Effect of irradiation on the expression of E-cadherin and β-catenin in early and late radiation sequelae of the urinary bladder and its modulation by NF-κB inhibitor thalidomide

Radiotherapy is an essential modality of cancer treatment that is applied to around 50% of human cancer patients in the form of a primary, adjuvant or neoadjuvant therapy [1, 2]. Despite high precision tumor targeting achieved by modern radiotherapy (by intensity-modulated radiation therapy, image-guided treatments, etc.), normal tissues adjacent to the tumor are exposed to a radiation dose which is associated with adverse effects [3, 4]. An organ at risk during pelvic radiotherapy is the urinary bladder and two phases of an inflammatory response are reported to occur after exposure [5]. Early symptoms are described as early radiation cystitis (ERC) and they are characterized by an increased miction rate correlating with a reduction in urinary bladder storage capacity, dysuria, nycturia and hematuria. ERC occurs days to weeks after the start of radiotherapy and usually resolves within several weeks, treated only symptomatically to alleviate clinical symptoms [5‐7]. After a dose-dependent, symptom-free latency period, which can last for several months or up to decades, a progressive and irreversible chronic phase occurs, often referred to as late radiogenic sequelae (LRS) of the bladder (also referred to as hemorrhagic cystitis or radiation cystitis), for which no standard treatment guidelines are available [5‐8]. Besides clinical symptoms similar to the ERC, the late phase is characterized by fibrotic changes of the bladder wall leading to an enormous loss of bladder volume and urinary incontinence due to detrusor dysfunction [9].

Currently, the only possibility to decrease the severity of LRS is a reduction of the radiation dose, resulting in lower tumor control and survival rate [10]. In order to increase the patient’s quality of life with the same tumor control rate, treatment methods for LRS are urgently needed. LRS in the urinary bladder are directly modified by severity and duration of early response reactions, known as consequential late effects (CLE) [11]. Presumably, radiation-induced impairment of the protective barrier function of the urothelial wall leads to additional injury of the underlying tissue by permeating urine, which increases the severity of the late effects [8, 12, 13]. The adherens junction (AJ) proteins E‑cadherin and β‑catenin are important for epithelial barrier function [14, 15]. E‑cadherin is a calcium-dependent cell–cell adhesion molecule, which is linked by β‑catenin and other proteins to the actin cytoskeleton [14]. Radiation-induced adverse effects on the E‑cadherin/β-catenin complex have already been discovered in different tissues [16]. In addition, E‑cadherin/β-catenin are involved in fibrotic diseases of, for example, lung or kidney [14].

Present treatment methods for ERC did not show any ameliorating effects on LRS, but current research data suggested that the nuclear factor-kappa B (NF-κB) pathway is involved in the inflammatory processes [17]. Functional studies in the same mouse model as in the current study assessed the time course of ERC and LRS by transurethral cystotonometry and immunohistochemical analyses and showed a biphasic activation of the NF-κB proteins p50 and p65 during ERC and p50 reactivation in the late phase [17]. Furthermore, the systemic administration of the NF-κB inhibitor thalidomide during the early phase had promising beneficial effects on ERC and the incidence and severity of the LRS [17]. It reduced NF-κB activation and shifted the ED50 value for early radiation cystitis and late radiation sequelae to higher doses [17]. Mitigating effects of thalidomide in radiation-induced lung fibrosis had been discovered as well [19].

The aim of the present study was to investigate as a follow-up to our previous study [17] the effects of irradiation on the AJ proteins E‑cadherin and β‑catenin and their modulation by the NF-κB inhibitor thalidomide. The urothelial expression of E‑cadherin and β‑catenin was investigated in single-dose experiments in the same mice with a follow-up of 360 days.

The present study investigated the effect of irradiation on the expression of E‑cadherin and β‑catenin in the urothelial wall and its modulation by NF-kB inhibitor thalidomide in a mouse model.

