Minocycline and photodynamic priming significantly improve chemotherapy efficacy in heterotypic spheroids of pancreatic ductal adenocarcinoma

Abstract

The prognosis for patients with advanced-stage pancreatic ductal adenocarcinoma (PDAC) remains dismal. It is generally accepted that combination cancer therapies offer the most promise such as Folforinox despite their associated high toxicity. This study addresses the issue of chemoresistance by introducing a complementary dual priming approach to attenuate the DNA repair mechanism and to improve the efficacy of the Top1 inhibitor. The result is a regimen that integrates drug-repurposing and nanotechnology using 3 clinically relevant FDA-approved agents (1) Top1 inhibitor (irinotecan) at subcytotoxic doses (2) benzoporphyrin derivative (BPD) as a photoactive molecule for photodynamic priming (PDP) to improve the delivery of irinotecan within the cancer cell and (3) minocycline priming (MNP) to modulate DNA repair enzyme Tdp1 activity. We demonstrate in heterotypic 3D cancer models that incorporate cancer cells and pancreatic cancer-associated fibroblasts that simultaneous targeting of Tdp1 and Top1 were significantly more effective by employing MNP and photoactivatable multi-inhibitor liposomes encapsulating BPD and irinotecan compared to monotherapies or a cocktail of dual or triple-agents. These data are encouraging and warrant further work in appropriate animal models to evolve improved therapeutic regimens.

Graphical Abstract

1. Introduction

Pancreatic ductal adenocarcinoma (PDAC) remains one of the most challenging malignancies due to its aggressive nature and tumor microenvironment [1-4]. Standard single-agent chemotherapy has not improved the lifespan of patients [5]. The effect of chemotherapy is minimal and drug resistance is still a major obstacle among other factors that negatively impact patient survival [6, 7]. In this regard, nanomedicine appears as an emerging field to enhance the efficacy and safety of these toxic chemo drugs. For instance, nanoliposomal irinotecan ONIVYDE^® (nal-IRI), is approved for treating metastatic PDAC and has shown superior anti-tumor efficacy over free irinotecan in several in vivo tumor models, including pancreatic, colorectal, breast, gastric, lung, and cervical cancers [1,8-11]. Moreover, NAPOLI 3 (open-label phase 3 study; employs NALIRIFOX a combination of nal-IRI with fluorouracil, oxaliplatin, and leucovorin) demonstrated modest improvement in overall survival (11.1 months) compared with Gemcitabine + nab-paclitaxel (Gem+NabP; 9.2 months), in previously untreated patients with metastatic pancreatic ductal adenocarcinoma (NCT04083235) [12, 13]. Despite the clinically meaningful gains, the high toxicities associated with these agents make this regimen adversely affect the quality of life. It is, therefore, intriguing to devise new transformative approaches to improve chemotherapy's efficacy while minimizing systemic toxicities [14].

Irinotecan, a pro-drug for carboxylesterase-generated active metabolite (SN-38), stabilizes topoisomerase I (Top1) and DNA adducts by binding to the top1-nicked DNA complex which results in irreversible double-strand breaks promoting cell death [10]. The efficacy of Top1 targeted therapies is mainly limited due to the deregulation of metabolic pathways and changes in enzymes that participate in sabotaging the drug effects. The DNA repair enzyme, tyrosyl-DNA phosphodiesterase 1 (Tdp1; a key repair factor for damage caused by the Top1 poison), resolves the Top1-DNA cleavable complexes to allow DNA replication and cell proliferation. Tdp1 inhibition may therefore increase DNA strand breakage and hypersensitivity to Top1 targeting therapies such as nal-IRI [15, 16]. Clinically the role of Tdp1 and its response to Top1 targeted therapies or the survival benefit is unclear. The upregulation of Tdp1 and Top1 from normal to tumor tissue suggests that these proteins play a significant role in the development or maintenance of tumor cells [15-18]. Some studies have shown the potential of minocycline (MNO), a broad-spectrum antibiotic, to regulate the activity of tdp1 in cancer cells [19-20]. Pommier (SID: 144206734) and Huang et al confirms a reduction in the Tdp1 expression with MNO as does the data in the current manuscript. However, the detailed mechanism of action remains unclear and will be the subject of the next set of studies. In this regard, combination therapies that target multiple pathways are generally accepted to overcome potential resistance [21]. Huang et al. demonstrated this by repurposing minocycline in combination with irinotecan [20]. The combination of irinotecan with minocycline reduced micro-metastases to improve overall survival in an orthotopic xenograft mouse model of human ovarian carcinomatosis [20]. Several ongoing clinical trials are testing the ability of MNO to reduce side effects associated with chemotherapy (e.g., NCT02297412 (breast cancer), NCT01693523 (pancreatic cancer), and NCT02055963 (head and neck cancer)).

