Light-Inducible Spatio-Temporal Control of TLR4 and NF-κB-Gluc Reporter in Human Pancreatic Cell Line

Augmented Toll-like receptor 4 (TLR4) expression was found in nearly 70% of patients with pancreatic adenocarcinoma, which is correlated with increased tumorigenesis and progression. In this study, we engineered a new light-oxygen-voltage-sensing (LOV) domain-based optogenetic cell line (opto-TLR4 PANC-1) that enables time-resolved activation of the NF-κB and extracellular-signal regulated kinases (ERK)1/2 signalling pathway upon blue light-sensitive homodimerisation of the TLR4-LOV fusion protein. Continuous stimulation with light indicated strong p65 and ERK1/2 phosphorylation even after 24 h, whereas brief light exposure peaked at 8 h and reached the ground level 24 h post-illumination. The cell line further allows a voltage-dependent TLR4 activation, which can be continuously monitored, turned on by light or off in the dark. Using this cell line, we performed different phenotypic cell-based assays with 2D and 3D cultures, with the aim of controlling cellular activity with spatial and temporal precision. Light exposure enhanced cell attachment, the formation and extension of invadopodia, and cell migration in 3D spheroid cultures, but no significant changes in proliferation or viability could be detected. We conclude that the opto-TLR4 PANC-1 cell line is an ideal tool for investigating the underlying molecular mechanisms of TLR4, thereby providing strategies for new therapeutic options.


Introduction
Pancreatic adenocarcinoma is a devastating malignancy and accounts for 4.5% of all deaths caused by cancer worldwide, making it the seventh leading cause of global cancer deaths. The 5-year survival rate for people with pancreatic cancer is about 8% on average and depends on whether the cancer is detected at an early (32%) or late stage (3%) [1]. Pancreatic adenocarcinoma is one of the greatest challenges in oncology due to the late diagnosis and the lack of curative therapeutic strategies [2]. Toll-like receptors (TLRs) promote pancreatic carcinogenesis and epithelial-mesenchymal transition and, therefore, represent a promising target class for cancer therapy [3,4]. TLRs are a family of evolutionarily conserved pattern-recognition receptors (PRR) that play a key role in the innate immune response and largely determine the development of the adaptive immune response. TLRs are class-type I transmembrane proteins composed of an extracellular domain (ectodomain), a transmembrane region, and an intracellular domain. The ectodomain consists of leucinerich repeats and two to four evolutionarily conserved cysteine residues, which recognize and bind to evolutionarily conserved molecular motifs in pathogen-associated molecular patterns (PAMPs) of microorganisms or damage-associated molecular patterns (DAMPs) of damaged tissues to activate an innate and adaptive immune response [4][5][6].
To date , there are 11 members of the TLR family identified, of which TLR1, TLR2,  TLR4, TLR5, TLR6 and TLR11 are located on the cell surface, and TLR3, TLR7, TLR8 and  TLR9 are localised to the endosomal/lysosomal compartment [7]. Regarding pancreatic adenocarcinoma, recent studies show that 69% of patients with pancreatic cancer display an elevated expression of TLR4, which is correlated with an increased activation of the nuclear factor kappa B (NF-κB), phosphoinositol-3 kinase/AKT and mitogen-associated protein kinase signalling pathways, all being involved in tumourigenesis and tumour progression [7,8]. For example, the canonical NF-κB signalling pathway regulates the expression of genes involved in angiogenesis, metastasis, cell survival, cell proliferation, inflammatory processes, and immune modulation. Therefore, diminishment or even blockage of TLR4-related pathways such as the NF-κB signalling pathway are promising strategies for possible targeted therapy.
Activation or inactivation of TLR4 by using agonists, antagonists, activators and/or inhibitors possess only limited spatial and temporal control of the underlying signalling pathway and lack the ability to either continuously change or terminate signalling [9][10][11][12].
Optogenetics is an innovative technique that makes use of genetic and optical methods to control molecular processes, cellular signals and animal behaviours spatially and temporally. In optogenetics, light-sensitive protein domains of microbial or plant photoreceptors that are capable of light-controlled inter-or intramolecular interactions are integrated into effector proteins. Consequently, light induction allows activation, inactivation, localisation, or stabilisation/destabilisation of signalling pathways, depending on the protein type and set-up [13,14]. The currently available domain repertoire allows for blue- [15][16][17], red- [18,19], or green- [20] light-induced formation of protein complexes.
Here, we engineered a novel optogenetic controllable cell line (opto-TLR4 PANC-1) in which the TLR4 can be switched on by blue light and off in the dark by fusing the light-oxygen-voltage sensing (LOV) domain to the TLR4.
