Nongenomic Stimulatory Effect of T3 on Calcium Dynamics in GnRH Neurons via Integrin αVβ3

Abstract

Many clinical studies have identified correlations between thyroid dysfunction and reproductive issues, yet the underlying mechanisms behind this interaction remain poorly understood. In this study, we investigated the effect of triiodothyronine (T3) on the activity of gonadotropin-releasing hormone (GnRH) neurons, a key regulator of the central reproductive axis. Dual labeling confirmed that GnRH neurons express thyroid receptor (TR)α and integrin αVβ3 receptors mediating genomic and nongenomic effects of thyroid hormones, respectively. Using calcium imaging in an ex vivo model, we show that T3 induces a rapid and sustained increase of calcium oscillation frequency in GnRH neurons. No change in response was detected after application of T4. The T3 stimulatory effect was not inhibited by a TR-specific antagonist (1-850) but was mimicked by membrane-impermeable T3-BSA, indicating a mechanism independent of nuclear TR signaling. In contrast, the blockade of membrane αVβ3 integrins (with cilengitide) prevented the T3-induced increase in GnRH neurons calcium peak oscillation frequency. Further investigation using modulators of intracellular calcium and calcium entry revealed that binding to αVβ3 integrin can induce distinct calcium responses depending on the ligand, with T3 triggering a complex response involving multiple channels and calcium sources, possibly with compensatory mechanisms. In sum, these results demonstrate for the first time a direct effect of thyroid hormones on GnRH neuronal activity, with T3 stimulating calcium oscillations through the nongenomic αVβ3 integrin pathway. Understanding this thyroid-reproductive axis interaction will help clarify the mechanisms linking thyroid dysfunction to reproductive disorders and pave the way for targeted therapeutic interventions.

Central to the hypothalamic-pituitary-gonadal axis, gonadotropin-releasing hormone (GnRH) neurons play a key role in coordinating reproductive function. These hypothalamic neurons drive the release of gonadotropins (luteinizing hormone and follicle-stimulating hormone) from the pituitary gland, which in turn regulate gonadal activity and steroid secretion. Kisspeptin neurons are critical for GnRH neuronal function (1). However, GnRH neurons are highly sensitive to internal physiological factors, such as hormonal feedback mechanisms, metabolic status, and stress (2-4), integrating such signals to ensure the appropriate release of GnRH for a pre-ovulatory surge of luteinizing hormone (5). Disruptions in these internal signals can impair GnRH function (2, 3). In addition, the reproductive axis is influenced by external environmental cues, including circadian and photoperiodic changes in seasonal species and humans (6-8). Among the hormonal signals influencing the hypothalamic-pituitary-gonadal axis, thyroid hormones have emerged as critical modulators (9-12).

Many clinical studies have highlighted the relationship between thyroid disorders and reproductive dysfunctions, with aberrant thyroid hormone levels, such as those seen in hypothyroidism and hyperthyroidism, being associated with subfertility, altered puberty timing, polycystic ovarian syndrome (PCOS), and other reproductive issues (10, 13-17). Although numerous clinical studies have highlighted the comorbidity between thyroid and reproductive disorders, the functional mechanisms through which thyroid hormones interact with key players of the reproductive axis remain largely unexplored. Thyroid hormones, specifically thyroxine (T4) and its bioactive form triiodothyronine (T3) are vital for maintaining the body's homeostasis and functioning. They play essential roles in neuronal development and differentiation (18, 19), regulation of energy metabolism (20, 21), protein synthesis (22), and seasonal adaptations (23) through genomic (24, 25) and nongenomic mechanisms (26). Genomic action of thyroid hormones involves binding to nuclear receptors, directly regulating gene expression. A nongenomic action can be a rapid process binding to plasma membrane integrin αVβ3 receptors (22, 27).

We recently showed that GnRH neurons express thyroid hormone receptors including integrin αVβ3 receptors (28). In this study, we explore the molecular mechanisms by which T3 modulates the activity of GnRH neurons, using calcium imaging in an ex vivo model. Unraveling these signaling pathways will not only enhance our understanding of how thyroid disorders impact fertility but could also pave the way for more targeted therapeutic strategies in treating thyroid-related reproductive dysfunctions.

Material and Methods

Ex Vivo GnRH Neurons Maintained in Explants

All animal procedures were approved by the by the National Institute of Neurological Disorders and Stroke (NINDS) Animal Care and Use Committee, and all experiments were conducted at NINDS, in accordance with National Institutes of Health (NIH) guidelines. Explants were made as previously described (29). Embryos 11.5 days old (E11.5, unsexed) were collected from the uterus of timed-pregnant NIH Swiss female mice euthanized with CO₂. The nasal pits were dissected in Gey’s Balanced Solution (Life Technologies Inc., Grand Island, NY) supplemented with glucose (Sigma, St. Louis, MO), plated on permanox coverslips (Nunc, Rochester, NY) and adhered with a chicken plasma (Cocalico Biologicals, Reamstown, PA) thrombin (Sigma) clot. Explants were kept in a humidified incubator at 37 °C, 5% CO₂ and fed with serum-free medium (SFM) (30). On day 3 in vitro (3 div), fresh SFM containing 80 uM fluorodeoxyuridine (Sigma) was applied for 2 days to inhibit proliferation of non-GnRH cells and then fresh SFM was applied every 3 days until use. Explants with more than 50 GnRH cells that had migrated into the explant periphery, were selected for calcium imaging experiments and used at 7 to 8 divisions unless otherwise stated.

