P-glycoprotein Targeted and Near-infrared Light-guided Depletion of Chemoresistant Tumors

Abstract

Drug resistance remains a formidable challenge to cancer therapy. P-glycoprotein (Pgp) contributes to multidrug resistance in numerous cancers by preventing accumulation of anticancer drugs in cancer cells. Strategies to overcome this resistance have been vigorously sought for over 3 decades, yet clinical solutions do not exist. The main reason for the failure is lack of cancer specificity of small-molecule Pgp inhibitors, thus causing severe toxicity in normal tissues. In this study, Pgp-targeted photodynamic therapy (PDT) was developed to achieve superior cancer specificity through antibody targeting plus locoregional light activation. Thus, a Pgp monoclonal antibody was chemically modified with IR700, a porphyrin photosensitizer. In vitro studies showed that the antibody-photosensitizer conjugates specifically bind to Pgp-expressing drug resistant cancer cells, and caused dramatic cytotoxicity upon irradiation with a near infrared light. We then tested our Pgp-targeted approach in mouse xenograft models of chemoresistant ovarian cancer and head and neck cancer. In both models, targeted PDT produced rapid tumor shrinkage, and significantly prolonged survival of tumor-bearing mice. We conclude that our targeted PDT approach produces molecularly targeted and spatially selective ablation of chemoresistant tumors, and thereby provides an effective approach to overcome Pgp-mediated multidrug resistance in cancer, where conventional approaches have failed.

1. Introduction

Drug resistance causes treatment failure and death in the majority of patients with metastatic cancer [1–3]. P-glycoprotein (Pgp) is a primary cause for drug resistance in cancer therapy [4, 5]. Most anticancer drugs meet their pharmacological targets and carry out their anticancer actions within cancer cells. Pgp, as well as other ATP-binding cassette (ABC) transporters, is overexpressed in the plasma membrane of cancer cells and medicates efflux of structurally diverse and mechanistically unrelated anticancer agents out of cancer cells [6]. This reduces cytotoxic effects of these agents and causes multidrug resistance (MDR). Pgp was first identified as responsible for drug resistance in ovarian cancer (OvCa) patients [7]. Since then, Pgp overexpression has been found to be associated with treatment failure in more than half of human cancers, including ovary, colon and liver cancers, as well as leukemia and lymphoma [8–11].

Strategies to overcome this resistance have been actively sought for more than three decades, and three generations of small-molecule inhibitors of Pgp have been developed to sensitize MDR cancer cells [9, 12]. However, they have yet to reach the oncology clinic [12–15]. Lack of cancer specificity, thereby causing unacceptable toxic effects in healthy tissues, is a main hurdle that has prevented successful translation of the small-molecule Pgp inhibitors [13, 15, 16]. Pgp plays pivotal physiological roles in the body’s defense against toxic compounds [13]. Extratumoral distribution of small-molecule inhibitors has caused dramatic toxicity in vital organs such as the brain [17], causing failure of the clinical trials testing the inhibitors [18]. The status quo of targeting Pgp is still on the pathway to identify new small-molecule inhibitors [16, 19]. However, the intrinsic limitation will not be solved by making new generation of inhibitors. Although some nanocarriers are developed to overcome the Pgp-mediated chemoresistance [20], the results from animal studies haven’t been translated into clinical studies. Indeed, all FDA-approved cancer nanomedicines are substrates for Pgp and their anticancer actions are limited by Pgp-mediated MDR [21, 22]. In conclusion, a dramatic overhaul is warranted in the strategies to overcome Pgp-mediated MDR.

Targeted PDT provides a highly specific approach to treat cancers by combining antibody-based cancer targeting and localized light activation of the drug, photosensitizer (PS) [23]. Thus, it can solve the intrinsic limitation of small-molecule Pgp inhibitor approach: low cancer specificity. In targeted PDT process, PS is conjugated to a cancer-selective antibody and is activated with a near-infrared (NIR) light at the tumor site to induce rapid cancer cell destruction [23]. Targeted PDT using cetuximab-IR700 to target epidermal growth factor receptor [23] is currently being tested in a phase I/II trial for the treatment of head and neck cancers (NCT02422979). Pgp functions as efflux pump primarily at the plasma surface of MDR cancer cells, and its surface expression level is proportional to drug resistance capability of cancer cells [24, 25]. Thus, Pgp is an excellent target for targeted PDT, and this approach can provide high cancer specificity through antibody targeting plus locoregional light activation, ultimately overcoming the intrinsic limitation of small-molecule Pgp inhibitor approach.

In this study, we aim to develop Pgp-targeted PDT with high cancer specificity for treating drug resistant cancers. Thus, we prepare antibody-photosensitizer conjugates (APCs) using an anti-Pgp monoclonal antibody (Pab), and test their specificity and phototoxicity using cell culture models. Then, their tumor delivery and anticancer efficacy are examined using two mouse xenograft models of drug resistant cancers.

2. Results

2.1. Chemistry and Pgp specificity of Pab-IR700

We prepared the Pgp-specific APCs by incubation of anti-Pgp monoclonal antibody 15D3 (Pab) with IR700-NHS at molar ratio of 1:4 in phosphate buffer (pH 8.0) for 1 h followed by gel filtration. The resultant antibody conjugates (Pab-IR700) contain approximately 2 molecules of IR700 in each antibody. The purity of Pab-IR700 conjugates was first examined by SEC-HPLC analyses (Figure 1A). Pab-IR700 eluted slightly earlier than Pab (Figure 1A) with the retention time of 7.9 and 8.0 min, respectively, and both eluted much earlier than free IR700 with the retention time of 12.5 min, indicating successful conjugation and purification of Pab-IR700.

Figure 1. Chemistry and specificity of Pab-IR700.

(A) SEC-HPLC analysis of IR700, Pab, and Pab-IR700 detected with UV-Vis detectors at 689 nm and/or 280 nm. (B) Flow cytometry analysis of three Pgp-expressing cell lines after treatment with Pab-IR700 or IgG-IR700; (C) Confocal images of Pab-640R distribution in three Pgp positive and GFP-expressing cell lines. Nucleus was labeled with Hoechst 33342 (blue). Scale bar, 20 μm.