Photodynamic therapy (PDT) is a photochemistry-based modality approved for several cancer and non-cancer pathologies [22, 23]. It has emerged as a promising adjunct to other cancer treatments (surgery, chemotherapy, and radiotherapy) and has been demonstrated to have the potential to manage PDAC in pre-clinical and clinical settings [22, 24, 25]. Photodynamic priming (PDP), a natural fallout of PDT, modulates the microenvironment to render tumors more responsive to drug delivery. PDP enhances vascular and cellular permeabilization, depletes stromal content, induces tumor immunogenicity by releasing inflammatory cytokines and improving T-cell infiltration, and circumvents overlapping toxicities when multiple drug regimens are used [23, 26-28]. Light-activable nanoliposomes allow co-delivery of multiple therapeutic payloads to the target site, thus providing exceptional photo-triggered release of secondary agents reducing normal toxicity and improving the treatment outcome [29-33]. Photoactivable nanoliposomal-based PDT has been shown to activate the apoptotic pathway, damage drug efflux pumps such as the ATP-binding cassette (ABC) transporters, and potentiate chemotherapy treatment in pancreatic cancer cells [34, 35].

In this study, we address chemoresistance by developing mechanism-based combination regimens that integrate drug-repurposing strategy with optically activated nanotechnology that allows the minimization of chemo agent toxicity. Our approach employs three clinically relevant and FDA-approved agents with no overlapping toxicities (1) antibiotic minocycline (MNO) (2) photoactive molecule benzoporphyrin derivative (BPD) and (3) Top1 inhibitor (irinotecan; IRI) at subcytotoxic doses (Fig. 1). We introduce complementary dual priming and chemotherapy to deliver a 1-2-3 punch to cancer cells. The first punch attenuates Tdp1, a protein that is over-expressed on pancreatic cancer cells and drives cancer cell proliferation. The second punch leverages PDP by delivering engineered nanoliposomal formulation to modulate the spheroid and trigger chemotherapy release from PMIL resulting in enhanced intracellular concentration of IRI. The third punch enhances inhibition of Top1 (Fig. 1). Individually these three punches are ineffective, however, an appropriate sequential combination of the three synergistically impacts the viability of PDAC+PCAF spheroid.

Fig. 1:

(a) Chemical structures of minocycline (MNO), 20:0 Lyso-PC benzoporphyrin derivative (BPD-PC) and irinotecan (IRI) (b) Conceptual representation and mechanism of PDAC tumor cell death following antibiotic priming and chemotherapy. (c) Minocycline prevents the Tdp1 to counteract the effect of the Top1 inhibitor. Photodynamic priming employing PMIL causes damage to drug efflux transporters prior to triggering chemotherapy release, enhancing the intracellular delivery and efficacy of Top1 inhibition.

2. Materials and Methods

2.1. Materials

All lipids, 1-arachidoyl-2-hydroxy-sn-glycero-3-phosphocholine (20:0 lyso PC), 1,2-distearoyl-sn-glycero-3-phosphocholine (18:0 PC (DSPC)), 1,2-dioleoyl-sn-glycero-3-phospho-(1'-rac-glycerol) (sodium salt) (DOPG), cholesterol, 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy(polyethyleneglycol)-2000 (DSPE-mPEG-2000) were obtained from Avanti Polar Lipids, and Verteporfin (Benzoporphyrin derivative; BPD) was purchased from US Pharmacopeia. Irinotecan (IRI) and minocycline (MNO) were purchased from LC Laboratories and MilliporeSigma. The carboxylesterase enzyme 2 (CES2) activity in PDAC spheroids was evaluated by following the hydrolysis of para-nitrophenyl acetate (p-NPA) by UV-vis absorption spectroscopy as described previously [36].

2.2. PMIL synthesis and characterizations

Photoactivatable multi-inhibitor liposome (PMIL) containing lipidated benzoporphyrin derivative (BPD-PC) photosensitizer, and Top1 inhibitor IRI was developed by our established method [29]. Briefly, prior to liposomal preparation, the photosensitizer BPD was anchored to the 1-arachidoyl-2-hydroxy-sn-glycero-3-phosphocholine (20:0 Lyso PC) through Steglich esterification [24, 37, 38]. The product 20:0 Lyso PC-BPD (lipidated-BPD) was purified as described previously [24, 37, 38].

Thin-film hydration and extrusion method is used for the preparation of all liposomal nanoconstructs with a similar lipid composition. The lipids 18:0 PC (DSPC) (790.14 g/mol), Cholesterol (386.65 g/mol), and DSPE-mPEG-2000 (2803.79 g/mol) were mixed at a ratio of 53.2:45.9:0.3 mol% with 200 nmoles of lipid-anchored (BPD-PC; 1249.72 g/mol). Dried lipid films were hydrated in 250 mM ammonium sulfate solution ((NH₄)₂SO₄). These multilamellar vesicles were subjected to freeze-thaw cycles and sequentially extruded through polycarbonate membranes using a mini-extruder system (Avanti Polar Lipids, Inc.) to prepare small unilamellar vesicles. The external aqueous phase was exchanged on a size exclusion column packed with Sepharose CL-4B (Sigma-Aldrich) pre-equilibrated with HEPES-buffered dextrose (5 mM HEPES, 5% dextrose, pH 6.5). PMIL were prepared by entrapping Irinotecan Hydrochloride Hydrate (IRI.HCl. 3H₂O; 677.18 g/mol) into the photoactivatable liposomes [29]. DSPE-mPEG₂₀₀₀ and DOPG micelles were formed by replacing the chloroform with HEPES-buffered dextrose and were added to the suspension of the pre-formed liposomes (IRI entrapped) to allow post-insertion of the DSPE-mPEG2000 and DOPG micelles to PMIL. Unentrapped IRI was subsequently removed using dialysis (Spectra/Por Float-A-Lyzer G2 Dialysis; MWCO 100 kDa) overnight at 4°C, against 1L of HEPES-buffered saline (5 mM HEPES, 145 mM NaCl, pH 6.5). Liposomes entrapped with IRI only (L[IRI]) were also prepared and purified similarly.

