The Multitarget Ligand 3-Iodothyronamine Modulates β-Adrenergic Receptor 2 Signaling

3-Iodothyronamine (3-T₁AM) is an amine with structural similarities to thyroid hormones [1]. 3-T₁AM modifies several important physiological parameters in rodents (for reviews see Zucchi et al. [2]and Piehl et al. [3]).

In vitro, the receptor target of 3-T₁AM was assigned to the trace amine-associated receptor 1 (TAAR1) [4]. Interaction of 3-T₁AM with TAAR1 activates the G_s/adenylyl cyclase signaling pathways [4]. Paradoxically, several 3-T₁AM-induced physiological effects still persist in mTaar1 knockout mice, suggesting further receptor targets in vivo. In this context, previous studies using turkey erythrocytes indirectly suggested β-adrenergic receptors as potential targets for thyronamines [3, 5]. Furthermore, several findings indicate that 3-T₁AM binds with high affinity to the locus coeruleus [6].

3-T₁AM has been proposed to be a multitarget ligand (reviewed by Zucchi et al. [2]), interacting also with a member of nonselective Ca²⁺ entry channels such as the transient receptor potential melastatin channel 8 (TRPM8), which is also known as menthol receptor [7]. In the eye, functional expression of TRPM8 and other thermosensitive TRP isoforms such as the TRP vanilloid 1 channel (capsaicin receptor) have been identified in corneal neurons as well as in corneal endothelial cells [8, 9]. Both are temperature-sensitive TRP channels occurring in corneal tissue layers and cells [10]. Interestingly, TRPV1, in connection with a putative modulation by thyronamines via G protein-coupled receptors (GPCRs), may be a potential drug target in reducing inflammatory symptoms in eye and conjunctiva diseases. It is known that TRPV1 can be stimulated during exposure to the same hypertonic conditions identified in tear samples obtained from dry eye patients or patients with conjunctivitis [11, 12].

The present study was undertaken to elucidate a potential role of β-adrenergic receptor 1 (ADRB1) and 2 (ADRB2) as new targets for 3-T₁AM. We also explored a putative link between 3-T₁AM actions on Ca²⁺ channels and these adrenergic receptors, which are able to specifically modulate the aforementioned TRP channels. The human conjunctival epithelial cell line (IOBA-NHC) was utilized due to the increased expression of β-adrenergic receptors, which has also been observed in the conjunctiva and eye [13, 14, 15].

All full-length β-adrenergic receptors were cloned from genomic human DNA into the eukaryotic expression vector pcDps. The β₁- (ADRB1, NM_000684.2) and β₂-adrenergic receptors (ADRB2, NM_000024.5) were N-terminally tagged with a hemagglutinin (YPYDVPDYA) epitope. Plasmids were sequenced and verified with BigDye-terminator sequencing (PerkinElmer Inc., Waltham, Mass., USA) using an automatic sequencer (ABI 3710xl; Applied Biosystems, Foster City, Calif., USA).

For the cyclic AMP (cAMP) assay and MAPK characterization, HEK293 cells were cultured in MEM Earle's media containing L-glutamine supplemented with 5% FBS and nonessential amino acids (Biochrom AG, Berlin, Germany) in a humidified 5% CO₂ incubator at 37°C. For functional assays, transfections were performed as described previously [16]. The IOBA-NHC cell line was used as a cell model for human conjunctival epithelial cells. Cells were grown in DMEM/HAMs F12 1:1 supplemented with 10% FBS, 1 μg/ml insulin, 5 μg/ml hydrocortisone and antibiotics in a humidified 5% CO₂ incubator at 37°C [17].