The E‑cadherin/β-catenin complex is crucial for epithelial barrier integrity and studies in intestinal epithelium of mice and in the dermis of a mouse footpad model suggest that radiation can harm AJ [16, 22]. In the mouse footpad model two distinctive phases of the damage response were identified [16]. An early phase with loss of E‑cadherin mediated cell contact, and a regenerative phase, during which Wnt/β-catenin signaling was activated [16]. For the mouse urothelium a biphasic early radiation response phase from day 0–15 (early phase 1) and from day 16–30 (early phase 2), followed by a late response phase is well characterized [8, 17, 18, 23, 24], but information about changes in AJ protein expression is not available. Therefore, we investigated the effect of irradiation on E‑cadherin and β‑catenin in the urothelium during ERC and LRS. In contrast to skin epidermis, urothelium is not characterized by a fast cell turn-over and no breakdown of cell contacts is described during the early radiation response [8]. However, it is known that total cell number of the urothelium decreases in the early phase 1 to a first nadir on day 13, followed by slight recovery at the beginning of the early phase 2 and a second and more pronounced nadir at the end of the early phase [25]. The most pronounced cell loss is observed in the superficial layer with a progressive reduction of umbrella cells accompanied by a loss of luminal uroplakin III until the end of the early phase [8]. Results of our study suggest a compensatory induction of AJ to maintain and re-establish urothelial barrier integrity and counteract urothelial cell loss and damage to the permeability layer. A first increase in both AJ occurred during the first early phase followed by a stronger upregulation in the second early phase. The time course of E‑cadherin and β‑catenin upregulation is comparable to the time course of the functional bladder response [17]. In our previous study in the same mice functional bladder impairment reached a peak in the ERC around day 18 to 21 post irradiation [17]. E‑cadherin expression reached a peak around day 18 to 21, like the peak in early functional bladder response [17]. The β‑catenin expression showed a peak around day 24 to 27. Our results are consistent with human studies that report elevated E‑cadherin levels in the bladder urothelium after long-term exposure to low-dose ionizing radiation as well as in patients with interstitial cystitis [26, 27].

Parallel to the direct loss of urothelial cells, an inflammatory response is induced by irradiation that contributes to the clinical manifestation of ERC [17, 23, 25]. NF-κB signaling is involved in this inflammatory response [17]. In our previous study, we demonstrated a biphasic activation of the NF-κB proteins p50 and p65 in the urothelium of the same mice as in the current study during the early radiation response [17]. P50/p65 heterodimers are known for the induction of inflammatory genes [28, 29]. The NF-κB-inhibitor thalidomide can reduce activation of p50 and p65 as demonstrated in our previous study [17]. Early administration of thalidomide lowered the activation of p50 and p65, reduced the incidence of animals with functional bladder impairment in the early phase and significantly increased the ED50 dose for ERC [17]. The therapeutic effect of thalidomide was most pronounced in the THA1–15 group [17]. In our current study thalidomide treatment in the early phase 1 (THA1–15) resulted in an increased expression of E‑cadherin and β‑catenin at the cell membrane. Especially at the beginning of the early phase 1 a significantly steeper rise was observed and this effect was more pronounced for β‑catenin. In the early phase 2, a significantly higher and extended plateau was noticed in comparison to the irradiation control. Multifaceted crosstalk between NF-κB and Wnt/β-catenin as well as NF-κB and E‑cadherin has been described in the context of inflammation, proliferation and neoplastic disease [30‐42]. Direct or indirect inhibition of NF-κB can upregulate E‑cadherin and co-localization with β‑catenin [31‐36, 38]. Also, COX‑2 can influence E‑cadherin expression through NF-κB and irradiation-induced upregulation of COX‑2 in the ERC has been shown in the same mouse model in an older study [23, 36]. It seems that inhibition of NF-κB represses acute inflammatory signals on E‑cadherin/β-catenin and therefore indirectly enhances the postulated compensatory effects to stabilize urothelial barrier integrity, for which adherens junction assembly is crucial [14, 15]. Consistent with our findings, radiation-induced dermatitis was ameliorated in a rodent model by reducing E‑cadherin/β-catenin degradation by genistein or dasatinib treatment [16]. Consistent with functional data of our previous study, administration of thalidomide in the early phase 2 did not achieve the same significant modulation of E‑cadherin and β‑catenin, most likely because the initial inflammatory cascade is not interrupted. Repression of the initial of inflammation via NF-κB inhibition and subsequent elevation of E‑cadherin has also been reported in the treatment of pulmonary fibrosis [35]. Further, an early treatment time point was also essential to achieve an ameliorating effect on the early bladder response with the human keratinocyte growth factor palifermin in the same mouse model [43].

At the beginning of the late phase clinical signs of ERC have resolved. However, a depletion of total urothelial cell number, especially superficial umbrella cells and reduced luminal uroplakin III is still present at that time, indicating long lasting radiation-induced damage to the urothelium with low compensatory potential [8, 18, 23, 25]. Restoration of the urothelium occurs around day 180 and ongoing proliferative activity results in urothelial hyperplasia at the end of the late phase [25]. In our study, only E‑cadherin but not β‑catenin showed significantly higher expression levels at the beginning of the late phase until day 180 compared to the non-irradiated control. Thereafter a radiation-induced upregulation was seen for both AJ proteins until the end of the late phase. Functional data from our previous study demonstrated a mean complication-free survival of 166 days prior to clinical onset of LRS [17].

