Photodynamic therapy and the role of heat shock protein 70

Abstract

Photodynamic therapy (PDT) is an effective treatment for superficial epithelial skin cancers and is also being utilised for skin rejuvenation. PDT with porphyrin-derived photosensitisers may be capable of inducing rapid cytochrome c release initiating apoptotic cascade via an activation of different caspases. Hsp70 has been demonstrated to directly bind to the caspase recruitment domain (CARD) of apoptotic-protease activating factor 1 (Apaf-1), thereby preventing the recruitment of oligomerisation of Apaf-1 and association of Apaf-1 with procaspase 9. Further, cytoplasmic Hsp70 rapidly translocates to the cell surface where it stabilises damaged membranes and preserves their integrity depending on the PDT dose. The induced cell surface expression and release of Hsp70 and its relevance for tumor response or skin rejuvenation is not fully understood, but seems to be of interest for monitoring, predicting or optimising treatment regimens. This review aims to summarise the current knowledge on PDT mediated cell signalling.

• photodynamic therapy (PDT)
• 5-aminolevulinic acid (ALA)
• tumour response
• skin rejuvenation
• heat shock protein (Hsp) 70

Introduction

Photodynamic therapy (PDT) is a promising treatment for certain types of cancer and other non-malignant conditions, which involves the administration of a photosensitising agent (one of the porphyrin-based compounds is 5-aminolevulinic acid (ALA) which is a precursor of protoporphyrin IX (PpIX)). The photosensitiser accumulates in more rapidly proliferating cells causing injury to various cutaneous targets (Figure 1A and B). Subsequent exposure to photoactivating light induces the generation of reactive oxygen species, particularly singlet oxygen, with following apoptotic or necrotic cell death [1], [2]. Furthermore, photosensitisers are often taken up with some selectivity by malignant or dysplastic tissues and light can be targeted to the appropriate tissue [1] limiting the damage to the surrounding normal tissue.

Figure 1. Porphyrin metabolism under physiologic conditions (A) and in tumour cells during ALA-based PDT (B).

The absorption spectrum of porphyrins are in the Soret band ranging from 360 nm to 400 nm as well as in Q-bands between 500 nm and 635 nm. The usually used red light with a wavelength of 630 nm has a maximal tissue penetration of approximately 2–3 mm [3–5] (Figure 2). Non-coherent light sources, emitting light in a broader arrangement of wavelengths, can improve the efficacy of the treatment due to the absorption in several wavelengths. Wavelengths of 700 nm to 800 nm reach cells as deep as 1 to 2 cm under the epidermal surface [6–8]. Light at wavelengths longer than 850 nm does not lead to sufficient photochemical response in deeper tissue sections due to the weak absorption of photosensitisers of these wavelengths.

Figure 2. Absorption and fluorescence curves for Protoporphyrin IX.

PDT has been used as an effective organ preserving treatment for actinic keratoses, Bowen's disease and superficial basal cell carcinomas (BCCs) [4], [9]. PDT is used because of its minimal invasiveness, easy repeatability and favourable aesthetic results. Other advantages of PDT in dermatology include the simultaneous treatment of multiple tumours and the relatively short healing time. One of the most important side effects is considerable pain. Apart from the therapy of cancers, PDT is also being utilised for skin rejuvenation [3], [4].

Numerous studies have been attempted to identify subcellular PDT targets. These efforts have successfully identified mitochondria, lysosomes and various cytoplasmic membranes among the primary targets of PDT-mediated damage, depending on the particular photosensitiser and dose. Furthermore, the sensitivity of cells to PDT varies between cell types such as kerantinocytes or fibroblasts. PDT also activates several signal transduction pathways such as protein kinase cascades, cytokine expression, expression of early response genes, induction of glucose-regulated proteins and heat shock proteins leading to apoptosis [9], [10].

In PDT-induced apoptosis, cytoplasmic Hsp70 rapidly translocates to the cell surface after treatment, depending on the PDT dose. Moreover, the translocation of Hsp70 is related to the changes in mitochondrial transmembrane potential. With non-lethal PDT-induced surface stress, Hsp70 also translocates to the cell surface, but with a slower rate and a lower final surface-concentration [9], [10].

The induced cell surface expression and release of Hsp70 including its relevance for tumour response and skin rejuvenation is not fully understood [9–11] but may be of interest for monitoring, predicting or optimising treatment regimes. For this reason we were interested in giving an overview of the working mechanism of PDT, the role of HSPs during PDT as a non-thermal therapy with the foci of tumour eradication and skin rejuvenation.

