Rationale of hyperthermia for radio(chemo)therapy and immune responses in patients with bladder cancer: Biological concepts, clinical data, interdisciplinary treatment decisions and biological tumour imaging

Abstract

Bladder cancer, the most common tumour of the urinary tract, ranks fifth among all tumour entities. While local treatment or intravesical instillation of bacillus Calmette–Guerin (BCG) provides a treatment option for non-muscle invasive bladder cancer of low grade, surgery or radio(chemo)therapy (RT) are frequently applied in high grade tumours. It remains a matter of debate whether surgery or RT is superior with respect to clinical outcome and quality of life. Surgical resection of bladder cancer can be limited by acute side effects, whereas, RT, which offers a non-invasive treatment option with organ- and functional conservation, can cause long-term side effects. Bladder toxicity by RT mainly depends on the total irradiation dose, fraction size and tumour volume. Therefore, novel approaches are needed to improve clinical outcome. Local tumour hyperthermia is currently used either as an ablation therapy or in combination with RT to enhance anti-tumour effects. In combination with RT an increase of the temperature in the bladder stimulates the local blood flow and as a result can improve the oxygenation state of the tumour, which in turn enhances radiation-induced DNA damage and drug toxicity. Hyperthermia at high temperatures can also directly kill cells, particularly in tumour areas which are poorly perfused, hypoxic or have a low tissue pH. This review summarises current knowledge relating to the role of hyperthermia in RT to treat bladder cancer, the induction and manifestation of immunological responses induced by hyperthermia, and the utilisation of the stress proteins as tumour-specific targets for tumour detection and monitoring of therapeutic outcome.

Keywords:

• Adaptive immunity
• cytotoxicity
• heat shock proteins
• hyperthermia
• innate immunity
• molecular imaging

Biology of the urinary bladder and radiation toxicity

Due to the specialised biology of the urinary bladder, late side effects can occur after therapeutic irradiation. The urinary bladder consists of a mucosa with transitional epithelium, connective tissue, nerve fibres and three muscle layers [1]. The internal lining of the bladder wall which is termed urothelium consists of 3–7 layers of transitional cells that reside on the basal membrane of the extracellular matrix that contains collagen and adhesive glycoproteins, for example. The basal membrane includes actively proliferating cells and umbrella-like luminal cells. The lamina propria, which consists of connective tissue and smooth muscle fibres, is located below the basal membrane. The muscle cell layers converge at the bladder neck to form an inner, middle and outer muscle cell layer, whereby the outer layer extends to the female urethra or male prostate which are responsible for the necessary sphincter functions.

The underlying mechanisms of bladder toxicity after radio(chemo)therapy (RT) are presently not fully understood. The urothelium with its low proliferative capacity (8–12 days turnover rate) is not considered a classical radiation-sensitive tissue. However, approximately 3 months after irradiation urothelial cells show first signs of radiation-induced damage including nuclear irregularities and cellular oedema. At 6–12 months after irradiation a dose-dependent increase in the proliferative activity can be observed in urothelial cells [1]. Although changes in the turnover rate of bladder epithelium cells and cell death cannot be considered as precursors of late radiation side effects, morphological changes such as the appearance of immature cells, changes in tight junctions and normal polysaccharide layers have been found to correlate with the onset of irritative clinical symptoms such as urinary frequency and urgency.

Regarding the vasculature, proliferation of endothelial cells can be observed at 6 months, and vascular fibrosis can develop 6–12 months after RT [1]. Apart from the endothelial system, the bladder musculature also appears to be sensitive towards ionising irradiation. Shortly after RT smooth muscles can be replaced by fibroblasts and thus result in an increased collagen deposition which reduces the bladder capacity [2,3]. Late bladder fibrosis appears to be associated with vascular ischaemia of the bladder wall. Therefore, the endothelial rather than urothelial system should be considered as a target for late RT-induced damage that typically occurs many years after exposure to irradiation [4,5]. Despite open questions with respect to urinary frequency or other sequelae as end points to characterise radiation-induced side effects, radiation biology teaches that bladder toxicity is a rather late event. Early damage does not necessary correlate with late sequelae. Acute bladder side effects which typically occur days and weeks after radiotherapy are more likely due to cell death, ulceration and increased mitotic activity of epithelial cells [6,7]. Changes in the reservoir function of the bladder appear not to be related to direct damage of nerve cells of the bladder but rather to epithelial necrosis which can favour the development of secondary infections [1–3]. Regarding these findings the question arises about the role of the dose and fractionation of irradiation in bladder toxicity. Early data demonstrated that the α/β ratio of bladder tissue ranges between 7 Gy and 8 Gy [8]. Bentzen and Thames focussed on the evaluation of the sensitivity of fractionated irradiation on late sequelae in the bladder. Their experimental data demonstrated that the α/β ratio of this tissue was lower (5.8 Gy), and thus ranges between early and late responding tissues [9,10].

