An optimized protocol with a stepwise approach to identify specific nuclear receptor ligands from cultured mammalian cells

Summary Here, we describe an optimized protocol to identify specific nuclear receptor ligands. First, to rule out any compound interference with luciferase activity per se, we describe an in vitro assay assessing potential inhibition or activation of luciferase enzymatic activity. Second, to comply with EMA and FDA guidelines to mitigate drug-drug interactions, we detail assays assessing constitutive androstane receptor (CAR) and pregnane X receptor (PXR) activation ability. Finally, to minimize off-target detection effects, we describe the use of mammalian one- (or two-) hybrid systems. For complete details on the use and execution of this protocol, please refer to Hering et al. (2018).


SUMMARY
Here, we describe an optimized protocol to identify specific nuclear receptor ligands. First, to rule out any compound interference with luciferase activity per se, we describe an in vitro assay assessing potential inhibition or activation of luciferase enzymatic activity. Second, to comply with EMA and FDA guidelines to mitigate drug-drug interactions, we detail assays assessing constitutive androstane receptor (CAR) and pregnane X receptor (PXR) activation ability. Finally, to minimize off-target detection effects, we describe the use of mammalian one-(or two-) hybrid systems. For complete details on the use and execution of this protocol, please refer to Hering et al. (2018).

BEFORE YOU BEGIN
The protocol below describes the specific steps when using the human hepatocellular carcinoma HepG2 cell line as a model system, but has also been applied to IHH (immortalized human hepatocyte) and HEK293 (human embryonic kidney) cell lines. This protocol requires a luminometer system and an epifluorescence microscope.

Preparation of solutions for the in vitro recombinant luciferase assay
Timing: 15 min 1. Luciferase assay solution [LAS] is prepared as described below and the pH is check and adjust to 7.2 if needed. The LAS solution can be stored for 4 weeks at 4 C.
2. Prepare luciferase inhibitor II solution (2 mM in DMSO), aliquot and store protected from light at À20 C (solution can be store up to 1 year at À20 C). The LAS solution can be stored for several weeks at 4 C but must be at 20 C-22 C for the assay.

Preparation of solution and material for the all transcriptional assays
Timing: 2 days 3. Charcoal-dextran stripped fetal calf serum (FCS) (troubleshooting 1) a. Prepare a suspension of 5% activated charcoal in 500 mL ultra-pure water. b. Centrifuge at 4,000 g for 5 min and wash 3 times the pellet with distillated water c. Centrifuge at 4,000 g for 5 min, discard water. d. Add 2.5 g dextran and 500 mL FCS and incubate 16 h at 4 C under agitation. e. Centrifuge at 4,000 g for 20 min at 22 C f. Sterilize the solution by filtering it twice on 0.22 mm filters. g. Inactivate complement proteins by heating the charcoal-dextran stripped FCS (CD-FCS) at 56 C for 45 min. h. Aliquot (50 mL) and store at À20 C 4. Cell growth, transfected cells seeding and cell treatment media (see Tables below for composition). a. Cell growth (CG) medium contains 10% FCS and is used for cell culture, seeding and transfection. b. Transfected cell seeding (TCS) medium contains 10% of CD-FCS and is used to seed transfected cells. c. Cell treatment (CT) medium contains 1% of CD-FCS and is used for drug testing.  (Kuncharoenwirat et al., 2020) and impacts also on the transcriptional activity of some other NRs such as the estrogen receptor (Rajendran et al., 1987).
Preparation of solutions and materials for the GFP-CAR nuclear translocation assay Timing: 2 days 5. Prepare charcoal-dextran stripped fetal calf serum (FCS) and media as described above. 6. Cool down sterile 1 mL pipette tips and 2 mL cryo-tubes for 16 h at À20 C. 7. Thaw Matrigel for 16 h at 4 C. 8. Make 1.2 mL aliquots of Matrigel into cooled 2 mL cryo-tubes using cooled pipette tips and store them at À20 C. 9. Cool down (À20 C) 5 mL pipettes and 12-well plates before adding Matrigel. 10. Prepare 4% paraformaldehyde solution in PBS.
a. Warm 800 mL of PBS at 60 C and add 40 g of paraformaldehyde powder. Some compounds may inhibit luciferase enzyme activity, leading to an artifactual decrease or increase of the bioluminescent reaction in transfected cells. An increased signal detection may be due to the stabilization of the enzyme by the tested compound, causing its accumulation in the cell. Upon addition of a large excess of luciferin to the cell lysate, the inhibitory compound is displaced and this generates a higher than expected bioluminescent signal independently of transcriptional regulation and thus incompatible with RN ligand identification (Thorne et al., 2010). Thus, this first step allows the selection of the appropriate reporter gene assay [Luciferase or Secreted Embryonic Alkaline Phosphatase (SEAP)] for the next steps (CAR/PXR off target evaluation and targeted NR assay). ] or 1 mL 1003 tested drug solution (final concentration from 10 mM to 100 nM for example) iii. 10 mL of 100 mM luciferin solution (prepare in step 4). b. Shake the plate (directly in the luminometer if possible) and collect light signal with a luminometer (acquisition time 1 s).
CRITICAL: Since luciferin is light-sensitive, the tube containing the luciferin solution must be wrapped in an aluminum foil and the solution must be added to wells just prior reading.
If the tested compounds act on luciferase activity, all further experiments must be carried out using the SEAP reporter construct [p(UAS) 4 -tk-SEAP].