PANC-1 is a human pancreatic cancer cell line isolated from a pancreatic carcinoma of ductal origin. It is characterized with a high metastatic potential and poor differentiation abilities. Above all, it is known as a suitable transfection host [21,22]. The LOV domain, isolated from the Vaucheria frigida aureochrome 1 (yellow-green algae), noncovalently binds a flavin chromophore, which can be stimulated by blue light (λmax ≈ 470 nm) absorption to initiate a photochemical reaction that results in the formation of a covalent adduct between a conserved cysteine and the flavin ring [23]. This results in a change in protein conformation allowing for dimerisation of the LOV domains [24][25][26][27] and activation of the TLR4-LOV fusion protein, a process fully reversible in the dark. TLR4 activation and real-time detection of the underlying signalling pathways can be confirmed with the stable integrated NF-κB-Gluc reporter system. This opto-TLR4 PANC-1 reporter cell line offers a novel tool for analysing TLR4 signalling pathway and the concomitant genotypic and phenotypic effects in a time-resolved manner, and enables a new high-content analysis (HCA) approach for the screening of TLR4 signalling modulators.

Design of a Light Inducible TLR4 Pancreatic Cell Line with Stable NF-κB-Gluc Reporter
Previously, blue light-inducible activation of several members of the receptor tyrosine kinases (Opto-RTKs) have been successfully generated by fusing light-sensitive protein domains to the far C-terminus of the RTK [15]. In this study, a similar strategy was initially employed to generate a pancreatic cell line where the photosensitive protein domain (LOV) was fused to the C-terminal end of the Toll-like receptor 4 (TLR4) to allow a spatially and temporally precise dimerisation of the TLR4 and activation of the signalling cascades.
We first designed a lentiviral reporter construct with five tandem repeats of the NF-κB transcription responsive elements (NF-κB-TRE), Gaussia Luciferase (Gluc), CMV-dTomato, Hygromycin B (HYG B), and a full-length TLR4 (TLR4 in pEZ-Lv195 vector) where the cDNA of the LOV domain was inserted in-frame at the 3 end of TLR4 ( Figure 1A and Figure S1). Cells expressing these constructs should respond to blue light and the respective endogenous downstream signalling of NF-κB and should be quantifiable by measuring the Gaussia Luciferase ( Figure 1B). To verify this assumption, we transiently co-transfected either TLR4 or TLR4-LOV with the NF-κB-TRE-Gluc reporter into HeLa, 293Ta, and PANC-1 and measured Gluc activity 6 h post LPS (100 ng/mL) or light (470 nm, 7 V, 7.5 min) treatment, respectively, or left untreated (w/o). As depicted in Figure 1C-E, light exposure induced high NF-κB activation in cells transfected with TLR4-LOV, but not in cells with TLR4. In comparison, LPS stimulated NF-κB activation in both TLR4 and TLR4-LOV transfected cell lines. and (E) PANC-1 were transiently co-transfected with either TLR4 or TLR4-LOV with the NF-κB-TRE-Gluc reporter, and Gluc was measured 6 h post LPS (100 ng/mL) or light (470 nm, 7 V, 7.5 min) treatment, respectively, or left untreated (w/o). Post-ANOVA multiple comparisons relative to the control were performed using Dunnett's test. Error bar = SD. n = 6; ** p < 0.01, *** p < 0.001.

Light Induces TLR4-Dependent NF-κB and MAP Kinase Pathway Activation in Opto-TLR4 PANC-1 Cell Line
To test the light-inducible TLR4-LOV dimerisation/activation and subsequent signal transduction (NF-κB-Gluc), the selected clone, opto-TLR4 PANC-1, was pre-incubated overnight in serum-free culture medium and exposed to light at 470 nm for 0-60 min using the Amuza LED Array System, followed by incubation in the dark ( Figure 3A). After 6 h, NF-κB signalling was determined by measuring Gluc. As depicted in Figure 3A, light exposure of opto-TLR4 PANC-1 at 470 nm and 7 V indicated a time-dependent Gluc activity peaking after 7.5 min (2.5-fold). Importantly, only opto-TLR4 PANC-1 showed Gluc activity after light exposure compared to PANC-1 NF-κB (reporter), but both cell lines could be stimulated with LPS ( Figure 3A). Additionally, no change in cell viability (PrestoBlue) could be observed ( Figure 3B). To investigate whether prolonged exposure or the change in voltage has an impact on TLR4-LOV/ NF-κB-Gluc reporter activity and cell viability, cells were continuously illuminated at 470 nm for 12 h at 7 V ( Figure 3C,D) or stimulated for 7.5 min at 4.5-, 7-, 10-and 13-volts, followed by incubation for 6 h in the dark ( Figure 3E,F). Subsequently, promoter activity and cell viability were measured. Interestingly, prolonged light exposure (12 h) or increased voltage (10 and 13.5 V) did not boost reporter activity compared to 7 V and 7.5 min; however, light intensity at lower voltage (4.5 V) indicated reduced reporter activity. In contrast, an applied voltage higher than 7 V already diminished cell viability significantly after 6 h. This observed effect is most likely due to accumulating heat and the factor of time ( Figure 3F). To exclude that light at 7 V impairs cell viability, cells were exposed to light for 24 h and apoptosis and necrosis of cells were measured using annexin V and 7-AAD. No significant toxicity was detected ( Figure 3G).

