The THP-1 (Human Acute Monocytic Leukemia) cell line was obtained from American Type Culture Collection (ATCC TIB-202; Manassas, VA). THP-1 cells, which originate from the blood of a boy with acute monocytic leukemia [11Tsuchiya S, Yamabe M, Yamaguchi Y, Kobayashi Y, Konno T, Tada K. Establishment and characterization of a human acute monocytic leukemia cell line (THP-1) Int J Cancer 1980; 26: 171-6.], were cultured in RPMI 1640 (ATCC) medium supplemented with 10% fetal bovine serum (FBS, Thermo Scientific HyClone, Logan, UT, USA), 50 U/ml penicillin and 50 μg/ml streptomycin (Invitrogen, CA), and 0.05 mM 2-Mercaptoethanol (Invitrogen). The cell cultures were maintained in a 37°C incubator with 5% CO₂ and under a humidified atmosphere.

A library of pharmacologically active compounds (LOPAC), containing 1280 compounds with known pharmacological actives, was purchased from Sigma (St. Louis, MO). Lipopolysaccharides (LPS), dobutamine hydrochloride, ellipticine, histamine dihydrochloride, Bay 11-7085, PD 169316, BTO-1, and IRAK-1/4 Inhibitor I were obtained from Sigma. CCG-2046 was purchased from Tocris Bioscience (Ellisville, MO).

The secreted Human TNF-α in the cell culture medium of THP-1 cells was measured using an assay kit (Cisbio Bioassays, Bedford, MA, USA) based on the HTRF format (Fig. 1). HTRF [12Degorce F, Card A, Soh S, Trinquet E, Knapik GP, Xie B. HTRF: A technology tailored for drug discovery - a review of theoretical aspects and recent applications Curr Chem Genom 2009; 3: 22-32., 13Harbert C, Marshall J, Soh S, Steger K. Development of a HTRF kinase assay for determination of Syk activity Curr Chem Genom 2008; 1: 20-6.] combines standard FRET (fluorescence resonance energy transfer) technology with time-resolved fluorescent measurement, which could eliminate short-lived background fluorescence. For the assay optimization, THP-1 cells suspended in culture medium were dispensed at 2000, 3000 or 4000 cells/5 µL/well in 1536-well white solid plates (Greiner Bio-One North America, Monroe, NC) using a Bioraptr dispensing system (Aurora Discovery, CA, USA). Twenty-three nL of LPS, a known TNF-α inducer [4Martich GD, Danner RL, Ceska M, Suffredini AF. Detection of interleukin 8 and tumor necrosis factor in normal humans after intravenous endotoxin: the effect of antiinflammatory agents J Exp Med 1991; 173: 1021-4.], at final concentration ranging from 0.56 ng/ml to 9.2 µg/ml was transferred to the assay plate by a pintool work station (Kalypsys, San Diego, CA). The assay plates were incubated at 37°C for 5, 17, or 24 hr, followed by addition of 5 µL mixture of cryptate labeled and XL665 labeled anti-TNFα antibodies. After the assay plates were incubated at room temperature (RT) for 3 hr, the plates were measured by a plate reader in a HTRF detection mode.

Fig. (1) Principle of HTRF-based TNF-α assay. The assay was designed using time resolved fluorescence resonance energy transfer (TRFRET) technology. In this assay, two anti-TNF-α antibodies are labeled with europium Cryptate and XL665, respectively. In the absence of TNF-α, there is no FRET between these two labeled antibodies, which produces fluorescent signal at 615 nm (Fig. 1A). When TNF-α is present, these two labeled antibodies are brought into close proximity of each other, which allows FRET to occur between the two fluorophores, europium Cryptate and XL665 (Fig. 1B). This reaction produces a fluorescent signal at 665 nm. The increase of FRET levels is proportional to the concentrations of TNF-α.