To evaluate the binding specificity of Pab-IR700 to Pgp-expressing cancer cells, we performed flow cytometry and confocal laser scanning microscopy (CLSM) analyses after staining 3T3-MDR1, NCI-ADR^Res, and KB-8-5-11 cells with the antibody conjugates. After incubation with 10 μg/mL of Pab-IR700 and control IgG-IR700 at 4°C for 30 min, the fluorescence of the cells was acquired with flow cytometry. As shown in the histograms of Figure 1B, only the cells treated with Pab-IR700 showed high level of fluorescence while those treated with IgG-IR700 only showed background signal which was similar to the PBS group. In order to avoid phototoxicity in live cell imaging process, we replaced IR700 in the antibody conjugates with a far-red dye CF™640R (640R), which is not photoactive, for confocal microscopy experiments. We then investigated the intracellular distribution of Pab-640R in both Pgp-expressing cell lines by CLSM. After 4-h incubation, the 640R fluorescence was detected in the cell membrane of these cell lines where Pgp is located, and some were observed in the vesicles in the cells, indicating Pgp-mediated endocytosis (Figure 1C). Cellular uptake of Pab conjugates was also examined in the control cell line 3T3 that does not express Pgp. Neither Pgp-targeted nor control antibody conjugates bound to the 3T3 cell membranes as shown in the results from flow cytometry (Supplementary Figure 1A) and CLSM (Supplementary Figure 1B).

2.2. In vitro phototoxicity of Pab-IR700

We examined Pgp-specific phototoxicity of Pab-IR700 in Pgp-expressing cells. After incubation with Pab-IR700 followed by light irradiation at the light dose of 5 J/cm², 3T3-MDR1 cells exhibit cellular swelling, bleb formation, and rupture of vesicles, indicating quick cell death (Figure 2A). The same treatments did not change the morphology of Pgp-negative 3T3 cells (Figure 2A), indicting Pgp-specific cell killing of this targeted PDT procedures. We then evaluated dose-dependent phototoxicity of Pab-IR700 with Alamar Blue assay in both Pgp transfected 3T3-MDR1 cells and Pgp-expressing MDR cell lines NCI-ADR^Res and KB-8–5-11. As shown in Figure 2B, Pab-IR700 caused dramatic cytotoxicity towards 3T3-MDR1 cells upon NIR irradiation (5 J/cm²) with the IC₅₀ value of approximately 0.58 μg/mL. Similar phototoxicity was observed with the IC₅₀ values of 0.96 and 1.34 μg/mL in MDR KB-8–5-11 and NCI-ADR^Res cells, respectively (Supplementary Figure 2). Pgp-targeted PDT did not affect cell viability of 3T3 cells up to the Pab-IR700 concentration of 10 μg/mL; PDT with IgG-IR700 didn’t hurt 3T3-MDR1 cells either (Figure 2B). We further tested phototoxicity of Pab-IR700 with Alamar Blue assay in primary human hepatocytes that express Pgp at a basal level [26]. As shown in Supplementary Figure 3, Pab-IR700 did not damage primary human hepatocytes up to the Pab-IR700 dose of 10 μg/mL, and produced minor toxicity at the dose of 20 μg/mL. This result may indicate low side effects when our Pgp-targeted PDT approach is translated to clinical studies.

Figure 2.

Phototoxicity of Pab-IR700. (A) Microscopic observation of 3T3 and 3T3-MDR1 cells treated with Pab-IR700 followed by NIR irradiation. Scale bar, 50 μm. (B) Dose-dependent phototoxicity of Pab-IR700 and IgG-IR700 in 3T3 and 3T3-MDR1 cells after NIR irradiation at 5 J/cm². Data are means ± SD (n = 3, *** p < 0.001). (C) Light dose-dependent phototoxicity of Pab-IR700 in 3T3-MDR1 cells. Data are means ± SD (n = 3). (D) Live/Dead staining of NCI-ADR^Res cells after photokilling with Pab-IR700 and IgG-IR700. (E) Live/Dead staining of KB-8–5-11 cells after photokilling with Pab-IR700 and IgG-IR700. Scale bar, 50 μm. (F) PI staining in mixed 3T3-GFP and 3T3-MDR1 cells after Pab-IR700-mediated PDT. Scale bar, 40 μm. (G) 3T3-MDR1 cells were irradiated with a narrow-beam light through a pinhole after Pab-IR700 treatment. Only cells irradiated by the NIR light in the center area marked with a yellow circle showed cell death. Scale bar, 500 μm. (H) Flow cytometry analysis of 3T3-MDR1 cells after targeted PDT followed by Annexin V-FITC and PI staining.

We further examined light dose-dependent phototoxicity of Pab-IR700 in 3T3-MDR1 cells. Different light doses were given by varying light irradiation time. As shown in Figure 2C and Supplementary Figure 4, we observed clear light dose dependence of phototoxicity within the APC concentration range of 0.3–10 μg/mL. The IC₅₀ value of Pab-IR700 decreased from 3.75 to 0.16 μg/mL when the light dose increased from 1.25 to 11.25 J/cm² (Figure 2C and Supplementary Figure 5). In addition, there was no significant difference of the phototoxicity when incubation time before light treatment increased beyond 2 h (Supplementary Figure 6).

We also performed Live/Dead cell staining assay to further examine phototoxicity of Pab-IR700. The Live/Dead stain images in Figures 2D and 2E showed cell death occurred 4 h after light treatment, indicating that cell death was induced quickly after irradiation. IgG-IR700 treatment showed no effects after light irradiation, indicating Pab-IR700 produced a rapid and specific cell death in NCI-ADR^Res and KB-8–5-11 cell lines upon light irradiation (Figures 2D and 2E). To further demonstrate Pgp specificity of our targeted approach, we co-cultured 3T3-MDR1 and 3T3-GFP cells, treated them with Pab-IR700 followed by light treatment. Four hours after light treatment, cells were stained with a membrane-impermeant nuclear dye propidium iodide (PI). Images in Figure 2F showed PI staining only in 3T3-MDR1 cells but not in 3T3-GFP cells, indicating Pab-IR700 mediated photokilling is specific to Pgp expression. Light specificity was also confirmed by irradiating 3T3-MDR1 cells with a narrow-beam NIR light through a pinhole after Pab-IR700 treatment. As shown in Figure 2G, only cells irradiated by the NIR light in the center area marked with a yellow circle showed substantial cell death while the cells outside the pinhole area showed no cell damage. Overall, Pab-IR700 produced Pgp-specific and light-dependent cytotoxicity towards Pgp-expressing chemoresistant cancer cells.

