 110 J/cm² (EMT-6) or 150 J/cm² (Meth-A tumors). A 5 mm core diameter liquid light guide (2000A, Luminex) was also used to apply the light. The primary observation was tumor regrowth, as evaluated by visual inspection thrice weekly and mice with no evidence of tumor regrowth for 90 days were considered complete cures. In another experiment within the same study, PDT was compared with 35 Gy of X-ray radiotherapy at 3.33 Gy/min using a Philips RT250 radiotherapy system (250 kVp, 0.5 mm Cu). The study concluded that PDT of the subcutaneous tumors was able to generate tumor-specific sensitized immune cells that provided a prolonged anti-tumor immunity.

In another immune competent model of colon cancer, vascular PDT was performed on subcutaneous CT26WT tumors or CT26.CL25 tumors expressing the immunogenic tumor antigen β-galactosidase by intravenous injection of 1 mg/kg of BPD dissolved in 5% dextrose. [30] The tumors were irradiated using a 1W 690 nm diode laser (B&W Tek Inc., Newark) at 120 J/cm² fluence and an irradiance of 100 mW/cm², with a PS-light interval of 15 minutes. Interestingly, 70% of PDT treated CT26.CL25 tumor bearing mice re-challenged with contralateral subcutaneous CT26.CL25 cells rejected tumor growth and the remaining 30% of the mice were found to escape immune destruction because of a loss of expression of the immunogenic β-galactosidase.

PS constructs have also been directly injected intratumorally for PDT-based combination regimens. Mice bearing primary right flank tumors at a volume of 100 mm³ were treated with a PDT-indoleamine 2,3-dioxygenase (IDO) inhibitor immunotherapy combination regimen at a chlorin-based photosensitizing ligand dose of 20 μmol/kg injected intratumorally. [31] Following a PS-light interval of 12 hours, the primary right flank tumor was irradiated with 650 nm light at a fluence of 90 J/cm² and an irradiance of 100 mW/cm² and a secondary untreated tumor was monitored for immune-based control. Irradiation of the nanoconstruct containing both the chlorin-based ligand and the IDO inhibitor in the primary tumor was the most effective at controlling the growth of a contralateral secondary tumor that was not exposed to either the construct or light.

In a subcutaneous PDX model of non-small cell lung cancer, mice were intravenously injected with a novel PS1 PDT agent, or with Photofrin at doses of 1.5 μmol/kg or 6 mg/kg, respectively. [36] Following a PS-light interval of 24 hours, the subcutaneous PDX tumors were irradiated with 665 nm light for the mice injected with PS1, or with 630 nm light for the mice injected with Photofrin. Photoirradiation was performed using tunable dye lasers pumped by an argon-ion laser (Spectra-Physics) and Plexiglas holders that construct light exposure to only 4-5 mm. In comparison with the approved photosensitizer Photofrin, PDT using PS1 was more effective at inducing tumor necrosis and apoptosis and prolonged the time-duration before PDX tumor regrowth.

As mentioned earlier, subcutaneous xenografts are particularly advantageous for studying activatable therapies such as PDT due to their accessibility and relatively feasible photoirradiation. Although the skin (~500 μm) in such models is not a significant barrier for near infrared light penetration, skin phototoxicity can be a concern for non-targeted PDT regimens and can impair the accuracy of tumor volume measurements. From experience in our lab, we found that skin phototoxicity and scarring following non-targeted PDT often resolves within a week.[5] It must also be noted that PDT of subcutaneous tumors often does not represent the toxicity profile of the regimen if performed in vivo, and thus PDT doses that may provide complete control of a subcutaneous tumor may not be tolerated orthotopically at sensitive anatomical sites, such as the liver, pancreas and brain.[13] [42] This adds a layer of complexity for selecting PDT dosimetry parameters either in orthotopic models or in patients, which cannot reliably be guided by PDT doses that are effective in subcutaneous models.