Clinical data and current treatment recommendations

Radiation of bladder cancer with RT provides a non-invasive treatment approach which appears to be superior, especially for elderly patients. However, elderly patients frequently have urinary problems prior to the start of RT, which must be kept in mind when scoring RT-induced toxicity. Optimisation of RT includes a reduction of the irradiated volume of the bladder using innovative target volume concepts and the implementation of novel image-guidance techniques which allow adaption of the radiation dose to changes in organ and tumour motion as well as biological tumour imaging [11]. In this regard, novel combinations of magnetic resonance imaging and radiotherapy might further improve the therapeutic window [12]. Biologically, stimulation of the patient’s immune system against the tumour, and combinations of radiotherapy with cytostatic drugs, small molecules and hyperthermia are promising strategies to improve clinical outcome. Transurethral resection of the bladder tumour (TURB) followed by RT provides another alternative to primary cystectomy in high-risk superficial and muscle-invasive bladder cancer, with the advantage that the bladder can be maintained [13]. This organ-preserving therapy is successful in approximately 75% of cases, with a 5-year overall survival (OS) rate of 50–60% [14,15]. In this concept, cystectomy is reserved for patients with incomplete response or local relapse. Positive determining factors for bladder maintenance are: 1) tumour size of <5 cm, 2) early tumour stage, 3) complete TURB, 4) absence of urethral obstruction, and 5) no evidence of pelvic lymph node metastases [14]. The strongest prognostic factor for OS is complete visible and microscopic tumour resection [14].

The standard treatment procedure of bladder cancer is dependent on the actual tumour stage. Superficial high-risk transitional cell carcinomas Ta/T1 G3 and Tis show high local relapse rates as well as progression to higher muscle-invasive T2 stages during the follow-up period after sole transurethral resection. For this reason, the European Association of Urology decided to perform intravesical immune-instillation of bacillus Calmette–Guérin (BCG) after complete TURB. Only in case of a relapse was salvage cystectomy necessary. Muscle-invasive carcinomas ≥ T2 are mostly treated with cystectomy and pelvic lymphadenectomy and, dependent on the pathological T and N stage, adjuvant chemotherapy such as cisplatine and gemcitabine treatment has to be included. An alternative to radical cystectomy in patients with bladder cancer is a trimodality treatment which includes radical TUR and radiochemotherapy. With this treatment regimen comparable survival rates after 5 and 10 years can be achieved and approximately 80% of these patients benefit from bladder maintenance [14].

Radiotherapy is carried out with intensity-modulated radiotherapy (IMRT) or is 3D-planned on a linear accelerator. The bladder and pelvic lymph nodes are treated up to a total dose of 45–50 Gy with single doses of 1.8–2.0 Gy. After that a sequential boost up to a dose of 60 Gy or more is given to the bladder. Within the first five weeks the irradiation field includes the pelvic lymph nodes. In order to spare radiation dose to the small intestine, all irradiations should be carried out with a full bladder. After a radiation-induced shrinkage of the field, the patient should present with an empty bladder to get a minimum dose to the healthy tissue. Dose escalation and hyperfractionated irradiation can improve response and survival rates including OS [16]. However, radiation therapy alone does not induce comparable response rates like cystectomy. In both T2 and T3/4a bladder cancer patients, the 5-year OS of radiotherapy alone was inferior to surgery [17,18]. The combination of radiotherapy with chemotherapy (cisplatin, carboplatin, mitomycin C, 5-fluorouracil) significantly improved remission rates and OS compared to that which is achieved by surgical resection. As an advantage, this combined therapeutic approach enabled bladder maintenance in approximately 80% of the patients [14,19]. Best results in a single institutional study were obtained when radiotherapy was combined with combination cisplatin and 5-fluorouracil [20]. Similar radiosensitising effects as cisplatin can be achieved with gemcitabine [21], paclitaxel plus/minus carboplatin [22,23] and capecitabine [24].