We further performed a flow cytometry analysis to determine the death pathways of Pgp-targeted PDT using a Dead Cell Apoptosis Kit with Annexin V-FITC and PI. As shown in Figure 2H, 8 h after the targeted PDT procedures, 22.3% of the cells were at early apoptosis state (Annexin V⁺ and PI^-), and 59.3% of them were at late apoptosis/secondary necrosis phases (Annexin V⁺ and PI⁺). A small portion of the cells (6.5%) were at necrosis state (Annexin V^- and PI⁺). Based on the data, we concluded that Pgp-expressing cells undergo both apoptosis and necrosis pathways after Pgp-targeted PDT, with apoptosis as the major death mechanism.

2.3. Pab-IR700 caused phototoxicity in tumor spheroids of drug resistant cells

We examined cellular uptake and phototoxicity of Pab-IR700 in tumor spheroids, which recapitulate some key features of the in vivo microenvironment, including hypoxia and the presence of ECM [27–29]. In vivo microenvironment has a significant impact on cancer responses to PDT. For example, PDT is a type of oxygen-dependent therapy, and thus clinical relevance is often questionable when PDT effects are studied in monolayer cell cultures that are under well-oxygenated conditions.

We first used Pab-640R to examine penetration of the antibody conjugates in tumor spheroids using live cell imaging. NCI-ADR^Res cells were cultured to form 3D spheroids and were then incubated with Pab-640R for 24 h. The distribution of Pab-640R in NCI-ADR^Res spheroids were detected by CLSM (Figure 3A). The Pab-640R was able to penetrate the spheroids as the fluorescence of 640R was detected deeply inside the spheroid. The 640R fluorescence within the spheroids was further analyzed with Image J. As shown in Figure 3B, Pab-640R showed an extensive distribution inside the spheroid and the intensity of 640R was in a depth-dependent manner. The spheroids were further digested and intracellular 640R fluorescence was analyzed by flow cytometry (Figure 3C). Cells treated with Pab-640R showed higher 640R fluorescence than those treated with the control IgG-640R, indicating the Pab conjugates not only bind to the Pgp-expressing cells but also penetrate the spheroids. IgG-640R has the same size and surface property as Pab-640R, and should possess the same penetration capability. However, due to minimal cellular uptake, it does not show substantial accumulation in the tumor spheroids. To further confirm the distribution of Pab-640R in NCI-ADR^Res spheroids, we further sectioned the spheroids and then imaged the sections. The fluorescent images in Supplementary Figure 7A and its quantitative distribution data in Supplementary Figure 7B confirmed the excellent delivery of the Pab conjugates in 3D tumor spheroids of NCI-ADR^Res cells.

Figure 3.

Penetration and phototoxicity of antibody conjugates in tumor spheroids. (A) CLSM images of NCI-ADR^Res spheroids after incubation with Pab-640R and IgG-640R overnight. Scale bar, 100 μm. (B) Quantitative distribution of Pab-640R and IgG-640R in NCI-ADR^Res spheroids (***p < 0.001). (C) Flow cytometry analysis of the digested cells of NCI-ADR^Res spheroids after treatment with Pab-640R or IgG-640R. (D) PI staining in NCI-ADR^Res-GFP spheroids after Pab-IR700-mediated PDT. Scale bar, 300 μm. (E) Mean fluorescence intensity of GFP and PI signals inside the NCI-ADR^Res-GFP spheroids after 48- and 96-h light treatment. Data are means ± SD. (n=3, ### p < 0.001).

We further examined phototoxicity of Pab-IR700 in tumor spheroids of GFP-expressing NCI-ADR^Res cells after incubation with Pab-IR700 (20 μg/mL) and NIR irradiation (10 J/cm²). The APC and light doses were selected based on previous studies that investigated PDT in tumor spheroids [30–32]. In this experiment, the GFP fluorescence was used as a live signal while PI stain served as dead signal. As shown in Figures 3D and 3E, NCI-ADR^Res tumor spheroids after targeted PDT procedures showed stronger PI stain and weaker GFP signal than the control spheroids, indicating the substantial cell death and reduced cell viability of the spheroids. As shown in Supplementary Figure 7C, treatment with Pab-IR700 plus light irradiation disrupted the structure of NCI-ADR^Res spheroids and suppressed their growth up to 28 days.

2.4. In vivo Pgp-targeted PDT in a subcutaneous tumor model of chemoresistant OvCa

To examine anticancer activity of our Pgp-targeted PDT approach in vivo, a mouse xenograft model of chemoresistant OvCa was established by subcutaneous injection of 5 × 10⁶ NCI-ADR^Res-Luc-GFP cells to Balb/c nude mice bilaterally. The treatment scheme was shown in Figure 4A. Ten days after tumor inoculation, Pab-IR700 (300 μg) was injected intravenously to the nude mice and accumulation of IR700 was imaged with an IVIS Lumina III system two days after injection. The fluorescent images in Figure 4B demonstrated excellent accumulation of Pab-IR700 in NCI-ADR^Res tumors. Tumor tissues of some mice were collected and sectioned. The distribution of Pab-IR700 in tumor sections was further examined with CLSM, and the confocal images in Figure 4C and Supplementary Figure 8A confirmed tumor uptake of Pab-IR700. Two days after IV injection of Pab-IR700, NCI-ADR^Res-Luc-GFP tumors on the right flank were irradiated with a 690 nm NIR light and tumors on the left side were not irradiated to serve as controls. This light dose of 50 J/cm² was selected based on previous studies using antibody-IR700 conjugates for animal studies [23, 33, 34]. Tumor loading was monitored using bioluminescence imaging (BLI) with IVIS. As shown in Figures 4B and 4D, all the left-flank tumors that were not irradiated showed steady growth in both PBS and Pab-IR700 groups. There was a dramatic decrease in tumor loading in the right-flank tumors that were irradiated in the Pab-IR700 group, while those in PBS group showed similar growth to the control left-flank tumors, indicating that only NIR irradiation does not suppress tumor growth. Starting from Day 16, the right-flank tumors in the Pab-IR700 group showed tumor recurrence, but a delayed tumor growth was observed when compared to the left-flank tumors in the same group. The mice treated with Pab-IR700 were sacrificed and the tumor tissues were collected in Day 19. The image in Supplementary Figure 8B shows dramatic tumor shrinkage on the irradiated right-flank tumors compared to the non-irradiated left-flank ones. We collected some tumor tissues at Day 14 (2 days after light irradiation) and analyzed the tissue samples by immunohistochemistry. As shown in Figure 4E, apoptotic cancer cells were observed in the irradiated tumor tissues in Pab-IR700 group using TUNEL assay and the complexity of tumor tissues was disrupted as revealed with Hematoxylin and Eosin (H&E) staining. We also observed a decreased GFP fluorescence in these tissues, indicating death of NCI-ADR^Res-Luc-GFP tumor cells caused by Pgp-targeted PDT.