Molecular mechanism of PDT

In the case of PDT-induced apoptosis, translocated cytoplasmic Hsp70 to the cell surface has been shown to have a crucial role in tumour treatment response [9], [10]. Additionally, it initiates cell reparation and wound healing leading ultimately to collagen remodelling [12], [13].

HSPs belong to a large family of protein chaperones involved in assisting protein folding and unfolding in cells. In addition to the induction by hyperthermia, HSPs are activated by a wide range of physical, chemical and biological stresses including various laser therapies or PDT [14–23].

Increasing levels of HSPs enhance the ability of cells to deal with the resultant accumulation of abnormally folded proteins, either facilitating the refolding of damaged proteins or participating in the synthesis of new proteins to replace those irreparably damaged. The most important HSPs are the family members of Hsp70 (72 kDa = Hsp72 and 73 kDa = Hsp73) and Hsp47. Both, Hsp72 and Hsp73 are present in the cytoplasm and nucleus of keratinocytes, fibroblasts and adipocytes. Hsp73 is synthesised constitutively in all mammalian cells and therefore is often referred to as the ‘constitutive Hsp70’. The synthesis of Hsp72 is usually restricted to the cell experiencing stress and therefore often is referred to as the ‘inducible Hsp70’. The two proteins exhibit extremely high sequence homology (95%) and similar biochemical properties [19]. The inducible response involves transcriptional activation mediated by heat shock factor (HSF) which binds to a specific heat shock element in the heat shock gene leading to the up-regulation of HSP after PDT [21–23].

Hsp47 is up-regulated at 42°C and higher temperatures [17] and is the only known stress molecule able to bind collagen [24]. It is found in the endoplasmic reticulum of cells producing collagen type I [25] and is actively involved in collagen type I biosynthesis [26]. Expression of Hsp47 is age related and its up-regulation by heat in older organisms is reduced [27]. Various conditions lead to the involvement of Hsp47 in the collagen metabolic pathway resulting in the modulation of collagen I synthesis rate [28–30]. Hsp47 was not up-regulated in in vitro fibroblasts after conventional PDT compared with a PDT that caused collagen damage consistent with the hypothesis that this collagen chaperone is usually up-regulated by modalities known to damage collagen such as laser hyperthermia [31], [32].

The molecular machinery that drives the apoptotic programme consists of a family of cysteine proteases, caspases, which cleave their substrates at specific aspartic acid residues [33]. In normal cells, caspases are constitutively expressed as inactive single polypeptide chains, procaspases, and their activation requires specific proteolytic cleavage [34], [35]. Active caspases can amplify apoptotic events by their ability to cleave their own precursor forms as well as those of other caspases [36], [37]. Mitochondria have a crucial function in initiating the cascade of caspase activation in response to different apoptotic signals [38]. Disruption of the outer mitochondrial membrane by apoptotic stimuli results in the release of cytochrome c into the cytoplasm [39]. Cytochrome c binds to the cytosolic apoptotic-protease activating factor 1 (Apaf-1), promoting Apaf-1-mediated activation of caspase 9 in an adenosine triphosphate/deoxyadenosine diphosphate (ATP/dADP)-dependent manner [40], [41]. Active caspase 9 can subsequently amplify the caspase cascade by its ability to process its own proenzyme as well as the effector caspase 3 [42]. Association of procaspase 9 with Apaf-1 is an essential step in its initial activation and is mediated by the CARD sequence located at the amino termini of both proteins [43] (Figure 3).

Figure 3. Apoptotic pathway induced by PDT.

Hsp70 as well as Hsp90 have been demonstrated to directly bind to the caspase recruitment domain (CARD) of Apaf-1, thereby preventing the oligomerisation of Apaf-1 and association of Apaf-1 with procaspase 9 [44–46]. Treatment of cells with diverse DNA-damaging agents dissociates the Hsp90–Apaf-1 complex relieving the inhibition of procaspase-9 activation. Release of cytochrome c from mitochondria by apoptotic signals induces ATP/dADP-dependent formation of the oligomeric Apaf-1-caspase-9 apoptosome, thereby activating caspase 3 [4], [47–49]. Furthermore, Hsp90 binding to Apaf-1 induces a potential conformational change in Apaf-1 and inhibits cytochrome c-ATP/dADP-mediated oligomerisation of Apaf-1, thereby exerting a dominant-negative effect on the function of Apaf-1 to activate procaspase-9 [44]. In receptor-mediated apoptosis, caspase 8 is mobilised before cytochrome c release in both normal and transformed cell types during PDT-induced apotosis.

