Targeted downmodulation of OXTR and Cx43 in the myometrium by SERTIs as a novel tocolytic strategy

Abstract

Approximately 10% of US births deliver preterm before 37 weeks of completed gestation. Premature infants are at risk for life-long debilitating morbidities and death, and spontaneous preterm labor explains 50% of preterm births. There are currently no FDA-approved tocolytics, and this research explores the therapeutic potential of ‘selective endoplasmic reticulum translocation inhibitors’ (SERTIs) for downmodulating contractile-associated proteins in the myometrium. The lead compound in this study, TK047, is an analog of cyclotriazadisulfonamide (CADA), which we found to reduce plasma membrane localization of oxytocin receptor and connexin 43 in human uterine smooth muscle cells, decrease oxytocin-induced calcium influx, and limit Cx43 gap junction channel activity. TK047 imparts strong negative inotropic effects in human and mouse myometrium, and does not exhibit cellular toxicity or affect mouse fetal viability at therapeutic doses. CADA analogs have already been shown to be effective in other disease models, and here they present as a potential first-in-class tocolytic option through the concurrent downmodulation of two contractile-associated proteins essential to parturition.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s44494-026-00005-x.

Introduction

Approximately 15 million preterm births occur annually worldwide [1, 2]. Preterm infants that survive often experience life-long heath complications, including learning disabilities, cerebral palsy, and heart/respiratory problems [3]. In 2016, 10% of US births were premature, a number that has changed little in decades because there remain no FDA-approved drugs that reliably halt spontaneous preterm labor allowing pregnancies to go to term [4, 5]. Here we investigate the ability of new class of compounds that downmodulate ‘contractile-associated proteins’ (CAPs) in the myometrium in order to provide effective tocolysis.

The myometrium is unlike any other muscle in the body in that it must remain entirely quiescent for 40 weeks during pregnancy, only to change phenotype during parturition to ensure productive labor. These unique requirements placed on the myometrium mean that the mechanisms that mediate contraction are exceptional when compared to other muscle types. To this point, the myometrium exhibits a non-canonical response to cyclic nucleotides by relaxing to nitric oxide in a cGMP-independent manner [6, 7], and it rapidly desensitizes to ‘protein kinase A’ stimulation [8], limiting the therapeutic potential for β2-adrenergic receptor agonists. These idiosyncrasies of the myometrium highlight the failure of repurposed smooth muscle drugs, such as ritodrine and nifedipine, which do not improve fetal outcomes and can elicit cardiac arrythmias, pulmonary edema, and hypotensive crisis [9–11]. Even drugs that were designed specifically to enhance myometrium quiescence, such as the oxytocin receptor (OXTR) antagonist, atosiban, do not prolong pregnancy or improve neonatal outcomes compared to placebo [12]. This unexpected result speaks to a larger question concerning the limitations of targeting single proteins or pathways as part of an effective tocolytic strategy.

Critical to the initiation and maintenance of labor are OXTR and the connexin 43 (Cx43) [13]. OXTR is a G-protein coupled receptor (G_αq) that is stimulated by the neurohormone, oxytocin (OT), and whose transcript levels increase 300-fold during parturition [14]. OXTR activation initiates an influx of Ca²⁺ into uterine smooth muscle cells, driving contraction of the uterus. Cx43 gap junction channels (GJCs) complement OXTR in that they electrically couple myocytes, allowing for the propagation of depolarization events [15, 16]. The cable properties of Cx43 are particularly germane to active labor as the uterus is denervated during pregnancy [17], and we have recently determined that Cx43 expression is dysregulated in the myometrium of women who experience preterm labor [18].

The failure of monotherapies, drugs that target single proteins or pathways, to reliably prevent early birth after onset of labor highlights the need for a more robust approach to tocolysis. The concurrent targeting of two or more myometrial contractile/quiescent pathways, called combination tocolysis, has been attempted with some success [19–22]. Our laboratory’s previous ex vivo analysis of human myometrium has found that the concurrent use of drugs that increase nitric oxide availability, while also inhibiting the formation/activity of Cx43, impart significant negative inotropic effects [18]. Recently, it has also been found that the simultaneous inhibition of OXTR and other calcium-handling pathways results in delayed delivery in a mouse model [21]. With this in mind, we seek to explore new tocolytic approaches through the concurrent downmodulation of multiple CAPs.

A relatively new class of small molecules called ‘selective endoplasmic reticulum translocation inhibitors’ (SERTIs), are currently being explored for their therapeutic potential [23, 24]. SERTIs cause protein degradation by selectively binding to an amino acid sequence of the nascent peptide chain during its translocation across the endoplasmic reticulum (ER) membrane. Most proteins enter the interior lumen of the ER by crossing its membrane during translation, a process known as co-translational translocation [25–33]. Many nascent membrane proteins are targeted to Sec61 translocon [34] by means of a signal sequence or cleavable signal peptide (SP) [35], the latter usually comprising the first 15–45 amino acids at the N terminus of the nascent protein (preprotein). SPs have a wide variety of primary structures, but the sequence is specific for every protein, and it is unique for each species The most protein-selective SERTIs are CADA compounds, analogs of cyclotriazadisulfonamide, which was shown to inhibit replication of human/simian immunodeficiency virus by selectively downmodulating plasma membrane expression of cell-surface ‘cluster of differentiation 4’ (CD4) [24, 36–41]. More recently, CADA has been found to downmodulate several other proteins, including immune checkpoint protein PD-1. CADA compound KZR-261 was tested in a clinical trial for cancer patients with solid tumors and found to be safe at doses that prevented tumor growth. [42]

Recently reported Cryo-EM structures [42, 43] show that CADA compounds and several less selective translocation inhibitors bind to the Sec61 channel across the ER membrane. In one study, CADA decreased the expression of only five human T-cell membrane proteins by more than 50% [44], suggesting that it is displaced by most proteins. CADA binds to the SP of certain translating nascent proteins, diverting them into the cytoplasm, where they are degraded proteolytically [45–47]. Many proteins lacking N-terminal signal peptides also employ the Sec61 pathway [25, 26, 41] The first transmembrane domain (TMD) of connexins evidently acts as a signal anchor, directing the translating nascent chain to the Sec61 channel [48]. GPCRs that lack an N-terminal signal peptide, such as OXTR, are targeted to the ER membrane via their 1 st TMD. This domain may be inserted by the ER membrane protein complex (EMC). In the latter case, Sec61 has been reported as necessary for ER membrane insertion of the second TMD [49]. Accordingly, we postulated that CADA compounds might be capable of impairing ER membrane insertion of both Cx43 and OXTR.

Here we investigate the ability of a novel CADA compound to provide tocolysis by downmodulating both OXTR and Cx43 in human and mouse myometrium. The long-term goal of this research is to employ translational approaches to develop novel therapeutic entities that target critical CAPs and pathways of the myometrium to halt preterm labor and prevent preterm birth.