CAR/PXR activation assay
Timing: 5 days This step allows to comply with EMA and FDA recommendations (guideline to mitigate drug-drug interaction highly regulated by CAR and PXR) by determining the ligand-dependent activation of the xenobiotic receptors PXR (Pregnane X Receptor, NR1I2) and CAR (Constitutive Androstane Receptor, NR1I3) by the tested drugs (Berthier et al., 2012, Hering et al., 2018. CRITICAL: If more testing conditions are needed, increase the number of wells at day 1 and pool transfected cells before seeding into 96-well plate to have a homogenous transfected cell population. The human isoform hCAR1 is constitutively active in reporter gene assays, thus drugs must be tested alone to detect inverse agonist activity and in combination with the hCAR1 inverse agonist PK11195 in order to blunt its high basal activity and detect agonist activity.

Timing: 4 Days
This step allows the identification of drugs able to activate the xenobiotic nuclear receptor CAR independently of direct ligand-receptor interaction, but rather by triggering its nuclear translocation through signal transduction pathway(s) activation (Koike et al., 2007). It requires an epifluorescence microscope (excitation filter 485 G 20 nm / emission filter 530 G 25 nm).
11. Day 1: Seed 2.5310 6 HepG2 cells in 6 cm-diameter plate in CG medium and incubate for 16 h at 37 C in a humidified atmosphere and 5% CO 2 . 12. Day 2: Transfect cells with the peGFP-hCAR1 expression vector.
a. Prepare the transfection mix (troubleshooting 3): i. DNA mix: dilute 5.5 mg peGFP-hCAR1 expression vector in 250 mL sterile 150 mM NaCl. Vortex gently and spin down briefly using a bench top centrifuge. ii. JetPEI mix: dilute 11 mL JetPEI transfection reagent in 250 mL 150 mM sterile NaCl. Vortex gently and spin down briefly using a bench top centrifuge. iii. Add JetPEI mix to DNA mix, vortex the solution immediately and spin down briefly. iv. Incubate the mix at 20 C-22 C for 20 min. b. Replace medium by 5 mL fresh CG medium. c. Add the transfection mix on the cells and incubate for 24 h at 37 C. d. Pre-cool for 16 h 24-well plate and pipette tip box at À20 C 13. Day 3: Seed transfected cells into phenol red-free Matrigel-coated 24-well plate: ll OPEN ACCESS a. Pre-coat plate with phenol red-free Matrigel: (troubleshooting 5) i. Place pre-cooled 24-well plate, pipette tip box and 10 mL cold (4 C) phenol red-free DMEM on ice. ii. Thaw Matrigel aliquots on ice (2 mL for one full 24-well plate), mix gently and dilute 2 mL in 4 mL cold phenol red-free DMEM. iii. Add 200 mL diluted Matrigel in each well and incubate plate at 20 C-22 C for 1h b. Split transfected cells.
i. Wash cells with pre-warmed PBS (37 C) ii. Add 500 mL trypsin-EDTA solution and incubate for 10 min at 37 C.
iii. Add 5 mL TCS medium, resuspend cells with a Combitipsâ, count cells and dilute in TCS medium to reach 100,000 to 200,000 cells per mL. iv. Remove excess Matrigel and wash carefully wells with pre-warmed PBS. CRITICAL: In contrast to endogenous hepatic CAR, overexpression of eGFP-hCAR1 in cancer cell lines give rise to an artifactual nuclear localization of the exogenous fusion protein (Figure 1 A, B and C). Thereby, in order to study a potential ''compound-dependent'' nuclear translocation, it is mandatory to first reversibly sequestrate the fusion protein into the cytoplasm. To do so, we developed the above described method based on culture plates pre-coated with Matrigel ( Figures 1D-1F). The unliganded-hCAR1 cytosolic-sequestration by the Matrigel is under the control of MEK1/2 kinases activation (for details, see ''troubleshooting 5'' section of this protocol).
Targeted nuclear receptor assay: The NR1D1 mammalian two-hybrid assay as an example

