Redox signalling and homeostasis are ubiquitous to all life on Earth, involving a complex and dynamic cell-dependent system that acts as both sensor and effector of cellular environmental changes to coordinate response to stimuli (including during tissue regeneration, wound infection, or cancer). CP and PAL are potent sources of RONS that directly affect cellular redox homeostasis, however redox responses can also be secondary, tertiary, etc., to the initial CP exposure [
2,
15]. Essentially, all therapeutic applications of CP are classified into the field of applied redox biology [
19]. This has been exemplified by scavenging of CP/PAL-derived RONS with the antioxidant N-acetylcysteine (NAC) that unanimously reduced or ablated DNA damage, cell cycle arrest, ERS, autophagy and apoptosis to CP/PAL in several
in vitro studies in cancer and prevented wound healing and antioxidant response in normal cells [
101,
108,
109,
118,
119,
120,
121,
122,
123,
124,
125,
126,
127,
128,
129,
130,
131,
132,
133,
134,
135,
136,
137,
138].
Cells can either be directly treated with CP while in media or indirectly treated with PAL, usually PAM or PAW. This leads to accumulation of RONS, which dissolve from the plasma/gas phase into the liquid phase, or are generated by a complex series of secondary reactions (
Figure 1) [
139,
140,
141,
142]. In cases of direct CP treatment of media containing cells, shorter-lived highly reactive oxidant species would also be pertinent in redox signalling. However, RONS such as NO
• and
•OH are too reactive to diffuse further than a µm range in the liquid phase, but may form again through the decomposition of ONOO
- [
139]. Hence the most commonly measured RONS are aqueous H
2O
2, NO
2 - and NO
3- as markers of RONS generation, due to their long lifetime. Research to date has found that ultraviolet photons play a negligible role in eukaryotic cell responses to CP, with CP-derived UV dose being an order of magnitude below the minimal erythema dose and the UV not even being able to reach living skin cells [
143]. However, CP-derived UV photons have also been shown to generate minor amounts of atomic H
+,
•OH and NO
• through photolysis of other RONS [
23,
144], which experiments with antioxidants cannot discriminate.
Epithelial cells are the most affected by CP and illicit a host of molecular responses over time. Generally, oxidative stress leads to protein unfolding and cessation of protein synthesis (ERS/UPR), cell cycle arrest (G2/M phase), mitochondrial dysregulation and post-translational modifications (predominantly phosphorylation via kinases) following the first few hours of CP exposure [
145]. Cell cycle arrest and protein synthesis cessation/unfolding are instated quickly to preserve cell survival early after exposure, while DNA repair occurs later. The nuclear factor erythroid-2 related factor 2 (Nrf2) antioxidant response to oxidative stress is the most prominent cell response occurring immediately following to even days after CP exposure to condition cells against redox stress, eventually restoring cell function and proliferation [
145]. In this way, light CP exposure can promote faster healing through redox eustress; levels of redox imbalance that condition cells to altered redox (usually oxidative) state, while returning to redox homeostasis. CP can also elicit autophagy, the cellular process of breaking down damaged components and organelles in the lysosome [
146], and eventually apoptosis in cancer cells by amplifying redox stress [
94,
102,
134,
135]. Several distinct protein signalling pathways regulate these processes and can cross-communicate to affect cell survival and proliferation [
147]. For example, CP activated ERK1/2 MAPK and Akt, the latter leading to significant activation of NFκB phosphorylation to promote cell proliferation [
148]. Interestingly, nearly the opposite has occurred in cancer cells. CP caused apoptosis through simultaneously inhibiting Akt/mammalian target of rapamycin (mTOR) while exciting JNK and p38 MAPK signalling pathways, leading to autophagy, activation of p53-mediated activation of proapoptotic caspases, while inactivating Akt also inhibited NFκB to reduce tumour cell proliferation [
149,
150]. CP application in cancer therapy commonly leads to cell cycle arrest among different cancer types, and eventually autophagy and apoptosis [
151]. These two examples demonstrate the hormesis approach to killing malignant cells while also promoting the healing of healthy tissue. This section will discuss modulations of several cell signalling pathways affected by CP and PAL in the context of wound healing and cancer.