TLR4 Activation Can Be Reversibly Turned on or Off by Light Illumination or Incubation in the Dark in the Opto-TLR4 PANC-1 Cell Line
LOV domain was described as a reversible photochromic switch, which can be activated by blue light, and deactivated by near-UV light, or in the dark [28]. To check that this photochromic switch and the subsequent activation of the TLR4 is reversible, the opto-TLR4 PANC-1 cells were alternately illuminated and incubated in the dark, and NF-κB and ERK signalling was measured ( Figure 5). More precisely, the opto-TLR4 PANC-1 were cultivated and starved for 24 h with serum-free culture medium and exposed to light at 470 nm for 7.5 min or treated with 100 ng/mL LPS and incubated for 24 h in dark before cells were again stimulated with light or LPS, respectively. NF-κB signalling was determined by measuring Gluc 6 h and 24 h after each stimulation ( Figure 5A,B). As depicted in Figure 5B, NF-κB signalling already increased significantly 6 h after light exposure, peaking after 24 h due to accumulation of Gluc in the medium. Re-stimulation of LPS showed a more pronounced increase in NF-κB activity than re-illumination at 470 nm (T48h). However, NF-κB expression after the third light treatment already aligned to the NF-κB expression level induced by LPS (T72h) and, evidently, exceeded it after the fourth stimulation (T96h).  . Comparative time-curve and re-activation analysis of opto-TLR4 PANC-1 cell line. Cells were exposed to light at 470 nm for 7.5 min or treated with 100 ng/mL LPS and incubated for 24 h in the dark before cells were again stimulated with light or LPS, respectively. Gluc was measured 6 h and 24 h after each stimulation. (A) Timeline for opto-TLR4 PANC-1 cell seeding, treatment, and imaging. (B) Gluc activity. Western blots of whole-cell extracts of opto-TLR4 PANC-1 cell line after light exposure (470 nm) for 7.5 min at timepoints 0 and 24 prior to extraction of proteins. Blots were probed with antibodies against (C) phospho-p65, total p65, or (D) p-ERK1/2, ERK1/2, and vinculin as loading control. Cells were treated continuously with LPS (100 ng/mL) or with light (470 nm) for 24 h and cell extracts were blotted and probed with antibodies against (E,F) phosphor-p65 p65, and (G) phospho-ERK1/2, ERK1/2, and vinculin as loading control. Post-ANOVA multiple comparisons relative to the control were performed using Dunnett's test. Error bar = SD. n = 6; * p < 0.05; *** p < 0.001.
Next, cells were treated as before and protein expression of phospho-p65 and phospho-ERK1/2 was analysed 0.5 h, 3 h, 8 h, and 24 h after the first stimulation with light or LPS, respectively, and 8 h after the second stimulation (32 h). As shown in Figure 5C-G, phosphorylation of both proteins p65 and ERK1/2 could be turned on by light and off in the dark, clearly demonstrating the reversibility of the photoreaction of the LOV domain and, corresponding, TLR4 signalling pathways ( Figure 5C,D). When cells were exposed for 7.5 min, p65 and ERK1/2 phosphorylation increased already after 0.5 h, peaking at 8 h for p65 and 0.5 h for ERK1/2, and reached the ground-level 24 h post illumination. In comparison, continuous stimulation with light or LPS indicated strong p65 and ERK 1/2 phosphorylation even after 24 h ( Figure 5E-G).