A particularly pronounced surface Hsp70 expression was found on apoptotic cells and even more striking on detached dying cells 18 h post-PDT [10]. The cell surface expression of Hsp70 on keratinocytes and fibroblasts seems be to dose-dependent, persisting at a similar level between 15 and 120 min [50] up to 6 h post-PDT. Hsp70 expression was not detectable in control samples exposed to the photosensitiser or light source alone [10]. Under non-lethal PDT-induced surface stress, Hsp70 also translocated to the cell surface, but with a slower rate and a lower final surface concentration [47]. The reason for this rapid emergence of surface-expressed Hsp70 remains to be elucidated. It is possible that photo-oxidative lesions in the membrane of PDT-treated cells cause the inner membrane sections to invert to the outer surface [51] and expose HSP molecules associated with proteins or lipids lining the inner leaflet of the cell membrane. On the other hand, surface localised HSPs may have a role in the efforts of cells to stabilise damaged membranes and preserve their integrity [52]. When Hsp70 is expressed at the cell surface, the COOH-terminal domain is exposed whereas the ATP-binding site and NH₂-terminal domain remain unexposed within the membrane [10], [53]. The COOH-terminal domain is involved in the interaction of HSP with its chaperoned client peptides that may serve as a source of tumour rejection antigens [54], [55]. Hence, altered or damaged cell proteins are more likely to be found attached to the substrate-binding domain of HSP than normal non-damaged polypeptides [56] and could become abundantly exposed to immune surveillance elements following PDT. The level of Hsp70 expression in cancer cells was found to correlate with their immunogenicity [57] whereas specifically membrane expressed Hsp70 on such cells was shown to elicit tumour immunity and also correlate with sensitivity to allogeneic natural killer cells [58–61]. Further, the opsonisation of PDT-treated cells is reduced in the presence of Hsp70-specific antibodies [62]. Hsp70 directly binds complement proteins [63]. Hence, the opsonisation of PDT-treated cells by complement proteins may occur by direct binding to surface-expressed Hsp70 that were recognised as altered self targets by the complement. Moreover, various types of immune cells including macrophages, neutrophils, dendritic, natural killer and B cells have specific receptors for complement proteins as well as for HSPs [64] and are prone to use them to attack cells that are surface-expressing HSPs and/or are opsonised by complement proteins. In addition to the elevated expression of Hsp70 by cancer cells themselves [60], [61], [65], [66]. Tumours like squamous cell carcinomas contain significant numbers of tumour-associated leukocytes or are newly infiltrated because PDT induces a massive infiltration of these inflammatory cells [67]. Hsp70 released from PDT-treated tumour cells is captured by macrophages triggering in these cells toll-like receptor-based signal transduction (TLR2 and TLR4) activity with downstream activation of NF-κB, resulting in the production of inflammatory cytokine tumour necrosis factor α [1]. PDT inhibits Hsp90 from stabilising client proteins modulating tumour responsiveness following PDT [68].

Furthermore, PDT induces matrix metalloproteinases directly but also by an indirect paracrine loop mediated by soluble factors released by epidermal keratinocytes [69]. MMP1 initiates collagen breakdown and its gene expression was elevated 1 day post-PDT while returning to baseline levels within the following 24 h [70]. Prolyl-4-hydroxylase α, an enzyme involved in post-translational modification of collagen in fibroblasts, was significantly elevated 1 month after PDT [71]. It catalyses the hydroxylation of proline residues. This is required for collagen synthesis [72],[73] and correlates with Hsp47 [71].

PDT and the treatment of epithelial cancers

PDT is effective in the treatment of actinic keratoses, Bowen's disease and superficial BCCs, and has been approved in the USA for actinic keratoses since December 1999 [11], [74–76].

PDT is a non-thermal photochemical process and therefore does not cause deep tissue heating. This fact in combination with no or superficial eschar formation leads to excellent cosmetic outcomes. PDT is also considered a cost-effective therapy and may be an option in patients for whom surgery is not feasible [77].

However, in the therapy of BCCs, PDT is only recommended to patients with histological proven superficial and non-pigmented BCCs because two basic factors limit its application within cutaneous oncology [2]. First, the complete destruction of a tumour by PDT depends on a sufficiently high concentration and uniform distribution of PpIX in the malignant tissue. ALA is a hydrophilic molecule with a stratum corneum/water partition coefficient of 0.1 [78] and permeates poorly across intact skin [3]. Tumour cells have an increased metabolic rate and their ‘low ferrochelatase activity’ results in a reduced depletion of PPIX into haeme leading to a higher ALA content than in the surrounding normal cells.