The molar concentration of BPD-PC (ε687 nm = 34,895 M⁻¹ cm⁻¹) and IRI (ε384 nm = 21,835 M⁻¹ cm⁻¹) was determined by diluting liposomes in DMSO and measuring the absorption spectrum using UV–visible absorption spectrophotometry. Entrapment efficiency was determined in all preparations by quantitating IRI and comparing the resulting IRI/phospholipid final ratio to its initial ratio.

Hydrodynamic diameter (nm), polydispersity index (PDI), and ζ-potential (mV) of all liposomal nanoconstructs were measured through Zetasizer Nano ZS Dynamic Light Scattering Instrument (Malvern Instruments). Measurements were performed in triplicates and values were reported as mean and standard deviation.

2.3. Stability and photo-triggered release kinetics of IRI from PMIL

IRI release kinetics under dark conditions and photoinduced drug release were carried out using dialysis (Spectra/Por; 100 kDa cutoff) membranes against 10% fetal bovine serum under dark conditions in an incubator shaker at 37°C. A 690 nm laser (Modulight ML6600), was used for irradiation (25 J/cm², 150 mW/cm²) experiments. During dialysis, samples were collected periodically. IRI was quantified through spectrophotometry (ε384 nm = 21,835 M⁻¹ cm⁻¹).

2.4. Heterotypic spheroids formation and viability assessment

Heterotypic spheroids were developed as described previously [37, 39]. Briefly, human pancreatic ductal adenocarcinoma MIA PaCa-2 cells or AsPC-1 cells (ATCC) and pancreatic cancer-associated fibroblasts (PCAF; a kind gift from Dr. Diane Simeone) were cultured in Dulbecco’s Modified Eagle’s medium (DMEM), supplemented with 10% heat-inactivated Fetal Bovine Serum (FBS, Gibco, Thermo Fisher) and 100 U/mL penicillin, 100 μg/mL streptomycin (Mediatech) in Corning T75 cell culture flasks (Corning), and maintained in a humidified CO₂ atmosphere at 37°C. All cells were confirmed negative for mycoplasma when tested using the MycoAlert Plus mycoplasma kit (Lonza). The MIA PaCa-2 and AsPC-1 cell lines were transduced to express mCherry and PCAF cell line was transduced to express EGFP as previously described [39]. The fluorescence expression of the cell lines was confirmed previously by flow cytometry and fluorescence imaging [39]. Suspended heterotypic spheroids of PDAC (MIA PaCa-2 or AsPC-1 cells) and PCAF cells were developed in 96-well round bottom ultralow attachment plates (Perkin Elmer) at 37°C, at a density of 5 x 10³ cells per well (1:1 ratio) for 24 h. Spheroids were then treated with different therapeutics (as described below).

For viability assessment, the spheroids were imaged with an Operetta CLS high-content analysis system (Perkin Elmer, LIVE configuration), and fluorescence signals were recorded through a 0.16NA 5x air objective at λ_exc = 550 nm, λ_em = 570-650 nm (mCherry) and λ_exc = 475 nm, λ_em = 500-570 nm (EGFP). Images were analyzed via a custom Image J script (https://imagej.net/).

2.5. Immunoblotting

Top1, Tdp1, and CES2 protein expression levels in monotypic or heterotypic spheroids were evaluated using immunoblotting. Total protein lysates were collected from spheroids using RIPA buffer (Thermo Scientific) containing a 1% protease inhibitor cocktail (Calbiochem, Merck Millipore) and 1% phosphatase inhibitors 2 and 3 (MilliporeSigma). Lysates were centrifuged at 13,000 rpm at 4°C to remove cellular debris. Samples of 20 μg of protein were separated and transferred to polyvinylidene fluoride membranes. The membranes were blocked in a blocking buffer (1x TBS, 0.1% Tween 20) and placed in primary antibodies against Tdp1, Top1, or CES2 (Table S2). The next day primary antibody was detected using horseradish peroxidase–linked secondary antibody and visualized with the ChemiDoc^™ MP imaging system (BIO-RAD).

2.6. Confocal microscopy and flow cytometry detection of cellular uptake

Spheroids prepared as described in section 2.4 were incubated with PMIL. After 3, 6, or 12 h of incubation, spheroids were washed three times with 100 μl of respective serum-containing cellular media. Prior to imaging, cells were washed twice with 1XDPBS and stained. Hoechst^® 33342 was used to stain the nuclei of the cells before fluorescence imaging. Images were acquired using a confocal microscope (Olympus FluoView-1000 confocal microscope) through a 4X objective. The nuclei and BPD-PC were visualized using 405 nm (Hoechst and BPD) laser excitation, respectively, with appropriate filters (Hoechst: 425–475 nm; BPD-PC: 655–755 nm).