Materials and methods

Tissue collection

Myometrial biopsies were obtained with written informed-consent from women with pregnancies undergoing Cesarean section, as previously described [18]. Exclusion criteria in pregnant women include: uterine or generalized infection to include COVID-19, a maternal age of less than eighteen years, any history of drug abuse, co-morbid diagnoses such as HIV infection or AIDS, hepatitis C infection, uncontrolled diabetes, renal disease, and any use of steroids other than betamethasone (including topical use) during pregnancy. Patient demographics and concomitant drug therapies are provided in Supplementary Data 2. Immediately following collection, tissues were transported to the laboratory in ice-cold Uterine Transport Buffer containing, in mM: NaCl (120), KCl (5), CaCl₂ (0.5), KH₂PO₄ (0.587), NaHCO₃ (5), Na₂HPO₄ (0.589), MgCl₂ (2.5), dextrose (20), Tris Base (25), adjusted to pH 7.4. Tissues were dissected to isolate smooth muscle from fascia and adipose tissue, then either immediately employed in contractile experiments or snap frozen in liquid-nitrogen and stored in a vapor-phase freezer unit at − 150 °C.

CADA compound synthesis

Synthesis and characterization of compounds TK047, SG004, and SG007 can be found in an accompanying document (Supp. Data. 1) to this manuscript.

Cell culture

Primary (1°) uterine smooth muscle cells were isolated from freshly dissected term human myometrium using a gentleMACS™ dissociator, as previously described [50]. Uterine myocytes (CD31-) were column selected (Miltenyi Biotec:130-091-935, Auburn, CA) and are referred to as 1° pregnant human uterine smooth muscle cells (phUSMC). Cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) with 50 U/ml streptomycin, 50 µg/ml penicillin, 10% fetal bovine serum (FBS), and supplemented with estrogen (15 ng/ml) and progesterone (200 ng/ml). For all experiments, cells were growth-arrested for 48 h prior to experimentation. Cells subjected to growth-arrest were cultured as described above, but in the absence of FBS and the addition of 1× Insulin-Transferrin-Selenium (1:100, Cat: 41400045, ThermoFischer, Waltham, MA). All cells were cultured in a balanced oxygen (95%:5% O₂:CO₂) incubator at 37 °C.

Calbryte™ Ca²⁺ permeability assay

1° cultured (p3-p6, pool of 8 term nonlaboring patients, n=5 replicates for experiments) phUSMC or telemorized phUSMC⁵⁴⁵⁴ (p5-p15, n=3 replicates for experiments) were seeded in a flat bottom 96-well microplate at ~ 7,500 cells/well (675,076, Greiner Bio One, Kremsmünster, Austria). Upon 80–90% confluency cells were switched to growth-arrest media for 48 h. Cells were treated with varying doses of TK047 (10–30 µM, or DMSO vehicle control < 1% final volume) for periods between 4 and 48 h, then rinsed 2× in KREBS buffer containing, in mM: NaCl (118), KCl (4.75), CaCl₂ (2.5), KH₂PO₄ (1.2), NaHCO₃ (25), MgCl₂ (1.2), dextrose (20), adjusted to pH 7.4, with the addition of 0.04% Pluronic® F-127 and 1 mM probenecid (anion transporter inhibitor), followed by 5 µM Calbryte™ 520 AM (Cat. 20650—AAT Bioquest, Sunnyvale, CA) for 1 h. Following incubation period cells were rinsed again (3x), baseline fluorescence was recorded (ex/em 495/540 nm) for ~ 20 s, then imaged according to experimental protocol. For analysis, ΔCa_i²⁺ was calculated by comparing relative changes in fluorescence between control and treatment cells following exposure to 100 nM OT (solubilized in reaction buffer). For all experiments using 100 nM OT, this was done to elicit a maximal OT response [51, 52].

Membrane potential assay

1° cultured phUSMC were prepared as described above for the Calbryte™ assay. Following growth-arrest, cells were treated for 48 h with 10 µM TK047 (or DMSO vehicle control <1% final volume). Cells were rinsed three times with KREBS and loaded with FluoVolt™ for 30 min (Cat: F10488, ThermoFisher, Waltham, MA), followed by three rinses with KREBS. Following a 5-min baseline recording, cells were exposed to 100 nM OT (solubilized in reaction buffer) and changes in membrane potential were recorded for 20 additional minutes. Each well was normalized to itself by dividing each time point by the baseline average so that the data is presented as change relative to pre OT treatment.

Contractile studies

In a temperature controlled (37 °C) organ bath (DMT 820MS, Danish Myo Technology, Hinnerup, Denmark), strips of myometrium (~ 0.5 × 15 mm) from the superior portion of the transverse incision were clip-mounted to a force transducer and stretched to L₀ then held at 1 g of final tension [18, 50, 53]. Following stable isometric tension, all tissues were challenged with 60 mM KCL for 3 min, returned to standard KREBS, then allowed to equilibrate for up to one hour. Some tissues were then exposed to 8 nM OT (established EC₅₀ [54]) depending on experimental parameters. Control tissues were exposed to a volumetric equivalent of drug solvent (DMSO, < 1% final volume). Area under the curve (AUC), peak tension, and contractile frequency were determined using LabChart (version 8.1.12, Win10, ADInstruments., Colorado Springs, CO).

For ‘acute’ and ‘2 h’ studies, freshly dissected tissues were mounted isometrically at 1 g of tension and maintained at 37 °C in an organ bath with Krebs buffer organ bath bubbled with balanced oxygen (95%:5% O₂:CO₂). “Spontaneous” contractions are defined as tissue exposed to 60 mM KCL for 3 min, with no use of additional agonist to stimulate contractions. “OT-induced” contraction received the aforementioned KCL treatment, followed by 8 nM OT after ~ 1 h. For’24 h’ studies, tissues were mounted as described above, stimulated using the “OT-induced” protocol, then following 20-min of sustained, steady contractions (baseline OT response), the Krebs bath was replaced with ‘phUSMC cell culture media’ +20 mM HEPES, then the individual organ bath chambers were detached from the apparatus and transferred to a balanced oxygen (95%:5% O₂:CO₂) incubator at 37 °C. Tissue was maintained at 1 g of tension during this period. The following day the media was replaced with Krebs buffer and each organ bath chamber was reconnected to the apparatus for additional recording.

For organ bath experiments each ‘n’ is a unique human patient or mouse from which the values of 1–3 myometrial strips for each condition were averaged. The last 3 contractions of each dosing or control period were analyzed to determine AUC and peak tension. Peak tension is defined as the delta (maximum–minimum) for each given contraction. Each strip is normalized to its own baseline prior to dosing, then to the average of the control strips to account for rundown (controls are volumetric equivalents of DMSO not to exceed 1% of final volume). The average of all control strips for each ‘n’ was set to a nominal value of ‘100’ so that data is presented as a “% change from control.” Contractile frequency, presented as ‘contractions per unit time’ (CPUT), is defined as the number of contractions in the final 15 min of the dosing interval divided by the number of contractions during the 15 min period just prior to dosing.

