Dormant cancer cells accumulate high protoporphyrin IX levels and are sensitive to 5-aminolevulinic acid-based photodynamic therapy

Photodynamic therapy (PDT) and diagnosis (PDD) using 5-aminolevulinic acid (ALA) to drive the production of an intracellular photosensitizer, protoporphyrin IX (PpIX), are in common clinical use. However, the tendency to accumulate PpIX is not well understood. Patients with cancer can develop recurrent metastatic disease with latency periods. This pause can be explained by cancer dormancy. Here we created uniformly sized PC-3 prostate cancer spheroids using a 3D culture plate (EZSPHERE). We demonstrated that cancer cells exhibited dormancy in a cell density-dependent manner not only in spheroids but also in 2D culture. Dormant cancer cells accumulated high PpIX levels and were sensitive to ALA-PDT. In dormant cancer cells, transporter expressions of PEPT1, ALA importer, and ABCB6, an intermediate porphyrin transporter, were upregulated and that of ABCG2, a PpIX exporter, was downregulated. PpIX accumulation and ALA-PDT cytotoxicity were enhanced by G0/G1-phase arrestors in non-dormant cancer cells. Our results demonstrate that ALA-PDT would be an effective approach for dormant cancer cells and can be enhanced by combining with a cell-growth inhibitor.

Scientific RepoRts | 6:36478 | DOI: 10.1038/srep36478 Tumor cells in the human body interact with neighboring cells and are present in a high cell density. However, cancer cells grown in a 2D monolayer culture are at a low density and lose tissue-specific properties 19,20 . A 3D-cultured cancer spheroid achieves much higher cell density than 2D culture and reproduces several parameters of the tumor microenvironment, including oxygen and nutrient gradients and the development of dormant tumor regions 15,19,21,22 . Furthermore, high cell density leads to contact inhibition. Since Contact inhibition suppresses cell growth 23 , cancer spheroids possibly enter the dormant state. Dormant cancer cells are insensitive to most treatments, although there are no reports between those cells and ALA-PDT effect. We speculated that  Cells were incubated with glucose and FBS-free medium for 24 h, after which 100 μ M 2-NBDG was added, followed by incubation for 25 min at 37 °C. (g) Fluorescence intensity of 2-NBDG in the 2D culture. (h) 2-NBDG uptake in the 3D culture was analyzed by confocal fluorescence microscopy. (i) Fluorescence intensity of 2-NBDG in the 3D culture. (j) Re-growth of degraded spheroids in the 3D culture compared with that in the 2D culture. Cells were incubated for 4 days to form S500 spheroids and for the same period for the porphyrin metabolism after cell dormancy and ALA treatment are simultaneously regulated in cancer spheroids in a 3D culture. In this study, we tested this hypothesis.

Results
Cancer spheroids were formed in EZSPHERE 3D culture plates. An aliquot of 5 × 10 5 PC-3 prostate cancer cells was seeded in every 35-mm dish of the EZSPHERE 3D culture plate. During seeding, the cell number was the same in each microwell of the culture dish. Four days after seeding, approximately 2,300 spheroids had formed in each dish. The average diameter of the spheroids was 189 ± 30.1 μ m (Fig. 1a). Hoechst staining of spheroids revealed a cell density higher than that in a 2D culture and a spheroid thickness of approximately 80 μ m (Fig. 1b). This thickness is sufficient for oxygen diffusion because spheroids above 200 μ m in diameter experience hypoxic conditions at the core 22 . Similarly, we formed two other sizes of spheroids by reducing the number of seeded cells (Fig. 1c). To create different cell density conditions in a 2D culture, cells were seeded in every 35-mm flat dish to a specified cell density and cultured for 4 days (Fig. 1d). The cellular protein amount has a positive relationship with cell density (Fig. 1e). We used 3D cultured spheroids and 2D cultured cells for subsequent experiments.