Flow cytometry was used for the quantification of intracellular uptake after incubations as described previously [39]. Briefly, spheroids were incubated with PMIL, after 1, 3, 6, or 12 h of incubation, spheroids were collected in 1.5 mL tubes and washed with PBS (1000g, 5 min) in the dark. The fluorescence intensity of cell-associated BPD-PC was measured using the BD FACSAriaTM II flow cytometer (BD Biosciences^®). Ten thousand events were recorded and gated for each group using a 405 nm laser and a 610 nm dichroic long-pass filter for BPD-PC. Median BPD-PC emission was quantified using FlowJo^® software (version 10, BD). Data are presented as mean ± SD from six biological replicates for each group.

2.7. Immunofluorescence, image acquisition, and analysis

To assess the impact of minocycline priming (MNP) or photodynamic priming (PDP), heterotypic spheroids of MIA PaCa-2 + PCAF and AsPC-1 + PCAF were incubated as described in sections B and C. After 1, 3, 6, or 12 h of treatment, spheroids were washed three times with serum-containing cellular media. PDAC spheroids were then rinsed with PBS and processed for 5 μm frozen sections. Sections were fixed and incubated overnight at 4°C with antibodies directed against Tdp1, γH2AX, ABCG2, or P-gp, (Table S2). After washing the secondary antibodies (Table S2) were applied. Stained sections were imaged on a confocal microscope (Olympus FluoView-1000 confocal microscope) through a 20X objective, with appropriate filters. Image analysis of fluorescent markers was done on Cellprofiler (4.2.5) [47]. Individual cells were identified using primary object identification. From this point, the intensity of different channels (e.g., DAPI, FITC, and Cy5) was taken at the regions of interest and identified to be cells. The observations were based on the examination of 3 sections from at least 5 spheroids from two experiments.

2.8. L[IRI] treatment without/with Minocycline priming

Dose-dependent response was established with heterotypic spheroids of MIA PaCa-2 + PCAF or AsPC-1+ PCAF that were incubated with MNO, L[IRI], or MNO+ L[IRI], at varying concentrations (0 μM to 1000–3000 μM). After 24 h of incubation, spheroids were washed three times with serum-containing cellular media. The spheroids were imaged 72 h post-treatment and viabilities were determined as described in section 2.4.

To assess MNP impact, heterotypic spheroids of MIA PaCa-2 + PCAF or AsPC-1 + PCAF were incubated with 50 μM of MNO for 1, 3, 6, or 12 h. Following incubation with MNO, spheroids were washed three times with serum-containing cellular media. These spheroids were collected for immunofluorescence (Tdp1; modulation after MNP and γH2AX; DNA damage MNP), image acquisition, and analysis or viability assessments.

For MNO primed L[IRI] treatment response, spheroids were incubated with MNO (50 μM) for 1, 3, 6, or 12 h. Following incubation with MNO, spheroids were washed three times with serum-containing cellular media. L[IRI] (10 μM) was added at this stage in combination with MNO at varying concentrations (0 μM to 300–3000 μM) and maintained in a humidified CO₂ atmosphere at 37°C for 24 h. Spheroids were then washed three times with serum-containing cellular media and imaged 72 h post-treatment, for viability assessment, as described in section 2.4.

2.9. Minocycline and photodynamic priming in combination with irinotecan therapy

Dual primed (MNP and PDP) combination regimen employing PMIL (MNP-PMIL-PDP), PDAC+PCAF spheroids were incubated with L[IRI] (10 μM) or L[BPD-PC] (0.3 μM BPD-PC equivalent) or PMIL (10 μM IRI equivalent, 0.3 μM BPD-PC equivalent) nanoconstructs with/without MNP (50 μM as described in section 2.8 for 1 h). After 6 h of incubation with L[IRI] or L[BPD-PC] or PMIL, spheroids were washed three times with serum-containing cellular media and irradiated using 690 nm laser (Modulight ML6600), with a light dose of 25 J/cm² at an irradiance of 150 mW/cm² for PDP treatment group only. The dark controls were washed with media and maintained in a humidified CO₂ atmosphere at 37°C. Following PDP spheroids were maintained in a dark incubator with a humidified CO₂ atmosphere at 37°C. 72 h post-PDP/dark, spheroids were imaged, and spheroid viabilities were found, as described above.

3. Results

3.1. Minocycline priming sensitizes the heterotypic PDAC spheroids for low-dose chemotherapy

Irinotecan (IRI) is FDA-approved for colorectal cancer and pancreatic carcinoma. However, the effect of IRI, like other Top1 inhibitor therapies can be counteracted by the enzymatic activity of Tdp1. In this study, we are proposing an approach that has significantly increased the effectiveness of low-dose IRI in heterotypic spheroids of PDAC. Heterotypic tumor spheroids are promising in vitro platforms for assessing drug delivery and efficacy in a cellular microenvironment that is clinically relevant. The heterogeneity, cell-to-cell, and cellular-matrix interactions, tight junctions, and chemical gradients make spheroids a suitable platform for testing drug penetration more likely to solid tumor tissues [40-42]. Tumor spheroids developed in this study comprised PDAC cells (AsPC-1; metastatic origin or MIA PaCa-2; primary cancer origin) with pancreatic cancer-associated fibroblasts (PCAF) (see Fig. S1 and S2).