Clinical and biological rationale of the implementation of hyperthermia in the treatment of bladder cancer

The success of organ-preserving RT in bladder cancer can be further increased by the integration of regional deep hyperthermia as a further treatment modality. The increase of the temperature within the tumour above 40 °C (up to 44 °C) stimulates the local blood flow and thus improves tumour oxygenation and thus enhances radiation-induced DNA damage and also drug cytotoxicity [25]. Elevated temperatures also can cause lethal effects on tumour cells, particularly in tumour areas which are poorly perfused, hypoxic, or have a low tissue pH. Furthermore, other indirect effects of hyperthermia have been described including the induction of secondary cell death by vascular damage within the tumour, surface expression and release of stress-induced danger-associated molecules including heat shock proteins (HSPs), and the stimulation of immune competent effector cells [26].

Biologically, hyperthermia above a temperature of approximately 42.5 °C can induce direct tumour cell killing. Depending on the mode of cell death (i.e. primary and secondary apoptosis, necrosis, mitotic catastrophe) dying tumour cells either can induce anti- or pro-inflammatory effects mediated through the adaptive and innate immune system [27–30]. It becomes more and more evident that long-term tumour control and prevention of distant metastases highly depends on the capacity of the host’s immune system [31–33].

Adaptive immunity and hyperthermia

The role of hyperthermia on adaptive immunity is temperature dependent. Some studies have reported hyperthermia to induce immunogenicity in human colon adenocarcinomas, human lung cancer and melanoma cell lines by increasing the accessibility of novel tumour-associated antigens [31–33] and Hsp70 on the cell membrane of tumour cells and in the extracellular milieu [34–38], whereas others indicated that hyperthermia can reduce tumour immunogenicity [39]. It has been shown that temperatures below 42.5 °C mostly do not affect the expression of major histocompatibility class I antigens (MHC class I), whereas temperatures above 45 °C result in a reduced cell surface expression of MHC class I antigens on human tumour cell lines that might negatively affect T cell-mediated immunity [40–42]. Contradictory reports also exist with respect to the role of hyperthermia on the presentation of exogenous antigens by MHC class II molecules and its co-stimulatory capacity [43–45]. The activation of responding T cells by antigen presenting cells (APCs) takes place in the draining lymph nodes, and thus the migratory activity of APCs and lymphocytes also have to be considered. Hyperthermia has been shown to increase the migratory activity of APCs and pro-inflammatory T cells, which is beneficial for an effective anti-tumour immune response [46–48]. In line with these findings, bladder cancer-derived exosomes that express Hsp70 have been shown to stimulate cytotoxic T lymphocyte responses in vitro [49]. Hyperthermia also has been shown to promote the migratory capacity of anti-inflammatory, immunoregulatory T cells (T_regs), which attenuate protective anti-tumour immunity of cytotoxic T and natural killer (NK) cells. T_regs belong to a subset of CD4⁺ T cells which constitutively express the α chain of the IL-2 receptor (CD25) [50–52], and the intracellular transcription factor Forkhead box p3 (Foxp3) [53]. The lack of these cells triggers autoimmune destruction of a variety of tissues [54,55], and they are regarded as being responsible for peripheral self-tolerance [50,53,56]. Interestingly, CD4⁺CD25⁺ immunoregulatory T (T_reg) cells are frequently elevated in cancer patients [57,58] and thus prevent the induction of protective anti-tumour immunity [59–66].

The cytolytic activity of cytotoxic CD8⁺ T cells involves Fas/Fas ligand (FasL)-induced cell death and/or the release of cytolytic granules containing perforin and granzymes [67]. Hyperthermia has been reported to increase FasL (CD95L) expression on cytolytic T cells and thereby enhances their tumour killing activity [68]. On the other hand there is a potential influence of hyperthermia on the expression of cytoprotective molecules such as Hsp70 in the cytosol of tumour cells. Intracellular Hsp70, which is strongly induced by hyperthermia, is known to inhibit certain apoptotic pathways. However, at present no evidence exists that Hsp70 blocks the perforin/granzyme pathway [40]. A dual targeting of members of the HSP70 and HSP90 families has been shown to promote tumour cell death and enhances anti-tumour effects in bladder cancer [69].