3. Data Analysis and Interpretation

The assessment of treatment efficacy for PDT is not necessarily a trivial task, as it is subject to the study design and hypothesis. Therefore, this section will discuss several methods for longitudinal monitoring of treatment efficacy (section 3.1), and intermediary response monitoring using sophisticated functional imaging modalities (section 3.2).

3.1. Longitudinal Monitoring of Treatment Efficacy

A key advantage of using subcutaneous tumor xenograft models is the ability to longitudinally monitor tumors in response to treatment using simple measurements. The three most common modalities used to longitudinal monitor tumor progression include caliper measurements, ultrasound imaging and bioluminescence imaging. Of these three, caliper imaging is the most frequently used due to its technical simplicity and non-perturbative approach, although it can be subject to experimenter bias and is recommended only under blinded conditions. Caliper measurements are not recommended for tumors below 2 mm in diameter and also cannot account for variations in the thickness of the skin, which can be problematic for proinflammatory treatments like PDT. Ultrasound imaging is more sensitive than caliper measurements and is therefore recommended for monitoring smaller tumors. The unavailability and expense of sophisticated ultrasound imaging instruments may be a concern. Manual selection of tumor tissue during image analysis can also be subject to experimenter bias and requires experience to accurately determine tumor margins. Whole mouse imaging systems used to detect bioluminescence signals from tumors may also not be available to all research facilities. Although non-invasive and highly sensitive, the less common bioluminescence imaging has multiple drawbacks. These weaknesses including the need for stable transfection of bioluminescence enzymes (e.g. luciferase), the non-linear bioluminescence transmission through the tissue, the need for an exogenous probe (e.g. luminol) and the dependence on ATP, which varies with the types of treatments administered.

In the study by Spring et al., tumor volumes were calculated using the hemi-ellipsoid equation and the three dimensions of the tumors were measured longitudinally using calipers.[7] The tumor volumes reached 50 mm³ within 18 days after implantation, after which the PDT-based combination treatment using a photoactivable multi-inhibitor nanoliposome (PMIL) was initiated. The PMIL contained both the PS BPD and anti-metastatic and anti-angiogenic small molecule inhibitor, XL184.[7] Following PDT-based combination therapy, subcutaneous tumor volume was monitored for up to 47 days (Figure 1a). The biggest difference between the test treatment arm using PMIL and the control arms was observed at 37 days, where the fractional residual tumor volume was significantly lower than all the control arms with no treatment, free XL184, nanoparticle formulation (NP[XL148]), PDT using liposomal BPD (L[BPD]), or a mixture of L[BPD]and NP[XL184] (Figure 1b). Importantly, treatment using the PMIL significantly reduced liver and lymph node metastases more than any other treatment arm, demonstrating the power of co-encapsulated therapies. Metastases were quantified in the liver and lymph nodes using a quantitative reverse transcription-polymerase chain reaction (qRT-PCR) assay that measures human and mouse glyceraldehyde 3-phosphate dehydrogenase (GAPDH) housekeeping genes.

Figure 1.

a) Growth rates of subcutaneous AsPC-1 tumors and fractional residual tumors at 19 days following treatment with the c-MET/VEGFR-2 inhibitor (XL184), free or in a nanoparticle (NP); with PDT using liposomal BPD (L[BPD]); with combinations of BPD and XL184; and with co-encapsulated XL184 nanoparticles within BPD liposomes (PMIL). (Figure adapted from Spring B.Q. et al. (2016) Nat. Nanotechnol. 11:378-387 doi:10.1038/nnano.2015.311)

In the study by Lu et al. looking into the combined effects of a nanoconstruct containing a PS and the IDO inhibitor (IDOi@TBC-Hf), a subcutaneous contralateral model was used to monitor the immune-based control of untreated tumors.[31] PDT of the primary tumor in combination with IDO inhibition lead to tumor antigen presentation and T cell proliferation, that lead to the destruction of the contralateral tumor (Figure 2a). The volumes of the primary right flank and secondary left flank untreated tumors were monitored longitudinally using calipers, and tumor sections were analyzed for infiltrating leukocytes using immunohistochemistry. It was found that the untreated secondary tumors did not grow when the primary tumors were treated with the nanoconstruct containing the PS and the IDO inhibitor with photoirradiation (IDOi@TBC-Hf; Figure 2b).