4.1. Keap1-Nrf2-Antioxidant Response Pathways
Both cancerous and non-cancerous cells are armed with antioxidant response mechanisms to maintain redox homeostasis and affect survival, proliferation and migration [
145]. In fact, proteomic analysis revealed that the response to oxidative stress by Nrf2 is the most prominent cellular response to CP exposure in epithelial tissues [
145]. The Nrf2 pathway acts as a redox-sensitive activator of the antioxidant response element (ARE). Under normal cellular homeostasis, Nrf2 rests in the cytosol bound to Kelch-like ECH-associated protein 1 (Keap1), keeping Nrf2 inactive and enabling E3 ubiquitin ligase-dependent proteasomal degradation [
152]. Keap1 has canonically acted as the redox stress sensor, as thiols in multiple Keap1 cysteine residues are sensitive to CP-derived RONS [
153]. When these cysteines are oxidised, Nrf2 is released, translocates to the nucleus, ligates to small musculoaponeurotic fibrosarcoma (sMAF) proteins on DNA and transcribes ARE genes [
152]. The ARE encodes a host of antioxidant proteins and enzymes to scavenge RONS that may harm vital cellular components, including GSH, GSH metabolism-related enzymes GSH reductase (GR), GSH peroxidase (Gpx), GSH S-transferase (GST), γ-glutamylcysteine ligase catalytic/modifier (Gclc/m) subunits, antioxidant enzymes superoxide dismutase (SOD), catalase, haem oxygenase 1 (HO1), thioredoxin (Trx), thioredoxin reductase (TrxR), sulfiredoxin (Srx), and multifunctional stress response enzymes UDP-glucuronyl transferase (UGT) and NADPH quinine oxidoreductase 1 (NQO1) until redox homeostasis is returned [
154].
Clear evidence shows that redox signalling is essential to wound healing [
155]. Chronic wounds occur in parallel with chronic redox stress, which amplifies inflammation, apoptosis and impedes neovascularisation [
155]. As a consequence, therapeutic treatments that modulate redox status in interstitial tissue and cells are coming into focus to remedy chronic wounds [
155], including CP induction of antioxidant response mechanisms via Nrf2 signalling (
Figure 2). When exposed to CP, keratinocytes (HaCaT) increased HO1, catalase, NQO1, GST, and Gclc/m transcription via Nrf2 activation and translocation, concurrent with the release of proangiogenic chemokine VEGFA, growth factors HBEGF, CSF2, PTGS2, and inflammatory cell cytokine interleukin 6 (IL6), which promote re-epithelialisation and wound repair [
156,
157,
158,
159]. While Nrf2 disassociates from Keap1, Keap1 also alters cytoskeletal arrangement, regulating E-cadherin and F-actin filamentation and critical during tissue regeneration [
160,
161]. In conjunction with reducing connexin 43 (Cx43) expression to relax cell-cell focal adhesion in skin cells [
160], CP increases keratinocyte motility to re-epithelialise wounds. The role of Keap1 in response to CP has so far been limited to its role as a redox switch for Nrf2 and as a cytoskeletal junction protein. Yet, Keap1 is involved in several regulatory roles in the cell ranging from protein degradation, promoting NFκB p65 translocation for cell survival, cell cycle progression and p62-mediated autophagy [
162]. Nevertheless, there is growing momentum towards studying Keap1 beyond its canonical function as inhibitor of Nrf2, but this is yet to be explored with regards to CP treatment.
When translated into mice studies, direct CP treatment of acute full-thickness wounds accelerated wound healing, neutrophil and macrophage infiltration and TNFα, TGFβ and IL-1β transcription
in vivo, and promoted HO1 and NQO1 transcription in dermal fibroblasts and epidermal keratinocytes via Nrf2
ex vivo [
160,
163]. THP1 monocytes also activate Nrf2 and upregulate HO1 transcription in response to CP exposure [
164,
165]. Studies to date have also shown that RONS, like ONOO
-, can also activate Nrf2 indirectly via the PI3K/Akt pathway [
166,
167], and upon overexposing keratinocytes to CP at levels that significantly damage lipids, proteins and DNA, Nrf2 is suppressed as a result of Akt degradation [
109]. Similarly, inhibiting JNK (but not ERK or p38) attenuated CP-mediated HO-1 induction [
123], indicating that CP-derived RONS may cross activate Nrf2 through JNK, although the mechanism for this effect is yet to be determined.