Cell viability/cell toxicity assays

For both cell viability and toxicity assays 1° phUSMC myocytes (as described above) were seeded at ~7500 cells/per well in a 96-well plate, grown to ~ 90% confluency, then growth-arrested for 48 h followed by treatment with 1–30 µM TK047 (or DMSO vehicle control < 1% final volume) for an additional 48 h. Following treatment, visual changes in cell morphology was confirmed at 10× with a phase contrast microscope. 2,3-Bis(2-methoxy-4-nitro-5-sulfophenyl)−2H-tetrazolium-5-carboxanilide sodium salt (XTT) cell viability assay (Cat. X12223, ThermoFisher, Waltham, MA) or lactate dehydrogenase (LDH) cytotoxicity fluorescent assay (Cat: C20303, ThermoFisher, Waltham, MA) were used according to manufacturer protocols, which includes were positive and negative controls, for maximal and null response. For both assays a volumetric equivalent of vehicle (DMSO) was used.

Immunofluorescence/protein localization

Immunofluorescence and cell localization analyses were conducted in 1° phUSMC myocytes seeded in 35 mm glass bottom dishes (P35GCol-1.5–14-C, MatTek, MA), growth-arrested for 48 h, followed by treatment with 10 µM of TK047, SG004, SG007 (or DMSO vehicle control < 1% final volume) for an additional 48 h. Following treatment, cells were fixed in 4% paraformaldehyde, and, after blocking overnight, were labeled with anti-Cx43 antibody (1:200, Cat. Ab11370, Abcam, Cambridge, MA) or anti-OXTR antibody (1:50, Cat: Ab300443, Abcam, Cambridge, MA). After incubation with goat anti-rabbit IgG (H+L) Tetramethylrhodamine (TRITC) conjugate (1:100, Cat: A16101, Invitrogen, Carlsbad, CA), fixed dishes were treated with wheat germ agglutinin (WGA)-conjugated Alexa 488 (1:100, Cat: W11261, Carlsbad, CA) to identify cell boundaries, and Hoechst 33,258 pentahydrate bis-benzamide (1:50,000, Cat: H1398, Invitrogen, Carlsbad, CA) to indefinity nuclear boundaries. Treated dishes were imaged on a Keyence BZX-710 fluorescent microscope with an oil-immersion 60× objective.

Image processing

Image analyses were performed in ImageJ’s Fiji software on 3 channel tiff images (DAPI, TRITC, GFP). A FIJI macro was designed to identify nuclei, designate cell exclusion zones, and form cell masks for TRITC analysis of cell compartments of interest. The output integrated density readings identified Cx43 or OXTR presence in the plasma membrane and cytosol (nuclei excluded). Triplicate dishes (n=3) were analyzed with 3 fields of view assessed per plate (values averaged).

Fluorescent recovery after photobleaching (FRAP) 1° phUSMCs (p6-p9, from a singular patient, n=6 replicates for experiments) were seeded onto 35 mm glass bottom dish (P35GCol-1.5–14-C, MatTek, MA) and grown to 80% confluency. Following 48-h growth arrest, cells were treated for 24 h with 10 µM TK047 (or DMSO vehicle control < 1% final volume), and compared against positive control cells treated with 30 µM 18α-Glycyrrhetinic acid (18αGA; Cx43 GJC inhibitor, G8503, Sigma Aldrich St. Louis, MO—or DMSO vehicle control < 1% final volume) for 15 min. Cells were rinsed twice with PBS, then loaded with KREBS containing both 25 mM HEPES and 10 µM of the fluorescent indicator 5-Carboxyfluorescein Diacetate, Acetoxymethyl Ester (5-CFDA, AM) for 15 min at 37 °C (C1354, Invitrogen Carlsbad, CA). Cells were rinsed twice with PBS to remove the 5-CFDA, AM and subsequently imaged

FRAP processing

Cells were imaged using the Leica Stellaris 5 Confocal Microscope under 40× immersion objective with a fixed solid-state diode laser at 488 nm. The intensity of the fluorescent signal was measured between 500 and 560 nm (emission band-pass filter). The Leica Application Suite (LAS) AF FRAP wizard was used to locate neighboring cells with a singular (target) cell of interest and define the region of interest within the target cell. The FRAP parameters were as follows: 5 pre-bleach scans at a minimal time of 26 ms iterations, 40 bleach scans at 100% laser power with zoom mode initiated and minimal time applied, lastly followed for recovery using 60 post-bleach scans at minimal time. All data were extracted as csv files and corrected for photobleaching during post-bleach collection. Fluorescent intensity was corrected by normalizing post-photobleaching measurements to the mean fluorescent intensity of the 5 pre-bleach values (F_i). From there, each normalized fluorescent intensity value at time (F_t) was compared relative to the fluorescent intensity immediately after photobleaching (F₀) to generate a normalized recovery curve over time, F_norm(t) function:

$$F_{\mathit{norm}}\left(t\right)=\frac{F\left(t\right)-F_0}{1-F_0}$$ 1

as previously described and modified [55–57]. The percent of mobile fraction was subsequently determined from the normalized recovery curves and the terminal fluorescence was extracted from the final five post-bleach frames.

Statistical analysis

Human and mouse ex vivo tissue experiments were performed with at least three biologic replicates, where each ‘n’ constitutes a unique patient or animal. Cell work was performed as described, with each ‘n’ constituting an individual well or dish, with technical replicates or multiple reads per well averaged. t-test were unpaired and one/two-tailed, as appropriate. All error bars on graphs displayed as SEM. All data were analyzed using Prism (Graphpad Software, San Diego, CA). For all analyses, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p ≤ 0.0001.

Results

Transmembrane domains permit ER targeting of Cx43 and OXTR

Based of the ability of CADA compounds to downmodulate the expression of certain nascent membrane proteins by targeting the translocon by means of a signal peptide, we elected to evaluate 20 novel CADA derivatives developed in-house for their tocolytic potential via the downmodulation Cx43 and OXTR in human uterine smooth muscle (data not shown). SG004, SG007 and TK047 were selected for further evaluation. Within their initial TMD’s, both proteins of interest contain the requisite signal peptide sequences needed for targeting by CADA compounds (Fig. 1). The N-terminus of the Cx43 protein is cytoplasmic and the 1 st TMD of connexins apparently acts as a signal anchor, directing the translating nascent chain to the Sec61 channel in the ER membrane [48, 58]. While CADA compounds bind Sec61, they are likely displaced by most Sec61 client proteins [43], allowing translocation and insertion of their TMDs into the ER membrane. CADA compounds can recognize polar amino acids in certain signal sequences, impairing protein translocation, which promotes cytoplasmic degradation [41]. Unlike Cx43, the soluble N-termini of most GPCRs are extracellular. Those that that lack an N-terminal signal peptide are targeted to the ER membrane via their 1 st TMD, which is inserted by the ER membrane protein complex [49] with the soluble N-terminus toward the ER lumen. According to a new model [59], short or negatively charged soluble domains of GPCRs are inserted into the lumen by the EMC. Nascent chains with neutral or positively charged soluble domains are passed to alternative insertases, including Sec61. As seen in Fig. 1, the N-terminal soluble domain of human (and mouse) OXTR has more negative than positive residues, suggesting that their 2nd TMDs engage Sec61. In any case, both the 1 st and 2nd domains contain polar residues. We propose that CADA compounds may bind one or more polar amino acids in the first of second TMD and block or alter the topology of insertion of OXTR, as well as Cx43. Blocked translocation by CADA compounds cause degradation and it is also known that multi-pass proteins with incorrect topology can undergo degradation [49].