Timing: 5 days
This step allows the identification of potentially specific ligands for the targeted NR. In this protocol, we describe as an example the procedure allowing the identification of REV-ERBa (NR1D1) ligands using a mammalian two-hybrid system (for full description, see Hering et al., 2018). Briefly, REV-ERBa interacts with NCoR1 to repress target gene expression. The two-hybrid system is based on this interaction, thus cells are co-transfected with pGal4-hREV-ERBa and with pVP16-NCoR. The  17. Day 4: Cell treatment is processed in triplicate using CT medium (prepare 400 mL for each condition). a. Prepare 0.1% (v/v) DMSO (negative control). b. Prepare 10 mM GSK 4112 (positive control). c. Prepare each tested drug (final DMSO concentration 0.1% v/v) alone d. Prepare each tested drug in combination with 10 mM GSK 4112 to evaluate the potential antagonist properties of compounds. e. Wash cells with pre-warmed PBS (37 C) and treat cells for 24 h with 100 mL of prepared solutions. 18. Day 5: Determination of reporter activities using a luminometer plate reader (troubleshooting 4): a. For Luciferase assay use the Dual-Gloâ Luciferase Assay System using instructions provided by the manufacturer. b. For SEAP assay use the Great EscAPeä SEAP Chemiluminescence Kit 2.0 using instructions provided by the manufacturer. For signal normalization, use the ''in lab'' protein quantification method.
Note: Mammalian two-hybrid systems should be preferred when the targeted NR possesses a repressive activity, such as REV-ERBs (NR1D1 and NR1D2). However, mammalian one-hybrid system (with only Gal4-NR and UAS-driven reporter vectors) should be chosen when the NR target has a ligand-dependent transactivation activity.

EXPECTED OUTCOMES
Classically, NR ligands (agonists, antagonists, inverse agonists) are identified either through their ability to promote the recruitment of transcription comodulators (coactivator or corepressor) in acellular methods such as FRET, or using cellular models such as reporter gene assays using notably luciferase enzymes. Although amenable to high throughput screening procedures, several pitfalls are frequently encountered but rarely taken into account when resorting to these latter assays. They are the ability of potential ligands to (1) interfere directly with luciferase enzymatic activity; or (2) to activate the nuclear xenobiotic receptors CAR (constitutive androstane receptor) and PXR (pregnane X receptor), which exert pleiotropic effects when activated (Daujat-Chavanieu and Gerbal-Chaloin, 2020).
Here we describe an optimized procedure allowing to by-pass these pitfalls, thus facilitating the identification of more specific NR ligands. This procedure is implemented in 4 steps, first by ruling out any interference of tested compounds with luciferase activity using an in vitro recombinant luciferase assay (Figure 2A). Second, to comply with EMA and FDA recommendations (guideline to mitigate drug-drug interaction highly regulated by CAR and PXR), two sets of assays assessing the ability of compounds to activate CAR and PXR either by direct binding ( Figure 2B) or through activation of signal transduction pathways ( Figure 2C). Finally, we propose to reduce off-target detection effects by using NR targeted mammalian one-(or two-) hybrid systems ( Figure 2D).

QUANTIFICATION AND STATISTICAL ANALYSIS
In vitro luciferase assay The compounds are tested in triplicates, and experiments are run twice. For each molecule, luciferase activity is compared to the control and statistically significant differences are determined using one-way Anova. If the luciferase activity in the presence of 10 mM compound is below 85% of the control, we recommend to process all transcriptional assays using the SEAP reporter system. In a second step, the FC for each compound is compared to the FC induced by the reference molecule. Statistically significant differences are determined using one-way Anova.