Cancer cells show dormancy depending on cell density in 2D culture and on spheroid formation.
Furthermore, we investigated the effects of cell density and spheroid formation on cell dormancy. Cancer dormancy is characterized by no proliferation, no death, metabolic suppression, and recovery of active status. First, we assessed cell proliferation in both the 2D culture and 3D culture. Ki-67, a cell proliferation marker, was downregulated in a manner dependent on cell density and spheroid size (Fig. 2a,b and Supplementary Figures S1a,b). In contrast, p21, a regulator of cell cycle progression in the G1 and S phases, was upregulated in spheroids. Bromodeoxyuridine (BrdU) is a thymidine analog used for monitoring cell proliferation. BrdU-positive cells were decreased in the same manner (Fig. 2c,d). These results suggested that cell proliferation was suppressed at high cell density. However, little cell death was observed in spheroids, even at their maximum sizes (Fig. 2e). To investigate metabolic activity, we measured the uptake amount of 2-NBDG, a fluorescent glucose analog. 2-NBDG fluorescence decreased with cell density in the 2D culture and spheroids ( Fig. 2f-i). The decreased proliferation was reversible; when the spheroids were reduced by pipetting and re-plated in a flat plate 2D culture, the cells demonstrated a recovery of the proliferation rate to control levels with a slight delay (Fig. 2j). Altogether, these results indicated that spheroids or cells cultured under a high cell density in 2D demonstrated dormancy. We accordingly used them as models for dormant cancer cells.

PpIX accumulation increases in dormant cancer cells. To investigate the effect of cell dormancy on
PpIX accumulation after ALA treatment, we evaluated the accumulation under different cell density conditions in the 2D culture. PpIX accumulation revealed a positive relationship with cell density (Fig. 3a,b). The PpIX amount increased 23-fold between the lowest cell density in the 2D culture and S500 spheroids. In contrast, PpIX accumulation decreased under hypoxic conditions. Thus, hypoxia was not involved with intracellular PpIX increase in dormant cancer cells. Similarly, PpIX accumulation revealed a positive relationship with spheroid size (Fig. 3c,d). These changes were inferred to have been caused by cell dormancy. Altogether, these results suggest that PpIX accumulation increases with cell dormancy under normoxic conditions. ALA-PDT cytotoxicity was induced in dormant cancer cells. ALA-PDT can lead to cell death via the necrosis or apoptosis pathway and is a highly effective form of therapy for treating superficial basal cell carcinomas. We investigated ALA-PDT cytotoxicity under different cell-dormancy conditions. Cell viability in the 2D culture was reduced by ALA-PDT in a cell density-dependent manner ( Fig. 4a and Table 1). Cell viability of spheroids was also reduced in a size-dependent manner ( Fig. 4b and Table 2). These results are in good agreement with the increase in PpIX accumulation (Fig. 3b,d). They indicate that ALA-PDT cytotoxicity was induced by the increase in PpIX accumulation in a cell dormancy-dependent manner. PEPT1 and ABCB6 were upregulated depending on cell dormancy, whereas ABCG2 was downregulated. We assessed the protein expression of transporters involved in porphyrin metabolism. In our previous studies, we demonstrated that PEPT1, ABCB6, and ABCG2 were key players in regulating intracellular PpIX after ALA treatment. In summary, PEPT1 overexpression or ABCG2 suppressing caused increase of intracellular PpIX accumulation 24 . Other studies showed that ABCB6 overexpression induced increase of PpIX 12 and ABCB6 suppressing induced decrease of PpIX 25 respectively. Therefore, we focused on these 3 transporters. PEPT1 and ABCB6 were upregulated in a cell density-dependent manner in the 2D culture and size-dependent manner in the 3D culture ( Fig. 5a-c; Supplementary Figures S2a,b). Furthermore, to examine the contribution of ABCB6 on dormant cancer cells, we used RNAi technology to suppress the expression of ABCB6 in the high cell density-2D cultured cells (Fig. 5d,e). The expressions of ABCB6 were suppressed by siRNA (Fig. 5d), and led to decreased PpIX accumulation (Fig. 5e). Since siABCB6 reduced PpIX accumulation in dormant cancer cells, it is suggested that increased expression of ABCB6 contributed to PpIX accumulation increase in dormant cancer cells. These results strongly suggest that ABCB6 is upregulated and affected PpIX accumulation in dormant cancer cells. In contrast, ABCG2 was downregulated in the same manner. Expression of neither hypoxia-inducible factors (HIF)-1α nor -2α was observed in S500, which was the maximum size of spheroids (Supplementary 2D culture. After 4 days, cell numbers in the 2D culture and spheroids were confirmed to be similar. Then, cells were again placed in the 2D culture at 4.2 × 10 3 cells/cm 2 . Cells were counted by trypan blue staining. n = 3. Bars represent standard deviations (SD). Figure S2c). These data also suggest 3D spheroids didn't have the hypoxic area. mRNA expression of vascular endothelial growth factor (VEGF) was upregulated in a cell density-dependent manner (Supplementary Figure  S2d). Altogether, these results suggest that cell dormancy affects not only one but multiple genes to enhance porphyrin production.