Fig. 1.

Predicted CADA-binding signal sequences for Cx43 and OXTR. A CADA compounds SG004, SG007 and TK047 are predicted to bind to TMD1 of human and mouse Cx43. B They are expected to bind to either TMD1 or TMD2 of OXTR (underlineD. The N-terminal soluble domain of human OXTR contains 4 negative residues and 3 positive (net charge − 1). The same residue of mouse OXTR contains 5 negative residues and 2 positive (net charge − 3). Polar residues are shown in color: (magenta, neutral); (red, positive); (blue, negative), while the soluble domain lies upstream of TMD1. 3D structures (right) acquired from Protein Data Bank (OXTR—6TPK [60], Cx43—7F93 [61]), and modeled in PyMol 3.1 [62]. Cx43, a hexamer, is presented as a gap junction channel (GJC)

TK047 identified as lead CADA compound

3 structurally similar CADA compounds, TK047, SG004, and SG007 (Fig. 2C), were tested for their ability down modulate plasma membrane (PM) expression of OXTR and Cx43. 1° pregnant human uterine muscle cells (phUSMC, n=3, p4-8) were cultured in glass-bottomed dished, growth-arrested for 48 h, then treated with 10 µM of either TK047, SG004, SG007, or volumetric equivalent of DMSO for an additional 48 h. Fixed dishes were probed with antibodies for Cx43 or OXTR, TRITC labeled, then imaged at 60× magnification. Relative to control, TK047, SG004 and SG007 reduced PM localization of OXTR by − 6.98% ± 2.43 (p = 0.0455), − 9.64% ± 15.60; (p=0.5700) and − 28.99% ± 1.22 (p = < 0.0001) respectively, while with increasing cytosolic concentrations of OXTR by +214.30% ± 30.71; (p = 0.0022), +28.31% ± 46.17 (p = 0.5728), and +37.49% ± 10.64 (p = 0. 0244; Fig. 2B, left panel). Further, TK047, SG004 and SG007 reduced PM localization of Cx43 by − 48.19% ± 9.67 (p = 0.0380), − 10.44% ± 8.00 (p = 0.2617) and − 36.85% ± 11.99 (p = 0.0372), respectively, while with fluctuating cytosolic concentrations of Cx43 by +141.10% ± 90.85 (p = 0.1953), − 42.38% ± 20.58; (p = 0.1086), and +21.73% ± 22.72 (p = 0.3929; Fig. 2B, right panel). The corresponding decrease in PM expression of Cx43 and OXTR and increase in cytosolic localization intimates that the CADA compounds are preventing proper insertion and trafficking of the proteins to the PM. TK047 was selected for further inquiry due to the robust decrease in PM localization of the proteins and increased cytosolic presence.

Fig. 2.

CADA compounds downmodulate OXTR and Cx43. 3 novel CADA compounds, SG004, SG007 and TK047, were investigated for their tocolytic potential. 1° phUSMC (n = 3) were exposed to 10 µM of each compound, or DMSO control, for 48 h following growth arrest. A Immunofluorescent images with overlayed DAPI; nucleus [blue]/TRITC; POI [red]/WGA; plasma membrane [green], TRITC only, and cytosolic/PM ROI masks B left panel—OXTR localization to the PM was decreased by − 6.98% (TK047) − 9.64% (SG004) and − 28.99% (SG007), and rose in the cytosol by +214.30% (TK047) +28.31% (SG004) and +37.49% (SG007). Average value of control wells set to nominal value of ‘1’ and treatment values are presented as fold change relative to control. Right panel—Cx43 localization to the PM was decreased by − 48.19% (TK047) − 10.44% (SG004) and − 36.85% (SG007), and fluctuated in the cytosol by +141.10% (TK047) − 42.38% (SG004) and +21.73% (SG007). C Despite similar structures, each CADA compound exhibits unique selectivity for OXTR and Cx43

Identifying optimal TK047 concentration

To validate the optimal concentration of TK047, ‘cell viability’ and ‘cell toxicity’ assays were run. 1° phUSMCs aggregated from 8 term pregnant patients were treated with TK047 at ½ log doses from 1 to 30 for an additional 48 h (n=5 per condition), or received DMSO (volumetric equivalent). A colorimetric redox cell viability XTT assay determined that at all tested concentrations (1 µM p = 0.2403, 3 µM p = 0.1491, 10 µM p = 0.1849, and 30 µM p = 0.6536) TK047 treatment did not diminish cell viability relative to control (Fig. 3A). Similarly, a LDH cytotoxicity assay was run under the same condition determined that TK047 exposure for 48 h does not induce cytotoxicity, relative to control, at concentrations of 1–10 µM (1 µM p = 0.1799, 3 µM p = 0.1879, 10 µM p = 0.1088), with a small, but significant increase in cytotoxicity at 30 µM (p = 0.0446; Fig. 3B). A secondary visual confirmation of these finding using 10× phase-contrast brightfield images indicated that treatment does not alter apparent cell health, except at 30 µM TK047, which induces anomalous membrane morphologies and some cell lifting consistent with the LDH cytotoxicity findings at this single concentration (Fig. 3C). Based on these data we elected to use a maximal concentration of TK047 of 10 µM for most tissue and cell work.

Fig. 3.

TK047 does not affect cell viability or induce cytotoxicity at effective concentrations. A Using a ‘cell viability’ screen, phUSMC (n = 5) were exposed to 1–30 µM TK047 for 48 h. At all tested concentrations TK047 did not adversely affect cell viability. B Under the same experimental conditions, a cytotoxicity screen determined that TK047 is not cytotoxic; however, there was a small but significant increase in toxicity at 30 µM. C 100× phase contrast microscopy validates cell viability/cytotoxicity findings with visual morphologies present at 30 µM (inset, blue arrow)

TK047 is safe in a mouse model of pregnancy

To determine the safety of TK047 in a whole animal model, C57BL/6J mice (Charles River Laboratories, Sacramento, CA) were bred over night from 16:00 to 8:00 (9–16 weeks) and received daily intraperitoneal injections of TK047 (5 mg/kg/day) for 3 days from ‘day post coitum’ (d.p.c.) 11–13 (mid-gestation, roughly aligning with gestational weeks 23–27 in humans, a key window for tocolysis, and also corresponds to critical organogenesis window in C57BL/6 [63]), then allowed to carry to term. Pup weight post-delivery on day 1 (PD1) averaged 1.33 ± 0.12 (g) for control mice (DMSO 50 µl/day, n=3), and 1.30 ± 0.11 (g) for pups from TK047-treated mothers (n=4), which was not significantly different (p = 0.8893; Fig. 4A). By PD21, pups from control animals averaged 8.46 ± 1.12 (g), while TK047 pups were 8.54 ± 0.76 (g), which did not represent a significant difference (p = 0.8929; Fig. 4B). Similarly, litter size was not significantly different, with the average number of pups for control animals at 9.66 ± 1.15, and 9.25 ± 0.5 for TK047-treated mothers (p = 0.5385; Fig. 4C). In both treatment and control groups, all pups were born alive on gestational day 19, while one pup from the control group and two from the treatment group not surviving until day 21. In non-pregnant adult mice treated similarly for 3 days with TK047 (n=6; 5 mg/kg/day) or DMSO (n=3), then euthanized within 6 h of the final injection, there was an average decrease in liver weight of by 17.6% ± 10.8 (Fig. 3D); however, this was not significantly different than control liver weight from DMSO treated animals (p = 0.069). In all treatment and control groups there was no visual indication of stress (beyond handling and injections), or variance in behavior of mother or pups, nor were there any gross level anatomical observations indicating possible teratogenic or toxic properties of the compound.