Furthermore, only NF-κB protein p50 but not p65 was activated during the late phase in our previous study [17]. P50 homodimers are associated with the upregulation of driving factors for chronic inflammation [17, 28, 44‐48]. Thalidomide treatment during this late phase (THA150–180) achieved no relevant treatment effect in our previous functional study and it seems that thalidomide has no long-lasting effects on chronic NF-κB activation [17]. Thalidomide treatment during the late phase (THA150–180) had also no significant effect on E‑cadherin expression and only minor effects on β‑catenin with an earlier rise on day 180 and a higher expression on day 300 than in the irradiated control.

However, treatment with thalidomide during the early radiation response revealed a significant therapeutic effect on LRS with an extended median complication-free survival time of 300 days and a significantly higher ED50 dose [17]. Severity of radiation response in early phase 2 has been shown to correlate significantly with the development of a late response and is itself influenced by the severity of the first acute wave, suggesting a significant consequential component [18]. Depletion of the bladder urothelium and damage to the permeability barrier is most pronounced at the end of the early phase, resulting in leakage of hyperosmolar urine into deeper subepithelial connective tissues as well as muscle layers [8]. In our study, early thalidomide treatment (THA1–15) resulted in an extended upregulation at the end of the early phase 2 for both AJ proteins. This effect was especially pronounced for E‑cadherin. Furthermore, a significant increase of β‑catenin in the beginning late phase from day 60 to 180 was observed, whereas no significant effect was seen for E‑cadherin. We suggest that thalidomide is indirectly improving the urothelial barrier function by inhibiting negatively interfering effects of NF-κB on AJ, resulting in enhanced expression of membranous E‑cadherin and β‑catenin at the end of the early response phase. Further, it seems that at the beginning of the late phase, thalidomide additionally influences β‑catenin independently from E‑cadherin. It is suggested that thalidomide potentially supports recovery and regeneration of the urothelium in the beginning late phase by upregulation of Wnt/β-catenin. Wnt/β-catenin and Hippo signaling is also reported in the recovery phase of radiation-induced dermatitis [16]. Furthermore, the pronounced increase of β‑catenin in the early phase 1 in the THA1–15 group might indicate active Wnt/β-catenin additionally to upregulation of AJ proteins. Unfortunately, nuclear β‑catenin was not visualized in our study. It seems difficult to assess nuclear β‑catenin in the urothelium by immunohistochemistry. Other studies in normal, inflamed and neoplastic urothelium were also not able to detect nuclear β‑catenin staining and β‑catenin remained membranous or cytoplasmic [49, 50]. Future studies are needed to confirm activated Wnt/β-catenin in the recovery phase of the urothelium.

Limitations of this study were that alterations in AJ protein expression was only assessed by immunohistochemistry in FFPE tissue specimens. Further, DMSO was used to dissolve the non-water-soluble thalidomide for intraperitoneal injection as recommended by the manufacturer. DMSO itself has anti-inflammatory potential via NF-κB inhibition [51]. Therefore, we cannot exclude a small additional effect from the dissolvent itself on the therapeutic effect of thalidomide. Further, the treatment effect of thalidomide in irradiated animals was evaluated only in comparison to an irradiation control. We did not analyze whether constitutively expressed E‑cadherin and β‑catenin in the non-irradiated control is altered by thalidomide because we did not expect an effect in the absence of any NF-κB activation. However, to better understand the complex interactions between thalidomide and the multifaceted signaling pathways that are involved in the pathogenesis of radiation-induced urinary bladder dysfunction, further studies including modern high-throughput methods such as RNA microarray technology, next generation sequencing, mass spectrometry or western blot analyses are warranted.

The present study revealed that compensatory mechanisms during early radiation cystitis (ERC) and late radiation sequelae (LRS) increase the E‑cadherin and β‑catenin expression to maintain or re-establish the integrity of the urothelial barrier function. Early administration of thalidomide improves these mechanisms by inhibiting the nuclear factor-kappa B (NF-κB) signaling cascade and its interfering effects. As a result, a steeper rise of both adherens junction (AJ) proteins was found in the first few days post irradiation together with a higher and extended expression in the important second early phase, which is crucial for the severity of LRS. Findings of the current and the previous study indicate that thalidomide has to be administered at the beginning of the early phase to achieve beneficial effects. Future studies are needed to confirm activated Wnt/β-catenin in the recovery phase of the urothelium.