Figure 4.

Pab-IR700 mediated PDT in a mouse xenograft model of chemoresistant OvCa tumors. (A) The scheme of the animal study. (B) IVIS imaging of tumor uptake and targeted PDT effects of Pab-IR700. Fluorescence imaging (FI) showed that Pab-IR700 accumulated in NCI tumor with a good tumor/background ratio 48 h post Pab-IR700 injection. BLI showed that the tumor loading in two flanks of the mice were similar before light treatment. After light treatment, tumor loading of the light-treated right-flank tumors in Pab-IR700 group were reduced dramatically, while the non-irradiated left-flank tumors in the Pab-IR700 group and the bilateral tumors in PBS group showed steady growth. (C) Confocal images of tumor sections show intratumoral Pab-IR700 distribution 2 days post Pab-IR700 injection (on Day 12). Blue: DAPI; Green: NCI-ADR^Res-GFP-Luc; Red: Pab-IR700. Scale bar, 100 μm. (D) Luciferase activity of NCI-ADR^Res-GFP-Luc tumors was quantified with BLI. Data are means ± SD (n=5, *** p < 0.001). (E) Histological analysis of NCI-ADR^Res-GFP-Luc tumors using H&E staining and TUNEL assay. Blue: DAPI; Green: NCI-ADR^Res-GFP-Luc; Red: Alexa Fluor 594 labeled dUTP (TUNEL Positive cells). Scale bar, 100 μm.

To assess potential side effects of our targeted PDT approach, we examined biodistribution of Pab-IR700 in NCI-ADR^Res xenograft-bearing mice. Two days after IV administration of Pab-IR700 (300 μg), NCI-ADR^Res tumor-bearing mice were sacrificed, and the tumors and other main organs, including the brain, heart, lung, liver, spleen, and kidney, were collected and imaged with IVIS. As shown in Supplementary Figure 8C, Pab-IR700 accumulated in the tumor, liver, and kidney, but were largely absent in brain, heart, lung, and spleen. The Pab-IR700 level in the tumor was about 2-fold than those in the liver and kidney (Supplementary Figure 8D). To further assess potential off-target effects of our targeted PDT approach, we examined the possible damage to the liver and kidney after targeted PDT procedures. H&E staining of the liver and kidney sections indicted no morphological change in these organs (Supplementary Figure 8E). We also measured the levels of aspartate transaminase (AST) and alanine transaminase (ALT) in the blood of the mice 48 h after light treatment. Increased AST and ALT levels were not observed in PDT-treated mice (Supplementary Figure 8F), indicating that targeted PDT didn’t damage the liver or muscle in vivo. All of these results indicate targeted PDT produces minimal off-target effects and is a highly cancer specific procedure.

2.5. In vivo Pgp-targeted PDT in chemoresistant tumor model of head and neck cancer

We further evaluated anticancer effects of our Pgp-targeted PDT approach in a mouse tumor model of chemoresistant head and neck cancer. This model was established by injection of 2 × 10⁶ KB-8–5-11-GFP-Luc cells into the floor of the mouth of each mouse via an extra-oral approach. Five days after tumor inoculation, the mice were injected intravenously with Pab-IR700 or control IgG-IR700 (300 μg), and the IR700 distribution was examined with IVIS two days after injection. Pab-IR700 showed substantially higher tumor accumulation than IgG-IR700 in this orthotopic tumor model (Figure 5A). We quantified the IR700 fluorescence intensity at tumor tissues in Figure 5C, and the IR700 intensity in Pab-IR700 group was 6.9 folds than IgG-IR700 group, confirming Pgp-targeted tumor delivery of Pab-IR700. At day 7, the tumors were irradiated with NIR according to a schedule shown in Figure 5B, tumor loading was monitored by detecting the luciferase activity using BLI (Figure 5A and Supplementary Figure 9A) and tumor size was measured by a caliper. Mice treated with PBS or IgG-IR700 plus NIR irradiation were used as controls. The results showed that Pab-IR700 based targeted PDT significantly suppressed the tumor growth, while administration with IgG-IR700 followed by NIR irradiation had no effects on the tumor growth when compared with the PBS group (Figure 5D). The median survival time of the PBS group was 14.5 days, while that of the Pab-IR700 group was 20.5 days (Figure 5E), indicating that Pgp-targeted PDT extended the survival of mice bearing MDR tumor. Pgp-targeted PDT did not reduce the body weight of mice (Figure 5F), indicating no significant side effects of this treatment to mice. We collected the tumor tissues at Day 9 for histological observation. Images from H&E staining showed that targeted PDT disrupted the complexity of tumor tissues (Figure 5G). Using TUNEL assay, we also observed apoptotic tumor cells after targeted PDT (Figure 5E). In addition, the images also revealed that targeted PDT reduced the GFP-positive tumor cell population (Figures 5G and Supplementary Figure 9B). Zoomed in Images in Figure 5H clearly showed that the TUNEL-positive cancer cells lose their GFP expression and cell morphology.

Figure 5.

