Paraptosis and Photodynamic Therapy: A Progress Report

Abstract

Photodamage to the endoplasmic reticulum (ER) can initiate a death pathway termed paraptosis. The ‘canonical’ model of paraptosis, initiated by certain drugs and other stimuli, requires a brief interval of protein synthesis, involves the action of MAP kinases and can be followed by biochemical markers. The latter include changes in expression of AIP-1/Alix and IGF-1R proteins and translocation of HMGB-1 from nucleus to plasma membrane. There is also a report indicating that an enhanced level of autophagy can impair death by paraptosis. The pathway to paraptosis follows the canonical pathway when ER photodamage is minor (<LD₅₀). When the extent of ER photodamage approaches LD₉₀ levels, there are deviations from the ‘canonical’ pathway: interfering with protein synthesis does not prevent paraptosis nor does a brief chilling of cells after irradiation, MAP kinases are not involved, and stimulation of autophagy was not cytoprotective. We had previously speculated that ER protein crosslinking might potentiate paraptosis [8] but this appears to be incorrect. At higher PDT doses, substantial cross-linking of a typical ER protein (BiP, binding immunoglobin protein, an HSP chaperone) was detected and paraptosis was impaired. This may relate to decreased mobility of cross-linked proteins. Other pathways to cell death were then observed.

Graphical Abstract

Morphology of apoptosis 24 h after ER photodamage to OVCAR-5 cells in vitro

INTRODUCTION

Paraptosis is a death pathway characterized by appearance of multiple vacuoles that gradually fill the cytoplasm. This can occur after a variety of external stimuli including chemotherapy [1–7] and ER photodamage [8–10]. The potential relevance of paraptosis as a route to cancer control is indicated by studies showing that cell lines with an impaired apoptotic program can be eradicated by paraptosis [2,3]. ‘Canonical’ paraptosis is associated with activation of MAP kinases and requires a brief period of new protein synthesis [1]. Several biochemical signals have been identified as paraptosis markers, i.e., relocation of HMGB1 from nucleus to plasma membrane [11] and altered expression of IGF-1R and AIP-1/Alix proteins [3]. Changes in cell morphology can be followed by microscopic examination [7,8].

An example of ‘non-canonical’ paraptosis is illustrated by the effect of taxol, an anti-tumor agent that promotes tubulin polymerization and blocks the progression of mitosis. Paraptosis initiated by taxol does not require either MAP kinases or protein synthesis [12]. Other anomalous effects were observed when glioma cells were exposed to the staurosporin analog NIM811 [13]. Autophagy, initiated by rapamycin, antagonized paraptosis but inhibition of protein synthesis by cycloheximide promoted cell death. Another group examined effects of staurosporin on cervical cancer cells and observed a cytoprotective effect of cycloheximide [7].

ER photodamage appears to initiate canonical and non-canonical pathways to paraptosis. At <LD₅₀ PDT doses, paraptosis was impaired and viability preserved by MAPK antagonists, but these effects were not observed at higher (~LD₉₀) PDT doses [14]. As the light dose was further increased, ER protein cross-linking became more prominent, paraptosis was impaired and cell death involved apoptosis and necrosis [14]. In this study, we examined additional elements of the non-canonical form of paraptosis initiated by ER photodamage, using OVCAR-5 cells in cell culture.

MATERIALS AND METHODS

Chemicals and supplies.

BPD (benzoporphyrin derivative, Verteporfin, Visudyne) was purchased from VWR (Cat No 1711461). Solutions were prepared in DMSO. Cyclosporin A (CsA) and other reagents were obtained from Sigma-Aldrich and Calbiochem Corp. and were of the highest available purity.

Cell culture and clonogenic assays.

OVCAR-5 cells were maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum. The procedure for estimating photokilling by clonogenic assay has been described [15]. Viability was assessed by colony counting using an Oxford Optronix CelCount device. All experiments were performed in triplicate.