HO1, an integral product of Nrf2, catabolises haem into biliverdin, releasing Fe
2+ and carbon monoxide (CO), with biliverdin being further degraded to bilirubin by biliverdin reductase [
168]. Bilirubin, CO and ferritin upregulation by Fe
2+ protect cells through RONS scavenging to inhibit apoptosis and inflammation [
169]. This in concert with increased GSH synthesis seem to support Akt activation promoting cell survival by inducing Nrf2/ARE and inhibiting p53 and Bax/Bcl2-dependent caspase-mediated apoptosis. Moreover, p53 performs a biphasic function in Nrf2 modulation; low p53 activity enhances Nrf2 expression to foster cell survival, whereas high p53 activity suppresses Nrf2 to promote cell death [
170]. In this sense, the p53/Nrf2 axis acts as a focal point in deciding the fate of the cell, and may be a mechanism by which CP can be applied therapeutically on a hormesis basis.
In many cancer cells, Keap1 and Nrf2 gene mutations constitutively increase their activity [
171], acting as counterbalance to promote survival in the higher endogenous RONS environment resulting from malignant metabolic and proliferative activity [
172]. Inhibiting HO1 activity or Nrf2/HO1 gene silencing results in significantly greater apoptosis by CP [
123]. Therefore, Nrf2/HO1 inhibition and CP adjunct therapy could synergise as a cancer therapy, similar to HO1 inhibition improving conventional chemo/radiotherapy and reducing tumour growth [
173], or synergistic activity of CP with chemo/radiotherapy
in vitro and
in vivo [
124,
130,
174,
175,
176,
177,
178,
179,
180]. Unfortunately, the absence of clinically safe HO1 inhibitors currently precludes testing, but may be grounds for future clinical research. Simultaneous treatment with CP and pharmacological inhibition of Trx also synergistically enhanced caspase 3-dependent apoptosis with some degree of lipid peroxidation-dependent ferroptosis (iron-dependent form of apoptosis linked to cytotoxic lipid peroxide accumulation) in glioblastoma cancer cells and slowed tumour growth in glioblastoma-bearing mice [
130]. As glioblastomas are incurable, and manifest as the most aggressive and deadly tumours of the brain, and are rapidly growing in incidence [
181], CP therapy could have a massive impact on improving survival and prognosis of these patients.
Recently, it was observed that CP activates Nrf2 concurrent with BTB and CNC homolog 1 (Bach1) in THP1 cells [
165], an antagonist of Nrf2 that competitively binds with sMAF proteins as a negative-feedback to ARE [
182]. This is in contrast to previous work showing that CP enhances Nrf2 and suppresses Bach1 transcription in HaCaT [
156]. The reason for these contrasting cellular responses remains unclear. Considering that Bach1 is degraded in cells under oxidative stress [
183], Nrf2 activity may paradoxically lead to upregulation of Bach1 in cancer cells [
184], and Bach1 an important oncogene that drives tumour metastasis [
184,
185], comparing the response of Bach1 to CP in normal and cancer cells should be investigated. Overall, Nrf2-ARE activation following CP treatment is integral to restoring redox homeostasis and cell survival with interplay with p53, JNK and Akt. Additionally, Nrf2 may only be half the story, as the canonical Nrf2 inhibitor, Keap1, also signals cytoskeletal protein rearrangement and cell junction proteins to promote wound healing. However, an increasing body of research shows that Keap1 also influences cell fate decisions [
162], which should be further investigated as it relates to CP.