Pab-IR700 mediated Pgp-targeted PDT in chemoresistant tumor model of head and neck cancer. (A) IVIS imaging of tumor uptake and targeted PDT effects of Pab-IR700. FI showed that Pab-IR700 accumulated in KB-8–5-11-GFP-Luc tumors at a level 6.93-folded higher than IgG-IR700 2 days post Pab-IR700 injection. BLI showed that tumor loading of the irradiated tumors in Pab-IR700 group were reduced dramatically, while the irradiated tumors in IgG-IR700 group and the tumors in PBS group showed steady growth. (B) The scheme of the animal study. (C) Quantitation of IR700 fluorescence in KB-8–5-11-GFP-Luc tumor sites with IVIS. Data are presented as mean ± SD (n = 8, *** p < 0.001). (D) Target specific tumor growth inhibition by Pab-targeted PDT for KB-8–5-11-GFP-Luc tumors. Suppression of tumor growth was observed after targeted PDT. Data are presented as mean ± SD (n = 8, **p < 0.005). (E) Kaplan-Meier survival curve of targeted PDT in KB-8–5-11-GFP-Luc tumor model (n = 8, ## p < 0.005). (F) Body weight of mice after Pab-IR700-mediated PDT from Day 4 to Day 15. Data are means ± SD, n=8. (G) Histological observation of KB-8–5-11-GFP-Luc tumors after H&E staining and TUNEL assay. Blue: DAPI; Green: KB-8–5-11-GFP-Luc; Red: Alexa Fluor 594 labeled dUTP (TUNEL Positive cells). Scale bar, 100 μm. (H) Zoomed in images of KB-8–5-11-GFP-Luc tumor sections after TUNEL assay. TUNEL positive cells showed substantial cellular swelling and GFP intensity decreased in the targeted PDT group.

3. Discussion

Pgp was identified as a key mediator of cancer MDR for more than three decades [7], and new evidence has continuously supported this notion [35]. However, Pgp is still undruggable because small-molecule inhibitors fail to limit their actions on tumoral Pgp and thus caused substantial toxicity in normal tissues where Pgp plays an important protection role [13, 36]. In this study, cancer specificity of Pgp-targeted PDT is achieved by combining antibody-based cancer targeting and localized light activation of the PS. The first layer of cancer targeting comes from antibody targeting. Pab-IR700 specifically bind to Pgp that is expressed in Pgp transfected 3T3 cells as well as Pgp-expressing drug resistant cell lines NCI-ADR^Res and KB-8–5-11 (Figure 1), but not to control 3T3 cells that do not express Pgp (Supplementary Figure 1). Accordingly, Pab-IR700 treatment followed by light irradiation caused dramatic cytotoxicity in Pgp-expressing 3T3-MDR1, NCI-ADR^Res, and KB-8–5-11 cells, while the same treatments did not hurt these control cell lines that do not express Pgp (Figure 2 and Supplementary Figure 2). In mouse xenograft models of drug resistant OvCa and head and neck cancer, Pab-IR700 showed excellent delivery to drug resistant tumors after systemic administration (Figures 4 and 5), and its tumor accumulation was superior to control IgG-IR700 in the MDR tumor model (Figure 5). All these results indicated that cancer specificity can be enhanced using antibody targeting approach. Cancer specificity can be even enhanced through locoregional light irradiation on the tumor sites. Thus, sole treatment of Pab-IR700 did not cause toxicity to 3T3-MDR1 cells without light irradiation (Figures 2A and 2G). In xenograft mice bearing bilateral drug resistant tumors, only the tumors on the right flank were exposed to the NIR light after IV administration of Pab-IR700. Thereafter, the BLI result showed marked reduction in tumor loading in the right-flank tumors, while the left-flank tumors showed no response even Pab-IR700 was delivered to these tumors (Figure 4). These results clearly showed that in-tumor light irradiation can be combined with antibody targeting to enhance cancer specificity of the therapy.

The status quo of targeting Pgp is still on the pathway to discover new small-molecule inhibitors, however, clinical development of these inhibitors has stalled as the underlying causes of failure have not been addressed successfully [37]. Using macromolecular ligands to target Pgp may provide a superior approach to small-molecule methods. The function of Pgp is to pump the drugs out of cancer cells; instead, by using macromolecular ligands, the transport direction can be essentially reversed to mediate tumor uptake of the drugs that are linked to the ligands. Only a few studies used antibody targeting to overcome Pgp-mediated cancer MDR. In one study, Pseudomonas toxin was conjugated to an anti-Pgp antibody MRK-16, and the antibody-drug conjugates specifically killed Pgp-expressing MDR cells [38]. In another study, a Pgp antibody was conjugated to a lentiviral vector and the modified virus was successfully delivered to Pgp-expressing tumor in a mouse model of metastatic melanoma after systemic administration [39]. Although these methods showed Pgp-specific delivery, they may cause off-target toxicity in normal tissues that express Pgp, such as blood-brain barrier. Thus, it is necessary to combine with a secondary approach to enhance cancer specificity for targeting Pgp. In this study, we combine antibody targeting method with light triggered approach to enhance cancer specificity. The results with in vitro and animal models clearly show superior cancer specificity was achieved using our combination approach. Thus, this strategy may become a general approach to target surface proteins that play important roles, both in pathogenesis in diseased tissues and physiologic function in normal tissues.

Cancer targeting has been limited by availability of cancer-specific membrane markers, and most of cancer targeted delivery systems utilize limited receptors that are overexpressed in tumors, including EGFRs, integrins, and folate receptors [40]. For example, EGFR-targeted therapy is effective in treating some cancers such as non-small cell lung cancer and colorectal cancer; however, no EGFR therapies are currently approved for treatment of OvCa [41]. On the other hand, Pgp is widely expressed in OvCa, ranging from 40% to 92.8% across various studies [42–47]; importantly, its expression is correlated with treatment response or survival outcome [45–50]. Thus, Pgp-targeted PDT may have advantage over EGFR-targeted therapy in treating chemoresistant OvCa, especially in the patients that have been treated with Taxol, the first-line chemotherapy for OvCa and a substrate of Pgp. Another advantage of Pgp-targeted PDT is its low potential causing off-target effects. Firstly, compared to other tumor targets such as EGFR and integrins, Pgp has narrower presence in normal tissues, mainly in elimination organs such as the liver and kidney, and barrier tissues such as blood-brain barrier [37]. In addition, the Pgp antibody (15D3) used in this study does not inhibit Pgp function or cause antibody-dependent cytotoxicity [51]. Thus, even Pab-IR700 binds to Pgp in normal tissues, it would not affect Pgp endogenous function or cause cytotoxicity towards normal cells. In the contrast, the antibody Cetuximab used for EGFR-targeted PDT is an inhibitory antibody [23]. EGFR is expressed in normal tissues of epithelial, mesenchymal and neuronal origin and plays a major role in normal cellular processes such as stimulation of epidermal growth and acceleration of wound healing. Thus, administration of Cetuximab suppresses normal tissue proliferation and differentiation and causes toxicity in normal tissues, including dermatologic and gastrointestinal toxicities [52–54].