Fig. 4.

TK047 is safe in pregnant mice. A pregnant C57BL/6 J mice (n = 4 treated, n = 3 control) were injected with 5 mg/kg/day TK047 from d.p.c. 11–13. Pup weight was not significantly different between treatment and control groups at either day 1 or B day 21 after birth. C TK047 treatment did not affect litter size, and D did not alter liver weight in treated animals

TK047 reduces OT-mediated Ca²⁺ influx in phUSMC

Given the critical role OT plays in uterine smooth muscle Ca²⁺ influx following OXTR activation, as well as the importance of signal propagation between myocytes via Cx43 gap junction activity, we next sought to determine if TK047 altered Ca²⁺ handling in uterine smooth muscle cells. First, 1° phUSMC cells were treated with TK047 from 10 to 30 µM in ½ log increments for 4 h (n=3 per condition) to establish a half maximal effective concentration (EC₅₀) of TK047. Cells were then stimulated with 100 nM OT and peak fluorescence was determined over 90 s. There was a significant decrease in Ca²⁺ uptake (ΔCa_i²⁺), as measured by a reduction in fluorescent signal, relative to control beginning at 300 nM TK047 (p = 0.0355), and at all concentrations at or above 3 µM TK047 (0.01 µM p = 0.3563, 0.03 µM p = 0.1362, 0.1 µM p = 0.1617, 0.3 µM p = 0.0355, 1 µM p = 0.1047, 3 µM p = 0.0419, 10 µM p = 0.0148, 30 µM p = 0.0016), with an approximate concentration for for Ca²⁺ response (EC₅₀) between 10 and 30 µM (Fig. 5A). Acute pretreatment with TK047 for 10 min did not alter Ca²⁺ fluorescence relative to control (data not shown).

Fig. 5.

TK047 reduces Ca²⁺ influx in phUSMC. A phUSMC (n = 3) were pretreated with 10 nm-30 µM for 4 h then OT-stimulated (100 nM) in the presence of a fluorescent Ca²⁺ indicator (DMSO ctrl avg. set to 100). At concentrations of 300 nM, 3 µM,10 µM and 30 µM TK047, Ca²⁺ influx was significantly reduced. B When treated with 10 µM TK047 for 48 h (n = 5), Ca²⁺ influx (AUC) was reduced by ~ 14%. The OXTR antagonist, nolasiban (600 nM, red inset) produced a similar reduction in Ca²⁺ influx. C Peak fluorescent values following OT administration from panel ‘B’ reveal a reduction in peak fluorescence by 26% and a decrease in AUC of 20%

Based on these dose response data, 1° phUSMCs were treated with 10 µM TK047 for 48 h (n=5 per condition). Following baseline recording, cells were again exposed to 100 nM OT and fluorescence was recorded at ~ 1 s intervals for 60 s. Peak fluorescence was achieved ~ 8 s following OT administration, with the control cells exhibiting a 2.46-fold ± 0.36 increase in Ca²⁺ response over baseline, and a ~14% decrease in TK047 treated cells for a 1.87-fold ±0.37 increase over baseline (Fig. 5B). When cells were treated with 10 µM TK047 for 24 h (n=3), peak fluorescence was similar to control cells pretreated for 30-min with the OXTR antagonist, nolasiban (600 nM [64]; Fig. 5B, inset). TK047 treated cells experienced a reduction in the ‘area under the curve’ (AUC) by 20.32% ± 6.8 (p = 0.0087) and a reduction in peak height by 26.25% ± 12.1 (p = 0.0316), relative to control cells (Fig. 5C).

TK047 alters membrane potential following OT-stimulation in phUSMC

To determine the greater overall effect of TK047 on membrane potential (Ψ), 1° phUSMCs were loaded with the fluorescent indicator dye, Fluovolt™ (F10488, ThermoFisher, Waltham, MA), for 30 min, then rinsed. Baseline membrane potential was recorded for 5 min, then cells were stimulated with 100 nM OT and recorded for an additional 20 min. (Fig. 6A). During baseline period and following OT-administration, TK047 treated cells qualitatively exhibited far more sporadic depolarization events (inset—blue), varying in magnitude and frequency relative to control. In the 60-s following OT-administration, control cells exhibited an average peak height of 104.2% ± 1.65 relative to TK047 treated (Fig. 6B), which resolved to approximate baseline levels within 5-min of OT-stimulation, while the membrane potential of the TK047 cells was far more unpredictable and varied throughout the entire observation period.

Fig. 6.

TK047 alters membrane potential in phUSMC following OT-stimulation. A Using a membrane potential assay, 1° phUSMC (n=5) were exposed to 10 µM TK047 for 48 h, or DMSO control, then stimulated with 100 nM OT (arrow) following a 5-min baseline recording. The membrane potential of TK047-treated cells (blue) was far more variable and sporadic than control cells (grey) throughout the recording. Data presented as mean (sold/dashed lines) with SEM confidence intervals (shadeD and control baseline set to nominal value of ‘1.’ Inset graphs represent individual traces during the experimental period. B Control cells exhibited a peak depolarization value of 104% over TK047-treated cells (#) during the 60 s following OT-administration

TK047 minimizes gap junction intercellular communications

In order to determine if Cx43 GJC activity was functionally altered by TK047 we employed ‘Fluorescent Recovery After Photobleaching’ (FRAP). 1° phUSMC were treated with either 10 µM TK047 for 24 h or vehicle (vol. equivalent DMSO), then loaded with 5-CFDA, AM and imaged using a confocal laser scanning microscope. A 30 µm 18αGA (EC₅₀) was used as a positive control to demonstrate Cx43 GJC inhibition [57]. Target cells were photobleached for 40 iterations, resulting in an eliminated fluorescence intensity within the ‘region of interest’ (ROI). This enabled diffusion of unbleached dye from neighboring cells into the bleached ROI via GJC communication. Fluorescence recovery within the target cell was monitored over time and plotted as three distinct assessments. First, following photobleaching cells treated with either TK047, 18αGA, or vehicle, fluorescent recovery was recorded for 27.5 s (Fig. 7A) to establish relative recovery of fluorescent signal. Compared to control, cells treated with TK047 and 18αGA (n = 6 dishes, 6 FRAP recordings per ‘n’) observed a decreased mobile fraction by 19.98% ± 3.3 (p = 0.0003) and 23.95% ± 2.4 (p = <0.0001), respectively (Fig. 7B). There was no significant difference between TK047 and the positive control, 18αGA (p = 0.3531). Terminal fluorescence (TF) is defined as the average fluorescent values of the last 5 post-bleach intensity measurements and correlates to GJC activity. Relative to control, diminished terminal fluorescent values were observed in both TK047 and 18αGA treatments, where TF decreased by − 53.4% (17,690 ± 2982.9; p = 0.0002) and − 52.0% (19,455 ± 2470.3; p = < 0.0001), respectively (Fig. 7C). There was no statistical difference between terminal fluorescent intensity of TK047 and 18αGA treatments (p = 0.6583).

