Screening a small molecule library to identify inhibitors of NF-κB inducing kinase and pro-labor genes in human placenta

Abstract

The non-canonical NF-κB signaling (RelB/p52) pathway drives pro-labor genes in the human placenta, including corticotropin-releasing hormone (CRH) and cyclooxygenase-2 (COX-2), making this a potential therapeutic target to delay onset of labor. Here we sought to identify small molecule compounds from a pre-existing chemical library of orally active drugs that can inhibit this NF-κB signaling, and in turn, human placental CRH and COX-2 production. We used a cell-based assay coupled with a dual-luciferase reporter system to perform an in vitro screening of a small molecule library of 1,120 compounds for inhibition of the non-canonical NF-κB pathway. Cell toxicity studies and drug efflux transport MRP1 assays were used to further characterize the lead compounds. We have found that 14 drugs have selective inhibitory activity against lymphotoxin beta complex-induced activation of RelB/p52 in HEK293T cells, several of which also inhibited expression of CRH and COX-2 in human term trophoblast. We identified sulfapyridine and propranolol with activity against CRH and COX-2 that deserve further study. These drugs could serve as the basis for development of orally active drugs to affect length of gestation, first in an animal model, and then in clinical trials to prevent preterm birth during human pregnancy.

Conflict of interest statement

The authors declare that they have no competing interests.

Figures

Figure 1

Schematic presentation of strategy of drug screening. (A) Initial screening. HEK293T cells were transiently transfected with a mixture of firefly luciferase (FL) reporter and renilla luciferase (RL) reporter vectors. 48 hours later, the transfectants were exposed to individual compound (drug) for 2 hours, and in turn, lymph toxin- a1b2 (LT-α1β2) for 4 hours prior to dual-luciferase assay. Each chemical compound was run in triplicates and the averages of FL and RL in triplicates were obtained, respectively. With normalization to RL, the individual compound with ≥90% of FL inhibitory effect compared with the vehicle control was considered as a hit for inhibition of the non-canonical NF-κB signaling activity. (B) Secondary screening. HEK293T cells were transiently transfected with a mixture of vectors as described in (A). 48 hours later, the transfectants were exposed to individual hits from the initial screening, and in turn, TPA for 4 hours prior to dual-luciferase assay. Each chemical compound was run in triplicates and the averages of FL and RL in triplicates were obtained, respectively. With normalization to RL, the individual compound with ≤10% of FL inhibitory effect compared with the vehicle control was considered as the specific inhibitor for the non-canonical NF-κB signaling activity. (C) Screening of kinase inhibitors. HEK293T cells were transiently transfected with a mammalian expression vector for ectopic expression of an individual kinase (K), which was followed by a mixture of vectors as described in (A). Then the transfectants were treated with LT-α1β2 (for NIK and IKKα) or TPA (for IKKβ). Each experiment was run in triplicates. With normalization to RL activity, the individual compound with de-repressed FL activity compared with the empty vector control (p < 0.05) was considered as a candidate for inhibition of this kinase. (D) Representative gel pattern and analysis were obtained from 3 independent term placentas by Western blotting with abbreviations described in Table 1. The cropped gels from different Western blot analysis were divided by white space with full-length blots shown in Supplementary Fig. 1. Western blot results were quantified with use of ImageJ. The number below each lane representing relative protein levels of CRH or COX-2 treated with the individual drug relative to DMSO with normalization to β-actin was derived from the following formula: (peak area of CRH or COX-2 of each drug)/(peak area of β-actin of the same lane) of/(peak area of CRH or COX-2 of DMSO)/(peak area of β-actin of the same lane) × 100%. The relative abundance of CRH or COX-2 in the lane of DMSO was considered 100%.

Figure 2

Assays of efflux transporter MRP1 and cell viability. (A) This ATPase assay was performed in the basic way as detailed in Materials and Methods. The bars indicate the average of OD (595 nm) against each individual hit with error bars representing the standard deviation from experiments performed in 3 independent term placentas. *p < 0.01. (B) Primary term cytotrophoblast were treated with the individual compound and cell viability was measured with use of the CellTiter-Blue cell viability assay kit (Promega). The bars indicate the average ration of OD 560/590 (nm) against each individual hit with error bars representing the standard deviation from experiments performed in 3 independent term placentas. *p < 0.01.

Figure 3

Co-localization of NIK and sulfapyridine or propranolol in HEK293 cell. (A) HEK293 cells were treated with Alexa Fluor 647 cadaverine (AlexaFluor-647) or Alexa Fluor 647 cadaverine-labeled sulfapyridine (AlexaFluor-647-Sul) for 24 hr, followed by IF staining. (B) HEK293 cells were treated with AlexaFluor-647 or Alexa Fluor 647 cadaverine-labeled propranolol (AlexaFluor-647-Pro) for 24 hr, followed by IF staining. Each experiment was repeated three times. Arrows indicate co-localization.

Figure 4

Dose-dependent effects of sulfapyridine or propranolol on COX-2 and CRH in primary cytotrophoblast. Human primary cytotrophoblast were treated with sulfapyridine (A) or propranolol (B) at concentrations as indicated for 24 hrs. The total cell lysates were subjected to Western blot analysis (N = 3 individual experiments). The full-length blots are shown in Supplementary Fig. 3. Western blot results were quantified with use of ImageJ (right panels). The relative protein levels of CRH or COX-2 treated with each drug at different concentrations relative to non-treatment (0) with normalization to β-actin were derived from the following formula: (peak area of CRH or COX-2 of each concentration)/(peak area of β-actin of the same lane) of/(peak area of CRH or COX-2 of non-treatment)/(peak area of β-actin of the same lane) × 100%. The relative abundance of CRH or COX-2 in the lane of non-treatment was considered 100%. *p < 0.01, **p < 0.05.