The TNF-α antagonist screening was described in Fig. 2. Briefly, THP-1 cells were dispensed at 3000 cells/4 µL/well in 1536-well white solid bottom plates, followed by an addition of 23 nL compound in DMSO soultion from LOPAC library or DMSO only with a pintool work station (Kalypsys). Each compound was screened at 7 final concentrations ranging from 2.9 nM to 46 µM, as the DMSO solutions of compound library was titrated at 1:5 ratio. After a 5 to 10 min incubation, 1 µL LPS in assay medium at 5 µg/mL LPS (final concentration of 1 µg/mL) was dispensed into the assay plates by a Bioraptr dispensing system. The assay plates were incubated for 17 hours at 37°C, followed by addition of 5 µL mixtures of cryptate labeled anti-TNFα antibody and XL665-labeled TNF-α antibody. After the assay plates were incubated at RT for 3 hr, the dual fluorescence emissions of 615 and 665 nm with a 320 nm excitation were measured using an Envision plate reader (Perkin Elmer, Shelton, CT). The results were expressed as the ratio of 665nm/615nm emissions. The signal (ratio) was reciprocally reduced with the increase of secreted TNF-α in the medium. There was no detergent in the assay buffer for this TNF-α assay. Thus, the cells in each well were kept intact and only the secreted TNF-α was detected.

Fig. (2) qHTS protocols for HTRF-based and AlphaLISA-based TNF-α assays. THP-1 cells were dispensed at 3000 cells/well in 1536-well plates. After 23 nL compound or DMSO only was added into the assay plates, 1 µL assay medium with or without 5 µg/mL LPS (final concentration, 1 µg/mL) was added into the assay plates. The assay plates were incubated for 17 hours at 37°C. For HTRF-based assay format, 5 µL anti-TNFα cryptate/XL665 mixtures were added into the assay plates. After the assay plates were incubated at RT for 3 hr, fluorescence intensity (320 nm excitation, 615 and 665 nm emission) was measured using an Envision plate reader (Perkin Elmer). For AlphaLISA-based assay format, 1.5 µL of the mixture containing acceptor beads and anti-TNF-α antibody was added into each well. The assay plates were incubated in the dark at RT for 1 hr, followed by the addition of 1.5 µL of donor bead. After the assay plates were incubated in the dark at RT for another 0.5 to 1 hr, the assay plates were read in the AlphaScreen mode on the Envision plate reader (Perkin Elmer).

After primary screening, the selected actives were cherry-picked and re-tested for the inhibition of LPS induced TNF-α production in 24 point titrations with final concentration ranging from 6 pM to 46 µM in a HTRF human TNF-α assay using the same protocol as described above except that the 24 point titrations were laid out within one 1536-well plate.

Furthermore, the powder samples of eight compounds confirmed from the above assay were re-ordered from commercial sources and tested in the same assay protocol as described above.

An alternative homogenous TNF-α assay based on PerkinElmer AlphaLISA [14Du Y, Khuri FR, Fu H. A homogenous luminescent proximity assay for 14-3-3 interactions with both phosphorylated and non-phosphorylated client peptides Curr Chem Genom 2008; 2: 40-7., 15Eglen RM, Reisine T, Roby P, et al. The use of AlphaScreen technology in HTS: current status Curr Chem Genom 2008; 1: 2-10.] assay format (Fig. 3) was used to confirm the active compounds identified from the primary screen of HTRF based assay. The cell seeding, compound addition and LPS treatment were the same as described above except the detection of TNF-α was using an AlphaLISA-based assay. Briefly, 1.5 µL of the mixture containing acceptor beads conjugated with anti-TNF-α antibody was added into each well after the assay plates were incubated at 37°C for 17 hr. The assay plates were incubated in the dark at RT for 1 hr, followed by the addition of 1.5 µL of donor bead coated with streptavidin which captures the biotinylated anti-TNF-α antibody. After the assay plates were incubated in the dark at RT for another 1 hr, the assay plates were read in the AlphaScreen mode on the Envision plate reader (Perkin Elmer).

