Hypericin in the Dark: Foe or Ally in Photodynamic Therapy?

Photosensitizers (PSs) in photodynamic therapy (PDT) are, in most cases, administered systemically with preferential accumulation in malignant tissues; however, exposure of non-malignant tissues to PS may also be clinically relevant, when PS molecules affect the pro-apoptotic cascade without illumination. Hypericin (Hyp) as PS and its derivatives have long been studied, regarding their photodynamic and photocytotoxic characteristics. Hyp and its derivatives have displayed light-activated antiproliferative and cytotoxic effects in many tumor cell lines without cytotoxicity in the dark. However, light-independent effects of Hyp have emerged. Contrary to the acclaimed Hyp minimal dark cytotoxicity and preferential accumulation in tumor cells, it was recently been shown that non-malignant and malignant cells uptake Hyp at a similar level. In addition, Hyp has displayed light-independent toxicity and anti-proliferative effects in a wide range of concentrations. There are multiple mechanisms underlying Hyp light-independent effects, and we are still missing many details about them. In this paper, we focus on Hyp light-independent effects at several sub-cellular levels—protein distribution and synthesis, organelle ultrastructure and function, and Hyp light-independent effects regarding reactive oxygen species (ROS). We summarize work from our laboratories and that of others to reveal an intricate network of the Hyp light-independent effects. We propose a schematic model of pro- and anti-apoptotic protein dynamics between cell organelles due to Hyp presence without illumination. Based on our model, Hyp can be explored as an adjuvant therapeutic drug in combination with chemo- or radiation cancer therapy.


Introduction
Photodynamic therapy (PDT) is now widely used for diagnostic and treatment purposes in case of cancer, inflammatory diseases, infections, and stimulating antitumor immunity [1][2][3]. Photosensitizers (PSs) used in PDT are usually administered systemically and show a preferential accumulation in malignant cells. Nevertheless, exposure of healthy non-malignant cells to PS remains potentially high, especially in the vasculature [4]. Tumor response to PDT is variable, ranging from high sensitivity to extreme resistance. The parameters determining tumor sensitivity to PDT are PS cellular distribution, tumor oxygenation, vascularity, and immunogenicity [5].
Hyp and its derivatives have been used in experimental PDT approaches for a long time [2,3,6,7]. They have displayed illumination-dependent, anti-proliferative, and cytotoxic effects in many tumor cell lines [8,9]. The type of HypPDT triggered cell death signaling pathway (apoptosis or necrosis) is determined by the sub-cellular Hyp accumulation sites in membranous organelles [7,9,10].
In both cell types under control conditions, the majority of Bcl2 localized in discrete foci, presumably mitochondria [29] and the ER and mitochondria contact sites. Incubation with Hyp changed Bcl2 distribution ( Figure 1A). The Bcl2 signal apparently decreased in foci outside the nucleus, and there was noticeable Bcl2 translocation into nuclei in U87 MG ( Figure 1) and HAEC cells [30]. Bcl2 translocation into nuclei can enable the on-set of apoptosis via the association with nuclear receptor Nur77, which plays a role in regulation of differentiation, proliferation, apoptosis, and survival in different cell types [35]. (B) PKCα phosphorylated on Thr638 (green), Bcl2 phosphorylated on Ser70 (green), and PKCδ phosphorylated on Ser645 (green). The mitochondria are stained with TMROS Orange (red) and nucleus with Hoechst 33342 (blue). The fluorescence images were acquired by LSM700 confocal microscope (Zeiss, Germany) with 63X oil objective (NA = 1.46). The fluorophores were excited and detected under these conditions: Hoechst (405/410-490 nm), Alexa 488 conjugated with secondary antibodies for visualization of Bcl2, PKC and PKC (488/500-550 nm), MitoTracker® Orange CMTM/Ros (555/590-630 nm). (C) The corresponding Western blot analysis results of selected proteins in U87 MG cells with and without Hyp treatment in the dark, or 1 h after irradiation (4 J/cm 2 ). The experimental protocols are described in detail in previous publications [25,28,30].