First, we evaluated the toxicity of minocycline (MNO) in these heterotypic spheroids of PDAC+PCAF. As depicted in Fig. S3, MNO remains non-toxic, when the spheroids were exposed at concentrations up to 300 μM and minimally toxic to only AsPC-1+PCAF spheroids at 3000 μM. The dose escalation curve for MNO is unexpected as studies have confirmed the anti-cancer action or IC₅₀ values of MNO at ≤ 200 μM against human cancer cell lines [20]. Similar non-responsiveness is also observed for the monotypic PDAC spheroids (Fig. S3).

To develop a multi-agent strategy, we took advantage of non-toxic minocycline priming (MNP) to reduce the dose and enhance the sensitivity of IRI in PDAC+PCAF spheroids. Therefore, we employ MNP (50 μM), prior to exposing the cells to a fixed dose of L[IRI] (liposome entrapped IRI, 10 μM) and increasing concentrations of MNO (ranging from 0 to 3000 μM). The selected IRI dose does not induce any toxicity in PDAC+PCAF spheroids (Fig. S3). With MNP, significant changes in spheroid viability were observed in a dose-dependent manner (Fig. 2), when compared to the no-priming group, demonstrating that MNP synergizes with L[IRI] in Tdp1-expressing PDAC+PCAF spheroids. It is interesting to observe here that the short MNP time (1 hour) was more efficient in significantly killing the cells in the heterotypic spheroids. Fig. 2 shows that the dose requirement of MNO to achieve 50% viability inhibition for spheroids is reduced significantly by 5.4-fold (IC₅₀ = 16 ± 3 μM) in MIA PaCa-2+PCAF spheroids (Fig. 2d) when compared to the no priming group and by 2.8-fold (IC₅₀ = 1359 ± 182 μM) in AsPC-1+PCAF spheroids (Fig. 2b), when compared to the 6-hour priming (as the absolute IC₅₀ is undefined here). However, similar concentrations of MNO (16 μM or 1359 μM), when exposed to PDAC + PCAF spheroids either alone or in combination with equal concentrations of L[IRI] were observed completely noncytotoxic (Fig. S3). This shows that an appropriate sequential dosing of the therapeutic agents appears as a smart strategy to minimize the effective dose and hence the side effects of chemotherapy. These results agree with our previous studies, where MNO overcame chemoresistance against metastatic disease [20].

Fig. 2: Dose and time-dependent chemosensitizing effect of minocycline priming (MNP).

Heterotypic spheroids of PDAC+PCAF were first primed with MNO (50 μM; for 0, 1, 3, and 6h) prior to the wash and the addition of L[IRI] + MNO. Relative viabilities of heterotypic spheroids (a & c) show that minocycline synergizes with L[IRI] (10 μM) in Tdp1-expressing heterotypic spheroids in a dose-dependent manner. (b & d) The IC₅₀ of MNO significantly reduces up to 5.4-fold as compared to the non-priming group. Results are Mean ± SD (n = 10–12) from 3 biological repeats. one-way ANOVA with a Tukey post-test; **** = P ≤ 0.0001 *** = P ≤ 0.001; ** = P ≤ 0.01; * = P ≤ 0.05).

3.2. Minocycline priming (MNP) decreases the expression of Tdp1

To evaluate whether MNP plays a key role in modulating Tdp1, we investigated the expression of Tdp1 along with Top1 proteins in tumor heterotypic spheroids (Fig. S4). The presence of the PCAF has elevated the protein level (both Tdp1 and Top1) in heterotypic spheroids, however, no correlation was observed between Top1 or Tdp1 levels and IRI sensitivity to these PDAC+PCAF spheroids. Resistance mechanisms such as the downregulation of proteins responsible for drug activation, effects, and transport are the common factors affecting the chemotherapy internalization and retention in cancer cells. Additionally lower conversion of the active metabolite from the prodrug also leads to reduced efficacy of IRI. Understanding these mechanisms for the residual disease is critical for designing appropriate dosimetry for combination therapies.

Tdp1 modulation in heterotypic spheroids by MNP at different time points (1, 3, 6, and 12 h) was observed by immunostaining and confocal imaging. Representative images (Fig. 3) show that MNP for 1 hour decreased the expression of Tdp1 in heterotypic spheroids of PDAC+PCAF but did not significantly modulate the γH2AX expression (the DNA damage marker). Quantitative analysis of Tdp1 immunofluorescence intensity revealed that at 1 h post-priming, MNP significantly reduced the Tdp1 fluorescence intensity by 77% in MIA PaCa-2+PCAF spheroids and by 73% in AsPC-1+PCAF spheroids as compared to no primed spheroids, and then appears to re-maintain its level. The impact of MNP on Tdp1 expression (Fig. 3) is relevant to MNP-based L[IRI] treatment of PDAC+PCAF spheroids (Fig. 2), where the 1-hour priming was observed to be efficient enough to reduce the IC₅₀ value, implying that Tdp1 activity may have suppressed the effectiveness of IRI in no priming group.

Figure 3: Minocycline priming modulates Tdp1 expression in heterotypic PDAC spheroids.