PpIX accumulation increased in proliferation-inhibited dormant-like cancer cells.
In the previous section, we revealed that PpIX accumulation increased and cancer cell proliferation decreased in a cell dormancy-dependent manner. We accordingly hypothesized that PpIX accumulation is increased in proliferation-inhibited dormant-like cancer cells. To test this hypothesis, cells were cultured with methotrexate (MTX) or cycloheximide (CHX), which are reported to be G0/G1-phase arrestors for PC-3 cells 26 . Cell density was prepared similarly between control and G0/G1-phase arrestors (Supplementary Figure S3a). Ki-67 was downregulated by drug additions (Fig. 6a and Supplementary Figure S3b). Expression levels of PEPT1 and ABCB6 were not changed; however, ABCG2 was downregulated by both the drugs. PpIX accumulation was increased approximately 4-fold by drug additions (Fig. 6b,c). Next, we investigated ALA-PDT cytotoxicity after visible light irradiation under these conditions. Cell viability was markedly reduced, which is a result in good agreement with the increased PpIX accumulation (Fig. 6d). However, remarkable cell deaths were not observed in ALA-group. Altogether, these results indicated that ALA-PDT cytotoxicity increased in dormant-like cancer cells.

Discussion
In this study, we first demonstrated that cancer cells exhibited dormancy, including the decrease of cell proliferation, dependent on cell density and spheroid size. Cell dormancy was characterized by no proliferation, no death, metabolic suppression, and recovery of active status. Second, we demonstrated an increase of PpIX accumulation after ALA treatment with the same dependency. We infer that PpIX accumulation increases with cell dormancy. Moreover, ALA-PDT cytotoxicity was induced and transporters involved in porphyrin metabolism were regulated in a cell dormancy-dependent manner in both the 2D culture and spheroids. PpIX accumulation was increased by incubation with ALA and G0/G1-phase arrestors and ALA-PDT cytotoxicity was also induced. To the best of our knowledge, this is the first study to demonstrate that dormant cancer cells accumulate high PpIX levels and are sensitive to ALA-PDT.
Many methods of 3D multicellular spheroid culture for biomedical research have been reported 19 . For example, matrigel culture, spinner flasks, and hanging drop are well-known methods for spheroid culture. In this study, we used the EZSPHERE 3D culture plate to form spheroids. This culture plate has been particularly used for embryonic stem cell research, and this is the first report to use the culture plate to form cancer spheroids [27][28][29] . This culture plate has two advantages. First, it allows the development of uniformly sized spheroids and the control of spheroid size by change in the number of cells seeded in the plate (Fig. 1c). Second, the culture medium can be exchanged with other media with the preservation of the spheroid 3D structure so that spheroids in the plate permit PpIX accumulation analysis by confocal fluorescence microscopic imaging (Fig. 3a,c). Finally, spheroids formed in the plate demonstrate the same phenotypes as described in previous reports 15,22 : inhibition of cell proliferation and VEGF upregulation (Supplementary Figure S2d).
Although understanding of porphyrin and heme biosynthesis has helped to advance the field of ALA-PDD and PDT, no well-defined mechanism of PpIX accumulation in cancer cells after ALA administration is known. In some studies, correlation between cell density and PpIX accumulation in the 2D culture has been suggested 30,31 . Georgakoudi et al. and Moan et al. reported that high-density populations produce substantially higher amounts of PpIX per cell than low-density populations but did not suggest a clear mechanism. In other studies, correlation in the 2D culture between cell proliferation and PpIX accumulation has been found 30,32,33 . Moan et al. have reported that the only cell cycle dependence of PpIX synthesis appears to be simple volume dependence in WiDr (human colon adenocarcinoma) cells. Given that cells in the G2 and M phases produce 1.9 times as much PpIX as cells in the G1 phase, the increase of PpIX accumulation could be explained by cell volume. In this study, PpIX accumulation also demonstrated a positive relationship with cell density in the 2D culture, and this relationship was observed even in the 3D culture (Fig. 3b,d). PpIX accumulation analysis of cells incubated with ALA and G0/G1-phase arrestors (methotrexate or cycloheximide) revealed that the PpIX amount was increased in PC-3 cells, although cell volume was unchanged (Fig. 6b). Thus, the effects of cell density on PpIX production may be cell line-dependent.