Light Induced TLR4 Activation Did Not Change Cell Viability, but Increased Cell Adhesion to, and Invasion into Extracellular Matrix
Since TLR4-mediated signalling has been reported to be responsible for cell survival, cancer progression, invasion, and metastasis [9,10,12], we next developed phenotypic assays in order to better understand the complex interplay of tumour progression and the TLR4 signalling pathway. First, we studied the role of light-induced TLR4 signalling in cell proliferation and survival in 2D cultures. Therefore, opto-TLR4 PANC-1 cells were allowed to attach to collagen type I coated wells for 24 h before cells were exposed to light, left in the dark, or were stimulated with LPS or PAR. Then, cell proliferation/viability was measured. As such, we observed that neither light exposure nor treatment with LPS or PAR influenced cell viability or proliferation significantly, as detected by light microscopy ( Figure 6A) or monitored in real time by means of measuring the changes in impedance over time ( Figure 6B). Next, cells were stimulated with LPS, left in the dark, or were exposed to light with additional PAR and TAK-242 treatment in different concentrations for 48 h, and cell viability was measured using PrestoBlue assay. As depicted in Figure 6C, light exposure or LPS treatment did not alter cell viability as long they were not treated with high PAR (3 or 10 µM) or TAK-242 (10-50 µM) concentrations.
To verify the influence of light-induced TLR4 on cell viability in a physiologically more relevant test system that better reproduces the three-dimensional (3D) complexity and growth behaviour of pancreatic carcinoma than the 2D cultures, we cultivated the opto-TLR4 PANC 1 cell line (2000 cells/well) as 3D spheroids, with an initial diameter of approximately 500 µm ( Figure 6D). After four days, spheroids were treated with PAR, LPS, or exposed to light for a further six days, and cell viability, as well as size of spheroids, were measured on days 0, 3, and 6. As depicted in Figure 6E-G, no significant difference in cell viability/metabolism and spheroid size (area) between treated (LPS or PAR), light-exposed, and non-treated spheroids could be detected on days 3 or 6 post treatment. Interestingly, both LPS-stimulated and light exposed spheroids (7.5 min twice a day) still indicated a significant increase in Gluc activity, whereas PAR treatment decreased the reporter activity compared to non-treated spheroids ( Figure 6H).
Having demonstrated that cell viability is only decreased when cells were treated with high concentrations of PAR or TAK-242 ( Figure 6C), but not due to LPS or lightinduced TLR4 activation ( Figure 6), we next examined whether TLR4 signalling triggers cell attachment to extracellular matrix (ECM), and/or invasion through ECM. Therefore, opto-TLR4 PANC 1 cells were left untreated (w/o), pre-treated with LPS or PAR, or illuminated with light for 6 h, and attachment and spreading on collagen type I coated wells detected in real-time for further 20 h ( Figure 7A). Both LPS stimulation and light-induction promoted cell attachment and spreading compared to non-treated cells, whereas PAR stimulation showed the opposite effect. To verify this data, cells were collected in non-adhering plates and were left untreated (dark), treated with LPS, or exposed to light for 7.5 min or 6 h with 5 µM or 25 µM TAK-242 or 3 µM PAR and plated on collagen type I coated wells 6 h later. As depicted in Figure 7B, LPS treatment and exposure to light (7.5 min or 6 h) significantly increased cell attachment after 30 min, whereas the TLR4/NF-κB inhibitors TAK-242 and PAR significantly decreased this effect. Since markers such as integrin β-1 or vimentin play an essential role in adhesion and invasiveness of pancreatic carcinoma cells [29], we analysed the relationship between light-induced TLR4 activation and integrin β-1 and vimentin expression. Opto-TLR4 PANC 1 cells were illuminated with light for 0-24 h and protein expression of integrin β-1, vimentin, E-cadherin, and β-actin (loading control) was analysed using western blotting. As shown in Figure 7C, light induction increased expression of integrin β-1 after 3 h significantly and that of vimentin slightly after 8 h. In comparison, E-cadherin expression was clearly diminished after 3 h and remained downregulated even after 24 h. plate) were either exposed to light for 72 h or kept in the dark. Phase contrast images were taken (A) and cell count (B) was measured every 24 h using the MiniMaxTM 300 Imaging Cytometer. (C) Cells were allowed to attach on ECIS array and 24 h later they were exposed to light, treated with 100 ng/mL LPS or 3 µM PAR, or left untreated for 48 h and changes in impedance measured in real-time. (D) Cells were non-treated, treated with 100 ng/mL LPS, or exposed to light (470 nm) +/− different concentrations of PAR or TAK-242 and cell viability measured 48 h later using PrestoBlue assay. (E) 2000 opto-TLR4 PANC-1 cells were cultured for 4 days as three-dimensional (3D) cell aggregates (spheroids) using non-adherent plates and processed for immunofluorescence microscopy using TLR4 antibody and Hoechst dye. Images were acquired (TLR4-Alexa-488: green, dTomato: red, Hoechst: blue) using a Leica TCS SP8 confocal laser scanning microscope. 