Polymerase Chain Reaction

The mRNA from single GnRH neurons in explants (at 7 div) and from adult GnRH-GFP slices (male and diestrus females) was converted to cDNA. Briefly, neurons were removed by suction, using a pulled glass capillary pipet and immediately placed in lysis buffer containing RNase inhibitor. cDNA synthesis from cells removed from explants was made by incubating cell content with reverse transcriptase mix (AMV + MMLV, Invitrogen, Waltham, MA) followed by 3′ poly(A) tailing using terminal transferase (Roche, Bâle, Switzerland) and cDNA amplification by polymerase chain reaction (PCR) (31). cDNA synthesis from cells removed from slices was made using Maxima™ H minus first strand cDNA synthesis kits (Thermo Scientific) following the protocol outlined for single cells. Each single-cell cDNA was tested for GnRH and B-tubulin transcript with PCR and stored at −20 °C until further use. PCRs for receptors that respond to thyroid hormones were run on 10 single GnRH neurons from explants and 10 single GnRH cells removed from slices (N = 5 female and N = 5 male). For all cells, cDNA was tested for each receptor. For cells from explants and from slices, the following protocol was used: 14.5 µL H20, 5 µL buffer, 0.5 µL 10 mM dNTP mix, 0.25 µL Platinum Superfi polymerase (Thermofisher Scientific), 1.25 µL reverse and forward primers to 1 µL of cDNA. PCR was performed as followed: 30 seconds at 95 °C, followed by 40 cycles of 15 seconds at 95 °C, 15 seconds 60 °C to 64 °C, 90 seconds 72 °C, followed by a 10-minute post-elongation at 72 °C. Amplified products and controls were run on a 1.5% agarose gel with loading dye (Thermofisher Scientific) for 10 minutes at 150 V and bands were visualized and imaged with a BioRad ChemiDoc imager (BioRad, Hercules, CA). Mouse brain cDNA and water were used as positive and negative controls, respectively. Sequences of primers used are listed in Table 1.

Table 1.

Primers for explants and for slices

Gene	Forward	Reverse	Product size
For explants
TRα1	GAGGGAGAACCCCCATCTCT	CCTGTGGCCTGAAGGGAAAT	350 bp
TRα2	GAAGACGACAGCAGTGAGGC	ACAGCGAGAAGTGTCACACAA	450 bp
TRβ	GCGCTCTGATCCGTGTTTTC	AGGCAGGCTTCAGACATTCC	235 bp
Integrin αV	AGCAGAAAAGAACAACTCTGCTG	CCAATGTGGGTGATCACGAG	280 bp
Integrin β3	ATGAATGCGCAGCACAGAGC	GTGGGGAGGGTGTCAGGAGG	222 bp
For slices
GnRH	CTGATGGCCGGCATTCTACTGC	CCAGAGCTCCTCGCAGATCCC	220 bp
TRα1	GAGGGAGAACCCCCATCTCT	CCTGTGGCCTGAAGGGAAAT	350 bp
TRβ	AGTGCAGTCTGAGTGGGGTA	TCCTTCTCCCCTTCCGTG AT	240 bp
TRβ nested	ACGAAGTCATCACCAGCTCC	CTGGGCTCTACAGCTTCACC	155 bp
Integrin αV	AGCAGAAAAGAACAACTCTGCTG	CCAATGTGGGTGATCACGAG	280 bp
Integrin β3	GTTCTCTCCGGAGGCTGAGG	AACTCAGAGGAGGTGGGGAG	303 bp

Immunocytochemistry

Explants were fixed in 4% formalin for 1 hour, rinsed with phosphate buffer saline (PBS) before incubation in 10% normal horse serum; 0.3% triton X-100 for 1 hour, then incubated for 48 hours with primary antibody (rabbit anti-TRα, ab53729, 1:300, RRID:AB_882974, Abcam, Cambridge, UK, or mouse anti-integrin αVβ3, LM609, ab190147, 1:200, RRID:AB_2925190) at 4 °C. Explants were then washed with PBS and incubated with biotinylated secondary antibody (RRID:AB_2340585, Jackson Immunoresearch, West Grove, PA) for 1 hour and reacted with streptavidin 555 (Thermofisher Scientific, Waltham, MA, RRID:AB_2571525) for TRα or using a tyramide kit (Akoya Bioscience, Marlborough, MA, NEL04A001KT) for αVβ3 to amplify the signal. After the first staining, explants were quickly fixed in 4% formaldehyde, rinsed and incubated with rabbit anti-GnRH antibody (SW-1 RRID:AB_2629221, 1:8000) for 1 night at 4 °C, washed and incubated with anti-rabbit Alexa Fluor 488 (1:100, Thermofisher Scientific, RRID:AB_2762833) for 1 hour the next day. After verification of the signal, explants were washed and mounted with anti-fade mounting solution (Fluoro-Gel, Electron Microscopy Science, Hatfield, PA).

