Experimental investigation of a combinational iron chelating protoporphyrin IX prodrug for fluorescence detection and photodynamic therapy

Photodynamic therapy (PDT) is based on the activation of a photosensitizer by visible light and the transfer of energy to molecular oxygen, creating reactive oxygen species (ROS) that induce oxidative stress and thus cell death [1]. 5-Aminolevulinic acid (ALA)-induced-PDT exploits the cell’s own innate heme biosynthesis pathway to generate the natural endogenous photosensitizer protoporphyrin IX (PpIX; the precursor to heme). Free heme inhibits further ALA synthesis, and this natural regulation step can be bypassed by purposefully adding exogenous ALA, which then drives the pathway to operate at maximal capacity. Cellular accumulation of PpIX occurs as the final step (the insertion of ferrous iron (Fe²⁺) into PpIX by ferrocheletase to form heme) is relatively slow [1]. This process occurs at a faster rate in malignant cells (which generally have upregulated and less well controlled heme biosynthesis) than in normal cells and so a temporally selective window occurs when red light (635 nm) can be applied to activate PpIX within the tumor, while sparing surrounding normal cells [1].

Within dermatology, a topical cream containing the small, soluble precursor to PpIX (ALA or its methyl ester, methyl-aminolevulinate (MAL; Metvix, Galderma, UK)) is applied to the area of skin to be treated [2]. Although effective treatment outcomes associated with excellent cosmesis have been demonstrated in licensed lesions (actinic keratosis, Bowen’s disease and superficial basal cell carcinoma (BCC)) when the disease remains superficial [3‐5] or alternatively in mild to moderate inflammatory acne [6], efforts continue to increase the efficacy and extend the applications of topical PpIX-PDT particularly to treat thicker nodular BCC or acrally located conditions [7]. Poor penetration into deeper layers can be improved clinically by employing more lipophilic ALA derivatives (e.g., MAL) [8‐10] or nano-emulsion formulations (e.g., ALA; Ameluz, Spirit Healthcare, UK) [5] and by performing skin pre-treatments [11‐13]. There is also interest in applying PDT beyond dermatology, in a wide range of oncological/non-oncological applications [14].

PpIX’s natural fluorescence can also be employed for photodetection of bladder cancer [15‐17] or fluorescence-guided surgical resection of high-grade glioma of the brain using orally administered ALA [18], as PpIX exhibits characteristic red fluorescence (at 635 nm and 700 nm) when excited by blue light (410 nm). Cells that have accumulated PpIX can therefore be identified through non-invasive fluorescence monitoring [19]. The more lipophilic hexyl ester derivative, HAL (Hexvix, Photocure, Norway) [20] doubled PpIX accumulation in a human bladder cancer photodetection pilot study, with PpIX being evenly distributed throughout the urothelial layers [15, 21]. A subsequent clinical study of HAL for urothelial cell carcinoma also demonstrated promising results [22], and photodetection of bladder cancer with HAL now has FDA and European approval [23]. Although this technique has increased tumor detection, the long-term impact on recurrence rates still remains unclear [24].

Using an iron chelating agent to reduce intracellular iron can inhibit the conversion of PpIX to heme, resulting in greater PpIX accumulation. ALA and hydroxypyridinone iron chelating agent co-administration has been previously demonstrated to successfully increase PpIX-PDT efficacy in vitro, in vivo, and during clinical application [25‐28]. 1,2-diethyl-3-hydroxypyridine-4-one (CP94) has been demonstrated to be the best iron chelator for this application as it diffuses through the cell membrane better than other clinically available iron chelators and has a greater selectively for iron than other transition metals [29, 30]. A novel-combined conjugate, AP2-18, which links ALA to CP94 via a cleavable ester (Scheme 1) has been derived [31] to overcome co-administration issues by releasing the two agents simultaneously following cellular uptake and ubiquitous cytosolic esterase release [32, 33]. Promising initial experimental findings of AP2-18 activity within human normal skin fibroblasts and skin cancer cells [34, 35] have been extended here as well as investigated within human cells of bladder origin for the first time to assess the feasibility of AP2-18 as an alternative to ALA/MAL administration in skin or HAL application in bladder for fluorescence photodetection and/or PDT treatment.