Deep penetrating tumours such as nodular BCCs can be reached theoretically by systemically administered ALA; however, prolonged cutaneous photosensitivity is the major drawback [76], [79]. The second limiting factor is the penetration depth reached by the different light sources (see indroduction) [1]. To resolve these problems, the tumour mass can be reduced before PDT, photosensitisers with an increased tissue penetration potential have to be developed, they have to be applied deeper in the tissue using special techniques or have to be used in combination with deep penetrating wavelengths.

For this reason, PDT has been combined with different laser modalities. Different lasers (Er:YAG-, CO_2-, flashlamp-pumped dye laser) are themselves effective in the ablation of non-malignant, pre-malignant, and malignant tumours [80–83] and can reduce the tumour mass before PDT as seen in nodular BCCs [84]. Combining Er:YAG laser ablation (600–1000 mJ at a frequency of 7 Hz) reducing tumour size below 2 mm with subsequent ALA methyl ester PDT was significantly more effective than the two treatments alone, defined at 3, 6 and 12 months with a final efficacy of 98.97% versus 94.85% (PDT only) and 91.75% (Er:YAG laser only) [85]. The combined method also provided the best aesthetic results compared to PDT only and Er:YAG laser only, although 67% of patients preferred solitary Er:YAG laser treatment over the PDT method (20%) and the combined treatment (13%) because of its simplicity [85].

Furthermore, the ALA permeation can be increased by different chemical enhancers like dimethylsulphoxide [79], [86], iontophoresis [87–91] or destruction of the cornified layer. The formation of PpIX was also found to be strongly temperature dependent [38] with more formation at higher temperatures. Another approach to the problem is the pre-treatment of the tumour with a fractionated laser before application of ALA to enhance its tissue penetration depth. ALA applied to Er:YAG laser-treated skin produced a significantly higher accumulation of PpIX within superficial skin and subcutaneous tumours as compared to the non-treated group measured by confocal laser scanning microscopy. The pre-treatment with the Er:YAG laser at 1.4–2.0 J/cm² enhanced the PpIX concentration within the BCCs at 2 h post-application. On the other hand, there was no significant difference in PpIX accumulation at 4 and 6 h between laser and non-laser treated BCCs [92].

PDT utilises the destructive power of reactive oxygen species generated via visible light irradiation of a photosensitive dye accumulated in the cancerous tissue/cells and up-regulates HSPs including Hsp70, Hsp47 and Hsp90. Simultaneously, PDT leads to the infiltration of various immune cells (e.g. neutrophils) into the treated site as well as to the activation of different immune phenomena, e.g. acute-phase response, complement cascade and production of cytokines/chemokines [93–95].

PDT and skin rejuvenation

Based on the good healing and anecdotal reports of the reversal of some of the signs of photodamage post-PDT, PDT is being used for skin rejuvenation although the exact mechanism of the improvements of the appearance of sun-damaged skin has not been fully elucidated. There are a few studies which could show an improvement of skin texture and tone as well as improvement in skin elasticity [4–6], [96]. The results are similar to the changes induced by topical retinoids occuring more rapidly [97–99]. PDT stimulated epidermal proliferation, resulting in an increase in epidermal thickness. This was parallelled by an up-regulation of Ki 67 and cytokeratin 16 2 to 7 days post-PDT. Also, procollagen I and III mRNA was up-regulated followed by the up-regulation of the corresponding proteins. As a marker of photodamage, elevated p53 might imply greater loss of normal barrier function resulting in a greater conversion of ALA to PpIX in the skin [71]. Because measurements of epidermal thickness and collagen I production returned to baseline values within a few months, maintenance therapy would be required to maintain clinical benefits. An increased amount of collagen could be due to the induction of Hsp70 and subsequent increase of Hsp47 such as after thermal ablative and thermal, non-ablative laser therapies [100–102], but has not been investigated until now.

Outlook

PDT is an effective treatment alternative for actinic keratoses, Bowen's disease and superficial BCCs as well as skin rejuvenation besides other more invasive therapy options.

In the case of PDT-induced apoptosis, translocated cytoplasmic Hsp70 to the cell surface has been shown to have a crucial role in treatment response [9], [10] due to its inhibitiory effect of both apoptotic and necrotic pathways [103]. Monitoring of Hsp70 expressions and other influencing factors such as skin cooling in clinical trials seem to be of interest to optimise treatment regimens for both indications.

In the future, combinations of PDT with fractionated laser regimen or other skin rejuvenation therapies could possibly improve clinical outcome in skin rejuvenation indications. Further, clinical studies comparing different light sources and photosensitisers seem to be important to optimise treatment outcome.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.