Calcium Imaging

Calcium imaging was performed as previously described (32). Briefly, explants maintained in the incubator, were loaded with 13.5 µM Calcium Green-1 (Life technologies, Carlsbad, CA) for 20 minutes then washed with SFM for 20 minutes. After washing, explants were mounted in a perfusion chamber (Warner Instruments LLC, Hamden, CT) kept in a 28 °C humid chamber and continuously perfused with a peristaltic pump (Instech Labs Inc, Plymouth Meeting, PA) at a rate of 300 to 500 µL/min. Fluorescent GnRH neurons were visualized at 20× (numerical aperture [NA], 0.75; working distance [WD], 1.0 mm) using an inverted microscope (Eclipse TE2000-E, Nikon, Tokyo, Japan) equipped with a charge-coupled device camera (QImaging, Surrey, Canada). Excitation wavelengths were provided with a medium-width excitation bandpass filter (465-495 nm) and emission was monitored through a 40 nm bandpass centered on 535 nm. For all experiments a 5 minutes control period in SFM was recorded followed by 5 minutes in an amino acid blocker (AAB) cocktail, 5 minutes of drug treatment (except for the initial concentration test where 3 × 5 minutes of T3 or T4 were applied) and a final AAB washout period (10 minutes or 15 minutes). Details of each calcium recording experimental design are provided in Table 2. Images were taken every 2 seconds for up to 35 minutes and the viability of the cells was tested at the end with a 40mM potassium chloride (KCl, 40mM) stimulation. Quantification was done by measuring the changes in levels of gray (optical density) in single GnRH neurons circled using iVision (Biovision Technologies, Exton, PA) software and calcium oscillations were analyzed in Matlab with frequency expressed in peaks per minute. At the end of the recording, explants were fixed in 4% formalin and stained for GnRH to verify identity of cells (Fig. 1).

Table 2.

Details of each calcium imaging experiments

Calcium imaging experiment	N exp	N cell	[Conc]	Function
T3 concentration gradient	5	96	T3: 3/30/300 nM
T4 concentration gradient	3	122	T4: 10/100/300 nM	T4 agonist
T3	4	119	T3: 30 nM
DRB	4	91	T3: 30 nM DRB: 150 µM	Inhibition of transcription
1-850	4	116	T3: 30 nM 1-850: 1 µM	TRα and TRβ antagonist
T3-BSA	4	82	T3-BSA: 30nM	Prevent T3 transport inside the cell
Cilengitide	4	99	T3: 30 nM Cil: 100 nM	Integrin ανβ3 and ανβ5 antagonist
Periostin	5	123	Peri: 20 nM	Integrin ανβ3 agonist
Thapsigargin (pre-incubation >30 minutes)	3	110	T3: 30 nM Thapsi: 1 µM	Inhibitor of SERCA pump
2-APB	3	137	T3: 30 nM 2-APB: 75 µM	IP3R and TRP inhibitor, activates TRPV
Nifedipine	3	172	T3: 30 nM NIF: 1 µM	L-type VGCC inhibitor

Abbreviations: DRB, 5,6-dichlorobenzimidazole 1-b-D-ribofuranoside; IP3R, inositol triphosphate receptor; SERCA, sarcoendoplasmic reticulum calcium ATPase; T3, triiodothyronine; T4, thyroxine.

Figure 1.

Recording of calcium oscillations in GnRH neurons in nasal explants. (A) The nasal region of E11.5 embryos, containing the nasal midline cartilage (NMC) and nasal pit epithelium (NPE), was dissected and cultured in vitro. GnRH neurons immunostained (blue-black) are shown that have migrated out of the main explant tissue into the periphery. (B) In situ, GnRH neurons are visualized using phase contrast microscopy and then calcium oscillations recorded using fluorescence (495 nm) microscopy. (C) Example of typical recording obtained from a single GnRH neuron showing calcium peaks (red ticks) that occurred during recording. Top lines indicate paradigm used: baseline period in SFM and AAB followed by application of exogenous drug(s) of interest and a KCl final stimulation.