5× magnification. Scale bars: 500 µm. (F-I) Spheroids were treated with 100 ng/mL LPS or 3 µM PAR, exposed to light (470 nm) for 7.5 min twice a day, or left untreated for 6 days and (F) cell viability was measured using PrestoBlue assay. (G) Spheroid area was measured on days 0, 3, and 6 by segmenting bright field images (H) using the FIJI Macro INSIDIA. Data were collected at least 3 times in octuplicates and normalised to the non-treated control. (I) Relative NF-κB promoterlinked Gaussia Luciferase activity in opto-TLR4 PANC-1 spheroids on day 6. Post-ANOVA multiple comparisons relative to the control were performed using Dunnett's test. Error bar = SD. n = 6; * p < 0.05; ** p < 0.01, *** p < 0.001. After demonstrating that light-induced activation of TLR4 signalling pathway promotes cell adhesion and the expression of proteins involved in tumour promotion and metastasis, we next examined whether TLR4 activation via light exposure also promotes tumour cell invasion through ECM. Therefore, spheroids of opto-TLR4 PANC 1 cells were generated and embedded in Geltrex/collagen type I matrices for 1 h. Subsequently, cells were stimulated with different concentrations of TAK-242 or PAR and exposed to light for 48 h ( Figure 7D) or 12 h ( Figure 7E). Consistent with the attachment assay ( Figure 7A,B), cell migration out of the spheroids and invasion into the ECM were enhanced when spheroids were LPS-treated or illuminated with light, and reduced when additionally treated with 10 or 50 µM TAK-242 and 0.3 or 3 µM PAR, respectively. High concentrations such as 100 µM TAK-242 or 30 µM PAR completely inhibited cell invasion ( Figure 7D-F). As expected, also Gluc activity was increased in opto-TLR4 PANC 1 spheroids embedded in Geltrex/collagen type I matrices and could be decreased when treated with TAK-242 or PAR compared to non-treated spheroids.

Discussion
LPS/TLR4 is involved in different signalling pathways such as NF-κB, AP-1, and IRF3 cascades, which are important for the innate immune response, cell survival, cell migration and many more physiological responses. In various types of cancer, deregulation of TLR4 is closely associated with tumourigenesis and cancer progression. The pro-carcinogenic activity of TLR4 in pancreatic carcinoma is still unclear [4]. Therefore, cell cultures that allow a temporal and spatial regulation of TLR4 and the underlying signalling pathways are a beneficial tool for research and drug development.
Hence, we set out to develop a novel optogenetic cell line that enables time-resolved activation of NF-κB signalling pathway upon blue light-sensitive homodimerisation of the TLR4-LOV fusion protein. The designed reporter construct NF-κB-TRE-Gluc, CMV-dTomato allows a simultaneous expression of the reporter genes dTomato and NF-κB-Gluc. The constitutive dTomato expression enables spatial detection of the cells, while the Gluc assay permits a quantitative and time-resolved monitoring of NF-κB activity in response to inflammatory stimuli. For the construction of the TLR4-LOV plasmid, the photosensitive protein domain (LOV) aureochrome-1, isolated from the Vaucheria frigida, was fused to the C-terminal end of the full-length human TLR4. Aureochrome-1 (AUREO1) has been described to function as a blue light-regulated transcription factor that has a LOV domain in the C-terminal region of a bZIP domain, an α-helical DNA-binding motif responsible for blue light-induced branching and for the development of a sex organ, respectively. Blue light activation of AUREO1, which exists as a monomer in reducing conditions, induces dimerisation, which subsequently increases affinity for the target DNA sequence [25,26]. By screening different LOV-sensing domains, Grusch et al. [15] clearly demonstrated that the domain found in AUREO1, in particular, resulted in homodimerisation and activation after low-intensity blue light exposure, when fused to the C-terminal end of receptor tyrosine kinases RTKs [15,27].
Transient co-transfection of either TLR4 or TLR4-LOV with the NF-κB-TRE-Gluc reporter into HeLa, 293Ta, and PANC-1 proved the functionality of the engineered constructs, as light exposure induced high NF-κB activation in cells transfected with TLR4-LOV, but not with TLR4. In contrast, LPS stimulation activated NF-κB activity in both, TLR4 and TLR4-LOV transfected cell lines.