Fig. (3) Principle of AlphaLISA-based TNF-α assay. The Alpha donor bead (blue color) is coated with streptavidin which captures the biotinylated anti-TNF-α antibody. The acceptor bead (orange color) is coated with TNF-α specific antibody. When TNF-α is present, the two coated beads are brought into proximity through binding to TNF-α. After excitation by laser at 680 nm, the singlet oxygen released by alpha donor bead travels to the nearby acceptor bead where it causes acceptor bead to emit fluorescence at 615 nm. The increase of AlphaLISA signal is proportional to the concentrations of TNF-α.

Cell viability was measured using a luciferase-coupled ATP quantitation assay (CellTiter-Glo^®, Promega, Madison, WI). THP-1 were dispensed at 3000 cells/well in 1536-well white/solid bottom assay plates using a FRD, followed by addition of compounds via a pintool work station. The assay plates were incubated for 17 hr at 37°C. At the end of the incubation period, 5 μL of CellTiter-Glo^® reagent was added, plates were incubated at RT for 30 minutes, and luminescence intensity determined in the luminescence mode using a ViewLux plate reader (PerkinElmer).

THP-1 cells were plated at the cell density of 4.8 × 10⁴ in 200 µl culture medium per well in a 96-well plate. Twenty-five µL culture medium with or without compound was added into each well, followed by addition of LPS at 1 µg/ml final concentration in culture. The final concentrations of the compounds in the wells ranged from 1.6 nM to 30 µM. After 17 hr treatment at 37°C, the cell culture supernatants were removed and measured for human TNF-α using human TNF-α immunoassay kit (R&D Systems, Minneapolis, MN). Briefly, 200 uL of sample or known standard (0-1000 pg/ml) was added to wells of a microplate which was pre-coated with a monoclonal antibody specific for TNF-α and incubated at RT for 2 hr. After washing away any unbound substances, an enzyme-linked polyclonal anti-TNF-α antibody was added and the plate incubated for 1 hr at RT. Following four washes, a substrate solution was added and incubated for 15-20 min, followed by the addition of a stop solution. The optical density of each well was determined at 450 nm with 570 nm as a reference filter using an EnVision plate reader.

The raw data was normalized to LPS (1 µg/mL, 100%) and assay medium with 0.1% DMSO (basal, 0%). The inhibition curves for each compound were analyzed using the non-linear regression analysis program in GraphPad Prism (Soft-ware).

Data normalization, correction and fitting of concentration response curves were performed as previously described [16Xia M, Huang R, Witt KL, et al. Compound cytotoxicity profiling using quantitative high-throughput screening Environ Health Perspect 2008; 116: 284-91.]. Briefly, raw results for each titration point was first normalized relative to the LPS control (1 µg/ml, 0%) and DMSO only wells (basal, -100%), and then corrected by applying a pattern correction algorithm using compound-free control plates (DMSO plates) to minimize the dispense and reading errors. Concentration-response titration points for each compound were fitted to the Hill equation yielding concentrations of half-maximal inhibition (IC₅₀) and maximal response (efficacy) values. Concentration response curves were classified into four major classes using the set of criteria listed in previous studies [17Inglese J, Auld DS, Jadhav A, et al. Quantitative high-throughput screening: a titration-based approach that efficiently identifies biological activities in large chemical libraries Proc Natl Acad Sci U S A 2006; 103: 11473-8.].

Compounds which showed inhibition in both the ratiometric and 665 nm readings, and had potency less than 5 µM and efficacy greater than 50% in the ratiometric reading were considered as active in the HTRF human TNF-α assay. These compounds were further prioritized based on their activity in the cell viability assay after 17 h compound treatment. Twenty-six active compounds that were not apparently cytotoxic (6 times more potent in the HTRF human TNF-α assay than that in the cell viability assay) were cherry-picked for confirmation and follow up studies.

We have optimized and validated a homogenous HTRF-based TNF-α assay in a 1536-well plate format that can be used to screen compounds to identify potential TNF-α inhibitors (Fig. 1). LPS, a known TNF-α stimulator, induced TNF-α production in a concentration-dependent manner after 17 hr incubation with the THP-1 cells (Fig. 4A). The EC₅₀ of LPS was 0.84 µg/ml, and the maximum induction of TNF-α production by LPS was more than 8-fold of the basal level.