Drugs

The AAB cocktail contained: (-)-Bicuculline methochloride (BIC; GABA_A receptor antagonist, 20 µM), D-(-)−2-amino-5-phosphonopentanoic acid (AP5; NMDA glutamatergic receptor antagonist, 10 µM), CNQX disodium salt (AMPA/kainate receptor antagonist, 10 µM) purchased from Cayman chemicals (Ann Arbor, MI). T3, 5,6-Dichlorobenzimidazole 1-b-D-ribofuranoside (DRB), thapsigargin, and nifedipine were purchased from Sigma, the cilengitide, 1-850 and L-Thyroxine (T4 agonist) from Cayman chemicals, the periostin from R&D systems (Minneapolis, MI), 2-APB from Tocris (Avonmouth, UK) and T3-BSA from Calbioreagent (Foster City, CA). The concentration used for each drug is specified in Table 2.

Analysis and Statistics

For each calcium imaging protocol, a minimum of 3 explants was used (82 cells minimum analyzed). A calcium peak was identified if the signal elevation was greater than the mean of the 5 previous and 5 next points plus a minimal value (30 AU, representing small variations in baseline). A cell was considered a fast responder if its calcium oscillation frequency was increased by more than 35% during the drug application period compared to its own baseline activity recorded during the preceding AAB period. A cell was considered a slow responder if its calcium oscillation did not increase during the drug application but later increased by more than 35% compared to its AAB baseline activity. Statistics tests were performed using GraphPad Prism. An ANOVA test was used to compare the frequency of oscillation between conditions (n = individual cells), followed by Tukey's multiple comparison test to determine P value between each group. Comparisons of proportion in responder types were done using Fisher exact tests. Results are presented as mean ± SEM and P < .5 was considered a significant difference and is indicated on the histograms by a different letter.

Results

T3 Increases Calcium Oscillations in GnRH Neurons in Explants

Due to the anatomical distribution of GnRH cells throughout the forebrain and their sparse number in any given 200 mm experimental slice, we used an organotypic model containing large numbers of GnRH cells to address the action of T3 on GnRH neuronal activity. We first assessed the expression of nuclear TR receptors in GnRH neurons using PCR in single GnRH neurons collected from explants and from male and diestrus females adult brain sections. We showed that GnRH neurons express transcripts for the TRα receptor but have undetectable levels of transcript for the TRβ receptor in most of GnRH neurons (Fig. 2A). Immunohistochemistry for TRα confirmed its nuclear expression in the majority of GnRH neurons in explants (Fig. 2B). Calcium imaging was employed in an ex vivo model of primary GnRH neurons. In all experiments, to assess the direct effect of T3 on GnRH neurons, we isolated GnRH neurons from GABAergic and glutamatergic inputs, known to be present in this ex vivo model (32), using an amino acid blocker cocktail (AAB). To determine the optimal T3 concentration to use, a concentration gradient (T3; (3nM-30nM-300nM) was used and the frequency of calcium oscillations in GnRH neurons was monitored. Incubation with T3 resulted in a significant increase in GnRH neuronal activity with a stronger stimulation observed during application of the 2 higher doses (30nM and 300nM) compared to the lower dose (3nM) (Fig. 2C, Table 3). Based on efficacy, the T3 30 nM concentration was used for all subsequent experiments examining the signaling pathway underlying its mode of action on calcium activity in GnRH neurons. Incubation for 5 minutes with T3, resulted in a long-lasting effect on the activity of GnRH neurons, with the frequency of calcium oscillations remaining significantly higher for at least 15 minutes after washing with AAB (Fig. 2D, Table 3). Subpopulations of GnRH neurons with different types of response to T3 were identified with 10.1% of GnRH neurons having a fast response to T3 (increase of frequency oscillations during the drug application period), 25.2% of neurons having a slow response (increase after the drug application), 40.3% of neurons having a combination of fast and slow response and 24.4% of nonresponders (Fig. 2D). In contrast to the activation produced in GnRH cells after T3 application, T4 produced no significant change in GnRH neuronal activity (Table 3).

Figure 2.

Dose-dependent and long-lasting effect of T3 on calcium oscillations in GnRH neurons. (A) Expression of TRα isotypes and TRβ in individual GnRH neurons from explants and adult brain slices (N = 4 for each are shown on graph)assessed by PCR, (B) photomicrographs of immunofluorescent staining showing the nuclear expression of TRα (red) in GnRH neurons (green). Scale bar = 10 µm. (C, D) Representative calcium recording in a single GnRH neuron (left) and corresponding histogram summarizing the average calcium oscillation frequency (peaks/min) in recorded GnRH neurons (right) in presence of (C) an increasing concentration gradient of T3 (3nM, 30nM, 300nM; N = 5; n = 96) and (D) 30nM T3 (N = 4; n = 119). All drugs were diluted in serum-free media containing a cocktail of amino acid blockers (AAB: 20µM BIC, 10µM CNQX, 10µM D-AP5). Each period corresponds to 5 minutes. Different letters indicate significantly different results in Tukey's multiple comparisons test (P < .05; see Table 2). Pie chart showing the proportion of GnRH neurons exhibiting a response-type to 30nM T3: fast = orange, slow = dotted, fast + slow = dotted orange, no response = gray. Traces are shown to illustrate the profile of response.

Table 3.