Using the lentiviral delivery system, the NF-κB-TRE-Gluc reporter was first stably integrated into human PANC-1 cells. The obtained PANC-1 NF-κB reporter clone allows for both, up-and downregulation of the NF-κB-dependent luciferase activity in a time-and dose-dependent manner after TNFα/LPS and or PAR/TAK-242 treatment, respectively.
Next, the TLR4-LOV fusion protein was stably transfected into the engineered PANC-1 NF-κB reporter cell line and demonstrated that TLR4 can be time and dose-dependent activated by blue light and turned off in the dark.
TLR4 is a type I membrane receptor primarily activated by LPS, a core component of the outer membrane of Gram-negative bacteria. The ectodomain of the TLR4 has been described to be important for both avoiding constitutive, ligand-independent receptor signalling [28], and providing ligand regulated recognition and dimerisation of TLR4. The latter causes TLR4 Toll/IL-1R (TIR) domains rearrangement, followed by the recruitment of signalling adapter proteins and activation of the corresponding signalling cascades [29][30][31]. Hence, we tested whether light-induced dimerisation via the LOV domains may also trigger the activation of the full-length receptor. We first screened stably transfected cell clones that exhibit TLR4 dependent NF-κB Gluc activation in response to blue light. The opto-TLR4 PANC-1 clone selected here allows a dose-dependent and temporally precise dimerisation and activation of the TLR4 and respective endogenous downstream signalling cascades. Furthermore, the cell line enables the TLR4 to be continuously turned on or off by light illumination or incubation in the dark, demonstrating the reversibility of the photoreaction of the LOV domain.
Additional to UV radiation, blue light has also been documented to influence cellular physiology in different cell lines [32,33]. Therefore, we tested whether blue light exposure causes cell toxicity or changes in cell behaviour. By investigating the effects of blue light on human dermal fibroblasts, Opländer et al. [34] demonstrated that blue light emitted at 410 or 420 nm showed dose-dependent toxicity and oxidative stress, while longer wavelengths (453, 480 nm) did not display any toxicity, but displayed effects on proliferation. At this point, using the specified settings, we could confirm that illumination with blue light at 470 nm (4.5 or 7 volts; 500 ms pulse width; 250 ms pulse interval; 100 µs repeat interval; 5.0 V amplitude) did not show cytotoxic effects, changes in cell morphology or cell proliferation, even when exposed continuously for more than 24 h. However, if the voltages were higher than 7 (10 V, 13,5 V) a dose-dependent decrease in cell viability could be detected, most likely due to the heating effect of the lamp.
In summary, compared to standard cell lines, the opto-TLR4 PANC-1 cell line offers additional features that are advantageous for research and drug development. The lightinducible cell line, in comparison to the ligand-inducible one, enables physiological relevant activation levels, comparable to LPS, but allows additionally a dosage, temporal, and spatial control of TLR4 activation and a real-time measurement of the NF-κB signalling via stable integrated Gaussia Luciferase. Through single or multiple light-induction or incubation of the cells in the dark, it is possible to continuously change or terminate cell signalling allowing for better understanding of the intracellular NF-κB pathway. This is not possible through receptor activation via ligands such as LPS, as instantaneous removal of the ligand requires additional procedure steps, such as cell washing, medium change, etc., that influence cell behaviour and physiology. Furthermore, ligands such as LPS are dependent on their co-receptors [29,30] but can also bind to transient receptor potential (TRP) channels that have recently been identified as non-TLR4 membrane-bound sensors of LPS [35]. Another problem with LPS is the potential induction of off-target effects caused by sample impurities or contamination with nucleic acids and proteins [36]. All these experimental uncertainties and additional processing steps can be avoided with the opto-TLR4 PANC-1 cell line, as light allows a ligand-free and specific control of the TLR4 without changing or disturbing the cellular environment.