Gene expressions were dramatically altered in a cell dormancy-dependent manner. To the best of our knowledge, this is the first report to demonstrate that PEPT1, ABCB6, and ABCG2 expression levels are regulated by Figure 4. Effect of ALA-PDT on cell viability. Cells were incubated with 1 mM ALA in a complete medium for 24 h and then exposed to 1080 mJ/cm 2 light for 5 min. Cell viability was determined on the next day by the MTT assay after treatment. (a) ALA-PDT in the 2D culture. Cells were incubated for 1 day before ALA treatment. n = 6. (b) ALA-PDT in spheroids. Cells were incubated for 3 days before ALA treatment. The cell density of "2D" is the same as that of "S500". n = 3. Bars represent standard deviation (SD) (  Table 2. Cell viability of spheroids. * The cell density of "2D" is the same as that of "S500". cell density or dormancy. ABCG2 was upregulated in gastric and ovarian cancer spheroids 34,35 . In our study, ABCG2 was downregulated, and one of the causes was suppression of cell proliferation. Because cells treated with G0/G1-phase arrestors expressed less ABCG2 (Fig. 6a). However, the final reason for PEPT1 and ABCB6 upregulation is unknown. In our previous study and others, PEPT1 and ABCB6 were upregulated under hypoxic conditions 11,[36][37][38] . However, we could not detect HIF expression in the 2D or 3D culture (Supplementary Figure S2c). Cells treated with CoCl 2 or DMOG, which are known as HIF inducers, did not demonstrate increased PpIX accumulation (data not shown). Furthermore, coproporphyrin III, which is eliminated from cells under hypoxic conditions after ALA treatment, was not detected in spheroids 11 . These results suggest that hypoxia is not a critical factor in PpIX accumulation increase. However, TNF-α and IFN-γ increase PEPT1 expression and activity 39 . Transporter expression may be upregulated by cytokine concentration in a cell density-dependent manner.
Recently, cancer spheroids have been reported to reflect heterogeneous tumor tissues more accurately in the 3D culture than in the 2D culture. Gene expression in spheroids is heterogeneous owing to exposure to oxygen and nutrients and to other physical and chemical stresses in different microenvironments 40,41 . Ki-67 positive cells were present at edges of spheroids, and the cells in the center were dormant [42][43][44] . We demonstrated that dormant cancer cells tended to accumulate high PpIX. Accordingly, we expected that PpIX levels in the cells in the center would be high and those at the edges would be low. However, we observed that the cells at the edges accumulated high PpIX and the cells in the center accumulated low PpIX (Fig. 3c). This observation may reflect two points. One point is the heterogeneity of micro-environmental ALA concentrations, given that cells at the edges readily take up ALA from the medium. The other point is the heterogeneity of cancer dormancy. In other words, the cells at the edges might be more dormant than those in the center in our spheroids which made by EZSPHERE. The distinct reason of PpIX heterogeneity in spheroids is unknown. On the other hand, PpIX levels of cells were also heterogeneous in the 2D culture. The PpIX level changed more than 2-fold between high-and low-PpIX-accumulating cells in the same culture dish (Fig. 3a). This phenomenon is in good agreement with BrdU-positive heterogeneity.
The results of this study indicate that ALA-based treatment would be an innovative therapy for dormant cancer cells, which are relatively insensitive to most chemotherapeutic drugs and radiation. In addition, these cells can cause tumor recurrence when they re-enter the cell cycle 13,15,16 . Accordingly, much research has been attempted to improve treatments for resistance. In this study, we demonstrated that dormant cancer cells tended to accumulate high PpIX and had higher sensitivity to ALA-PDT than non-dormant cells (Fig. 4a,b). Moreover, ALA-PDT cytotoxicity to non-dormant cells could be enhanced by combination with G0/G1-phase arrestors (Fig. 6d). In previous studies, combination with methotrexate (MTX) enhanced ALA-PDT in skin carcinoma and prostate cancer 45,46 . In brief, these authors argued that PpIX accumulation was induced by MTX and that one of the causes was the upregulation of coproporphyrinogen oxidase. From previous reports and the increase of PpIX accumulation in dormant cancer cells in this study, we propose that ALA-based diagnosis and therapy will be effective approaches for dormant cancer cells. However, the reason for high PpIX accumulation in those cells remains unclear. Further study of the PpIX accumulation mechanism in dormant cancer cells should contribute to ALA-based treatment.