Frequency of calcium oscillations in each experimental design and results of statistical test between periods

	Periods [5 minutes each]					P values between periods (P, Tukey's multiple comparison test)
T3 concentration	AAB	T3 3nM	T3 30nM	T3 300nM		btw P1 and P2	btw P2 and P3	btw P1 and P3
Mean peaks/min	0.92 ± 0.071	1.07 ± 0.076	1.4 ± 0.079	1.3 ± 0.09		.034	<.001	<.001
T4 concentration	AAB	T4 10 nM	T4 100 nM	T4 300 nM
	0.73 ± 0.043	0.88 ± 0.096	0.82 ± 0.199	0.65 ± 0.131		.4067	.737	>.999
T3	AAB	T3 30nM	AAB	AAB	AAB
	0.77 ± 0.067	1.12 ± 0.08	1.17 ± 0.07	1.09 ± 0.074	0.95 ± 0.069	<.001	.91	<.001
DRB	AAB/DRB	T3/DRB	AAB/DRB
	1.16 ± 0.088	1.66 ± 0.10	1.43 ± 0.092			<.001	.007	<.001
1-850	AAB/1-850	1-850/T3	AAB
	1.19 ± 0.09	1.62 ± 0.099	1.73 ± 0.095			<.001	.216	<.001
T3-BSA	AAB	T3-BSA	AAB
	0.86 ± 0.072	1.07 ± 0.085	1.12 ± 0.086			<.001	.74	<.001
Cilengitide	AAB/Cil	T3/Cil	AAB
	0.79 ± 0.07	0.78 ± 0.063	0.75 ± 0.063			.985	.838	.786
Periostin	AAB	peri	AAB
	0.73 ± 0.058	1.58 ± 0.077	0.9 ± 0.064			<.001	<.001	.013
Thapsigargin	AAB/thapsi	T3/thapsi	AAB
	0.69 ± 0.074	0.9 ± 0.078	1.08 ± 0.09			<.001	.002	<.001
2-APB	AAB/2-APB	T3/2-APB	AAB
	0.68 ± 0.07	1.64 ± 0.089	0.85 ± 0.066			<.001	<.001	.089
Nifedipine	AAB/nif	T3/nif	AAB
	0.78 ± 0.065	0.98 ± 0.067	1.02 ± 0.073			<.001	.585	<.001

Abbreviations: AAB, amino acid blocker; DRB, 5,6-dichlorobenzimidazole 1-b-D-ribofuranoside; T3, triiodothyronine; T4, thyroxine.

T3 Stimulatory Effect Is Independent of Transcription via Nuclear TR Receptors and Is Mediated at the Plasma Membrane

To investigate whether the effect of T3 on GnRH neuronal activity involves TRα or TRβ receptors, an antagonist for TR receptors (1-850, 1 µM), was applied prior to and during exposure to exogenous T3 and the frequency of calcium oscillations in GnRH cells evaluated. In presence of the antagonist, T3 still induced an increase in the calcium oscillation frequency in GnRH neurons (Fig. 3A). Since TRα and TRβ mediate the genomic action of T3 by acting as ligand-dependent transcription factors, we applied T3 in the presence of an RNA polymerase II transcriptional inhibitor (DRB, 150 µM). As with 1-850, in the presence of DRB, T3 still induced an increase in the calcium oscillation frequency of GnRH neurons, consistent with a nongenomic action of T3 (Fig. 3B). As further proof that the effect of T3 was independent of nuclear T3 receptors, we incubated GnRH neurons with a conjugated compound in which T3 was linked to bovine serum albumin (T3-BSA), thereby preventing it from entering the cell. Consistent with earlier results, a significant increase in the oscillation frequency of GnRH neurons was detected, with the stimulatory effect lasting at least 15 minutes after removal of T3 (Fig. 3C). These results indicate that the effect of T3 is mediated by a membrane receptor(s). Apart from a decrease in slow responders with DRB, no major changes in the proportion of cells exhibiting a response-type between these 3 conditions and T3 alone were observed (Table 4).

Figure 3.

T3 effect is independent of TR genomic pathway. (A-C) Representative calcium recording from a single GnRH neuron (left), proportion of cells exhibiting a response-type (middle), and corresponding histogram summarizing the average calcium oscillation frequency (peaks/min) in all GnRH neurons (right) in presence of (A) a TR antagonist (1µM 1-850; N = 4, n = 116), (B) a transcriptional inhibitor (15µM DRB; N = 4, n = 91) and (C) 30nM T3-BSA (N = 4, n = 82), preventing T3 entry into the cell. All drugs were diluted in serum-free media containing a cocktail of amino acid blockers (AAB: 20µM BIC, 10µM CNQX, 10 M D-AP5). Each period corresponds to 5 minutes. Different letters indicate significantly different results in Tukey's multiple comparisons test (P < .05; see Table 2).

Table 4.