To study the key roles of TLR4 signalling in cancer progression and metastasis, we next established different cellular models in which cell behaviour is under optical control. Using 2D cultures and free-floating 3D spheroids, we demonstrated that light-induced activation of TLR4/NF-κB has no significant effect on cell proliferation or toxicity, whereas blocking of TLR4 activity via TAK-242 or the NF-κB signalling pathway via PAR significantly inhibited cell viability and induced apoptosis, indicating that a basic expression of TLR4/NF-κB signalling is important for cell survival. Recent studies have shown that TLR4 overexpression not only counteracts apoptosis, but also promotes proliferation, invasion, and metastasis in several cancer cells [4,9,10,12,37]. In contrast, other studies have demonstrated that LPS-induced TLR4 signalling increases cell migration and adhesion to endothelial cells or ECMs, but no significant changes in apoptosis or proliferation have been documented [38]. Sun et al. [12] clearly proved that TLR4 expression was significantly enhanced in pancreatic cancer tissues and has an important role in tumorigenesis and progression of pancreatic cancer. Our results also indicated that LPS or light-induced TLR4 activation had no significant impact on cell proliferation, but the TLR4 antagonist TAK-242 was able to reduce proliferation and cell viability in a concentration-dependent manner. These results suggest that the present expression level of TLR4 in PANC-1 cell line, in addition to TLR4-independent mechanisms, triggers those signalling cascades that lead to a very high rate of proliferation in already unstimulated cells. Therefore, further stimulation remains ineffective. Nevertheless, light-induced TLR4 activation significantly increased the invasive potential and induced EMT of the opto-TLR4 PANC-1. Activation of TLR4 signalling by light illumination markedly increased cell attachment to collagen I in a time-dependent manner, whereas TLR4 or NF-κB inhibition showed the opposite effect. Additionally, activation of TLR signalling by light illumination for only 7.5 min profoundly upregulated integrin β-1 and the mesenchymal markers vimentin and decreased the expression of the epithelial marker E-cadherin. Moreover, light exposure of ECM-loaded opto-TLR4 PANC-1 spheroids enhanced invadopodia formation and cell invasion into ECM, whereas blocking of TLR4 and NF-κB signalling inhibited this effect in a concentration-dependent manner. This is consistent with a report employing pancreatic PCa cells, as well as other cancer cells, which suggested that TLR4/NF-κB activation enhances tumour invasion and metastasis [39][40][41].
Accordingly, this opto-TLR4 PANC 1 reporter cell line offers a novel tool for analysing the role of TLR4 signalling in tumour invasion and metastasis, thereby enhancing our understanding of the pathogenesis of pancreatic cells, and providing clues for developing new strategies against TLR4-mediated metastasis.

Cloning of Reporter and Opto-Tlr4-Lov Fusion Construct
The NF-κB-TRE-Gluc, Hygro reporter (THP, Vienna, Austria; CS-NF-κB-02) was designed with five tandem repeats of NF-κB transcription responsive elements (TREs; TGGGGACTTTCCGC), the Gluc, the dTomato fluorescence protein under the control of the CMV promoter (constitutive expression of dTomato) and Hygromycin B (resistance gene) as selection marker and cloning/synthesis of the DNA construct into the pEZ-Lv195 vector (GeneCopoeia TM , 1 Rockville, MD, USA) commissioned.
For light-induced TLR-4-LOV activation, cells were exposed to blue light using the Amuza (San Diego, CA, USA) LED array system 10335, consisting of the LEDA-B LED array with 96 LED elements, and the LAD-1 LED array driver for the definition of the corresponding parameters. Parameters used were wavelength: 470 nm; time: 2.5-60 min (followed by incubation of 6-24 h in the dark) or 6-48 h (w/o incubation in the dark); voltage: 4.5-13,5 V; delay: 0 µs; pulse width: 500 ms; pulse interval: 250 ms; repeat interval: 100 µs, number of repeats:1; amplitude: 5.0 V and the mode was hold constant.

Immunostaining
Cell monolayers or spheroids were chemically fixed with 3.8% paraformaldehyde (Thermo Fisher Scientific, Vienna, Austria; 15670799) for 30 min and permeabilized with 0.2% Triton-X-100 (Merck, Darmstadt, Germany; 11332481001) for 15 min or 30 min (spheroids). After washing, monolayers and spheroids were stained with α-TLR4, h-Toll, CD284 (US Biological, Salem, MA, USA; 042879) antibody, or for nuclear localisation with α-NF-κB p65 antibody (5 µg/mL; Abcam, Cambridge, USA; ab16502) and allowed to incubate at 4 • C overnight. Cells were washed with 1× PBS three times for 5 min and incubated for one hour at room temperature with the secondary antibody (Alexa Fluor 488 goat anti-rabbit IgG (H+L); Thermo Fisher Scientific, Vienna, Austria; A11008). A Hoechst 33342 solution (Cambrex Bioscience, Walkersville, MD, USA; PA-3014) diluted 1:1000 was added for 30 min at room temperature before cells and spheroids were finally washed with PBS and mounted for immunofluorescence confocal microscopy (Leica, Wetzlar, Germany; DMI6000B) to analyse TLR4 or p65 localisation. For quantification of nuclear localisation of NF-κB, fluorescence intensity of stained p65 in the cytoplasm and the nucleus were measured using ImageJ [42] and the mean ratio [nucleus/cytoplasm] was calculated.