Fig. 7.

Cx43 GJC function is reduced in the presence of TK047. Using FRAP, Cx43 GJC activity was determined in 1° phUSMC (n = 6, 6 read points per ‘n’) pretreated with either 10 µM TK047 for 24 h or 30 µM 18αGA 15 min, followed by photobleaching to determine kinetic recovery profiles of fluorescence. A Normalized fluorescence intensities of cell treatments after photobleaching; control (grey), TK047-treated (blue) 18αGA (tan). B TK047 treated cells exhibited a decrease in mobile fraction by − 19.98% relative to control. C Terminal fluorescence of TK047 treated cells experienced a decrease of − 62.8%

TK047 imparts negative inotropic effects on human myometrium

Our cell-based assays indicate that pretreatment with TK047 for a time-period of up to 48 h downmodulates OXTR and Cx43, alters both Ca²⁺ handling and membrane potential, yet does not pose concerns surrounding cell viability or toxicity. Here we investigate the effect of TK047 with human myometrium to determine if the compound alters whole muscle contractile dynamics.

Acute effects of TK047 on human myometrium: Using human myometrium in an ex vivo organ bath, we first sought to determine if an acute 10-min pretreatment with 10 µM TK047 in term non-laboring (TNL) myometrium (n = 5 patients, 1–3 strips per patient) altered contractile dynamics following ‘spontaneous’ or ‘OT-mediated’ (8 nM) contractions. To test the effect of TK047 on spontaneous contractions (OT-free), tissues were permitted to equilibrate for ~1 h following high K⁺ stimulation, then pretreated for 10-min with 10 µM TK047 (Fig. 8A, bottom panel; n = 5 patients, 1–3 strips per patient). Following this, AUC, peak tension, and ‘contractions per unit time’ (CPUT) were recorded for 20-min. Relative to control, AUC for TK047-treated spontaneous tissues was 97.52% ± 8.01 (p = 0.7604), peak tension was 113.4% ± 10.50 (p = 0.2187), and CPUT was 118.89% ± 9.8 (p = 0.0922; Fig. 8A, top panel). Next, to determine if acute 10-min pretreatment with 10 µM TK047 altered OT-stimulated contractile dynamics, myometrial strips (n=5 patients, 1–3 strips per patient) were again exposed to high K⁺, allowed to equilibrate for ~ 1 h, OT-stimulated (8 nM) for 20-min to establish baseline response to OT in control and treated tissues, washed out and allowed to re-equilibrate for 30-min, pretreated with 10 µM TK047 for 10-min, then received a second OT-administration (Fig. 8B, bottom panel). This approach allows for the comparative analysis ([TK047+OT]/OT) that accounts for each strip’s variable reaction to OT. Relative to control, following OT-stimulation, the AUC for TK047-pretreated tissues was 121.41 ± 8.61 (p = 0.0203), peak tension was 97.28 ± 9.47 (p = 0.9336), and CPUT was 104.267 ± 16.74 (p = 0.8052; Fig. 8B, top panel).

Fig. 8.

Acute effects of TK047 on human myometrial contraction. A Using an ex vivo organ bath, spontaneously contracting TNL human myometrium (n = 5) was pretreated with 10 µM TK047 for 10-min (bottom panel). Relative to control, TK047 did not significantly alter the AUC 97%, peak tension 113%, or CPUT 118% (top panel). B When myometrium (n = 5) was OT-stimulated (8 nM), washed out, pretreated with 10 µM TK047 for 10-min, then OT-stimulated again (bottom panel), the relative response to OT, compared to control, was significant for AUC 121%, but not peak tension 97% or CPUT 104% (top panel)

2 h and 24 h TK047 pre-treatment: Next we sought to establish if a 2 h pre-treatment with TK047 in human myometrium provides sufficient time to modulate protein expression via the predicted Sec61 pathway to exert observable effects on myometrial contractile dynamics. TNL myometrium (n = 4 patients, 1–3 strips per patient) was mounted in an organ bath as described above. Following high K⁺ and OT stimulation, tissues were washed and pretreated with 10 µM TK047 for two hours, then re-stimulated with OT (Fig. 9A – left panel). Of note, during this longer 2 h pretreatment, small perturbations, consistent with tachysystole, were consistently observed (Fig. 9A—inset), an observation that aligns with the variable seen in the membrane potential of TK047 treated cells (Fig. 6). The AUC of TK047-treated tissues was reduced to 36.59% ± 10.47 (p<0.0001) of control, while peak tension was reduced to 27.64% ± 6.73 (p < 0.0001). Tissues were then washed and allowed to recover for 1 h, then OT-stimulated again to determine if the effect of TK047 could be recovered. AUC returned to 49.35% ± 31.22 (p = 0.1296) of control, while peak tension only returned to 26.26% ± 6.80 (p = 0.8415) of control (Fig. 9A – right panel).

Fig. 9.

Effects of 2-h and 24-h TK047 on human myometrial contraction. A Using an ex vivo organ bath, TNL human myometrium (n = 5) was pretreated with 10 µM TK047 under 1 g of tension for 2-h (left panel). Relative to control, TK047 treated tissue experiences rapid fluctuations, consistent with tachysystole (inset), along with a significant decrease in AUC to 36% and peak tension to 27%. Following a 1-h recovery, AUC 49% and peak tension 26% did not significantly recover (right panel). B Myometrium (n = 5) was mounted and OT-simulated (8 nM) as before, then treated for 24 h with 10 µM TK047 under 1 g of tension (left panel). Following reapplication of OT, AUC fell by 98% of baseline, and peak tension 90%. Recovery after washout resulted in an increase of AUC to 10% of baseline, and peak tension to 17%

To determine if the negative inotropic effects of TK047 could be further amplified, TNL myometrium (n = 5 patients, 1–3 strips per patient) was pretreated with 10 µM TK047 for 24 h. On day one, prior to TK047 exposure, tissues were mounted and K⁺/OT stimulated in the same manner as the acute and 2 h experiments. Each tissue’s response to OT (AUC/peak force) was first recorded to provide a baseline. Krebs-bicarbonate buffer was then exchanged for HEPES-buffered (20 mM) media (+ P₂/E₄) ± 10 µM TK047, placed in a 37 °C incubator (5% CO₂) for 24 h, then returned to the recording apparatus (tension was maintained throughout experiment), and bath was again exchanged for Krebs-bicarbonate (minus TK047). Tissues were once again challenged with high K⁺, then OT-stimulated (8 nM) for 90-min. Relative to control, in response to OT simulation, the AUC of TK047-treated tissues decreased by 98.17% ± 25.50 (p = 0.0049), while the peak tension decreased by 90.21% ± 32.21 (p = 0.0116). Following washout, AUC returned to 10.10% ± 2.45 of control (p = 0.0098), while peak tension increased to 17.06% ± 4.468 (p = 0.0713), relative to control (Fig. 9B – right panel).