Protocols.

Cells were grown on 30 mm square cover slips in plastic dishes. Cells were photosensitized by incubation for 1 h at 37°C with 0.5 μM BPD. The medium was replaced and cells irradiated at 690 ± 10 nm at specified light doses. In some studies, 20 μM cycloheximide was present for 2 hr, or rapamycin (0.1 μM) for 4 hr before irradiation of photosensitized cells. Labeling of nuclei by Hö33342 was used to detect apoptotic morphology. If cells became detached during the 24 h incubation, they were collected by centrifugation and examined after labeling with Hö33342. To assess effects of CsA, OVCAR-5 cells were treated with a 10 μM concentration for 24 h.

Post-irradiation protocols.

Unless otherwise specified, cells were incubated for 24 h at 37°on coverslips in a humidified atmosphere of 5% CO₂ + air. For some studies, cultures were chilled to 10° for 20 min directly after irradiation, then incubated at 37° for 24 h.

Microscopy.

Images were acquired using a Nikon E-600 microscope fitted with a Rolera EM-CCD camera, with the aid of MetaMorph software (Molecular Devices, Sunnyvale, CA) as described previously [8–10].

Western blots.

Cross-linking of an ER chaperone protein (binding immunoglobulin protein, BiP/GRP 78) was determined on western blots 1 h after irradiation at which time cross-linking head reached a steady-state. Protein extracts were prepared with modified RIPA buffer: 50mM Tris-HCl, pH 7.4, 150mM NaCl, 0.25% deoxycholic acid, 1% NP-40, 1mM EDTA. 1mM PMSF, 1mM sodium orthovanadate, 1mM sodium fluoride, 1X protease inhibitor cocktail purchased from Roche (Cat # 11873580001). Immunoblot analysis of BiP was performed using a rabbit monoclonal antibody against BiP (Cell Signaling Tech. Cat.# 3177). An antibody for IGF-1R was obtained from Abcam, Cambridge MA.

RESULTS

Dose-response information

Clonogenic data showing dose-response information on effects of photodamage on survival of OVCAR-5 monolayers photosensitized with BPD are shown in Fig. 1. The shoulder on the curve is a reflection of the pro-survival effects of autophagy [15].

Figure 1.

Dose-response curve for OVCAR-5 cells photosensitized with BPD as described in the text. Each point represents the average of three determinations with the average values and standard deviations indicated.

Paraptosis and ER protein cross-linking

Fig. 2 shows the effects of different light doses on morphology of OVCAR-5 cells 24 hr after irradiation (upper panels) along with the labeling patterns produced by the nuclear probe Hö33342. Paraptotic morphology was observed until the level of photokilling began to exceed an LD₉₀ PDT dose. Cells then began showing evidence condensed of chromatin along with a more intense nuclear labeling by Hö33342. This provides evidence of apoptosis and necrosis at the higher PDT doses.

Figure 2.

Effects of varying light doses on morphology of OVCAR-5 cells 24 hr after irradiation with specified light doses. Survival data are derived from Fig. 1.

Microscopic examination of ‘floaters’ that appear at higher light doses revealed that 80-90% consisted of nuclei surrounded with a trace of cytoplasm (Fig. 3 panels a and c). Some nuclei showed indications of an apoptotic morphology (panels b and d). The latter was confirmed by Hö33342 labeling which indicated evidence of chromatin fragmentation, typical of apoptosis.

Figure 3.

Analysis of ‘floaters’ 24 h after irradiation at 180 mJ/sq cm showing morphology associated with apoptosis (left panels) and paraptosis (right panels). The latter represent 80-90% of the cell fragments examined. Legend: a,b = phase contrast images; c,d = Hö33342 labeled images.