Innate immunity and hyperthermia

NK cells are a subset of innate immune cells that express a number of inhibitory and activatory receptors (immunoglobulin type, C-type lectin, natural cytotoxicity receptors) to differentiate between self and non-self. They are considered as the first line of defence against viral and bacterial infections and also against cancer. According to the ‘missing-self’ theory, the absence of MHC class I molecules on tumour cells allows NK cells that express activating, immunoglobulin-type receptors to recognise these tumour cells that are immunologically invisible to cytotoxic T cells [70,71].

Hyperthermia has been found to increase the susceptibility of tumour cells to the cytolytic activity of NK and lymphokine-activated killer (LAK) cells [27]. At fever-range temperatures (≥39.5 °C), Dayanc and colleagues have reported that NK cell cytotoxicity is enhanced towards the non-classical MHC molecule MICA [28,29], and that this effect is dependent on the clustering of the activatory C-type lectin receptor NKG2D.

Already under physiological conditions, Hsp70 is frequently found on the plasma membrane of solid tumour cells [34], including bladder cancer cells (unpublished observation), whereas the corresponding normal tissues are always found to be Hsp70 membrane-negative [72,73]. A Hsp70 membrane-positive tumour phenotype was found to be associated with a high expression density of the lipid raft component globoyltriaosylceramide (Gb3), which serves as an anchorage lipid for Hsp70 in the plasma membrane of tumour cells [74,75]. Following stress, including elevated temperatures, the expression density of Hsp70 on the cell surface is further up-regulated. An incubation of NK cells with pro-inflammatory cytokines plus Hsp70 or peptides derived thereof [76–78] up-regulates the expression density of activatory C-type lectin receptors including NKG2C, CD94 and NKG2D on NK cells [79–81]. The tumour-specific membrane expression of Hsp70 on tumour cells renders them better targets for these pre-activated NK cells.

Another subset of the innate immune system involves monocytes/macrophages. Macrophages are highly versatile, multifunctional cells that phagocytose cellular debris, secrete a wide array of immunomodulatory cytokines, present antigens to T cells, and act as accessory cells in lymphocyte activation. Macrophages are sub-divided into M1 (classically activated, pro-inflammatory) or M2 (alternatively activated, anti-inflammatory) phenotypes. M1 macrophages are activated by lipopolysaccharide (LPS) and interferon (IFN)-γ to secrete bactericidal factors that promote T helper 1 (Th1) responses which are characterised by the secretion of IFN-γ and tumour necrosis factor (TNF)-α and the activation of CD8⁺ T cells. M2 macrophages are considered to be immunosuppressive and promote T helper 2 (Th2) responses which are characterised by the secretion of anti-inflammatory IL-4, -5, -6, -10 and -13, the induction of B cell activation, antibody class switching, and the production of antibodies [82]. Monocyte recruitment is initiated by chemoattractants that are secreted by both malignant and stromal cells in tumours including CCL2 (MCP-1), placental growth factor (PlGF) [83], CCL3 (MIP-1α), CCL4 (MIP-1β) and CCL5 (RANTES) [84]. Several clinical studies have shown a correlation between a high number of tumour-associated macrophages and an increased microvessel density [85–89], which is induced by angiogenesis-modulating factors such as vascular endothelial growth factor (VEGF), basic fibroblast growth factor (b-FGF), tumour necrosis factor-α (TNF-α), interleukin-1β (IL-1β), CXCL8, cyclooxygenase 2 (COX2), urokinase plasminogen activator (uPA), platelet-derived growth factor (PDGF) and matrix metalloproteinases (MMPs) 7, 9 and 12 [90]. A significant increase in the number of tumour-infiltrating macrophages and CD8⁺ T cells and a reduction in tumour volume have been observed in different tumour mouse models following exposure to hyperthermia [91–93].