Our targeted PDT procedures in this study only involve one cycle of APC administration and light irradiation, and has achieved quick ablation of chemoresistant tumors (Figures 4 and 5). Based on previous studies [55, 56], we expect that multiple cycles of our targeted PDT procedures will further enhance anticancer activity and reduce tumor relapse. In addition, we observed greater tumor response in the tumor region that was closer to the light source (Supplementary Figure 10C), indicating that tissue penetration of light limits tumor response of targeted PDT. Thus, our targeted PDT approach may achieve good responses in the tumors on and adjacent to the skin as well as the lining of internal organs and cavities. Head and neck cancer occurs near the skin and OvCa micrometastases are primarily disseminated on the lining within the peritoneal cavity [57, 58]. They are selected as the disease models in this study because of being accessible for light irradiation. Light delivery approaches need to be improved for treating other tumors that either are large in size or are metastasized. For example, flexible fiber-optic devices can deliver light to the tumors in deep tissues to produce PDT effects [57, 59]. More precise light delivery will be needed for PDT on the tumor regions that are close to normal tissues with accumulation of PSs such as the liver and kidney. For example, intraperitoneal light delivery approaches with narrow-beam lasers have been successfully applied for PDT of peritoneal OvCa tumors that are close to the liver [60–62]. In addition, many PSs including IR700 are NIR fluorescence dyes, and thus image guidance is often applied to enhance the precision of light irradiation. For example, APCs have been successfully used to image disseminated OvCa micrometastases and then to selectively eliminate them with targeted PDT in vivo [62]. Therefore, assisted by advanced light delivery approaches and image guidance, Pgp-targeted PDT can be applied to treat multiple cancer types where Pgp-mediated MDR poses a challenge to treatment.

Heterogeneous target expression in human tumors poses another challenge for targeted PDT. We observed uneven microdistribution of Pab-IR700 in NCI-ADR^Res tumor in xenograft mice (Supplementary Figure 10A) and heterogeneous Pgp expression in the same tumor tissue (Supplementary Figure 10B), likely indicating that heterogeneous intratumoral Pgp expression limited tumor response to targeted PDT. As a potential consequence, targeted PDT selects for tumor cells with low Pgp expression, resulting in tumor relapse. To overcome this limitation, Pgp-targeted PDT needs to be combined with conventional chemotherapy, and thereby our targeted PDT can specifically kill chemoresistant cancer cells while the chemotherapy agents preferentially kill chemosensitive cells. As a proof-of-concept experiment, we tested combination therapy of Pgp-targeted PDT with Taxol (a substrate for Pgp) in a co-culture model of KB-8–5-11 (Pgp-expressing and chemoresistant) and KB-3–1 (Pgp negative and chemosensitive) cells, which mimics human heterogeneous MDR tumors. The result in Supplementary Figure 11 showed that Pgp targeted PDT dramatically enhanced the Taxol’s toxicity and demonstrated a strong synergistic effects of these two modalities (Combination index is smaller than 1 when fraction affected is smaller than 0.75). This result indicated combination therapy of our targeted PDT with chemotherapy may be an effective approach for treating human tumors with heterogeneous Pgp expression.

Systemic tumor responses can be improved by harnessing tumor immunity induced by local PDT procedures. PDT can produce tumor immunity to execute anticancer effects on cancer cells that metastasize to remote sites [63–65]. It has been reported that local PDT induced cell apoptosis and necrosis, which boosted immune response in tumor sites by exposing tumor specific antigens to dendritic cells and thereafter activating cytotoxic T cells. The immune cells with tumor specific recognition are able to amplify and patrol in lymph nodes and blood circulation to identify and kill metastatic or relapsed tumor cells with the same antigen expression [63–65]. In addition, combination of PDT with checkpoint inhibitors elicited antitumor immunity and inhibited distant metastasis at a higher efficacy [66–68]. In light of emerging roles of PDT in advancing cancer immunotherapy, future studies will further unleash the tremendous therapeutic potentials of targeted PDT.

In conclusion, we developed a highly cancer specific PDT approach to overcome Pgp-mediated MDR, where conventional approaches have failed. The studies with in vitro and animal models clearly show that Pgp specific phototoxicity depends on binding of Pab-IR700 to Pgp on the cell membrane of MDR cancer cells and light activation of the APCs. Thus, using our Pgp-targeted PDT approach, superior cancer specificity can be achieved by combining antibody-based cancer targeting and localized light activation of the photosensitizer. This overcomes the main limitation of small-molecule Pgp inhibitor approach and thus is expected to have great translational importance in targeted cancer therapy.

4. Materials and Methods

4.1. Cell lines

Pgp-expressing 3T3-MDR1 is a mouse fibroblast cell line stably transfected with a cDNA coding for the human Pgp. It was obtained from Dr. Michael Gottesman’s laboratory at National Cancer Institute (NCI), and was maintained in DMEM cell culture medium (Corning Inc., Corning, NY, USA) supplemented with 10% fetal bovine serum (FBS, Sigma-Aldrich, St. Louis, USA), 400 IU/mL penicillin, 100 μg/mL streptomycin (Corning Inc.), and 60 ng/mL colchicine (Sigma-Aldrich). NCI/ADR^Res is an adriamycin-resistant OvCa cell line with high Pgp expression, while KB-8–5-11 is a multidrug resistant human KB carcinoma cell line independently selected with colchicine. Both cell lines were obtained from Dr. Gottesman’s lab at NCI, and were maintained in the same condition as 3T3-MDR1 cell line. GFP and/or firefly luciferase-expressing cell lines were constructed by transfection of the cells with reporter-encoding lentivirus (Biosettia, San Diego, CA, USA) according to a standard protocol provided by the vendor. Human hepatocytes derived from patient were obtained from Wake Forest Baptist Comprehensive Cancer Center Tumor Tissue and Pathology Shared Resource. Hepatocytes were maintained in Williams’ E medium (Gibco, Grand Island, NY, USA) containing Maintenance Supplement Pack (Gibco) up to 7 days; prior to Alamar Blue experiment, hepatocytes were plated in 96 plate well in Williams’ E medium containing Plating Supplement Pack (Gibco).