Under control conditions in both cell types, Bax distribution showed a diffused pattern with some localization in foci, corresponding to distributions in many cell types [31][32][33]. Hyp presence resulted in distinctive changes in Bax distribution. In U87 MG ( Figure 1A) and HCAEC [30], there was significant Bax translocation into distinct foci throughout the cell. In HCAEC, there was also a substantial Bax translocation into the nuclei. Bax translocation to mitochondria is often associated with an apoptosis, whereas translocation to the nucleus/ER has been indicated with necrosis, inflammation, and secondary apoptosis [36,37]. The Hyp light-independent effects on the Bcl2 proteins distribution are dependent on the actual intracellular concentration of Hyp. The intracellular Hyp concentration depends on the final concentration of Hyp in the cell media, incubation times and on the cell uptake rate [38,39]. Bcl2 and Bax translocation patterns in U87 MG and HCAEC cells due to Hyp in the dark may underlie a different cell response to Hyp in viability assays measured via flow cytometry with Annexin V and propidium iodide [30,39]. In U87 MG, Hyp did not affect cell The experimental protocols are described in detail in previous publications [25,28,30].
In both cell types under control conditions, the majority of Bcl2 localized in discrete foci, presumably mitochondria [29] and the ER and mitochondria contact sites. Incubation with Hyp changed Bcl2 distribution ( Figure 1A). The Bcl2 signal apparently decreased in foci outside the nucleus, and there was noticeable Bcl2 translocation into nuclei in U87 MG ( Figure 1) and HAEC cells [30]. Bcl2 translocation into nuclei can enable the on-set of apoptosis via the association with nuclear receptor Nur77, which plays a role in regulation of differentiation, proliferation, apoptosis, and survival in different cell types [35].
Under control conditions in both cell types, Bax distribution showed a diffused pattern with some localization in foci, corresponding to distributions in many cell types [31][32][33]. Hyp presence resulted in distinctive changes in Bax distribution. In U87 MG ( Figure 1A) and HCAEC [30], there was significant Bax translocation into distinct foci throughout the cell. In HCAEC, there was also a substantial Bax translocation into the nuclei. Bax translocation to mitochondria is often associated with an apoptosis, whereas translocation to the nucleus/ER has been indicated with necrosis, inflammation, and secondary apoptosis [36,37]. The Hyp light-independent effects on the Bcl2 proteins distribution are dependent on the actual intracellular concentration of Hyp. The intracellular Hyp concentration depends on the final concentration of Hyp in the cell media, incubation times and on the cell uptake rate [38,39]. Bcl2 and Bax translocation patterns in U87 MG and HCAEC cells due to Hyp in the dark may underlie a different cell response to Hyp in viability assays measured via flow cytometry with Annexin V and propidium iodide [30,39]. In U87 MG, Hyp did not affect cell viability [25,29], and in the HCAEC resulted in a significant decrease in viability from 90% to 50% [30]. Besides the translocation of Bcl2 and Bax, Hyp light-independent treatment also caused a decrease in the protein synthesis, as indicated by the Western blot analysis ( Figure 1C) [25].
The difference may be due to either different Bax distribution changes in U87 MG and HAEC (nucleus vs ER) or the fact that malignant U87 MG cells have additional survival mechanisms in response to Hyp, such as Bcl2 interaction with DNA repair protein Ku70 [40,41]. In addition, we have shown that the majority of Bcl2 protein present in U87 MG cells is phosphorylated at serine 70 (pBcl2S70) [25]. It has been shown that phosphorylation of Bcl2 at serine 70 is required for Bcl2's full and potent anti-apoptotic function [11,12]. The distribution of Bcl2 in the presence of Hyp ( Figure 1A) follows the same pattern as pBcl2S70 [25].