Normal human primary bladder epithelial cells and two types of human urinary bladder transitional cell carcinoma (RT-4, from a well differentiated tumor stage T2/grade G1 and aggressive RT-112, from a tumor grade G2) were cultured aseptically in 96-well plates for experimentation (primary bladder cells and RT-4, ATTC, Germany; RT-112, ECACC, UK). Additional experimentation was conducted with human primary epithelial squamous carcinoma cells (A431, ECACC, UK).

Initial investigations tested AP2-18 against the clinically approved prodrugs ALA and HAL (250, 500 or 1000 µM) ± the hydroxypyridinone iron-chelator, CP94. Fluorescence intensity (excitation λ = 405 ± 12 nm; detection λ = 635 ± 12 nm) was measured hourly (in arbitrary units) with a Bio-tek Synergy HT plate reader (Swindon, UK) as an indication of PpIX in RT-112 cells (n = 5) over 10 h at 37˚C following the generation of standard curves (Fig. 1a–c) to demonstrate that fluorescence was linearly aligned to synthetic PpIX (Sigma, UK) concentration. Further PpIX fluorescence experimentation utilized up to 1500 µM AP2-18 and monitored all three bladder cell types (n = 5) as well as skin cancer cells up to 6 h (a minimum of in triplicate on three separate occasions) with MAL replacing HAL in this cell type as MAL is utilized clinically within dermatology (unlike HAL). Background low light intensity conditions were employed in the laboratory as a precaution throughout all experimentation to limit potential photosensitizer photobleaching.

Photodynamic effects were investigated within all cell types after 6-h incubation with the respective prodrugs, to match earlier investigations and because PpIX accumulation observed appeared consistent up to this time point. Preliminary experimentation established that 500 µM ALA and 250 µM HAL were most effective in the bladder cell types. These concentrations were therefore adopted for continuity and because higher concentrations did not significantly increase RT-112 PpIX accumulation. For the evaluation of the new composite drug AP2-18, 250–1500 µM was employed in the three bladder cell types and ≤ 1500 µM in the skin cancer cells using a clinical Aktilite LED array (Galderma, UK) as the PDT light source. Bladder cells (n = 5) were irradiated 6 h after treatment (λ = 635 nm; dose = 30Jcm⁻²) and cell viability/death assessed 16 h later. The skin cancer cells were treated similarly except the clinical dermatology light dose of 37Jcm⁻² was employed with 500 µM MAL and 250 µM ALA (because these concentrations previously produced effective results in this cell type). Appropriate dark (drug only), blank (no drug), and positive (lysed) control groups were also conducted in parallel to the test groups.

Sixteen hours after PDT, the 96-well plates were centrifuged (900 rpm; 3 min), and the adherent cells utilized for the neutral red assay. As viable cells take up neutral red, the dye can be extracted, and its color intensity utilized to indicate how many survived [36]. A 7.5-µl neutral red solution was added to each well for 2 h. The supernatant was removed; the cells washed twice with phosphate buffered saline (PBS), and the neutral red dissolved with 200 µl solubility buffer (49 parts H₂O, 50 parts 99.9% ethanol and one part glacial acid). Absorbance as a measurement of cell viability was recorded at 540 nm. The number of surviving cells was presented in reference to the untreated positive control, which was set to 100%. The increase in cell death after PDT in comparison to the control cells was in all cases significantly higher, which confirmed the cytotoxic effect of PDT. Statistical analysis was conducted utilizing the Welch two sample t-test (within Microsoft Excel), with p-values < 0.05 being considered significant.