4.2. Synthesis of Pab-IR700

Mouse IgG1 isotype control (IgG) was purchased from R&D Systems, Inc. (Minneapolis, MN, USA). Anti-Pgp monoclonal antibody 15D3 (Pab) was produced in house using the hybridoma cell line from ATCC (Rockville, MD, USA) as described previously [69]. Briefly, hybridoma cells were initially cultured in DMEM media (Corning Inc.) containing 10% FBS (Sigma-Aldrich). The serum content was reduced by serial dilution until culturing in serum-free hybridoma medium (Thermo Fisher Scientific, Rockford, IL, USA). The media containing antibody was collected and the antibody was purified with a HiTrap Protein G HP column (GE Healthcare Life Sciences, Piscataway, NJ, USA). The identity and purity of the antibody was assessed by SDS-PAGE.

To prepare the APCs, Pab or control IgG was incubated with IR700-NHS at molar ratio of 1:4 in phosphate buffer (pH 8.0) for 1 h. The product of the conjugation was purified using a Zeba™ spin desalting column (40K MWCO, Thermo Fisher Scientific, Rockford, Illinois, USA). The protein concentration of the antibody conjugates was determined with BCA protein assay kit (Thermo Fisher Scientific), and the IR700 concentration was quantified by measurement of the absorption at 689 nm with the spectroscopy in order to estimate the number of IR700 molecules conjugated to each antibody molecule. The purity of Pab-IR700 conjugates was examined by SEC-HPLC. SEC-HPLC was performed using an UltiMate™ 3000 UHPLC system (Thermo Fisher Scientific) equipped with an AdvanceBio SEC-300A column (Agilent Technologies, Inc., Santa Clara, CA, USA). All the samples were monitored at the absorbance of 689 nm and 280 nm.

4.3. Flow cytometry

Immunostaining followed by flow cytometry was performed to detect target specificity of antibody conjugates. Cells were cultured overnight, were then trypsinized using 0.25% Trypsin, 0.1% EDTA (Corning, NY, USA), and suspended in PBS buffer. To examine the binding affinity and specificity of antibody conjugates, 1 × 10⁶ of live cells were first blocked with 10% goat serum at room temperature for 10 min, and then stained by 10 μg/mL of antibody conjugates at 4°C for 30 min. Fluorescence of the cells was acquired on an LSRFortessa flow cytometer (BD Bioscience, Franklin Lakes, NJ, USA). Ten thousand events of cells were analyzed and the data was processed using FlowJo software (FlowJo, Ashland, OR, USA).

4.4. Confocal microscopy

Cells were seeded on an 8-well Lab-Tek™ II Chambered Coverglass (Nalge Nunc, Rochester, NY, USA) and were cultured overnight. Cells were treated by 10 μg/mL of antibody conjugates at 37°C for 4 h. The cells were then washed with cold PBS twice, stained with Hoechst 33342 (Thermo Fisher Scientific), and visualized with a ZEISS LSM 710 confocal microscope (Carl Zeiss AG, Oberkochen, Germany).

4.5. In vitro phototoxicity studies

The phototoxicity of antibody conjugates was first examined with Live/Dead staining with a method described previously [70, 71]. Ten thousand cells were seeded in 96-well plates and were cultured overnight. Medium was replaced with the dose solution of 10 μg/mL Pab-IR700 or IgG-IR700. The cells were further incubated for 4 h at 37°C. After washing with PBS, the cells were irradiated with a 690 nm LED light (L690–66-60; Marubeni America Co., Santa Clara, CA, USA) at the light dose of 5 J/cm². Four hours after NIR irradiation, the cells were co-stained with Calcein AM (2 μM) and PI (5 μg/mL) at room temperature for 30 min, rinsed with PBS, and then imaged using a Cytation 5 Imaging Reader (BioTek, Winooski, VT, USA).

The phototoxicity of the nanoconjugates was then quantified using Alamar Blue assay according to a method described previously [71, 72]. Briefly, three thousand per well of cells were seeded in 96-well plates and cultured overnight. Subsequently, the cells were separately incubated with different concentrations of Pab-IR700 or IgG-IR700 for 4 h. Then, drug-containing medium was replaced with fresh medium, and cells were irradiated with the LED light (5 J/cm²). After 24 h, Alamar Blue reagent (Thermo Fisher Scientific) was added and incubated for 2 h. The fluorescence of the samples was then measured on a CYTATION 5 imaging reader (BioTeK) set at 540 nm excitation and 590 nm emission wavelengths.

To further examine Pgp specificity of photokilling with Pab-IR700, 3T3-MDR1 and 3T3-GFP cells were co-cultured at ratio of 1:1 at Lab-Tek™ II Chambered Coverglass. Cells were treated by 10 μg/mL of Pab-IR700 at 37°C for 4 h. After washing with PBS, cells were irradiated with the NIR light (5 J/cm²). After 4 h, the cells were stained with PI, and were then imaged with a ZEISS LSM 710 confocal microscope.

To study the effect of light dose on the photokilling of Pab-IR700, 3T3-MDR1 cells were incubated with Pab-IR700 at increasing concentrations for 4 h, and irradiated using the 690 nm LED light at the light dose of 1.25, 2.5, 3.75, 7.5, and 11.25 J/cm², respectively. After 24 h, cell viability was determined using Alamar Blue assay.

To further confirm the light dependence of the photokilling, 3T3-MDR1 cells were treated by 10 μg/mL of Pab-IR700 at 37°C for 4 h. After washing with PBS, cells were irradiated with a narrow-beam NIR light through a pinhole. Cells were then stained with Calcium AM and PI as described above.

For Annexin V and PI staining, 3T3-MDR1 cells were treated with Pab-IR700, and then irradiated with the 690 nm LED light as described above. At 8 h post irradiation, the cells were trypsinized, washed twice, and stained with FITC labeled Annexin V and PI (BD Bioscience) according to the manufacturer’s instruction. The fluorescence of the stained cells was detected with an LSRFortessa flow cytometer (BD Bioscience).