Variations in responder type in each condition compared to T3 alone

Vs T3 alone	DRB +T3	1-850 +T3	T3-BSA	Cil +T3	Peri	Thapsi +T3	2-APB +T3	Nif +T3
Fast	→ P = .15	→ P = .22	→ P = .82	→ P = .21	↑ ** P = .0096	→ P = .13	↑ * P = .017	→ P = .27
Slow	↓* P = .037	→ P = .54	→ P = .62	→ P = .64	↓↓ **** P < .0001	→ P = .31	↓ ** P = .0013	→ P = .57
Fast +slow	→ P > .99	→ P = .51	→ P = .66	↓↓ *** P = .0004	↑ P = .012*	→ P = .27	→ P = .32	↓* P = .015
No-resp	→ P = .43	→ P = .55	→ P = .42	↑↑**** P < .0001	→ P = .20	→ P = .24	→ P = .77	↑↑*** P = .0003

Abbreviations: AAB, amino acid blocker; cil, cilengitide; DRB, 5,6-dichlorobenzimidazole 1-b-D-ribofuranoside; Nif, nifedipine; Peri, periostin; T3, triiodothyronine; T4, thyroxine; Thapsi, thapsigargin.*P < .05, **P < .01, ***P < .001, ****P < .0001.

T3 Acts on Integrin αVβ3 Membrane Receptors to Stimulate Calcium Dynamics in GnRH Neurons

Integrin αVβ3 receptors have been reported to bind T3 and mediate the nongenomic effect of T3. PCR in single GnRH neurons showed expression of both subunits of the integrin αVβ3 receptors in GnRH cells from explants (co-expression in 4/6 neurons) and adult brain sections(co-expression in 7/10 neurons; Fig. 4A).Co-staining of integrin αVβ3 and GnRH confirmed the presence of this membrane receptor in GnRH neurons in 7div explants (Fig. 4B).To investigate if the effect of T3 could be mediated by αVβ3 receptors, we evaluated the frequency of calcium oscillations in GnRH neurons in explants pretreated with an antagonist for these receptors (cilengitide, 100nM), then challenged with T3. In presence of the antagonist, no increase in the calcium oscillation frequency in GnRH neurons was observed when T3 was applied (Fig. 4C). With cilengitide, the proportion of GnRH cells that did not respond to T3 increased by more than 50% compared to T3 alone, mostly due to a loss of fast + slow responders (Fig. 4C, Table 3). In contrast, incubation with periostin (20nM), an integrin αVβ3 agonist, resulted in a strong increase of calcium peaks in GnRH neurons. However, unlike application of T3 alone, the amplitude of response to periostin was larger with more fast responders, and a decrease in slow responders was observed resulting in calcium oscillations returning faster to pretreatment levels (although still higher than baseline due to the increase in fast + slow responders), immediately after removal of the agonist (Fig. 4D, Table 4).

Figure 4.

T3 effect is dependent on the integrin nongenomic pathway. (A) Expression of intergrin αV (ItgαV) and β3 (Itgβ3) isotypes in individual GnRH neurons from explants and adult brain slices (N = 4 for each are shown on graph) assessed by PCR. (B) Photomicrographs of immunofluorescent staining showing the expression of the integrin αVβ3 (red) in GnRH neurons (green). Scale bar = 10 µm. (C, D) Representative calcium recording in a single GnRH neuron (left), proportion of cells exhibiting a response-type (middle), and corresponding histogram summarizing the average calcium oscillation frequency (peaks/min) in all GnRH neurons (right) in presence of (C) an integrin αVβ3 antagonist (100 mM Cilengitide (cil); N = 4, n = 99) and (D) an integrin αVβ3 agonist (20 nM periostin (peri; N = 5, n = 123). All drugs were diluted in serum-free media containing a cocktail of amino acid blockers (AAB: 20µM BIC, 10µM CNQX, 10µM D-AP5). Each period corresponds to 5 minutes. Different letters indicate significantly different results in Tukey's multiple comparisons test (P < .05; see Table 2).

Fast and Slow Effects of T3 Involving Different Calcium Channels

Considering the difference in the response to periostin and T3 in the GnRH neurons, we next investigated the potential contribution of different calcium sources mobilized by the αVβ3 receptor, in mediating the fast (in the presence of T3) and long-lasting (“slow,” after removal of T3) response to T3. The contribution of internal calcium stores was tested using thapsigargin, an inhibitor of the sarcoendoplasmic reticulum calcium ATPase (SERCA) pump and 2-APB, an inositol triphosphate receptor (IP3R) antagonist. To assess the contribution of calcium entry, nifedipine, an inhibitor of L-type voltage-gated calcium channel previously reported to be involved in GnRH neurons calcium oscillations (32, 33), was used (Fig. 5A and 5B). With all 3 drugs an increase in the frequency of calcium oscillations was observed upon T3 application (Fig. 5A). However, in the presence of nifedipine, the response was attenuated (+0.20 peaks/min) due to significantly fewer cells responding to T3 (45.3% of nonresponders), Figs. 5A, 6A, Table 4). In contrast, the amplitude of response to T3 in the presence of 2-APB was much higher (+0.96 peaks/min) than with T3 alone (+0.35 peaks/min) with significantly more cells having a fast response and a strong stimulatory effect, indicating a potentiation of the effect (Fig. 6B, Table 4). We also noted that, while the slow response is conserved with thapsigargin and nifedipine, in the presence of 2-APB, the calcium oscillation frequency goes back to basal levels directly after removal of T3, mimicking the response to periostin (Fig. 5B, Table 4).