PDAC+PCAF spheroids were primed with MNO (50 μM). (a) Representative immunofluorescence imaging of DAPI (nuclear stain; blue), Tdp1 (green), and γH2AX (DNA damage marker; red), in heterotypic spheroids subjected to no treatment (NT) or MNP (1, 3, or 6 h). (b) Relative Tdp1 levels were found to be significantly lower in the MNP (1 h) group compared with the NT or other MNP groups (3, 6, or 12 h) in both PDAC + PCAF spheroids. Results are Mean ± SD (n = 3–9) from 2 biological repeats. Scale bar = 500 μm. One-way ANOVA with a Tukey post-test; **** = P ≤ 0.0001 *** = P ≤ 0.001; ** = P ≤ 0.01.

Although MNO may modulate Tdp1 expression similarly within the two PDAC+PCAF models, the sensitivity to Top1 inhibitor is markedly different and appears to be less effective in ASPC1+PCAF (see Figure 2) since IC₅₀ value obtained is 85-times more than in MIA PaCA 2+PCAF. This is not surprising since AsPC-1 cells were derived from the ascites of a metastatic PDAC patient whose disease had already shown resistance to both radiation and chemotherapy [26]. Therefore, higher tumorigenicity and chemoresistance are expected.

3.3. PMIL characterization, cellular uptake, and photo-modulation of multidrug efflux transporter proteins

The emerging trend of combining photodynamic therapy with chemotherapy employing multidrug nanoliposomal formulations provides spatio-temporally controlled therapeutic delivery to potentiate the local efficacy of a single treatment [31]. To utilize the full potential of the Top1 inhibitor for PDAC, a photoactivable multi-inhibitor liposome (PMIL) that co-encapsulates a lipidated benzoporphyrin derivative (BPD-PC) photosensitizer along with high payloads of IRI was prepared.

The physical characterizations of the nanoconstructs are summarized in Table S1. PMIL showed a loading efficiency greater than 90% with minimal size changes after IRI loading. Dynamic light scattering (DLS) showed monodispersed L[BPD-PC], L-[IRI], and PMIL constructs. The Z-average diameters and polydispersity index (PDI) of PMIL are 130 ± 2 nm and 0.04 ± 0.02, respectively. The physical stability of PMILs at different temperatures (4 or 37°C) in biological media containing 10% FBS or in HEPES saline was evaluated by measuring the hydrodynamic size and PDI changes. PMILs maintained their hydrodynamic diameter as well as a PDI < 0.1 for up to 7 days (Fig. S5 (a-b)), which highlights the great physical stability of the liposomes. The characteristic absorption band of irinotecan (at 384 nm) and BPD-PC (690 nm) are visible in the absorption spectra of PMIL nanoconstruct (Fig. S5c), demonstrating that co-encapsulation has not altered the characteristics of both agents. We further evaluated the photo-triggered IRI release from the PMIL against 10% FBS at 37°C. IRI was released from the PMIL without laser irradiation at a relatively slow rate (Fig. S5d). In the first 6 h, the total released IRI was only 6.5% under dark conditions. The lipid bilayer entrapping the lipid-conjugated photosensitizer BPD-PC acts to protect and minimize IRI release before laser irradiation. In contrast, the release of IRI was significantly accelerated under laser irradiation (690 nm laser at a fluence rate of 150 mW/cm² for 167 s). The cumulative release of encapsulated IRI (from 6 to 10 h) increased from 22.5% to 62.6% (in the photoactivated group), while only 8.1% of IRI was released from the dark control during the subsequent hours. Fig S5d shows that photoactivation of the PMIL increased the rate of IRI release by 2.8 times. These findings are also consistent with our prior work where we have shown that the mechanism of photo-triggered release of IRI from PMIL is exclusively a photochemical process and was restricted in the presence of singlet oxygen scavengers [24, 30].

Cellular uptake and distribution of PMIL were evaluated in monotopic and heterotypic PDAC spheroids. PDAC spheroids were cultured with media containing the PMIL, and BPD fluorescence was determined by confocal microscopy and FACS (Fig. S6 and S7). To assess the uptake of PMIL within the PDAC+PCAF spheroids, z-stack images of fluorescent BPD were obtained after fixing the spheroids at different time points. BPD-PC fluorescence was detectable in spheroids within the first 3 h, up to 8-fold when compared to the control, and then increasing gradually by 12 h post incubation (Fig. S6). However, there was no significant difference in fluorescence intensity between 6 h and 12 h post-incubation. The same conclusion was illustrated by the observation of median BPD signals by flow cytometry which indicated time-dependent cellular internalization of the PMIL (Fig. S6 and S7). Based on these results, 6 h PMIL incubation time point was selected for the following experiments. Of note, by 72 h, PDAC+PCAF spheroids did not indicate any necrosis confirming that in the absence of light PMIL are not toxic and spheroids still contained actively proliferating cells (data not shown).