4.6. Penetration and phototoxicity in tumor spheroids

NCI-ADR^Res-GFP spheroids were grown to study the photokilling of Pab-IR700 in 3D model with a method described previously [73]. Briefly, 1×10⁴ of NCI-ADR^Res-GFP cells in 200 μL medium per well were seeded into Corning 96 well clear round bottom ultra-low attachment microplate (Corning), and cultured for 5 days. To examine the penetration of the antibody conjugates, the spheroids were treated with 20 μg/mL of Pab-640R overnight. Subsequently, some spheroids were washed twice with the fresh medium and imaged using a ZEISS LSM 710 confocal microscope. Fluorescence intensity in the confocal images was quantified with ImageJ software. Other spheroids were digested to single cells and were analyzed for cellular fluorescent levels on an LSRFortessa flow cytometer (BD Bioscience).

To evaluate the phototoxicity, NCI-ADR^Res-GFP spheroids were treated with Pab-IR700 (20μg/mL) overnight. Then the spheroids were rinsed with fresh medium and irradiated with the LED light at the light dose of 10 J/cm². After 48 or 96 h, the spheroids were incubated in Calcein AM/PI solution at 37°C for 30 min. After washing, the spheroids were imaged using a Cytation 5 Imaging Reader (BioTeK). In a separate experiment, the growth of the spheroids was monitored with the CYTATION 5 imaging reader over the next 28 days after Pab-IR700 mediated PDT.

4.7. Animals

All animal experiments were done in accordance with a protocol approved by the Wake Forest Institutional Animal Care and Use Committee. Female Balb/c nude mice (4–6 weeks old) that were purchased from Charles River (Wilmington, MA, USA) were used in the animal studies.

4.8. In vivo Pgp-targeted PDT in a subcutaneous tumor model of chemoresistant OvCa

To establish subcutaneous OvCa models, 5 × 10⁶ NCI-ADR^Res-GFP-Luc cells were suspended in 0.1 ml PBS/Matrigel (BD Biosciences, CA, USA) (1/1, v/v) and inoculated into nude mice bilaterally. When tumors grew to about 60 mm³, mice were randomly allocated into two groups (n = 5) and were injected intravenously with saline and Pab-IR700 (300 μg). Fluorescence images were taken using an IVIS Imaging System for visualization of IR700 at 48 h post Pab-IR700 injection. Regions of interest for tumors were placed to obtain the fluoresce intensity of IR700. After that, the tumors on the right flank of all the mice were exposed to the 690 nm LED light at 25 mW/cm² irradiance and a total dose of 50 J/cm², while the tumors in the left flank were not irradiated, to serve as a control. This light dose was selected based on previous studies using antibody-IR700 conjugates for animal studies [23, 33, 34]. After light irradiation, tumor growth was measured using BLI and by a caliper twice per week. Tumor volume was calculated using the following formula: V = (L×W²)/2, where W is the width and L the length of the tumors measured. Blood, kidney, and liver were collected from the mice 48 h post light treatment. Kidney and liver were paraffin-embedded, sectioned and stained with H&E, while blood was placed in heparinized tubes, then isolated by centrifugation at 3,500 × g, 4°C for 15 min to isolate serum. The serum samples were submitted to the Animal Clinical Laboratory Core Facility at UNC School of Medicine for analysis of AST and ALT levels by automated chemical analyzer (VT 350, Ortho Clinical Diagnostics, Rochester, NY).

To examine biodistribution of Pab-IR700, 5 × 10⁶ NCI-ADR^Res-GFP-Luc cells were suspended in 0.1 ml PBS/Matrigel and inoculated into nude mice subcutaneous. Two weeks after tumor inoculation, Pab-IR700 (300 μg) was injected intravenously to the nude mice. The main organs, including brain, heart, lung, liver, spleen, and kidney, and tumor were harvest 2 days post injection, and the IR700 accumulation in different organs and tumor was imaged with IVIS Lumina III system.

4.9. In vivo Pgp-targeted PDT in orthotopic tumor model of chemoresistant head and neck cancer

To establish orthotropic head and neck cancer models, 2 × 10⁶ KB-8–5-11-GFP-Luc cells were suspended in 0.1 ml PBS/Matrigel (1/1, v/v) and injected into the floor of the mouth via an extra-oral approach. After 5 days, mice were randomly allocated into three groups (n = 8) and were injected intravenously with saline, IgG-IR700, and Pab-IR700 (300 μg), respectively. Fluorescence images were taken using an IVIS Imaging System for visualization of IR700 at 48 h post Pab-IR700 injection. After that, the tumors in the IgG-IR700 and Pab-IR700 groups were exposed to the 690 nm LED light at a total dose of 50 J/cm². After light irradiation, tumor growth was measured using BLI and by a caliper twice per week. Tumor volume was calculated using the following formula: V = (L×W²)/2, where W is the width and L the length of the tumors measured. Mice were euthanized by carbon dioxide inhalation if any tumor volume exceeds 1000 mm³. The body weight of each mouse will be used as a parameter to evaluate in vivo toxicity.

4.10. Immunohistochemical analysis

Two days after the NIR irradiation treatment, some animals were sacrificed and tumor tissues were excised for immunohistochemical analyses. The TUNEL assay was used to detect apoptotic tumor cells after systemic administration. Briefly, tumors were collected and fixed in freshly prepared 4% paraformaldehyde for 1 day. Tumor samples were paraffin-embedded, sectioned and stained with H&E. The TUNEL staining was performed with Click-it™ TUNEL Alexa Fluor™ 594 Imaging Assay Kit according to the manufacturer’s protocol (Thermo Fisher Scientific). Nuclei of tumor cells were stained with DAPI (Thermo Fisher Scientific). Images of the stained sections were taken with a ZEISS LSM 710 confocal microscope.

4.12. Statistical analysis

Quantitative data were expressed as mean ± SD. Means were compared using Student’s t test for two-sample comparison or one-way ANOVA followed by Tukey’s post-hoc analysis for multiple comparisons. P values <0.05 were considered statistically significant. Survival analysis was conducted with Kaplan-Meier curves, and their comparison was determined by Log-rank (Mantel-Cox) Test.