HEK293 cells (500,000 cells/well) were seeded in poly-L-lysine-coated 50-ml cell culture flasks. Twenty-four hours after seeding, cells were transiently transfected using Lipofectamine™ 2000 (Life Technologies). On the next day, cells were detached using Versene solution (Life Technologies, Darmstadt, Germany) and were transferred into fibronectin-coated 384-well plates (15,000 cells/well) and incubated for 24 h at 37°C with 5% CO₂. For receptor activation, a high-throughput dynamic mass redistribution (DMR) technology (Corning® Epic® system) was used [18] to determine direct effects of 3-T₁AM (Santa Cruz Biotechnology Inc., Dallas, Tex., USA; final concentration of DMSO 0.1%) on ADRB1 and ADRB2 following costimulation with isoprenaline (ISOP, Sigma Aldrich, St. Louis, Mo., USA, dissolved in H₂O).

G_s signaling was determined by measuring cAMP in transiently transfected HEK293 cells. For control purposes, we transfected cells with empty vector. All experiments were performed in comparison to ADRB1 and ADRB2 transfection. 3-T₁AM was dissolved in DMSO and used in a concentration of 0.1% for 10^-5M 3-T₁AM, which did not affect cAMP accumulation. Stimulation was performed 48 h after transfection as described previously [19]. Cells were incubated for 40 min without ligand or with 3-T₁AM (final concentration of DMSO 0.1%), ISOP or norepinephrine (NorEpi; Sigma Aldrich, St. Louis, Mo., USA; both dissolved in H₂O), or were costimulated with ISOP, or NorEpi and 3-T₁AM, or NorEpi with 3,3′,5-triiodothyronine (T₃; Sigma Aldrich; dissolved in HCl final concentration of 0.01%). Intracellular cAMP accumulation was determined as previously described [19].

MAPK activation was measured using a luciferase reporter gene assay (SRE-luc) (Promega, Fitchburg, Wis., USA). HEK293 cells were cotransfected with serum response element (SRE) plasmid DNA (pGL4.33), a reporter construct containing an SRE and the firefly luciferase reporter gene, and either receptor or empty vector plasmid DNA (mock). Two days after transfection, cells were incubated for 6 h with the respective substances (ISOP, NorEpi or 3-T₁AM) in supplement-free MEM at 37°C with 5% CO₂. Reactions were terminated by aspirating the media. Cells were lysed using 1× passive lysis buffer (Promega). Measurement was conducted with automatic luciferase substrate injection of 40 µl in a black 96-well plate using a Berthold Microplate Reader (Berthold Technologies GmbH & Co. KG, Bad Wildbad, Germany).

The intracellular Ca²⁺ concentration in IOBA-NHC cells was measured as previously reported [17, 19]. In brief, IOBA-NHC cells were loaded with 1 μM fura-2/AM for 30-40 min at 37°C. Fluorescence measurements were performed with a microscope (Olympus BW50WI) at room temperature using a Ringer-like solution containing 150 mM NaCl, 6 mM CsCl, 1.5 mM CaCl₂, 1 mM MgCl₂, 10 mM HEPES and 10 mM glucose (pH 7.4) (~300 mOsM). In addition, a digital-imaging system (TILL-Photonics, Munich, Germany) and TIDA software were used (HEKA-Electronics, Lamprecht, Germany). Fura-2 fluorescence was alternately excited at wavelengths of 340 and 380 nm and emission was measured at 510 nm. The fluorescence ratio (f_{340 nm}/f_{380 nm}) is a relative index of changes in [Ca²⁺]_i[20]. Results are presented as mean traces of f₃₄₀/f₃₈₀ ± SEM. Measurements lasted 10 min.

Data are shown as means ± SEM of independent experiments as indicated in the respective figure legends. For statistical analysis, one-way ANOVA was performed with Dunnett's multiple comparisons test with a threshold of statistical significance of p ≤ 0.05. Student's t test was used if the values were normally distributed in accordance with the Gaussian distribution (normality test). GraphPad Prism 6.0 (GraphPad software, San Diego, Calif., USA) was chosen for data analysis. SigmaPlot software version 12.5 (Systat, San Jose, Calif., USA) was used for the creation of particular diagrams.