Hyp Light-Independent Effects on the Distribution and Phosphorylation of Anti-Apoptotic PKCα and Pro-Apoptotic PKCδ
In various cell types including U87 MG, PKCα, and PKCδ isotypes exhibit an opposing effect in cell survival and apoptosis [12,[15][16][17]. A pro-apoptotic PKCδ has been shown to be a target of caspase-3, where anti-apoptotic PKCα inhibits apoptosis by phosphorylating Bcl2 [11,17]. Generally, inactive PKCs are considered to be cytoplasmic; upon activation by different signals, PKC translocate to the plasma membrane, to other membraneous organelles, and to the nucleus [11,42].
We have shown that the majority of PKCα present in U87 MG cells is already in a catalytically competent phosphorylated form pPKCα(Thr638) ( Figure 1B; and in [25]). This was in agreement with published works regarding the increased activity of PKCα in gliomas and glioma cell lines [11,12]. In addition, we have shown that the majority of pBcl2S70 protein present in U87 MG cells co-localizes with PKCα [25], suggesting that PKCα is likely one of the Bcl2 kinases in U87 MG cells.
Hyp was shown to co-localize with PKCα in U87 MG cells, and Hyp binding assays and molecular modeling indicated direct interactions between Hyp and PKCα [26,27]; however, Hyp presence did not affect PKCα distribution in U87 MG cells [25]. In the control cells and in the cells treated with Hyp without irradiation, the majority of PKCα was distributed evenly in the cytoplasm and at mitochondria ( Figure 1B; and in [25]). However, we have also shown that pretreatment with Hyp decreased PKCα translocation to the plasma membrane upon PMA treatment ( Figure 2 vs. Figure 6 in Dzurova et al. [25]) and increased cytoplasmic localization of PKCα. This finding suggests that Hyp competes for the PMA binding site at PKCα and prevents PKCα activation [25]. Dissociation constant K d for the Hyp binding to PKCα and to PKCδ was determined to be 111 nM and 94 nM, respectively [27]. These values are slightly lower than the binding of PMA (K d = 160 nM) for both isoforms, which reflects that Hyp can be a strong competitor with phorbol esters for the binding to these PKC isoforms [27]. Thus, we also investigated Hyp light-independent effects on the pro-apoptotic PKCδ distribution and phosphorylation in U87 MG cells [25,28].
We have shown the distribution of PKCδ and its phosphorylated form pPKCδ(S645) in the absence and the presence of the Hyp [28]. The S645 is considered a priming phosphorylation site, and it has been suggested that this modification converts PKCδ to its mature form [28,42,43]. There was no difference in non-specific PKCδ staining. Figure 2 shows a comparison of primed phosphorylated PKCα (pPKCα (Thr638)) and PKCδ(pPKCδ(S645)) forms in the presence of PKC activator (PMA), inhibitor (Gö6976) and Hyp, respectively [28]. The light-independent effect of Hyp on p(S645)PKCδ is similar to the PKC inhibitor Gö6976. In addition, we have identified that the p(S645)PKCδ signal localization in the presence of Hyp and Gö6976 is most likely GA-related compartments. Further, it was shown that Hyp is likely to follow the same ceramide's uptake pathway [28].
that Hyp competes for the PMA binding site at PKCα and prevents PKCα activation [25]. Dissociation constant Kd for the Hyp binding to PKCα and to PKCδ was determined to be 111 nM and 94 nM, respectively [27]. These values are slightly lower than the binding of PMA (Kd = 160 nM) for both isoforms, which reflects that Hyp can be a strong competitor with phorbol esters for the binding to these PKC isoforms [27]. Thus, we also investigated Hyp light-independent effects on the proapoptotic PKCδ distribution and phosphorylation in U87 MG cells [25,28].  The cleavage of PKCδ and the accumulation of the constitutively active fragment in the nuclei and in the Golgi apparatus (GA) are critical for triggering the nuclear fragmentation and for the ceramide-induced apoptosis [15,17].