Attachment Assay and Proliferation Assay
For the adhesion assay, opto-TLR4 PANC-1 cells were transferred to non-adhering plates (30,000 cells/ well; S-Bio, Hudson, NH, USA; MS-9096UZ) and immediately treated with LPS, or light and different concentrations TAK-242 or PAR.
For continuous measurement, the "electrical cell-substrate impedance sensing" (ECIS) model 9600Z was used. Therefore, 9W20idf PET plates (ibidi, Gräfelfing, Germany; 72098), precoated with 40 µg/mL neutralised rat tail collagen type I (Thermo Fisher Scientific, Vienna, Austria; A1048301), were overlaid with the pre-treated cells, and cell attachment and spreading were assessed by continuous resistance measurements for 20 h. For endpoint measurement, cells were plated and allowed to attach on 96-well cell culture plates (precoated with collagen I) for 30 min (attachment assay), rinsed twice with 1xPBS and adhering cells were automatically counted using the Mini Max 300 Imaging Cytometer (Molecular Devices, LLC, San Jose, CA, USA; 5022671).
For the proliferation assay, cells (3000 cells/well) were treated with light or left untreated for 0-72 h and cells were automatically counted using the Mini Max 300 Imaging Cytometer (Molecular Devices, LLC, San Jose, CA, USA; 5022671) every 24 h.

Cell Growth and Viability Assay
To measure cell growth/viability, opto-TLR4 PANC-1 cells were plated on 96-well plates for 24 h and were treated with different concentrations of PAR (Abcam, Cambridge, United Kingdom; ab120849), TAK-242 (Merck, Darmstadt, Germany; 614316-5MG) and/or illuminated with blue light for 48 h. Cell growth was detected in real-time by continuous resistance measurements for 48 h using the ECIS device (see above) or with PrestoBlue assay (Thermo Fisher Scientific, Vienna, Austria; A13262 see below) using the Spectra Max i3x Multiplate Reader and Transmitted Light (TL) detection cartridge (Molecular Devices, LLC, San Jose, CA, USA; 5022671). PrestoBlue was added to the wells to obtain a 1:10 dilution, incubated at 37 • C for 30-60 min and Ex/Em of 555/585 was measured using the SpectraMax i3x Multiplate Reader.

3D Area Measurement and Invadopodia Formation
Area measurements were performed using the MiniMax 300 Imaging Cytometer (Molecular Devices, LLC, San Jose, CA, USA; 5022671). Images of the spheroids were acquired before and after treatment to define the change in size. FIJI [43] and the FIJI macro INSIDIA [44] were used to achieve spheroid segmentation and to measure the area of the segmented spheroid images, as described previously [25]. The change of each individual spheroid, relative to the same spheroid before treatment (day 0), was calculated. Furthermore, the change in the area, relative to the geometric mean of the controls without treatment within the same experiment, was calculated on days 0, 3, 6 and 9. For the 3D invasion assay, 2000 opto-TLR4 PANC-1 cells were cultivated to spheroids in 200 µL serum reduced (1% FCS) cell culture medium using Prime Surface 96U plates (S-Bio, Hudson, NH, USA; MS-9096UZ) and incubated for 4 days at 37 • C and 5% CO 2 in a humidified atmosphere. The spheroids were embedded in extracellular matrix (ECM) consisting of equal amounts of growth medium, neutralised rat tail collagen type I (Thermo Fisher Scientific, Vienna, Austria; A1048301), and Geltrex (Thermo Fisher Scientific, Vienna, Austria; A1569601). Phase-contrast microscopy of spheroids +/− invadopodia at timepoint 0 (T0), 12 or 48 h after treatment (LPS, PAR, TAK-242, or light exposure at 470 nm for 12 h or 48 h) was performed using a Leica DMI6000B (Wetzlar, Germany) inverted microscope equipped with CTR6500 microscope drive control, DFC420C digital microscope camera with a 5 Megapixel CCD sensor and Leica Application Suite Version 3.8.0 software (Leica, Wetzlar, Germany). The number and length of invadopodia were calculated and graphically displayed using boxplot analysis.

Statistical Analysis
The data analysis for this paper was generated using the Real Statistics Resource Pack software (Release 7.6). Copyright (2013-2021) Charles Zaiontz. www.real-statistics.com (accessed on 11 June 2021). Post-ANOVA multiple comparisons relative to the control were performed using Dunnett's test.