Moreover, we quantified the intracellular levels of irinotecan within the spheroids after PMIL incubation (Fig. S8). IRI was extracted by homogenizing the spheroids and then quantified (IRI ug/g of cells) via liquid chromatography/tandem mass spectrometry (LC/MS-MS, Agilent). IRI showed a time-dependent uptake into PDAC+PCAF spheroids with most of the drug accumulation after 6 h of exposure to PMIL (as observed in Fig. S7 where BPD fluorescence is monitored). The amount of IRI increased up to 3.5-fold within the first 3 h post-incubation with PMIL. A dramatic decrease was observed following 12 h of incubation, probably due to the metabolization of IRI. However, the SN-38 levels within the spheroids were not detectable. This might represent the inability of the cells to metabolize the prodrug during this time frame (3-12 h PMIL exposure to spheroids). Prior studies employing MALDI imaging reported SN-38 was only observed once accumulated on the outside region of the spheroids, post-24 h of treatment [44]. In a similar study, Liu et. al. used serial trypsinization and LC/MS-MS, to measure the distribution of IRI and its metabolite in tumor spheroids, and SN-38 was quantifiable only after 12 h of exposure to the spheroids at its IC₅₀ concentration [45]. In our study the significant decrease in IRI amount (μg/g of the cells) at 12 h could be its conversion to its active metabolite SN-38, however, the lower amount of the metabolite can be a reason for the limited detection of the metabolite before 12 h. Increasing the initial concentrations of IRI (> 10 μM) in the spheroids may produce a detectable range of SN-38.

One of the most identified mechanisms of chemoresistance is the expression of ATP-binding cassette transporters (ABCG2) and multidrug resistance protein 1 (also known as P-glycoprotein, P-gp) that attenuate the drug response. Photodynamic modulation of these transmembrane transporters has been reported as a possible approach to reduce ABCG2-mediated resistance in PDAC+PCAF. However, free photosensitizers (e.g., BPD, PpIX) are also a substrate of ABCG2, therefore overexpression of ABCG2 potentially leads to less efficiency in the photodynamic action. To circumvent the latter, the phospholipidation approach of the photosensitizer (BPD-PC) and its nanoliposomal formulation has been shown to overcome to evade P-gp and ABCG2-mediated transport [46]. To demonstrate whether PDP has a significant influence on the expression of both transmembrane proteins, PDAC+PCAF spheroids harboring PMIL were first illuminated, fixed, and co-stained with antibodies to the established vesicular markers and compared with the fluorescence signal intensity of untreated spheroids. Fig. 4 demonstrates that noncytotoxic PDP (25 J/cm²) significantly decreased the ABCG2 and P-gp immunofluorescence signal in both PDAC+PCAF spheroids by approximately 2-fold within 3 h post-PDP thus demonstrating that PDP could overcome this pathway of resistance to IRI.

Fig. 4: ABCG2 and P-gp modulation in two heterotypic PDAC spheroids.

PDAC+PCAF were primed with PDP employing PMIL (25 J/cm²; 150 mW/cm²) (a) Representative immunofluorescence imaging of DAPI (blue), P-gp (green), and ABCG2 (red), in MIA PaCa-2 + PCAF (a) or AsPC-1+PCAF (b) subjected to no treatment (NT) or 3-hour post-PDP. (b) Relative P-gp and ABCG2 levels were found to be significantly lower in the 3-hour post-PDT group compared to the NT groups PDAC+PCAF spheroids. Scale bar = 500 μm.

3.4. Dual priming initiates cell death at sub cytotoxic doses of therapeutics

The potential of MNP to modulate the DNA repair enzyme expression, the impact of PDP priming on efflux transporters, and the triggered release of Top1 inhibitor from PMIL, provided us a compelling rationale to assess the combination of MNP, PDP, and IRI at nontoxic doses.

Cancer cell death in monotypic and heterotypic spheroids was evaluated. We selected doses for MNO (50 μM), PDP (25 J/cm²; 150 mW/cm²), and IRI (10 μM) that do not induce cancer cell death as a single agent (Fig. S9). After 6h of PMIL incubation, the spheroids were washed with fresh media (3X) and exposed to NIR light (690 nm). Of note, a dramatic decrease in cell viability (up to 65-85%) 72 h post-irradiation was only observed when dual priming (MNO-primed PMIL-PDP) was applied in these heterotypic PDAC models. The co-administration of single agents plus light MNO + L[BPD-PC] or L[BPD-PC] + L[IRI] (Fig. S9) does not induce any toxicity. The latter emphasizes that the triple combination where tumor spheroids are primed first by MNO and second by light gives superior outcomes of 3D heterotypic tumor nodule destruction at lower IRI concentrations. Clinically, 70 mg/m² (equivalent to 23.3 mg/kg) of liposomal IRI injection is used and 200 mg of MNO is administrated to the patient to achieve a peak plasma concentration (C_max) of 3–4 μg/mL [47]. Our study utilizes 10 μM L[IRI] concentration (1.2–1.4 μg/g cells at 6 h incubation, Fig. S8) and 50 μM of MNO (0.27 μg). The role of PDP here is crucial in enhancing the effectiveness of both agents while mitigating harm to healthy tissues using non-toxic concentrations. This is consistent with previous findings by us and others where a reduction of IRI concentration by photoactivation was reported [24,48].