4.2. ERS and UPR Signalling Pathway
Under redox imbalance, cells experience endoplasmic reticulum (ER) stress (ERS), causing elevated protein misfolding that impairs cell function. In an attempt to counteract this, the ER releases calcium to signal general suppression of protein synthesis, while upregulating a suite of chaperones that repair protein misfolding as the unfolded protein response (UPR) to restore homeostasis [
105]. There are three arms of the UPR that signal cells to increase their protein folding capability; (i) the inositol-requiring protein 1α (IRE1α), which uniquely splices X-box binding protein 1 (Xbp1) mRNA to express the Xbp1 transcription factor, (ii) protein kinase RNA-like ER kinase (PERK) that activates the eukaryotic initiation factor 2-alpha (eIF2α), and (iii) activating transcription factor 6 (ATF6), which gets cleaved in the Golgi apparatus to translocate to the nucleus [
105]. All three arms promote chaperones and recovery of ER biosynthesis activity to restore protein synthesis, while IRE1α and ATF6 also signal inflammatory responses, however only prolonged PERK activation can decide cell fate via C/EBP-homologous protein (CHOP) in two ways: Firstly, if the ER recovers, CHOP acts as negative feedback to PERK activity by promoting GADD34 and CReP binding to protein phosphatase 1 (PP1) which dephosphorylates (deactivates) eIF2α [
186,
187], or if ERS go unresolved, prolonged CHOP activation activates apoptosis [
188]. One of the integral ERS-associated proteins that signals ERS and initiation of the UPR is glucose-regulated protein 78 (GRP78). GRP78 is expressed on the ER membrane as an ERS sensor bound to PERK, IRE1α and ATF6, and acts as a chaperone that binds to misfolded proteins [
189]. Therefore, the UPR can be enacted by both normal and malignant cells in an attempt to restore cellular homeostasis, including in response to redox dysregulation.
Later, in airway epithelial cells, induction of the UPR in response to CP was confirmed, revealing significant upregulation of a number of ERS-associated proteins, including GRP78 [
145]. Importantly, apoptosis was absent in conditions of lower CP exposure in these studies [
145,
157]. While these studies have shown that CP can induce ERS/UPR, CP has also been shown to remediate cells already experiencing ERS. In a chemically-induced atopic dermatitis model
in vitro and
in vivo, elevated inflammation (upregulated TNFα, IL1β and CCL2, with decreased anti-inflammatory IL10), ERS/UPR, GRP78 expression and eventually CHOP-mediated apoptosis were observed, all of which were curtailed by CP treatment via induction of HO1 [
190]. HO1 induction as part of the response to CP indicates the role of Nrf2 in conditioning the cell for restoring redox homeostasis to cellular function [
163]. Altogether, these results show that inducing moderate ERS with CP may activate the UPR in a pro-survival response to redox eustress, while other cell signalling pathways promote wound healing (
Figure 3).
ChaC GSH-specific γ-glutamyl cyclotransferase 1 (CHAC1) degrades GSH, and thus plays a role in redox balance of the cell [
191] and is also overexpressed by prolonged activation of the PERK/IRE1α arm of the UPR [
192]. CHAC1 expression was not induced in keratinocytes treated with PAM activated for only 20 s with CP, but was transcriptionally induced up to 5-fold with PAM activated for 180 s [
157]. Therefore, the lack of cell death and no CHAC1 induction with lower CP activation showed PAM to be safe on skin cells [
157]. On the other hand, while the ERS/UPR may be exploited with CP to promote survival in healthy cells, prolonged redox dysregulation in cancer cells may also be possible through aberrant activation of the UPR. The expression of CHAC1 in breast and ovarian cancer cells and associated with significantly higher mortality in breast and ovarian cancer patients, implicating a possible role as a metastatic factor in these cancers [
193,
194]. Furthermore, CHAC1 was among the most upregulated genes in SCC and glioblastoma cells (>16 and 25-fold, respectively) following treatment with PAM [
195,
196] that was sizably larger than in keratinocytes [
157], indicating an intense degree of redox stress aggravating the UPR in cancer cells that led to apoptosis [
195]. In summary, while slightly elevated CHAC1 expression may increase malignancy of cancer cells [
193,
194], overexpressing CHAC1 with CP may kill cancer cells via prolonged PERK/IRE1α activation [
195]. Activation of the PERK/eIF2α and IRE1α/Xbp1 arms of the UPR, upregulation of GRP78 and eventually CHOP expression leading to apoptosis have also been observed in colorectal and melanoma cancer cells exposed to CP [
136,
197]. Neuroblastoma cells in response to CP exposure also activated eIF2α that led to stress granule formation (protein-RNA complexes that interrupt mRNA translation), in line with PERK-mediated UPR [
198]. To date, ATF6 activation has not shown involvement in CP-induced UPR, and was not activated in melanoma cells [
197]. The ERS/UPR is predominantly a result of elevated redox stress caused by CP, as NAC was able to prevent the UPR [
136]. Unfortunately, all evidence of CP inducing ERS and UPR to-date are
in vitro, while only one study showed CP could remediate already present ERS and UPR in a mouse atopic dermatitis-like
in vivo model [
190]. The relevance of the UPR has not yet been confirmed
in vivo with CP treatment of normal tissue, wounds or malignant tissue.