It has been shown that PKCα activity blocks the ROS production and, in return, that the Gö6976 inhibitor treatment is triggering a rapid increase in ROS in the mitochondria [14]. Mitochondria play an integrative role in controlling cell ROS production [44,45]. Further, it was also shown that, besides its anti-apoptotic activity, Bcl2 overexpression could protect the cells from the oxidative stress induced by a variety of oxidative insults [46].
In light of these facts, we investigated Hyp light-independent effects on the oxidative stress, and mitochondria structure and function ( Figure 3). We have shown the distribution of PKCδ and its phosphorylated form pPKCδ(S645) in the absence and the presence of the Hyp [28]. The S645 is considered a priming phosphorylation site, and it has been suggested that this modification converts PKCδ to its mature form [28,42,43]. There was no difference in non-specific PKCδ staining. Figure 2 shows a comparison of primed phosphorylated PKCα (pPKCα (Thr638)) and PKCδ(pPKCδ(S645)) forms in the presence of PKC activator (PMA), inhibitor (Gö6976) and Hyp, respectively [28]. The light-independent effect of Hyp on p(S645)PKCδ is similar to the PKC inhibitor Gö6976. In addition, we have identified that the p(S645)PKCδ signal localization in the presence of Hyp and Gö6976 is most likely GA-related compartments. Further, it was shown that Hyp is likely to follow the same ceramide's uptake pathway [28].
The cleavage of PKCδ and the accumulation of the constitutively active fragment in the nuclei and in the Golgi apparatus (GA) are critical for triggering the nuclear fragmentation and for the ceramide-induced apoptosis [15,17].
It has been shown that PKCα activity blocks the ROS production and, in return, that the Gö6976 inhibitor treatment is triggering a rapid increase in ROS in the mitochondria [14]. Mitochondria play an integrative role in controlling cell ROS production [44,45]. Further, it was also shown that, besides its anti-apoptotic activity, Bcl2 overexpression could protect the cells from the oxidative stress induced by a variety of oxidative insults [46].
In light of these facts, we investigated Hyp light-independent effects on the oxidative stress, and mitochondria structure and function ( Figure 3).

Hyp Light-Independent Effects on the Mitochondria Ultrastructure and Function
To better understand the Hyp light-independent effects, we investigated the Hyp effect on mitochondrial function and cell metabolism in malignant U87 MG and non-malignant HCAEC cells.
To monitor the Hyp effect on the mitochondria structure, we used organelle specific fluorescent

Hyp Light-Independent Effects on the Mitochondria Ultrastructure and Function
To better understand the Hyp light-independent effects, we investigated the Hyp effect on mitochondrial function and cell metabolism in malignant U87 MG and non-malignant HCAEC cells.
To monitor the Hyp effect on the mitochondria structure, we used organelle specific fluorescent dye MitoTracker Orange in the presence and absence of Hyp, PMA, and Gö6976 inhibitor ( Figure 3A) in the methanol fixed cells. Control cells and cells treated with PMA show a widespread mitochondrion network, which is also evident in the live cells stained with mitochondrial potential (∆Ψ m ) indicator Rhodamine123 ( Figure 3B). The presence of either Hyp or Gö6976 causes slight fragmentation of the mitochondrion ( Figure 3A).
In addition to slight fragmentation, the presence of Hyp also results in the ROS increase in both cell lines ( Figure 3B,C). However, the increase in ROS is not due to dissipation of the mitochondrial potential (∆Ψ m ). In contrast, Hyp results in slight hyperpolarization as it is indicated in Figure 3B by the increase in the Rho123 fluorescence intensity ( Figure 3B,C).