Methods
Cells and cell cultures. The human prostate cancer cell line PC-3 (provided by Dr. Inoue, Kochi University, Kochi, Japan) was maintained in an RPMI-1640 medium supplemented with 10% (v/v) FBS and 1% (v/v) ABAM. Cells were maintained under 5% CO 2 gas at 37 °C. Cell culture under hypoxic conditions (1% oxygen) was performed with an AnaeroPack-Kenki 5% (Mitsubishi Gas Chemical Co., Tokyo, Japan). 3D cell culture. EZSPHERE 3D cell culture plates were obtained from AGC Techno Glass Co., Ltd. (Tokyo, Japan) and used to culture cancer spheroids. Totals of 5.00 × 10 5 cells for S500, 2.50 × 10 5 cells for S250, and 1.25 × 10 5 cells for S125 were seeded with 3 ml medium on every 35-mm dish. After 2 days, 1 ml of old medium was carefully replaced with new medium. Four days after seeding, 2,300 spheroids were formed in each dish.
Confocal fluorescence microscopic imaging. Cells were implanted in flat normal cell-culture dishes for the 2D culture or in EZSPHERE for the 3D culture and incubated for 3 days. After cancer spheroid formation, cells were incubated with 1 mM ALA under 5% CO 2 at 37 °C in normoxia (21% oxygen) or hypoxia (1% oxygen) in the dark for 24 h. Then, the ALA-containing medium was removed, and cells were incubated with 1 μ M Hoechst 33342-containing medium for 20 min. Cells were rinsed twice with PBS and incubated in HBSS for confocal imaging. A Zeiss LSM 780 upright laser scanning confocal microscope (Carl Zeiss SAS, Jena, Germany) was used for live-cell imaging. The excitation was set at a wavelength of 405 nm for PpIX and Hoechst, 488 nm for BrdU and 458 nm for 2-NBDG. The emission was set at 620-700 nm for PpIX, 440-500 nm for Hoechst, 493-630 nm for BrdU, and 500-600 nm for 2-NBDG. Laser illumination was set at 2.0% power for all. Images were acquired using a 40× water immersion lens for PpIX, Hoechst, and 2-NBDG, and a 20× lens for BrdU. The images were analyzed with the ZEN software package of Zeiss.

Establishment of cancer cell lines transiently suppressing endogenous ABCB6 protein. Small
interfering RNA (siRNA) oligonucleotides and negative control scrambled oligonucleotides were custom synthesized by Dharmacon (Lafeyette, CO, USA) 11 with the following sequences: siRNA (ABCB6) sense 5′ -CUGUUUCGCUUCUACGACAtt-3′ antisense 5′ -UGUCGUAGAAGCGAAACAGtt-3′ Oligonucleotides were transfected into cells (16 × 10 3 cells/cm 2 in 2D culture) at a final concentration of 5 nM using DharmaFECTTM 4 (Dharmacon) according to the manufacturer's instructions. Cells were further incubated with 1 mM ALA for 24 h. mRNA and PpIX were extracted and measured as previously described 11 . Exposure of the cells to light-emitting diode (LED). Cells were incubated with 1 mM ALA under 5% CO 2 at 37 °C for 24 h. Cells were then exposed to LED irradiation for 5 min (630 nm, 1080 mJ/cm 2 ) by placement of the plate under an LED irradiation unit (provided by SBI Pharma CO., Ltd., Tokyo, Japan) as previously described 24 . This irradiation condition is almost equivalent to the clinical irradiation conditions. Cells were further incubated in the dark under 5% CO 2 at 37 °C for 24 h. Cell viability was then measured by the MTT assay or trypan blue staining as previously described 49 .