Fig. (4) (A) Time course of LPS-induced TNF-α production. THP-1 cells were treated with various LPS concentrations for 5, 17 and 24 hr. At the end of various time points, TNF-α production was measured in THP-1 cells using a homogenous HTRF-based TNF-α assay. Data are from a single experiment performed in quadruplicate, representative of several experiments. (B) Optimization of cell density. THP-1 cells were dispensed at 2k, 3k and 4k per well. After incubated with LPS for 17 hr, TNF-α production in the cells was measured. Data are from a single experiment performed in quadruplicate.

To determine the optimal incubation time, the cells were incubated with concentration titrations of LPS for 5 to 24 hr. The EC₅₀s of LPS are 2.52, 0.84 and 2.22 µg/ml for 5, 17 and 24 hr incubation, respectively. The signal-to-basal ratio under maximum LPS treatment was highest after 17 hr of incubation (8.7-fold), followed by 24 hr incubation (4.9-fold) and 5 hr incubation (4.6-fold). Therefore, the 17 hr incubation of LPS with THP-1 cells was chosen for compound treatment for the rest of the study (Fig. 4A).

The optimal cell density in a 1536-well plate format was determined by varying cell numbers in a homogenous HTRF TNF-α assay. The EC₅₀s of LPS were 0.70, 0.74 and 0.84 µg/ml for 2000, 3000, and 4000 cells per well, respectively. The signal-to-basal ratios under maximum LPS treatment after 17 hr incubation were 6.9-, 9.1-, and 9.1-fold for the cell density at 2000, 3000, and 4000 cells per well, respectively (Fig. 4B). 3000 and 4000 cells/well yielded slightly higher signal-to-basal ratios than 2000 cells/well. Thus, 3000 cells/well was chosen to prevent possible overcrowding of cells at the higher density of 4000 cells/well.

Thus, this assay is simple and robust and is suitable for high throughput screening of compound libraries. It has been miniaturized into the 1536-well plate format to increase screen throughput and to reduce reagents consumption.

The LOPAC library containing 1280 compounds with known pharmacological activities was screened to evaluate the HTRF TNF-α assay. The inhibitory effect of the compounds on TNF-α production was measured in the presence of 1 µg/ml (EC₈₀ value) of LPS in a quantitative high throughput screening (qHTS) platform. The average signal-to-basal ratio for LPS was 8.4 ± 1.3 (mean ± SD) and the average Z’ factor was 0.8 ± 0.06 for this primary screen, indicating a robust assay for compound library screen.

The compound potencies and efficacies were calculated from the primary results obtained from this LOPAC library screen using the HTRF-based TNF-α assay as it was described in the Methods section. The primary hits were selected as these compounds had IC₅₀ values less than 1 µM and inhibitory efficacy greater than 50% of control response in the HTRF TNF-α assay while they were inactive or 6 fold less potent in the cell viability assay. A total of 26 compounds met these criteria as potential inhibitors of TNF-α production and they were cherry-picked for the further follow-up studies.