To examine the Hyp effects on the mitochondrial function, we investigated the cellular bioenergetics in intact U87 MG ( Figure 3C) and HCAEC cells [30]. A detailed description of the method and protocols are in [30]. Figure 3C shows the oxygen consumption rate (OCR) measurements in U87 MG cells. Bioenergetics profiles reflect the U87 MG and HCAEC high proliferation and metabolic rates, respectively. U87 MG characteristics indicate a high proliferation and metabolic rates, and that substantial energy proportion originates from oxidative phosphorylation (OXPHOS) in addition to glycolysis. HCAEC characteristics indicates low proliferation and metabolic rates, and glycolysis as a dominant energy source [30].
Hyp significantly influenced the metabolism of U87 MG cells ( Figure 3C). U87 MG respiration significantly decreased, and Hyp also resulted in a proton leak decrease, which is in agreement with the observed hyperpolarization of mitochondrial potential ( Figure 3B,C). The Hyp presence in U87 MG cells seems to slow down overall cell metabolism, OXPHOS, and glycolysis. Hyp presence did not change overall the OCR profile in HCAEC, reflecting the glycolysis as a dominant energy source.
Based on work from our laboratories and that of others, we propose a schematic model of pro-and anti-apoptotic protein dynamics between cell organelles due to Hyp presence without illumination ( Figure 4). Based on our model, Hyp can be explored as an adjuvant therapeutic drug in combination with chemo-or radiation cancer therapy. mitochondrion network, which is also evident in the live cells stained with mitochondrial potential (ΔΨm) indicator Rhodamine123 ( Figure 3B). The presence of either Hyp or Gö6976 causes slight fragmentation of the mitochondrion ( Figure 3A). In addition to slight fragmentation, the presence of Hyp also results in the ROS increase in both cell lines ( Figure 3B,C). However, the increase in ROS is not due to dissipation of the mitochondrial potential (ΔΨm). In contrast, Hyp results in slight hyperpolarization as it is indicated in Figure 3B by the increase in the Rho123 fluorescence intensity ( Figure 3B,C).
To examine the Hyp effects on the mitochondrial function, we investigated the cellular bioenergetics in intact U87 MG ( Figure 3C) and HCAEC cells [30]. A detailed description of the method and protocols are in [30]. Figure 3C shows the oxygen consumption rate (OCR) measurements in U87 MG cells. Bioenergetics profiles reflect the U87 MG and HCAEC high proliferation and metabolic rates, respectively. U87 MG characteristics indicate a high proliferation and metabolic rates, and that substantial energy proportion originates from oxidative phosphorylation (OXPHOS) in addition to glycolysis. HCAEC characteristics indicates low proliferation and metabolic rates, and glycolysis as a dominant energy source [30].
Hyp significantly influenced the metabolism of U87 MG cells ( Figure 3C). U87 MG respiration significantly decreased, and Hyp also resulted in a proton leak decrease, which is in agreement with the observed hyperpolarization of mitochondrial potential ( Figure 3B,C). The Hyp presence in U87 MG cells seems to slow down overall cell metabolism, OXPHOS, and glycolysis. Hyp presence did not change overall the OCR profile in HCAEC, reflecting the glycolysis as a dominant energy source.
Based on work from our laboratories and that of others, we propose a schematic model of proand anti-apoptotic protein dynamics between cell organelles due to Hyp presence without illumination ( Figure 4). Based on our model, Hyp can be explored as an adjuvant therapeutic drug in combination with chemo-or radiation cancer therapy.

Conclusions
In conclusion, we have shown that Hyp has significant light-independent effects in malignant and non-malignant cells at several sub-cellular levels such as protein synthesis and distribution (Bcl2 family, PKC), mitochondria structure and function, and Hyp light-independent effects regarding ROS. Our findings suggest that Hyp without illumination can be explored as adjuvant therapeutic

Conclusions
In conclusion, we have shown that Hyp has significant light-independent effects in malignant and non-malignant cells at several sub-cellular levels such as protein synthesis and distribution (Bcl2 family, PKC), mitochondria structure and function, and Hyp light-independent effects regarding ROS. Our findings suggest that Hyp without illumination can be explored as adjuvant therapeutic drug in combination with chemo-or radiation cancer therapy.