Figure 5.

Inhibition of different calcium sources alters GnRH cell response to T3. (A) histograms summarizing the average calcium oscillation frequency (peaks/min) in all GnRH neurons (above), and pie charts showing the proportion of cells exhibiting a response-type (below), in presence of thapsigargin (thapsi, 1µM N = 3, n = 110), 2-APB (75µM N = 3, n = 127) and nifedipine (Nif, 1µM N = 3, n = 172). All drugs were diluted in serum-free media containing a cocktail of amino acid blockers (AAB: 20µM BIC, 10µM CNQX, 10µM D-AP5). Each period corresponds to 5 minutes. Different letters indicate significantly different results in Tukey's multiple comparisons test (P < .05; see Table 2). (B) Schematic indicating the function of each drug used in calcium imaging experiments (cf. Table 1). The blue cell represents a GnRH neuron, sharp arrows show agonist effect and calcium flux, and blunted arrows show inhibitory effect.

Figure 6.

Summary of the proportion of the different types of responders and amplitude of the response. (A) Stacked bar chart showing the proportion (percentage) of each responder-type for all conditions tested. (B) Histograms indicating the amplitude of the effect seen in each condition tested between the baseline period (P1) and the period of drug application (P2). Results are expressed in percentage delta change ([P2-P1]/P1*100).

Discussion

Despite clinical studies showing a clear link between thyroid dysfunctions and reproductive issues (9, 10, 12, 34, 35), the mechanism by which thyroid hormones influence reproductive function remain unclear.

In this work, we investigated the direct effect of T3, the main active thyroid hormone, on calcium transients in an ex vivo model containing numerous GnRH neurons. Using calcium imaging in these explants, we showed that T3 can have a direct effect on GnRH neurons, inducing a long-term increase in calcium oscillation frequency.

Thyroid hormone T3 interacts with multiple receptor types, the most well-characterized being nuclear thyroid receptors (TRα and TRβ). These receptors mediate the genomic effects of T3 by modulating gene transcription. Additionally, T3 can exert rapid, nongenomic actions through non-nuclear receptors, notably the integrin αVβ3 receptor located at the plasma membrane, which activates various intracellular signaling pathways (22). Although transcripts for TRα, TRβ, and the integrin aV subunit have been previously reported in GnRH neurons from adult mice using TRAPseq on preoptic tissue (36), we examined expression of these receptors and the integrin aV and β3 subunits in GnRH neurons from explants and adult male and diestrus female mice. Using PCR and immunocytochemistry, we confirmed the expression of TRα and integrin αVβ3 receptors in GnRH neurons both maintained in explants, consistent with microarray data (31) and in GnRH neurons from adult mouse brain. Notably, blocking transcription with a polymerase inhibitor or using a TR antagonist failed to prevent the stimulatory effect of T3 on GnRH neurons, suggesting that this effect does not rely on the action of T3 on its nuclear receptor. Similar stimulatory effects were observed using a T3-BSA conjugate, which prevents T3 from entering cells, indicating the role of membrane receptors in this process. Using an integrin αVβ3 antagonist, we demonstrated that blocking this receptor abolished the stimulatory effect of T3 on calcium oscillations in GnRH neurons. Together, these results indicate that T3 can modulate GnRH neuron calcium activity by acting via a nongenomic pathway through the αVβ3 integrin receptor.

Interestingly, periostin, an αVβ3 agonist (37), induced a strong stimulatory effect on GnRH neurons with a different pattern of activation than T3. While T3 induces a moderate but long-lasting increase in calcium oscillation frequency, periostin triggers a strong, fast response that returns almost to basal levels immediately after removal of the drug. This difference led us to hypothesize that the long-lasting stimulation by T3 might result from a combination of an initial rapid response followed by a slower, sustained response originating from different calcium sources. To explore this, we investigated the contribution of intracellular and extracellular calcium stores in mediating the response to T3. While the rapid stimulatory effect of T3 persisted despite modulators of intracellular or extracellular calcium stores, blocking with nifedipine reduced the number of responding cells. This suggests the involvement of external sources of calcium and VGCC in mediating the initial increase in calcium oscillation frequency induced by T3 binding to αVβ3.