Inotropic effects on myometrium of TK047-treated mice

To determine the effects of TK047 on myometrial contractile dynamics in a whole animal model, non-pregnant C57BL/6 J mice (n = 3) received daily intraperitoneal injections of either TK047 (5 mg/kg/day) or DMSO (50 µl), for four days. Mice were euthanized 1 h after the final dose, and the myometrium was mounted in an organ bath as described above (Fig. 10A). Following a 3-min high-K⁺ challenge, tissues were equilibrated for 30-min. In all control and TK047 treated mice, spontaneous contractions were achieved following high K⁺. For the 15 min prior to the addition of OT (baseline), CPUT in the TK047 treated mice were 132.6% ± 26.28 of control (p = 0.1415), which reached a peak level of 187.0% ± 30.82 (p = 0.0238) for the 15-min period directly following OT-stimulation (Fig. 10B). Despite the increase in CPUT in the TK047-treated mice due to the more sporadic nature of contractions (similar to what was observed in human tissue), AUC was not significantly affected. During the 5-min immediately preceding addition of OT, AUC was 86.01% ± 29.75 (p = 0.6627) of control. AUC was also measured during the 45-min period following OT (8 nM) administration (Fig. 9C). During the first 5-min AUC was 53.25% ± 46.89 (p = 0.3752) of control, from 0–15 min AUC was 94.33% ± 25.87 (p = 0.8373), from 15 to 30-min AUC was 119.70% ± 18.15 (p = 0.3392), and during the 15-min period following washout the AUC was 122.00% ± 21.37 (p = 0.3607) of control tissues.

Fig. 10.

Inotropic effects on myometrium of TK047-treated mice. A C57BL/6 J mice (n = 3 per condition) received daily intraperitoneal injections of either TK047 (5 mg/kg/day) or DMSO for 3 days. Following euthanasia, myometrium were mounted in an ex vivo organ bath then OT-stimulated (8 nM). The myometrium of TK047-treated mice exhibited significantly greater variability, with B CPUT being significantly higher than control following OT; however, C AUC was not significantly different during the 5, 15 or 30 min following OT, nor following washout

Discussion

Considering that approximately 1 of 10 pregnant women experience preterm labor or early delivery [65], developing effective tocolytic strategies to avoid early delivery is of paramount importance. Traditional “one-drug, one-target” monotherapies have failed to reliably prevent early birth, in part because of the unique nature of human uterine smooth muscle [6, 7, 53], and because high doses of most tocolytics cannot be used due to risks to mother and fetus [66]; therefore, we are currently investigating novel therapeutic approaches to address this enigma by simultaneously targeting multiple CAPs with a new class of therapeutic small molecules, called CADAs.

The ability of CADA compounds to prevent the trafficking of various integral proteins to the membrane via the Sec61 secretory pathway presents a unique tocolytic strategy [24, 36–41]. Combination tocolysis has shown promise by reducing the intensity of contractions [19, 67] and by preventing early delivery in an animal model [21], but questions remain surrounding its ultimate efficacy as a tocolytic approach in human patients. CADA compounds present as a unique extension to this method by reducing the trafficking of multiple integral membrane proteins with a single compound. This “one-drug, many-targets” approach may also provide more favorable pharmacokinetics during pregnancy, such as with simplified dosing schemes, and will reduce the number of drugs both mother and fetus are exposed to. TK047 is particularly compelling as it affects both OXTR and Cx43, two critical CAPs.

A evaluation of 20 CADA compounds derived in-house identified SG004, SG007 and TK047 for further analysis based on their ability to significantly downmodulate both OXTR [68] and Cx43 (Fig. 2B). Because there was no existing data on the toxicity of TK047, we conducted cell viability (Fig. 3A) and cytotoxicity (Fig. 3B) assays. Minimal adverse cell morphology and toxicity were only seen at the highest tested concentrations (Fig. 3C), providing us with a reasonable maximum dose to test the efficacy of the compound. We also tested the drug in pregnant and non-pregnant mice with TK047 at 5 mg/kg/day for 3 days during mid-gestation and found that there were no adverse effects on fetal viability, fetal weight, litter size, or maternal liver weight (Fig. 4). We selected this in vivo dose based on prior studies demonstrating that the structurally related CADA analog, VGD040, was pharmacologically active following once-daily intraperitoneal administration at 5 mg/kg, for four consecutive days, in mice without apparent toxicity [68]. This previously tolerated, biologically active dose provides a rational starting point for evaluating TK047 during pregnancy.

With the knowledge that 10 µM TK047 effectively downmodulates our proteins of interest at the plasma membrane, we next sought to determine the optimal concentration for cellular and whole tissue assays. When human uterine smooth muscle cells were exposed to 10 µM TK047 for as little as 4 h, OT-induced calcium influx dropped significantly (Fig. 5A), with similar effects following 48 h of preincubation (Fig. 5B, C). These data indicate that the effect of the drug on OT-modulated calcium influx can be detected within hours of exposure and remain as along as the drug is present. Of note, fluorescence following TK047 treatment is consistent with decreased Ca²⁺ influx; however, the measured signal may also reflect contributions from intracellular Ca release from the sarcoplasmic reticulum, which was not specifically controlled for in this study. As such, we elected to use 10 µM TK047 for all following cell and ex vivo tissue work to avoid potential cytotoxicity seen at higher concentrations.

Because OT-mediated calcium influx is not the sole mediator of muscle depolarization and contraction, we elected to determine if 48 h exposure to 10 µM TK047 affects membrane potential in human uterine smooth muscle cells. Interestingly, not only does TK047 exhibit a decreased peak response to OT (Fig. 6), but the treated cells show more erratic and variable depolarization events (Fig. 6A—insets). Expanding upon this, FRAP experiments in TK047-treated cells identified reduced Cx43 GJC communication, which further explain this observation (Fig. 7). These altered signaling pathways were functionally translated in both human (Fig. 9) and mouse (Fig. 10) whole myometrium causing rapid, uncoordinated depolarization events. This uterine tachysystole may result in less productive contractions during parturition. We hypothesized that this may be a result of TK047 altering calcium reuptake in the sarcoplasmic reticulum, but the use of thapsigargin [69], a sarcoplasmic reticulum Ca²⁺-ATPase inhibitor did not produce similar perturbations in the tissue (data not shown). These findings align with similar observations following dinoprostone (PGE2 agonist) administration, in which “irregular peak heights suggest frequent contractions without global synchronization.” [70] Clinically speaking, the American College of Obstetricians and Gynecologists does not recommend OT administration in women experiencing tachysystole [71], which is supported here in our observation that OT increased asynchronistic contractions in TK047 treated tissue (Figs. 9, 10) [21, 72].