Figure 2.

a) Irradiation of the nanoconstruct containing the chlorin-based PS and the IDO immune checkpoint inhibitor (IDOi@TBC-Hf) results in simultaneous tumor antigen release and sustained T cell proliferation, ultimately controlling untreated contralateral tumor. (Figure adapted with permission from Lu, K. et al., 2016. Journal of the American Chemical Society, 138(38), pp.12502-12510. Copyright 2018 American Chemical Society.)

In a clinically-relevant subcutaneous PDX model of NSCLC, PDT was explored using Photofrin and a novel photosensitizer PS1.[36] Following PDT, mice were monitored for 60 days and the tumor volumes were longitudinally measured. Biomolecular analysis of PDX tissue death was performed on the tumors using H&E staining and caspase3 immunohistochemistry, confirming that PDT induced apoptotic tissue damage. Figure 3 demonstrates that PDT using PS1 was more efficacious than with Photofrin in its capacity to control tumor regrowth, as enabled by an easily accessible tumor.

Figure 3.

Photographs of NSCLC PDX subcutaneous tumors in SCID mice (blue circle), before and 28 days after PDT using the novel PS PS1 or Photofrin. (Figure adapted from Nwogu C, et al. (2016) J. Surg. Res. 200:8-12 doi:10.1016/j.jss.2015.07.024)

Longitudinal tumor monitoring is facilitated by the accessibility of subcutaneous tumors and provides critical insights into dynamic responses to PDT, such as rates of tumor-regrowth and its negative impact on prognosis. Longitudinal tumor monitoring, however, informs only of the terminal effects of treatment on tumor volume, giving no information of mode of death or intermediate events that occur in response to PDT. By monitoring intermediary responses, interventions can be incorporated into the treatment regimens to potentiate the efficacy of the PDT-based regimen used. These will be discussed in the following section.

3.2. Intermediary Response Monitoring Using in vivo Functional Imaging

The accessibility of subcutaneous tumors has also been leveraged to combine sophisticated in vivo functional imaging techniques with therapy to guide the prediction of treatment success, which can ultimately be used to design interventions that help minimize treatment failures. Photoacoustic imaging (PAI) involves the pulsed laser excitation of intrinsic or exogenous chromophores with distinct absorption profiles to generate localized acoustic signals that propagate through the tissue and can be detected using an ultrasound probe.[43] The main advantage for PAI is the deep tissue imaging of a broad variety of chromophores that are otherwise undetectable, such as hemoglobin, using tunable lasers as the excitation source. A major strength for PAI is the capacity to selectively and quantitatively measure the relative proportions of oxyhemoglobin and deoxyhemoglobin to establish in vivo tissue oxygenation, which has proven to be powerful for oxygen-dependent regimens, such as PDT. Another powerful modality that provides valuable information for intermediary PDT responses is Blood Oxygenation Level Dependent – Magnetic Resonance Imaging (BOLD-MRI).[44] Conventional MRI imaging leverages deflections in the magnetic field in water and fat-containing tissue to provide structural and morphological information. BOLD-MRI is a type of functional MRI (fMRI) that integrates information on the differential magnetic properties of oxygenated and deoxygenated hemoglobin and thus can be used to determine relative changes in tissue oxygenation. High resolution micro-positron emission tomography (PET) imaging has also been used to report on tumor metabolic functionality for numerous treatment modalities and has implications in PDT response monitoring.[45] Micro-PET leverages radiolabeled probes with certain biological activities, such as a glucose conjugate of a radiolabeled probe, which informs of in vivo glucose consumption. Being endogenously radioactive, sensitivity of imaging is much greater than modalities requiring exogenous activation, such as fluorescence imaging. However, common PET probes used have a short half-life and the radioactivity of the PET probes used could also intrinsically induce tumor damage, confounding the results further. These modalities used to monitor intermediary responses to PDT, although powerful in their own right and provide functional information that is otherwise not attainable, they require expensive instruments, can take significant amounts of time to acquire, and oftentimes need dedicated facilities. However, the immediate clinical relevance of these modalities makes them highly valuable and worth investing in at the pre-clinical level, in order to serve as early indicators of the extent of PDT anti-tumor response.