Using DMR technology, we prescreened the potential effects of 3-T₁AM alone and in combination with ISOP on ADRB1 and ADRB2 signaling. A robust change in DMR following stimulation with 10^-6M ISOP for ADRB1 (fig. 1a) and ADRB2 (fig. 1b) in comparison to nonstimulated cells was registered. We also investigated the direct effect of 10^-5M 3-T₁AM on ADRB1 (fig. 1c) and ADRB2 (fig. 1d). The effect of 3-T₁AM on both receptors was small but statistically significant at several time points in comparison to nonstimulated cells (one-way ANOVA with a Dunnett's multiple comparisons test, p < 0.05). In addition, the modulatory effect of 3-T₁AM in the presence of 10^-6M ISOP was examined (p < 0.05; fig. 1a, b), with a shift in DMR for both receptors demonstrated in comparison to ISOP stimulation alone.

Stimulation of ADRB1 and ADRB2 with 10^-6M ISOP resulted in a robust increase of ADRB1- and ADRB2- mediated G_s signaling (fig. 2a). However, a high 3-T₁AM concentration of 10^-5M was not sufficient to activate these receptors (fig. 2a). Therefore, the observed effect in DMR was not compatible with G_s activation.

To test whether the observed signal in DMR was due to MAPK activation, stimulation of 10^-6M ISOP or 10^-5M 3-T₁AM was performed. A slight increase in MAPK signaling for ADRB1 with ISOP was observed, which was not found to be statistically significant (fig. 2b). In addition, no increase of MAPK signaling following 3-T₁AM stimulation was observed in comparison to basal levels (fig. 2b).

To determine a modulatory effect of 3-T₁AM on ADRB1 and ADRB2 on ISOP-induced signaling, we performed costimulation studies with a constant concentration of ISOP (10^-6M) and increasing concentrations of 3-T₁AM (10^-10-10^-5M) in G_s signaling (fig. 3).

Coincubation of ISOP and increasing concentrations of 3-T₁AM resulted in a modest effect for ADRB1, particularly at a concentration of 10^-7M 3-T₁AM (fig. 3a). The formation of cAMP at an ISOP concentration of 10^-6M was normalized to 1 and the fold change in cAMP formation induced by 3-T₁AM was indicated. A significant modulatory effect for ADRB2 signaling with a maximal enhancement of G_s signaling of twofold over mock-corrected 10^-6M ISOP signaling (fig. 3a) was demonstrated. Incubation of 3-T₁AM on mock-transfected cells did not result in cAMP formation (data not shown).

To further analyze whether this effect is reproducible with the endogenous ligand NorEpi, we performed costimulation studies with a concentration of 10^-4M NorEpi and increasing concentrations of 3-T₁AM (10^-10- 10^-5M; fig. 3b). For the endogenous ligand, which has a lower affinity at ADRB1 and ADRB2 in comparison to ISOP [21], the modulatory effects were much smaller (fig. 3b). Only a stimulatory effect for costimulation with 10^-9M 3-T₁AM was observed.

To rule out that the modulation of signaling is an artificial result of costimulation, we repeated the experiments by costimulation of 10^-4M NorEpi and increasing concentrations of T₃ (10^-10-10^-5M). Only for stimulation of ADRB1 with 3-T₁AM (at a concentration of 1 nM) was a significant difference compared to stimulation with NorEpi alone observed (fig. 3c).