The effects of TK047 appears to be reversible. Following washout after both 2 h and 24 h treatment, tissue contractile dynamics trended towards baseline. While there is limited data on the turnover rate of OXTR, when the receptor is internalized, it can return to the membrane within 4 h [72]. Further, Cx43 has been shown to have an approximate 2 h membrane half-life [73, 74], which might explain its rapid return to baseline levels [63]. Taken together, TK047 is affective at modulating OXTR and Cx43 protein expression and activity at 10 µM from 4 to 48 h, indicating that it may be useful for relatively acute and long term applications.

Limitations

We recognize that no experimental model can fully capture the complexity of the underlying biological system. For this study, the use of nonpregnant mice is a limitation in our 3 day in vivo treatment study (Fig. 10), as myometrial OXTR and Cx43 expression increase approximately 2.1-fold and 6.2-fold, respectively, in late pregnancy [75]; however, both targets are detectable in nonpregnant myometrium, making these animals informative for assessing baseline target modulation for tocolytic potential in pregnant mice. Further, in using both human and mouse tissues, TK047 responsiveness may differ between mouse and human myometrium because Sec61-dependent substrate recognition and downstream protein turnover can be influenced by species- and tissue-specific factors, even though the relevant targeting sequences are highly conserved. Accordingly, the divergent responses observed between human ex vivo and mouse in vivo studies may reflect biological differences in target processing or myometrial physiology rather than exposure alone. Also of note, we did not directly confirm Sec61 pathway interaction by TK047 in myometrial tissue as it was beyond the scope of this study; thus, its proposed mechanism in this system is inferred from the established activity of related SERTIs rather than demonstrated directly. That in mind, the CADA chemotype is well established as a substrate-selective Sec61 inhibitor by structural and functional studies [42, 43], and both GPCR and connexin biogenesis depend on ER membrane insertion pathways, supporting Sec61-dependent translocation inhibition as the most likely mechanism while not excluding contributions from altered trafficking or turnover [76, 77]. In regard to the use of cultured myometrial smooth muscle cells, we recognize that they may undergo phenotypic changes in vitro and not fully recapitulate intact myometrium; however, these cells are known to retain key myometrial markers, including OXTR and Cx43/GJA1 [78, 79] through multiple passages and remain a well-established model for mechanistic studies. We believe that use of matched controls in all experiments supports the validity of the observed relative treatment effects. Finally, the precise pharmacokinetic profile of TK047 are unknown, as are its effects on OXTR and Cx43 expression in other organs, such as the heart and brain, yet maternal and neonatal outcomes were not adversely affected following three days of treatment at 5 mg/kg/day, supporting the tolerability of this dosing regimen.

The dearth of available therapies to halt labor or delay preterm birth amplify the need for novel tocolytic approaches. TK047, and potentially other SERTIs/CADA compounds, represent a first-in-class approach to downmodulating multiple CAPs with a single compound (Fig. 11), providing the proven benefits of combination tocolysis, while minimizing the number of compounds mother and fetus may be exposed to. Given that multiple contraction-associated proteins are upregulated during the transition to labor, SERTI-based therapy may be most effective when initiated prophylactically or early in the course of preterm labor, before maximal accumulation of OXTR, Cx43, and related contractile proteins at the plasma membrane. Nevertheless, because these proteins undergo continual turnover, delayed downmodulation following treatment could still provide therapeutic benefit after labor-associated signaling has begun, although the effective intervention window will require further study. These data provide compelling evidence for the further testing and evaluation of TK047 and other CADA derivatives for their tocolytic potential. Many tocolytics have shown promise cellular and ex vivo models, but few reliably delay birth in an animal model or preterm labor. As such, because TK047 has preliminarily been shown to exhibit no adverse effects on pup size or development (Fig. 4), and future studies will investigate its ability delay birth in a sterile inflammation-based model of preterm labor.

Fig. 11.

Proposed mechanism of TK047-mediated tocolysis. CADA compounds likely downmodulate nascent proteins in the lumen of the endoplasmic reticulum via the Sec61 pathway by preventing their proper insertion into the membrane, followed by ubiquitination and cytosolic degradation. Affected proteins cannot be effectively trafficked to the outer plasma membrane. Here we have shown that TK047 downmodulates Cx43 and OXTR, imparting strong tocolytic effects. Created in BioRender. Barnett, S. (2026) https://BioRender.com/c44p087c44p087

Abbreviations

18αGA

18α-Glycyrrhetinic acid

5-CFDA, AM

5-Carboxyfluorescein diacetate, Acetoxymethyl ester

AUC

Area under the curve

CADA

Cyclotriazadisulfonamide

CAPs

Contractile-associated proteins

CD4

Cluster of differentiation 4

CPUT

Contractions per unit time

Cx43

Connexin 43

d.p.c.

Day post coitum

ΔCai2+

Ca²⁺ uptake

DMEM

Dulbecco’s modified eagle’s medium

EC50

Effective concentration

EMC

ER membrane protein complex

ER

Endoplasmic reticulum

FBS

Fetal bovine serum

FRAP

Fluorescent recovery after photobleaching

Gαq

G-protein coupled receptor

LAS

The leica application suite

LDH

Lactate dehydrogenase

OT

Oxytocin

OXTR

Oxytocin receptor

PD1

Post-delivery on day 1

PFA

Paraformaldehyde

phUSMC

Pregnant human uterine smooth muscle cells

PM

Plasma membrane

ROI

Region of interest

SP

Signal peptide

TF

Terminal fluorescence

TMD

Transmembrane domain

TRITC

Tetramethylrhodamine

XTT

2,3-Bis(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide sodium salt

Ψ

Membrane potential

Authors' contributions

SDB—Conceptualization, Methodology, Investigation, Formal analysis, Writing (original draft), Writing (review & editing), Supervision, Funding acquisition. IR—Methodology, Investigation, Formal analysis, Writing (review & editing), BB—Methodology, Investigation, Formal analysis, Writing (review & editing). TPK—Methodology. Writing (review & editing). SG—Methodology. Writing (review & editing). CW—Project administration, Investigation, Writing (review & editing). AI—Supervision, Writing (review & editing). ILOB—Conceptualization, Writing (review & editing), Supervision, Funding acquisition. TWB—Conceptualization, Methodology, Resources, Writing (original draft), Writing (review & editing), Funding acquisition.

Funding

Research supported by NIH R21 HD112011-01A1 (SDB) NIH R01 HDHD091114 (ILOB) NIH R21 HD108728 to (ILOB), and NIH R21 AI154254-01 (TWB).

Data availability

Data and materials related to this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

All human tissue collection was obtained in accordance with the Declaration of Helsinki and approved by the Institutional Review Board at the University of Nevada Biomedical Review Committee for the protection of human subjects (IRB #509108-26). All experiments were performed in accordance with the NIH guidance on the use of human tissues in research. All mouse work was approved by the University of Nevada, Reno, Institutional Animal Care and Use Committee.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.