To unravel the drastic and different response to the triple combination in the mono- and heterotypic 3D models, we evaluated the activity of carboxylesterase enzyme 2 (CES2), which is involved in the conversion of IRI into the active metabolite SN-38. The latter was correlated with MNP-PMIL-PDP efficacy where a linear trend was surprisingly found (R = 0.98, P = 0.0007; Fig. 6). In this regard, those models which exhibit the highest CES2 activity (MIA PaCa-2+PCAF spheroids) also have the most remarkable cell viability decreased (Fig. 6). AsPc-1 cells shows a low CES2 activity which is aligned with the poor response commented above. It is also interesting to observe that the presence of PCAF has elevated the protein level (Fig. S4), its activity, and MNP-PMIL-PDP efficacy in both heterotypic models. In PDT the light is directed and fibroblast death within and around the tumor milieu is one of the goals, as these cells supply vital survival factors for cancer cells. In fact, fibroblasts are held responsible for the desmoplasia in PDAC which is a major barrier in the treatment of the disease. Increased cross-linking and stiffening of the tissue, which restricts the infiltration of chemotherapy in tumor cells and contributes to chemoresistance. In this study we permeabilize the stroma, reduce the expression of TdP1 and reduce the ABCG transporters thus setting the stage for a more efficient treatment. The importance of these results lies in the fact that the MNO-primed PMIL-PDP could circumvent any treatment resistance that is typically conferred by the presence of PCAF. Photodynamic action serves as a complementary therapeutic option to treat cancers resistant to classical toxic chemotherapeutics. MNO may play a key role not only in modulating Tdp1 expression but also could be involved in the PCAF-cancer cell crosstalk [49]. For instance, some studies reported that MNO decreases the proliferation of the proinflammatory cytokine IL-6, the expression of VEGF, TGF-β and matrix metalloproteinase (MMP)-9 which are consistently implicated in the process of tumorigenesis [50-53].

Fig. 6: Analysis of PMIL sensitivity to carboxylesterase 2 (CES2) activity in pancreatic cancer cell lines.

(a) Relative viability of monotypic and heterotypic spheroids following MNP (MNO; 50 μM) for an hour before PMIL (0.3 μM BPD-PC equivalent; 10 μM IRI equivalent) incubation and PDP (25 J/cm²; 150 mW/cm²). (b) Pearson’s correlation analysis between relative viabilities of monotypic and heterotypic spheroids from MNP-PMIL-PDP groups and CES2 activity values in these spheroids. p-NP = para-nitrophenol. Results are Mean ± SD (n = 12–16) from 3 biological repeats. one-way ANOVA with a Tukey post-test; **** = P ≤ 0.0001 *** = P ≤ 0.001).

4. Conclusions

Combinations of treatment options in cancer therapeutics are a fairly well-accepted approach. In this study we address chemoresistance by developing mechanism-based combination regimens that integrate drug-repurposing strategy with optically activated nanotechnology that allows the minimization of chemo agent toxicity. Our approach employs three clinically relevant and FDA-approved agents with no overlapping toxicities (i) antibiotic minocycline (MNO) (ii) photoactive molecule benzoporphyrin derivative (BPD) and (iii) Top1 inhibitor (irinotecan; IRI) at sublethal doses. We introduce complementary dual (2-punch) priming by employing MNO to modulate the Tdp1-based DNA repair mechanism and photoactivation to improve the efficacy and delivery of the Top1 inhibitor.

Collectively, this in vitro study indicates that dual priming along with Top1 inhibition can enhance cancer cell death in heterotypic PDAC spheroids compared to the use of single agents in combination. Moreover, Therefore, it validates the strength of developing mechanistically cooperative combinations, where each therapeutic agent enhances and benefits from the other agent. The findings of this study using agents targeting different molecular pathways and their ability to reduce the concentration of the highly toxic chemotherapeutic irinotecan are encouraging. The tested triple therapy modality needs to be explored further in appropriate in vivo models.

Fig. 5: Dual priming increases IRI effectiveness in PDAC+PCAF spheroids at subcytotoxic doses.

(a) Schematic representation of the treatment regimen of heterotypic PDAC+PCAF spheroids followed by an imaging-based analysis of treatment response. Relative viabilities of heterotypic spheroids (b) (AsPC-1+PCAF) and (c) (MIA PaCa-2+PCAF) following MNP (MNO; 50 μM) for 1 h prior to the addition of PMIL (0.3 μM BPD-PC equivalent; 10 μM IRI) for 6 h and PDP (25 J/cm²; 150 mW/cm²) or L[IRI] (10 μM). Results are Mean ± SD (n = 8-12) from 2 biological repeats. One-way ANOVA with a Tukey post-test; **** = P ≤ 0.0001 *** = P ≤ 0.001.

Highlights.

• Single-agent therapy does not induce any survival improvement in PDAC patients

• Minocycline effectively targets the DNA repair mechanism

• Photodynamic priming effectively damages efflux transporter proteins and induces the safe delivery of irinotecan

• Minocycline and photodynamic therapy priming is a potent therapy to reduce the toxic dosage of irinotecan

We declare that this manuscript is original, has not been published before and is not currently being considered for publication elsewhere. We confirm that the manuscript has been read and approved by all named authors and that there are no other persons who satisfied the criteria for authorship but are not listed. We further confirm that the order of authors listed in the manuscript has been approved by all of us.