Further, the use of 2-APB, an IP3R and TRP channel blocker (38-40), led to a rapid and robust increase in calcium oscillation frequency in the presence of T3, similar to the response observed with periostin. At the concentrations used in our study, 2-APB not only acts as an antagonist but also as an activator of TRPV channels (41, 42), enhancing their response to various stimuli (including acids, cannabidiol, capsaicin, protons) (41-43). Moreover, TRPV channels are needed to induce periostin calcium response downstream of its integrin αVβ3 activation (37), suggesting that TRPV channels contribute to the high-amplitude but short-term activation mediated by αVβ3. Given the similarity between the response to periostin and to T3 in presence of 2-APB, we propose that TRPV activation may potentiate the T3 effect, triggering a signaling pathway similar to periostin stimulation. However, T3 alone appears to prioritize a different pathway, as its effect does not typically lead to such an acute response unless other channels are blocked. Other receptors might be involved in this type of T3 mediated response, directly stimulated by T3 or downstream of the integrin αVβ3 activation.

The differential mechanisms of calcium recruitment by T3 and periostin at the same receptor highlight potential variations in αVβ3 receptor conformation, binding site accessibility, or downstream signaling pathways. Additionally, in the presence of 2-APB, the long-term effect of T3 is absent, resembling the response observed with periostin, further reinforcing the similarity between both activation mechanisms. Notably, the only other condition where the slow response was reduced was by using the transcription inhibitor DRB, suggesting that the long-term effect of T3 involves a genomic mechanism. However, neither T3 receptor antagonists nor T3-BSA, altered this response. This raises 2 possibilities: either in addition to the T3 effect on integrin αVβ3 receptors, T3 acts through an alternative pathway to induce the slow response (which would be sensitive to 2-APB) or the binding of T3 to its receptor triggers a secondary mechanism and/or different signalization cascades, potentially involving transcription factors.

Integrin αVβ3 receptors participate in interactions between cells and the extracellular matrix as well as in synaptic plasticity (44, 45). It is plausible that the fast and slow responses to T3 have distinct physiological roles, with the rapid effect influencing neuronal activity and secretion, while the slower response contributes to structural remodeling. The role of T3 in modulating GnRH neuron activity remains unclear, but it is well established that T3 plays a critical role in brain development and seasonal reproductive adaptations (11, 18). As a mediator of environmental cues, T3 may act as a dynamic regulator of GnRH neuronal function.

Additionally, T3 is essential for energy metabolism balance (20, 21), suggesting that systemic T3 levels could provide information about metabolic status, allowing GnRH neuronal activity to be finely tuned according to bodily needs. Although T3 levels remain relatively constant in the circulation, there is a local production and regulation within the brain. Particularly, tanycytes, specialized glial cells lining the wall of the third ventricle, highly express deiodinase enzymes regulating concentrations of thyroid hormones (18). Due to their strategic location between the cerebrospinal fluid and the brain parenchyma, tanycytes are key sensors of metabolic and reproductive states (46), further supporting the role of T3 in neuroendocrine adaptation.

During seasonal changes and estrous cycles, tanycyte end-foot remodeling facilitates GnRH release (47, 48). It is possible that thyroid hormones, via αVβ3, contribute to these morphological changes, as αVβ3 signaling can be implicated in structural remodeling (37). This highlights the potential role of T3 as a sensing hormone, integrating internal and external cues to regulate neuroendocrine function dynamically.

With this study, we gain a better understanding of how T3 modulates GnRH neuronal activity, which is particularly relevant in the context of endocrine disruption, a growing concern in the 21st century. While the physiological role of T3 on GnRH neurons may be beneficial under normal conditions, thyroid imbalances could lead to detrimental effects on reproductive function. Given the increasing comorbidities between thyroid disorders and reproductive dysfunctions, this study paves the way for elucidating molecular mechanisms and developing targeted therapeutic strategies to address these interconnected health issues.

Glossary

Abbreviationsh

AAB

amino acid blocker

BIC

bicuculline methochloride

BSA

bovine serum albumin

DRB

5,6-dichlorobenzimidazole 1-b-D-ribofuranoside

GnRH

gonadotropin-releasing hormone

IP3R

inositol triphosphate receptor

PCR

polymerase chain reaction

SFM

serum-free medium

T3

triiodothyronine

T4

thyroxine

TR

thyroid receptor

Funding

This research was supported by the Intramural Research Program of the National Institutes of Health (NIH), National Institute of Neurological Disorders and Stroke (ZIA NS002824-34).The contributions of the NIH author(s) were made as part of their official duties as NIH federal employees, are in compliance with agency policy requirements, and are considered Works of the United States Government. However, the findings and conclusions presented in this paper are those of the author(s) and do not necessarily reflect the views of the NIH or the U.S. Department of Health and Human Services.

Author Contributions

C.Q. and S.W. conceptualized and designed experiments. C.Q., N.M., and A.B. performed experiments. C.Q., N.M., and S.W. analyzed data. C.Q. and S.W. wrote the manuscript. All authors contributed to the article and approved the submitted version.

Disclosures

The authors have nothing to declare.

Data Availability

Original data generated and analyzed during this study are included in this published article or in the data repositories listed in References.