To ascertain whether endogenously expressed adrenergic receptors respond to 3-T₁AM, we investigated IOBA-NHC cells [22] since β-adrenergic receptors are highly expressed in eye layers and cells [14, 23, 24]. It is also known that there is an association between voltage-dependent Ca²⁺ channels and adrenergic receptors [25]. In addition, there is also an association between adrenergic receptors and nonvoltage-dependent Ca²⁺ channels, such as the aforementioned TRP channels which are expressed in IOBA-NHC cells [17]. As shown in figure 4a and b, a NorEpi-induced Ca²⁺ influx was blocked by the nonselective β-adrenergic receptor antagonist timolol. More specifically, stimulation with 10^-6M NorEpi increased the fluorescence ratio f_{340 nm}/f_{380 nm} from 1.200 ± 0.001 to 1.209 ± 0.002 (p ≤ 0.05, n = 3) at 600 s (fig. 4a), whereas the NorEpi-induced Ca²⁺ increase was clearly suppressed in the presence of 10^-6M timolol to 1.200 ± 0.02 (p ≤ 0.05; n = 3; at 600 s; fig. 4b). Similarly, extracellular application of 10^-6M 3-T₁AM increased the fluorescence ratio from 1.199 ± 0.001 to 1.210 ± 0.002 (p < 0.01; n = 3; at 600 s) at 600 s (fig. 4c), which was also blocked by 10^-6M timolol (1.199 ± 0.002; p ≤ 0.01; n = 3; at 600 s; fig. 4d). The effect of 3-T₁AM in DMR might therefore be due to an effect of endogenously expressed TRP channels and ADRB1 or ADRB2.

3-T₁AM is an endogenous thyroid hormone derivative that activates TAAR1 (reviewed by Piehl et al. [3]). It has been suggested that 3-T₁AM is a multitarget ligand and affects further GPCRs or interacts with non-GPCR proteins [2]. In this study, DMR, cAMP assay and a MAPK activation assay were used to ascertain whether 3-T₁AM-induced cellular signaling is linked with ADRB1 and ADRB2. Furthermore, modulation of Ca²⁺ signaling by 3-T₁AM was evaluated using an IOBA-NHC cell line which has been described as a cell model for the investigation of human conjunctiva [22], also in an electrophysiological context [12, 17].

Our studies revealed that 3-T₁AM enhances G_s-mediated signaling of ISOP-stimulated ADRB2 as a modulator. This finding might be reasoned by positive allosteric effects, which means that binding of 3-T₁AM enhances binding and/or agonistic signaling properties (maximum of capacity) of ISOP. This may be explained by two different scenarios. In the case of so-called off-target effects, dimers constituting two interacting receptor protomers would mutually influence one another, and binding of 3-T₁AM at one protomer increases the capacity for ISOP-induced signaling at the second protomer. Several examples for such a signaling (or binding) modification in GPCR dimers have been reported (for review see Smith and Milligan [26]) and ADRB2 is known to constitute dimeric arrangements [27]. In the second potential scenario of ‘on-target' allosterism, 3-T₁AM would bind at an allosteric receptor site and improve, by spatial rearrangements of the monomeric receptor conformation, the signaling capacity for ISOP bound in the orthosteric binding site. Such a principle scenario of two different ligand-binding sites at a GPCR monomer has been previously suggested, for example at the lutropin receptor [28].

Interestingly, the observed effect of signaling inhibition at higher 3-T₁AM concentrations compared to lower concentrations (fig. 3) would be compatible with a two-binding site model for 3-T₁AM (low and high affinity), whereby one site would have a stimulating effect and the second an inhibitory influence. Such mechanisms have been described for the M₃ muscarinic acetylcholine receptor [29] or the β₁-adrenergic receptor [30].

It can be also hypothesized that the effect of a decreased G_s-mediated cAMP accumulation at increasing 3-T₁AM concentrations in costimulation experiments with ISOP or NorEpi (constant concentration) should be related to the activation of G_i. This would inhibit G_s-mediated signaling. It is known that β-adrenergic receptors are promiscuous for different G protein subtypes and that their differentiated activation can be dependent on the ligand variant or concentration [31].