Mallidi et al. monitored the oxygenation of subcutaneous glioblastoma tumors using PAI of oxyhemoglobin and deoxyhemoglobin as a measure of the degree of response to PDT.[5] By correlating the degree of post-PDT oxygenation with tumor reduction, the authors developed an algorithm that enabled the prediction to PDT response as a means of customizing therapy for maximal outcomes. Tumor growth was monitored using calipers and the ellipsoid equation was used to estimate tumor volume. Following PDT using liposomal BPD, PAI imaging of oxygenated and deoxygenated hemoglobin was performed using 750 nm and 850 nm wavelength illumination, respectively, to derive estimates of oxygen saturation before, immediately after, 6 h after and 24 h after PDT. Figure 4 shows a representation of the subcutaneous U87 tumors imaged using ultrasound imaging, oxygen saturation using PAI, H&E staining of necrotic tissue and immunofluorescence of hypoxic regions bound to pimonidazole.

Figure 4.

Representative images of structure using ultrasound imaging (A, E, I), oxygen saturation using PAI (B, F, J), necrotic regions using H&E staining (C, G, K), and hypoxic regions using immunofluorescence of pimonidazole (D, H, L). Scale bars are 5 mm for ultrasound and PAI images, and 1 mm for H&E and immunofluorescence images. (Figure reproduced from Mallidi, S. et al. (2015) Theranostics 5:289-301 doi:10.7150/thno.10155)

Ultrasound imaging was also used to longitudinally monitor subcutaneous and orthotopic PDX models of bladder cancer following a combined nanoparticle incorporating photodynamic, photothermal and chemotherapeutic treatment regimens. [37] The mice behavior was longitudinally monitored and 30 days after PDT, the PDX tumors were imaged using ultrasound imaging with and without microbubbles (5 x 10⁷ per mouse) to enhance the contrast of blood perfusing the tumors. Microbubbles used as ultrasound imaging contrast are mostly organic gases trapped in a lipid or protein-based shell.[46] Tumor volumes were monitored longitudinally every 2-3 days for up to 5 weeks. The study found that in the subcutaneous PDX model, the trimodal combination therapy using the nanoconstruct was more efficacious at controlling tumor growth over 30 days, than the monotherapies alone. In the orthotopic PDX model, trimodal combination therapy using the nanoconstruct resulted in a significantly thinner bladder wall containing the tumor, than with the free doxorubicin comparison, and a significant higher ratio of functional bladder area.

In another study, treatment response of subcutaneous prostate tumors to PDT with TOOKAD was monitored using blood oxygenation level-dependent contrast magnetic resonance imaging (BOLD-MRI).[6] BOLD-MRI was particularly relevant for monitoring the efficacy of this treatment modality, as PDT with TOOKAD is predominantly a vascular therapy leveraging peroxynitrite radicals. PDT induced a 25-40% localized reduction of BOLD-MRI signal, which suggested treatment significantly reduced blood flow to the irradiated regions (Figure 5b,c).