Cross-linking of the ER protein BiP became prominent only at relatively high light doses (Fig. 4). There was lower level of cross-linking at lower PDT doses where paraptosis was detected (Fig. 2). While BiP is only one of many ER proteins, these results are consistent with the proposal that immobilization of ER proteins by cross-linking interferes with the protein mobility necessary for vacuole formation. Studies carried out with another cell line also demonstrated impaired paraptosis when ER photodamage resulted in significant protein cross-linking [14].

Figure 4.

Cross-linking of the ER protein BiP after specified light doses. Cells were collected and analyzed 24 h after irradiation. This represents a typical blot pattern of three such experiments. The band designated ‘a’ represents a BiP dimer and ‘b’ indicates presence of higher oligomers.

Effects of cycloheximide and rapamycin

An example of canonical paraptosis is illustrated in this report by studies involving cyclosporin A. It has been reported that an interval of protein synthesis is required for the expression of paraptosis [1] although there are examples where this was not strictly necessary [12,13]. Bram et al reported that the mTOR antagonist rapamycin promoted autophagy which protected glioma cells from paraptosis initiated by a CsA analog [13]. This analog evokes the canonical model of paraptosis, e.g., a brief exposure to cycloheximide antagonized vacuole formation. Treatment with cycloheximide or rapamycin diminished paraptosis (Fig. 5). These agents offered only slight protection from photokilling at LD₅₀ PDT doses and none at an LD₉₀ PDT dose (Fig. 6). If misfolded or otherwise defective proteins are not ‘cleared’ it appears that cell death cannot be avoided. At PDT doses that exert a strong anti-tumor effect (~LD₉₀), treatment with either cycloheximide or rapamycin did not prevent paraptosis (Fig. 5) and IGF-1R expression was not enhanced, in contrast to the effect of a 24 hr exposure to 20 uM CsA (Fig. 7).

Figure 5.

Morphology of OVCAR-5 cells 24 h after specified PDT doses or treatment with 10 μM cyclosporin A (CsA). Left panels show cell morphology with no further additions. Center panels show the effect of a 2 h incubation with 10 μM cycloheximide prior to irradiation. Right panels indicate effects of a prior 4 hr incubation with 0.1 μM rapamycin. Panels a-c represent controls, d-f = LD₅₀ PDT dose, g-i = LD₉₀ PDT dose, j-l = CsA treatment.

Figure 6.

Clonogenic assays showing effects of cycloheximide (Cx) and rapamycin (Rapa) using the treatment schedule identified in the legend to Fig. 4.

Figure 7.

IGF-R1 expression by OVCAR-5 cells: left, untreated control; center, 60 min after an LD₉₀ PDT dose using BPD as the photosensitizing agent; right, after a 24 h exposure of cells to cyclosporin A.

Chilling cells for 20 min directly after irradiation impaired vacuole formation 24 h later, but only at low PDT doses (Fig. 8). The effect of cooling appears to mimic the response to cycloheximide relating to the likelihood that protein synthesis was briefly impaired. There may, however, be additional consequences involving other steps in the initiation of paraptosis.

Figure 8.

Effect of chilling cells to 10°C for 20 min directly after irradiation, then raising the incubation temperature to 37°for 24 h. Left panels were not chilled. Right panels show effects of the chilling procedure. PDT doses: a,b = 90 mJ/sq cm; c,d = 120 mJ/sq cm; e,f = 180 mJ/sq cm.

CONCLUSIONS

A prior study demonstrated that paraptosis was impaired by MAPK antagonists after low but not high levels of photodamage [14]. This corresponds to the ‘canonical’ route described previously [1–4]. As the level of ER photodamage increases to a level that can eradicate ~90% of the cell population, paraptosis becomes independent of MAPK activity and does not require an interval of protein synthesis. Greater levels of ER photodamage leads to ER protein cross-linking and impaired paraptosis, with death becoming associated with apoptosis and necrosis. These results indicate the presence of a separate route to paraptosis that bypasses the hallmarks of the ‘canonical’ pathway.