Furthermore, NorEpi as well as 3-T₁AM increased intracellular Ca²⁺ levels in IOBA-NHC cells, which could be blocked by the nonselective adrenergic receptor blocker timolol (fig. 4). This prompts the suggestion that 3-T₁AM is connected with Ca²⁺-permeable channels or intracellular store depletion processes in connection with adrenergic receptors, which was also reported in cardiomyocytes [32].

Since regulation of body temperature is more or less conducted by temperature-sensitive receptors, we postulate that 3-T₁AM somehow activates cold ‘receptors' such as the aforementioned menthol receptor TRPM8. In previous studies using the same cell line, we could delineate that the 3-T₁AM-induced Ca²⁺ effect was due to TRPM8 activation since the specific TRPM8 blocker BCTC abolished the 3-T₁AM effect [7]. Interestingly, timolol was observed to have the same effect, suggesting that 3-T₁AM binds to ADRBs, in particular ADRB2, which in turn activates Ca²⁺ channels such as TRPM8. A first assumption might be that 3-T₁AM effects are directly mediated via protein interactions between the GPCR and TRP channel (reviewed by Veldhuis et al. [33]), or indirectly via the β/γ-subunits of G_i/o[34], which can also be activated by the β-adrenergic receptors [31]. This hypothesis is supported by our observations in costimulation experiments (fig. 3), which might show activation of G_i at higher 3-T₁AM concentrations combined with constant ISOP or NorEpi levels (and therefore inhibition of G_s signaling). In consequence, the direct 3-T₁AM-mediated G_i-induced signaling should be investigated in future studies.

The established human conjunctival IOBA-NHC cell line was used as a relevant model of ocular surface cell biology as well as for investigation of electrophysiological investigation of TRP channels [10, 17]. Notably, an association between direct interaction of 3-T₁AM with TRPM8 and suppression of TRPV1 activity could be demonstrated in this cell line [7]. The results of the present study further indicate that the multitarget 3-T₁AM is a modulator of ADRB2 and mediates cellular and Ca²⁺ signaling. At this point, 3-T₁AM appears to play a potential role in both TRPM8- and TRPV1-mediated Ca²⁺ regulation and linked biological processes.

We demonstrate that 3-T₁AM-enhances ISOP-induced ADRB2 signaling via the G_s/adenylyl cyclase pathway. This suggests that ADRB2 is a GPCR target for 3-T₁AM and that 3-T₁AM might also have implications for tissues expressing ADRB2 such as the lung, skeletal muscles and skin [24]. In addition, there is evidence that ADRB is linked with Ca²⁺ channels in human conjunctival epithelial cells. This conclusion is warranted since the nonselective ADRB blocker timolol had an inhibitory effect on both NorEpi- and 3-T₁AM-induced Ca²⁺ increases. In summary, 3-T₁AM is a multitarget ligand and modulates ADRB-mediated signaling in human conjunctival epithelial cells. Therefore, thyronamines, in connection with ADRB, may provide novel compounds for suppressing Ca²⁺ channels such as TRPV1 since its activation is associated with inflammation processes occurring in conjunctivitis or dry eye syndrome. Our findings shed new light on the previously proposed complex spectrum of 3-T₁AM action [35] and point to a diversified functional profile of 3-T₁AM.

This work was supported by the Deutsche Forschungsgemeinschaft (DFG): Graduate College 1208 (Hormonal Regulation of Energy Metabolism, Body Weight and Growth) TP1 and TP3, KL2334/2-1 and DFG Priority Program SPP1629 Thyroid Trans Act BI 893/5-1, Me 1706/13-1, KO 922/16-1 and 922/17-1, STA 1265/1-1. Human conjunctival epithelial cells were a generous gift from the laboratory of Yolanda Diebold (University Institute of Applied Ophthalmobiology, University of Valladolid, Valladolid, Spain) and kindly provided within a collaboration by Friedrich Paulsen (Institute of Anatomy II, University of Erlangen-Nuremberg, Germany).

The authors have nothing to disclose.

J.D. and N.K. contributed equally to this work.