GFP-hCAR1 sub-cellular localization
Hoechst staining allows the detection of DNA/nuclei and is used to visualize the nuclear translocation of the GFP-hCAR1 fusion protein. As the transfection efficiency for HepG2 cells is generally low (z15%), transfected cells are spread sometimes unevenly in wells. The ''in batch'' transfection approach allows to uniformly dispatch transfected cells in wells upon splitting. It is mandatory to Figure 2. Illustration of activity profiles obtained at each step of the protocol A.
Step 1 (recombinant luciferase assay) allows to determine if tested compounds are interfering (red bar) or not (green bar) with luciferase activity per se (''A'' type molecules or''B'' type molecules respectively). B-type molecules should be tested with an alternate reporter assay such as the SEAP system. B.
Step 2 (CAR/PXR) will assess the effect of tested drugs on nuclear xenobiotic receptor activities. Molecules able to modulate their activities (red bar) must be excluded from the initial screening list. C.
Step 3 (CAR nuclear translocation) identifies drugs able to modulate the subcellular localization of the constitutively active hCAR1. Active compounds must be discarded. In the last step (step 4, targeted NR [D]), the response of the targeted NR to drugs allows the identification of highly specific hit(s) (red bar). Gray bars are reference compounds in each test.

OPEN ACCESS
thoroughly survey each well and take 2-3 pictures in 3 different zones of each well to analyze at least 15 to 20 GFP-stained cells. When the subcellular localization of the fusion protein in one compartment (nuclear or cytoplasmic) is observed for 80% of all GFP-stained cells, it is considered as representative. As this method is purely qualitative, no quantification is needed. When the GFP signal can be localized in the nucleus, and even if a signal can still be detected in the cytoplasm, the tested compound is considered as able to activate the CAR signaling pathway.

LIMITATIONS
In transcriptional assays, FC induced by the reference molecule can be low (from 31.5 to 2). In this case, the tested molecule is considered as an agonist when the FC is equal (or higher) to the reference FC. If the reference molecule induces a strong response (FC > 10) the tested molecule is considered as a potential agonist when it induces a FC close to 5 (50% of the reference).

Problem 1
The removal of hydrophobic and steroid-like molecules from the serum is mandatory to assay NR transcriptional activity. However, this step is time-consuming (2 days) and due to the multiple steps of the procedure, could be a potential source of bacterial contamination (before you begin step 3).

Potential solution
The ''home made'' CD-FCS serum can be replaced by a commercial charcoal-stripped FCS depleted from steroid-like molecules and endocrine disruptors. A source for such an alternative is provided in the KRT section.

Problem 2
The tested compound is not soluble in DMSO (steps 3 and 13-c-i)

Potential solution
If tested compound is solubilized in a vehicle other than DMSO, the vehicle must also be tested alone (control). The vehicle control (DMSO or else) is essential to provide a formal and quantitative assessment of compounds activity (especially for CAR, PXR and REV-ERBa transcriptional activity assays).

Problem 3
The JetPEI transfection reagent is compatible with several cell lines (HepG2, IHH and HEK293). The transfection efficiency for HepG2 is compatible for the successful completion of all steps of this protocol. It may however turn out to be low (around 15%-20% in our hand) for the GFP-hCAR subcellular localization assay depending on day-to-day technical variations (step 7, 12, and 15).

Potential solution
The HepG2 cell line is known to be hard to transfect. However, several transfection reagents are compatible with this cell line [Lipofectamine (Thermo Fischer Scientific) or Fugen HD (Promega)].
To the best of our knowledge, the transfection reagent does not affect the outcome of the various assays.

Potential solution
Long term storage of plasmids can be detrimental to transfection efficiency. If the problem persists, all plasmids should be re-amplified and purified.

OPEN ACCESS
STAR Protocols 2, 100658, September 17, 2021 Problem 5 The phenol red-free Matrigel pre-coating step is essential to reversibly sequestrate GFP-hCAR1 into the cytoplasm. However, the use of this material is expensive, time-consuming, and requires precooled materials. Matrigel batches quality/efficiency may also vary (step 13).

Potential solution
Matrigel reversible cytoplasmic sequestration of GFP CAR can be blunted by the MEK1/2 kinase inhibitor U0126 at 25 mM ( Figure 3A, Koike et al., 2007). Of note, Matrigel contains EGF (0.5-1.3 ng/mL depending on the batch). As EGF activates the MEK signaling pathway, the Matrigel pre-coating step can be replaced by a 30 min pre-treatment of transfected cells with (human) recombinant EGF (0.05-0.1 ng/mL) followed by a 2 h co-treatment with drugs ( Figure 3B).

RESOURCE AVAILABILITY
Lead contact Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Philippe Lefebvre (Philippe-claude.lefebvre@inserm.fr).

Materials availability
Plasmids generated in this study are available upon request to the lead contact.

Data and code availability
No data or code was generated in this study