Hypoxia-induced chemoattractants such as VEGF, endothelins and endothelial monocyte-activating polypeptide II (EMAP-II) which are present in hypoxic/necrotic areas attract tumour-associated macrophages [94] in different human tumours including breast [95], endometrium [96], ovary [97], bladder [86], colon [98] and tumours of the oral cavity [99]. An up-regulation of hypoxia-inducible transcription factors (primarily HIF-1 and HIF-2) by macrophages [100] induces the transcription of genes that regulate cell proliferation, metabolism, apoptosis and angiogenesis [101] and thus mediate tumour progression, and hypoxic tumours also tend to be more resistant to irradiation. Hyperthermia has been shown to sensitise hypoxic tumour cells to radiation treatment by improving tumour cell oxygenation and by enhancing direct tumour cell killing [102,103].

The existing data on the effects on the adaptive and innate immunity provide conflicting results. On the one hand hyperthermia can exert pro-inflammatory anti-tumour immune responses; on the other hand it can support migration of immune suppressive T_regs or support tumour growth and angiogenesis via switching of M1 to M2 macrophages. Novel studies that analyse both the tumour and its tumour microenvironment including the adaptive and innate immune system are necessary for a better understanding of the effects of tumour hyperthermia in bladder cancer and other cancer entities.

Hyperthermia-induced Hsp70 as a target for biological tumour detection and monitoring of outcome

Hsp70 which is overexpressed in many tumour entities including bladder cancer is also found as a membrane-bound molecule on tumour cells. Hyperthermia and RT have been found to further increase the density of cell-surface-expressed Hsp70 on tumour, but not normal cells [104]. Therefore, targeting of membrane-bound Hsp70 with either fluorescence-or radionuclide-labelled Hsp70 reagents offers a unique possibility to image the therapeutic outcome of hyperthermia and RT. A monoclonal antibody (cmHsp70.1) [105], a Fab fragment derived thereof [106] and a tumour penetrating peptide (TPP) [107,108] are presently available to detect membrane-bound Hsp70 on viable tumour cells in vitro and in vivo. Apart from Hsp70, IntegriSense, which is known to detect alpha v beta 3 integrin [109], and CXCR4, a G-protein-coupled receptor [110] can serve as tumour-associated biomarkers, respectively, for a large variety of different tumour entities including bladder cancer. Carbonic anhydrase IX [111,112], glycerophosphodiester phosphodiesterase domain-containing 3 (GDPD3) and sprouty-related EVH1 domain-containing 1 (SPRED-1), which regulates the Ras/MAPK pathway [113], have been identified as other candidate biomarkers. Although nuclear imaging of tumours has been shown to improve risk stratification in renal cancer [114], the presently available markers have not yet been validated in patients with bladder cancer.

The laboratory of Pockley et al. was among the first to show that Hsp70 is also present in the peripheral circulation of patients with inflammatory diseases [115]. Elevated Hsp70 levels in the circulation are also found in patients with different tumour entities including colorectal cancer [116], prostate cancer [117] and bladder cancer (unpublished). Our group has demonstrated that Hsp70 is actively released from tumours not as a free protein, but in lipid vesicles which show biophysical and biochemical properties of exosomes. It appeared that exosomes that are derived from tumours with an Hsp70 membrane-positive phenotype also present Hsp70 on their lipid surface, whereas exosomes derived from Hsp70 membrane-negative tumour cells did not present Hsp70 on their exosomal surface [35]. We could further demonstrate that the amount of Hsp70 in the circulation of tumour-bearing mice is associated with the viable tumour mass, and following therapy-induced tumour reduction, the Hsp70 serum levels also dropped [36]. By using our novel lipHsp70 ELISA which is able to quantitatively determine lipid-bound Hsp70 in the serum/plasma of human donors [37], we could show that the levels of Hsp70 in the circulation was significantly higher in tumour patients compared to healthy controls but also compared to patients with inflammatory diseases [38]. In the future, this assay will be used to correlate clinical outcome of RT plus/minus hyperthermia with soluble Hsp70 levels in liquid biopsies of tumour patients including those with bladder cancer. In line with our results, other groups propose to measure stress proteins and cytokines in the urine as well as in tumour sections as biomarkers to diagnose for bladder cancer [118,119].