4.4. PI3K/Akt Pathway
The phosphoinositide 3-kinase (PI3K)/protein kinase B (Akt) signalling cascade orchestrates diverse regulatory functions in cell survival, disease pathogenesis, angiogenesis and tumorigenic processes, and therefore it is an important pathway for both induction of wound healing and cancer intervention [
202,
236]. Healing of acute wounds normally occurs with basally elevated Akt/mTOR activity to promote cell growth, migration and angiogenesis to help revascularise the new tissue [
237,
238]. In contrast, chronic wounds including diabetic foot ulcers (DFU) and have been shown to exhibit impaired PI3K/Akt/mTOR signalling that impairs cell survival and reduces growth factor release, which would also stimulate surrounding tissue to heal the wound [
239].
PAM has demonstrably promoted keratinocyte proliferation
in vitro and
in vivo via promoting Wnt/β-catenin signalling and PI3K/Akt/mTOR [
128,
240,
241]. Akt inhibits glycogen synthase kinase 3β (GSK3β), which prevents β-catenin degradation and allows nuclear translocation to promote cyclin D expression and the consequent G1/S proliferative cell cycle phase in keratinocytes [
241]. Evidently, a hormesis approach to CP applies here, as the higher CP-derived RONS (H
2O
2 and possibly NO) exposures then reduced PI3K/Akt, ERK1/2 and β-catenin signalling back to untreated levels [
128]. Another axis is shown whereby CP-induced PI3K/Akt activation leads to downstream ERK1/2 and NFκB activation to also promote cell survival and proliferation, likely involving CP-derived NO [
128,
148,
240]. PI3K/Akt signalling also leads to NO synthase activation, which increases endogenous NO production [
242], followed by activation of protein kinase G and downstream ERK1/2 to promote proliferation [
243]. Therefore, not just CP-derived NO, but the stimulation of endogenous sources of NO is involved in promoting survival, proliferation and angiogenesis which is critical to wound healing.
Akt can also dampen pro-apoptotic p53 and Bax expression, leading to upregulation of anti-apoptotic Bcl2 and increased Nrf2 activity to improve cell survival [
163]. Skin-derived mesothelial cells exposed to PAM also had higher Akt expression and phosphorylation with dampened p53 activity to promote survival, while fibroblasts died through p53-dependent apoptosis [
244]. Further, CP treatment of skin cells also results in release of growth factors EGF and KGF, which promote autocrine skin cell proliferation, elevate MMP2/9 activity for ECM remodelling and release VEGF to recruit endothelial cells to aid in angiogenesis [
245]. Stimulation of PI3K/Akt by CP is summarised in
Figure 5.
4.4.1. Suppressing PI3K/Akt Signalling in Cancer Cells
The PI3K/Akt signalling pathway is commonly overexpressed through genetic alterations in cancer cells leading to metastases [
236]. In normal cells, active phosphatase and tensin homolog (PTEN) inhibits PI3K/Akt activation, whereas many types of cancers are PTEN-deficient, leading to aberrant PI3K/Akt signalling [
236,
246]. The use of CP in the destruction of cancer cells has been shown to occur with increased Akt degradation due to redox stress [
94,
102,
129,
134,
135,
177]. The first study in head and neck cancer (human SCC15 and mouse SCC7) cells found that CP increased apoptosis, concurrent with reduced Akt phosphorylation and expression through MUL1 encoded E3 ligase-mediated ubiquitination [
135]. Furthermore, tumour growth in mice model of (state cancer type) was slowed and eventually stopped with CP treatment, with increased MUL1 expression and decreased Akt phosphorylation confirmed immunohistochemically [
135]. MUL1 expression can also stimulate NFκB activity to inhibit apoptosis, and yet, this is interpreted to be a short-term effect in which cells eventually succumb to apoptosis during prolonged stress [
247]. Reduced Akt/mTOR expression has also been linked to reduced hypoxia-inducible factor 1α (HIF1α) expression to dampen proliferation [
102]. Despite this, direct CP treatment of glioblastoma cells attenuated proliferation and decreased total Akt protein expression, but increased Akt and ERK1/2 phosphorylation 24 hours after exposure [
248]. Although Akt degradation occurred in both situations, differences in the activation of Akt observed between different cell lines in these studies may be due to the PTEN status of cancer cells. The U-87 glioblastoma cell line that elevated Akt phosphorylation in response to CP is a PTEN-mutant, while the LN-18 cell line is a PTEN wild-type and did not experience elevated Akt phosphorylation [
248]. In view of this, the gene status of PTEN (among other tumour suppressor-related genes) may be a factor that affects CP therapy against cancers.