Figure 5.

a) TOOKAD-PDT of a subcutaneous prostate cancer xenograft mode with representative BOLD-MRI images demonstrating the time-dependent reduction (minutes) in signal, indicating a reduced blood flow to the tumor (b, c). With TOOKAD administration, photoirradiation alone had no impact on the BOLD-MRI signal (d, e). (Reprinted by permission from Springer Nature: Nature Medicine, Monitoring photodynamic therapy of solid tumors online by BOLD-contrast MRI, Gross S, Gilead A, Scherz A, Neeman M, Salomon Y. (2003))[6]

Fei et al. used MRI imaging combined with micro-PET to generate functional images, respectively to evaluate PDT responses in C3H mice implanted with RIF-1 tumors that were subject to PDT using a silicon phthalocyanine. [47] The authors generated 3D reconstructions of the whole mouse by overlaying structural information from MRI and functional information from micro-PET imaging. 48 prior to imaging, the PS was administered. The mice were then injected with the PET probe ¹⁸F-fluorodeoxyglucose (FDG) and 6 minutes later, the tumors were irradiated with 670 nm light. Micro-PET emission images were acquired for a total of 90 minutes and confirmed that the uptake of FDG was significantly reduced after irradiation, as compared to an untreated control, which informed of an immediate reduction in metabolic demand directly in response to PDT.

Using such sophisticated image modalities to probe the physiological and functional consequences of PDT not only helps identify the anti-tumor mechanisms induced (e.g. Thrombosis, oxygen depletion, cellular insult, attenuation of metabolism etc.), but helps provide an intermediary means for predicting efficacy and intervening to augment outcomes. The most common modalities used for intermediary response monitoring for PDT focus on oxygenation of the tumor and perfusion. Other emerging imaging modalities that could be informative for monitoring responsiveness to PDT include metabolic imaging.[48] As these modalities become better suited for deep-tissue imaging, their uses could extend to monitoring orthotopic tumors, which will be discussed in other chapters, and ultimately be extended to the clinic to better predict PDT treatment response in situ.

4. Discussion

The body of literature within PDT and the wider family of oncology fields demonstrates the power of subcutaneous xenograft models to rapidly evaluate the treatment outcomes new therapeutics and novel treatment regimens, both with regard to dosing strategies and combinatorial therapy. With specific consideration of PDT, the accessibility of subcutaneous tumors for photoirradiation and subsequent monitoring of physiological parameters, such as blood flow and oxygenation, proves to be an invaluable tool in the expedition of clinical translation of PDT.[6] Although extremely powerful in their own right, it cannot go unmentioned that subcutaneous models lack a critical component of recapitulating the PDT treatment response of the parent human disease, namely the anatomical site. Multiple studies have outlined the discrepancy between PDT treatment response of subcutaneous xenograft models and that of their orthotopic equivalent, mostly a decreased susceptibility to tumor destruction when moving into an orthotopic model.[7, 10] Multiple hypotheses account for this discrepancy, the most widely accepted being the more complex, pathophysiologically relevant interactions of the tumor xenograft with the appropriate respective anatomical environment, leading to variable responses to the treatment and varying patterns of metastases. This can also be accompanied by the recruitment of organ-specific partner cells when xenograft tumors are orthotopically generated, leading to a therapeutic response that is more representative of one in the clinical manifestation of the disease. Specifically, in the context of PDT, the discrepancy between subcutaneous and orthotopic tumor treatment responses is further complicated by the need for different light illumination strategies. Although fiber optic light delivery technologies have been advanced and optimized to serve the utility of lasers in bioimaging and phototherapy, accurate guidance of internal irradiation protocols of orthotopic tumors can be more technically involved.[49-51] Conversely, illumination of orthotopic tumors directly can be more effective for PDT when no light penetration barrier of intermediary tissue is present, as in the case for irradiating subcutaneous models. Regardless of the challenges and complexities they present, the utility of orthotopic xenograft models for evaluating PDT efficacy will be discussed in separate chapters of this book.