Clinical data on hyperthermia in oncology and potential rationale for bladder cancer patients

In clinical studies, hyperthermia has been combined with radiotherapy (RT) and/or chemotherapy as a radio- and chemo-sensitiser in different tumour types, including brain, head and neck, lung, breast, cervix, bladder, sarcoma and melanoma [120–123]. The Dutch Deep Hyperthermia Group published a significant increase of complete response rates when adding hyperthermia to radiotherapy in locally advanced bladder cancer > T1 (73% vs. 51%) [120]. In another prospective trial, Colombo et al. combined hyperthermia to intravesical mitomycin C instillation and could show a significantly lower recurrence rate after 2 years (17% vs. 58%) for intermediate- and high-risk superficial bladder cancer [124]. Wittlinger et al. reported on safety and effectiveness data for the quadrimodal treatment consisting of RT after TURB combined with regional hyperthermia therapy (RHT) in patients with T1 high-risk and T2 bladder cancer [125]. They found acute toxicity rates grade 3 and 4 in 20% and 9%, and late toxicity grades 3/4 of 24%, and thus therapy was rated as feasible and well tolerated. At 3 years local recurrence-free survival was 85%, OS was 80% with a bladder-preserving rate of 96% [125].

Figure 1. Schematic representation of beneficial anti-tumour effects induced by hyperthermia on the immune system and tumour microenvironment. Local tumour hyperthermia can increase tumour perfusion and thus reduce tumour hypoxia. Hyperthermia can induce the expression of adhesion molecules like selectins and ICAM-1 on tumour endothelial cells (ECs) which may facilitate leucocyte adherence. Hyperthermia can induce the expression of stress-induced ligands such as Hsp70 and MICA/B which are recognised by C-type lectin receptors on NK cells and CD8⁺ cytotoxic T cells, and thus can increase their cytolytic activity. High temperatures have also been shown to kill T regs which are known to repress immune responses. CD4⁺ as well as CD8⁺ T cells are attracted by an up-regulated expression of adhesion molecules on tumour blood vessels following exposure to heat. Hyperthermia can activate antigen presenting cells (APCs) and thus increase their expression of MHC class I/II molecules, CD40, CD86, CD80, and Toll-like receptors. The elevated expression and release of heat-induced tumour neo-antigens, either as free molecules or as exosomes which are cross-presented by APCs following uptake, induce an increased T cell signalling and an increased secretion of IFN-y by cytotoxic CD8⁺ T cells.

Regarding hyperthermia in bladder cancer, different systems for heating are used. Most centres use RHT utilising 70–120 MHz antennas. Minor centres work with intracavitary radiofrequency hyperthermia or intravesical therapy using heated perfusates. Geijsen et al. reported good feasibility of a combination of 70 MHz RHT and intravesical mitomycin C with low toxicity rates and excellent bladder temperatures in 20 patients [126]. Treatment consisted of six sessions once a week followed by four further maintenance sessions every 3 months. In RHT, antennas are positioned around the patient and deliver homogeneous temperatures to the defined organ [127]. Mostly, platinum-based chemotherapy is applied simultaneously during RHT, and RT starts within an interval of 60 min after RHT.

Conclusion

Local hyperthermia has been found to sensitise tumours against irradiation and chemotherapy in preclinical models [128,129]. Pilot feasibility trials focus on the role of regional hyperthermia [130] as a sensitiser for RT in the treatment of non-muscle-invasive bladder cancer [129,131,132] and other solid tumour entities [133–137]. Apart from its radiosensitising effects, hyperthermia has been found to increase the membrane density of Hsp70 on tumour cells, and thus might provide a therapy-induced biomarker for detection of tumours and monitoring of clinical outcome. Similar to RT [138], hyperthermia has also been found to impact the innate and adaptive immunity to fight distant cancer and metastasis [139,140]. Therefore, one might speculate that a combined treatment regimen consisting of RT and hyperthermia not only sensitises tumours towards radio(chemo)therapy but also might elicit immune stimulation.

Acknowledgements

We thank Anett Lange for editorial support.

Disclosure statement

Gabriele Multhoff’s laboratory is supported, in part, by grants from the Deutsche Forschungsgemeinschaft (DFG SFB824/2, INST 411/37-1 FUGG, INST 95/980-1 FUGG), Munich Advanced Photonics (MAP), the Bundesministerium für Forschung und Technologie (BMBF 02NUK038A, 01GU0823, DKTK-ROG), EU CELLEUROPE (EU315963), Helmholtz Zentrum München, German Research Center for Environmental Health (G-501390-001), and by multimmune GmbH, Munich. The authors alone are responsible for the content and preparation of this paper.