In addition to targeting PI3K/Akt in cancer, the signal transducer and activator of transcription 3 (STAT3) pathway counterbalances PI3K/Akt/mTOR signalling via PTEN and promotes tumourigenic proliferation, invasion, migration and angiogenesis in most cancers [
249]. When Akt is inhibited in PTEN-deficient cancer cells, there can be a large increase in STAT3 signalling that compensates [
250]. Therefore, anticancer therapeutic strategies that inhibit both PI3K/Akt and STAT3, particularly in PTEN-deficient cancers, could be more effective in halting tumorigenesis [
250]. CP and PAM at lethal exposures also led to the deactivation of STAT3, which together with the degradation of Akt signalling halted tumour growth and promoted caspase 3-driven apoptosis in osteosarcoma and pancreatic cancer cells [
94,
134,
177].
PAM has also demonstrated efficacy towards destroying cancer cells [
125,
251]. PAM-treated Human hepatoma (HepG2) cells had significantly decreased expression of Akt and mTOR, which led to decreased p62 (inhibitor of autophagy) and increased Beclin 1 and LC3-II expression to promote autophagosome formation [
251]. The autophagosome formation, autophagy and degradation of Akt/mTOR were all ablated by simultaneous addition of catalase and SOD, implicating extracellular O
2•-/H
2O
2 as the causative PAM-derived ROS [
251]. PAM also induced the expression of PTEN and consequently reduced Akt phosphorylation and deactivated NFκB, however it also sensitised cancer cells to TNF-related apoptosis-inducing ligand (TRAIL) and greatly enhanced caspase 8/3-driven apoptosis [
125]. Mechanistically, the increased PTEN activity was likely due to reduced expression of miRNA425 that inhibits PTEN translation [
125,
252]. TRAIL binds to death receptors to induce apoptosis and is highly selective towards cancer cells, but aberrant PI3K/Akt pathway activity common in cancer cells [
236] can also lead to TRAIL resistance [
253]. The importance of PTEN is accentuated by findings that PTEN deficiency increases the resistance of cancer cells to TRAIL [
253]. Therefore, these studies suggest that CP could restore TRAIL as a therapeutic option in these cancers, though this is yet to be discovered.
Whereas JNK and p38 MAPK promote p53 activity, the PI3K/Akt pathway in the cell inhibits p53 [
214]. Consequently, aberrant Akt activity promotes cancer cell survival through inhibition of p53 [
163,
214]. Treating oral SCC with PAM has led to enriched p53 expression and consequently p21-mediated cell cycle arrest, inhibition of angiogenesis, DNA repair and mTOR (PI3k/Akt) pathways and eventually caspase 8/9/3-driven apoptosis [
195,
254]. The ataxia telangiectasia mutated (ATM) pathway was also implicated in PAM- and direct CP-induced p53 activity [
220,
254] and although not cancer related, PAM selectively induced ATM expression in
Mycobacterium tuberculosis-infected macrophages [
255]. Additionally, the ATM pathway is associated with activity in a diverse array of other cell signalling pathways including PI3K/Akt/mTOR [
256]. Despite this, the ATM pathway response to CP and the relevance to anticancer effects of CP is still unclear. The multitude of consequences to deactivating PI3K/Akt signalling in cancer cells through cytotoxic CP exposure are summarised in
Figure 6.