Confocal laser scanning microscopy (Leica TCS SP8 with HyD; Milton Keynes, UK) was undertaken on all three bladder cell types in a separate but complimentary series of experiments. Sterile coverslips were placed in a 6-well plate and coated with attachment factor (Sigma, UK). Fifty thousand cells in 2 ml media were seeded onto the coverslips and incubated (37 °C/5% CO₂) overnight. Treatment was conducted as above before the cells were fixed. Growth media was removed, and 4% para-formaldehyde in PBS pH7 (PFA) added to just cover the cells for 10 min at room temperature. This was removed, and the coverslips washed twice for 5 min with PBS/0.2% triton to permeabilize the cells, followed by a further three washes with PBS to remove any traces of detergent/PFA. Antibody and subsequent incubation steps took place in a humidified chamber. Phalloidin 488 (Abcam, Cambridge, UK; diluted 1:1000 in PBS/0.2% triton for 1 h at room temperature and under light protection) and Hoechst 33,342 (Sigma, UK; final concentration of 1:10,000 for the last 15 min) stains were utilized to highlight f-actin and DNA respectively. Caspase-3 was also qualitatively visualized via immunohistochemistry conducted with anti-human cleaved caspase-3 and Alexa 546. Cells were blocked with 5% goat serum in PBS/0.2% triton for 1 h at room temperature. This was removed before incubation with rabbit anti-human cleaved caspase-3 (Cell Signaling Technology, London, UK; 1:400 in 5% goat serum in PBS/0.2% triton) overnight at 4˚C. The next day, the primary antibody was washed off thoroughly with PBS/0.2% triton and PBS before adding the secondary antibody (antibody goat anti rabbit Alexa546 (Invitrogen, ThermoFisher, Paisley, UK; 1:400 in PBS) and incubating for 4–6 h at room temperature. For counterstaining, phalloidin 488 and Hoechst 33,324 were added for the last 1 h and 15 min respectively, as described above. After the incubation steps, the coverslips were washed three times with PBS/0.2% triton and another three times with PBS before being mounted on microscope slides with ProLong Diamond Antifade (Invitrogen, ThermoFisher, Paisley, UK). For live cell images, 30,000 RT4 cells were seeded into a 35-mm diameter imaging µ-dish with a glass bottom (Ibidi, Graefelfing, Germany) and treated with 500 µM AP2-18 for 6 h (this concentration of AP2-18 was selected after considering the findings presented in Figs.1d–g and 2 in detail and concluding that the higher doses investigated did not convey additional benefit at this time point). Treatment media were replaced with RPMI/5% FBS/10% HEPES without phenol red. For single scan live cell imaging, the incubation box that houses the stage and microscope head was heated to 37˚C.

Previously intercepting heme synthesis through iron chelator addition significantly enhanced PpIX accumulation/PDT cytotoxicity in human skin cancer and normal fibroblasts [26, 28, 29]. This was not replicated in the normal and malignant bladder cells investigated here (Figs. 1g and 2), although it was demonstrated once again within human skin cancer cells (Figs. 3 and 4). The levels of fluorescence recorded were more dependent on the bladder cell type than on the treatment variations (Fig. 1g). Doubling prodrug concentrations in RT-112 cells did not double PpIX accumulation (Fig. 1d–f). RT-4 cells produced significantly higher levels of PpIX than other cell types (Fig. 1g), possibly because of their metabolic rate. These higher levels did not necessarily translate into significantly higher PDT cytotoxicity (Fig. 2). Instead, even the modest level of RT-112 PpIX accumulation already reduced the cell viability produced on irradiation by ~ 66%. Average RT-4 cell death was 85%. So, although PpIX accumulation was increased by nearly 200%, RT-4 cell death was only increased by 20%, indicating that in different cell types, different levels of cell damage may be required to induce cell death.

Iron chelator addition, either as separate CP94 or prodrug linked as AP2-18, did not seem to have any consequential influence, either on PpIX fluorescence (Fig. 1g) or bladder cancer cell death following PDT (Fig. 2). This general lack of iron chelator enhancement probably resulted in the new combinational prodrug AP2-18 not conveying any significant additional benefit over and above the existing prodrugs in these bladder cells (unlike observed in dermatological cells [34, 35]), although it was equally as effective as separate co-administration.

The normal bladder cells were also as susceptible to PDT-induced cytotoxicity as the cancer cells (Fig. 2; ~ 71%) which would not be desirable for clinical photodetection/PDT application. To our knowledge, there is no record of investigating normal primary bladder cells for PDT, with little PDT effect on immortal UROtsa cells and N1 fibroblasts PDT-resistance being reported instead [37‐39]. Here, ≤ 500 µM AP2-18 produced the least amount of PpIX in normal bladder cells (Fig. 1g) and correspondingly had the highest survival rate post-PDT (Fig. 2). As the same treatment parameters produced satisfactory RT-4 cancer cell cytotoxicity (Fig. 2), this may indicate that it is possible to improve the normal versus malignant contrast at this concentration.