The 26 cherry-picked compounds were re-tested in the HTRF-based TNF-α assay. The inhibitory activity of LPS-induced TNF-α production was confirmed in 25 of 26 selected compounds (Table 1) and the confirmation rate of was 96%. To eliminate the potential false positives due to the detection method used in the above assay, these 25 compounds were further evaluated in another homogeneous TNF-α assay that applied the AlphaLISA assay method (Fig. 3). The assay signal in the AlphaLISA assay was measured in the AlphaScreen mode instead of the HTRF detection mode in the HTRF TNF-α assay. The activities of all 25 compounds were confirmed in the AlphaLISA-based TNF-α assay with a concordance rate of 100%. The potency ranking of the compounds in these two assays was very similar. The IC₅₀ values of these compounds in the two assay formats correlated well with an R² of 0.93. Formoterol was the most potent compound among these confirmed compounds. Its IC50 values were 4 nM and 6 nM in both the HTRF-based assay and the AlphaLISA-based TNF-α assay, respectively (Table 1). Among these 25 compounds, 15 were known beta adrenergic receptor agonists including formoterol, fenoterol, alpha-methynorepinephrine, epinephrine, norepinephrine and isoproterenol. The IC₅₀ values of these beta adrenergic receptor agonists ranged from 4 nM (formoterol) to 3.5 µM (ritodrine). PD169316 (0.12 µM), a p38 MAPK kinase inhibitor, and NFκB pathway inhibitors, Bay 11-7082 (1.2 µM) and Bay 11-7085 (0.88 µM), were also found active in the inhibition of LPS induced TNF-α production. The IC₅₀ values of these 25 compounds from both the HTRF-based TNF-α assay and the AlphaLISA-based TNF-α assay are listed in Table 1. These data suggest that the two assays, the HTRF TNF-α assay and AlphaLISA TNF-α assay, are equally sensitive to inhibitors. These homogenous assays can be easily scaled up to screen a large compound library quickly and efficiently. They can also be used to counter-screen each other to eliminate the detection-related false positive compounds.

Table 1

Potency (µM, IC₅₀) and Efficacy (% in Parenthesis) of 25 Cherry-Picked Compounds in HTRF-Based TNF-α Assay, AlphaLISA-Based TNF-α Assay, and Cell Viability Assay

To further validate the TNF-α inhibitors identified from both the homogenous HTRF-based and the AlphaLISA-based TNF-α assays, eight compounds were selected based on potency and diversity of mechanism of action and purchased from commercial chemical vendors. The inhibitory effect of these compounds on LPS induced TNF-α production in THP-1 cells was determined using a traditional ELISA assay that is the gold standard for cytokine measurement. As shown in Table 2, all eight compounds inhibited LPS induced TNF-α production with IC₅₀s ranging from 10 nM to 1.95 µM. It was notable that the potencies of all the compounds tested in the ELISA-based assay were higher than the potencies obtained from the HTRF-based assay, with IC₅₀ shifts ranging from 2- to 7-fold (Table 2). For example, Bay 11-7085 had higher potency in the ELISA-based assay (IC₅₀, 0.293 µM) than in the HTRF-based assay (IC₅₀, 1.43 µM), with a 4.9-fold IC₅₀ shift (Fig. 5A). The IC₅₀ shifts of CCG-2046 (2.32 µM in the HTRF assay, and 0.66 µM in the ELISA assay, Fig. 5B), IRAK-1/4 inhibitor I (2.76 µM in HTRF assay, and 0.52 µM in ELISA assay, Fig. 5C), and PD 169316 (78.2 nM in HTRF assay, and 10.6 nM in ELISA assay, Fig. 5D) were 4-, 5- and 7-fold, respectively. The other four compounds, BTO-1, dobutamine, ellipticine and histamine, had 2 to 3-fold IC₅₀ shifts (Table 2). However, the ranking orders of the compound potencies were comparable between the two assay formats. The IC₅₀ values of these compounds in both the ELISA-based and HTRF-based assays correlated very well with an R² of 0.93. In addition, these eight compounds were also tested in a cell viability assay to evaluate the potential cytotoxicity because the cytotoxic compounds could show the false effect on the reduction of TNF-α production in the cell based assays. We found that all these compounds showed none or weak cytotoxic effect (Table 2). Taken together, these data suggest that the homogenous HTRF-based TNF-α assays can be used to identify small molecule compounds that inhibit TNF-α production.

Fig. (5) Concentration response curves of TNF-α inhibitors in a HTRF-based TNF-α assay (▼) and ELISA-based assay (■). TNF-α level was measured after THP-1 cells were incubated with various concentrations of Bay 11-7085 (A), CCG-2046 (B), IRAK-1/4 inhibitor (C) and PD 169316 (D) in the presence of 1 µg/ml LPS for 17 hr

Table 2

Comparison of Potency (µM, IC₅₀) of Commercial Ordered Compounds in ELISA-Based TNF-α, HTRF-Based TNF-α, and Cell Viability Assays