In skin cancer cells, iron chelation consistently, significantly (p < 0.001) improved PpIX accumulation (Fig. 3) when separate CP94 was co-administered with ALA/MAL but was most pronounced when the combinational iron chelating PpIX prodrug AP2-18 was employed. 500 µM AP2-18 significantly increased (p < 0.001) PpIX accumulation above that produced by all other treatment regimens, corresponding with previous findings [34, 35]. Following PDT (Fig. 4), although the highest AP2-18 concentration produced a visual trend towards greater cytotoxicity, this only reached statistical significance (p < 0.01) when considered against MAL + CP94. Increasing the AP2-18 dose further may have resolved this but generally iron chelation works best when the treatment parameters are limited and thus sub-optimal [35]. The standard prodrug regimens utilized were already highly effective (Fig. 4) killing ≥ 75% of skin cancer cells without any augmentation being necessary. Future work should therefore continue to test AP2-18 within dermatological cells and explore more limited/sub-optimal standard prodrug treatment regimens with shorter drug-light intervals to try to replicate the clinical scenario at the base of thicker nodular BCC tumors where the current licensed protocol is currently producing inadequate results (without repeat treatments).

Battah et al. [32] found that ALA and CP94 directly linked via an ester function (like AP2-18) was the least efficient in a series of conjugates that they synthesized but was slightly less hydrophilic than ALA and so should be slightly better at entering cells [32]. Hydrophilicity could be decreased further by extending the ester bridge between ALA and CP94 but this would also increase molecular size. Many experimental factors differed (cell type; culture conditions; light source), making direct comparison difficult. No additional iron was employed here for instance as this is not required for effective tissue culture, and it is also hard to anticipate the optimal cytosolic iron levels required [40, 41]. Adding excess Fe²⁺ would selectively favor the chelating compound, while quenching PpIX accumulation in the other groups without necessarily reflecting the cells’ natural environment/capabilities.

It is also important to consider the potential deleterious effect that adding ever increasing amounts of iron chelating compounds could have on this complex series of biochemical events. Once the heme biosynthesis pathway is operating in an optimal manner for the maximal PDT effect possible, with very limited ferrocheletase activity left remaining to inhibit through additional iron chelation, further sequestering of cellular free iron could result in reduced ROS production and thus less cell death on irradiation, as iron can catalyze ROS generation following PpIX-induced PDT [41].

Palasuberniam et al. [42] have determined that ferrochelatase activity is an important determinant of tumor response to iron chelation with deferoxamine (DFO). Knock down cell types without ferrochelatase activity were more sensitive to ALA-PDT, but this alteration completely abolished the enhancement effect of DFO iron chelation. Although the ferrochelatase activity of the specific cell types employed here has not been established, it is postulated that this factor may at least in part explain the differing findings observed between the human cells of bladder and skin origin. A431 epithelial squamous carcinoma cells are known to express high levels of epidermal growth factor receptor (EGFR) [43] and lactase dehydrogenase (indicating elevated glycolic metabolism) [44] and strong expression of the PPARGC1A gene (suggesting mitochondrial biogenesis and oxidative phosphorylation) [44]. Although ALA-PDT has been investigated in RT-112 bladder cancer cells previously [45], we are not currently aware of their ferrochelatase activity, and so this would be an interesting avenue of future research.

Confocal microscopy imaging also revealed that prodrug-induced PpIX accumulation alone already destabilized the actin cytoskeleton in bladder cancer cells (particularly RT-112) prior to PDT (Fig. 5). Iron depletion (which could result from more PpIX being available than normal to bind Fe²⁺) can indirectly lead to f-actin depolymerization [46] and again would reduce the potential potentiating effects of iron chelation. The primary bladder cells, however, managed to maintain their cytoskeletal integrity under prodrug treatment. PDT further damaged the cytoskeleton in bladder cancer cells. Visualization of caspase-3 release post-PDT indicated that both bladder cancer cell types were in the early stages of apoptosis, whereas absence of caspase-3 in normal bladder cells suggested cell death from primary necrosis.

Although all tested prodrugs raised PpIX levels significantly in treated cells and caused considerable cell death after PDT, none of the prodrugs ± CP94 iron chelation or the combined conjugate AP2-18 stood out as an optimal treatment regime in the bladder cells investigated. This is in direct contrast to the findings with human skin cancer cells, where both separate CP94 iron chelating supplementation to standard prodrug administration as well as the novel combinational iron chelating PpIX-prodrug AP2-18, conveyed significant advantage particularly in respect to PpIX accumulation. It is therefore concluded that the mechanism of action of PpIX-induced PDT is both complex and highly cell type dependent, and so potential enhancement techniques such as iron chelation cannot be universally applied without thorough prior investigation. These findings support further investigation of iron chelator enhancement of PpIX-induced PDT for dermatological applications, however, not bladder photodetection/PDT.