High-Throughput Screening and Initial SAR Studies Identify a Novel Submicromolar Potent Human cGAS Inhibitor

Abstract

Cyclic GMP-AMP synthase (cGAS) has emerged as a promising therapeutic target of several human diseases, including Alzheimer’s disease (AD) and related disorders. As a cytosolic DNA sensor, cGAS generates an innate immune response to promote neuroinflammation by producing an endogenous agonist of the stimulator of interferon genes (STING), 2′3′-cyclic GMP-AMP (cGAMP), which activates the cGAS–STING pathway. We have performed a high-throughput screening of a chemical library containing over 300 K small molecules at the Fisher Drug Discovery Resource Center (DDRC), Rockefeller University (RU), to identify multiple hit inhibitors of human (h)-cGAS. We used a modified Kinase Glo Luminescent Kinase assay, which was earlier developed at RU and later of the library using h-cGAS. The hit candidates bearing novel scaffolds are structurally diverse and exhibited in vitro activity in the low micromolar range. RU-0610270, a sulfonamide derivative, is one of the most potent hits (IC₅₀ = 1.88 μM), selected for hit expansion and structure–activity relationship (SAR) analysis. We synthesized RU-0610270 (listed as cpd 1) and new analogs and evaluated them in vitro against h-cGAS to identify cpd 6 (IC₅₀ = 0.66 μM) as the most potent hit analog. We further profiled cpd 6 and found that it modestly inhibited cGAMP levels by 29% at 30 μM in THP1 cells without detectable toxicity and by 76% at 100 μM, albeit with a moderate decrease (~20%) in cell viability. These results highlight a novel chemical series with promising in vitro activity, providing a starting point for the development of selective and potent human cGAS inhibitors for clinical use.

Graphical Abstract

INTRODUCTION

Cyclic GMP-AMP synthase (cGAS), a cytosolic double-stranded DNA (ds-DNA) sensor enzyme, has emerged as a promising therapeutic target for numerous human diseases, including autoimmune diseases, Alzheimer’s disease (AD), and other neurodegenerative diseases (NDs).^1-7 Upon its stabilization by ds-DNA, the 2:2 cGAS:dsDNA dimer catalyzes the condensation of ATP with GTP to produce 2′3′-cyclic GMP-AMP, cGAMP, which is a low nanomolar potent agonist of stimulator of interferon genes (STING) (Figure 1).⁸ This activates STING and the STING-dependent neuroinflammatory responses characterized by elevated TBK1 phosphorylation, IRF3 activation, and various chemokines and cytokines, such as IP10 (CXCL10) and type I interferons (IFN-I).^2,5 While activation of the cGAS-STING pathway through STING agonists can have positive antiviral effects and may sensitize tumors to immunotherapy, chronic activation of the pathway leads to sustained expression of interferon-inducible genes, which perpetuates inflammation and tissue damage.¹ In AD pathology, amyloid-β (Aβ) plaques and hyperphosphorylated Tau (pTau) induce oxidative mitochondrial stress and damage, causing the release of mitochondrial DNA (mtDNA) into the cytoplasm and activating STING similarly, as described above.^5,6 Chronic activation of this cGAS-STING-IFN pathway drives microglial dysfunction, promotes neurotoxic astrocyte activation, and accelerates synaptic loss and cognitive decline.^1,4,7

Figure 1.

cGAS activation regulates innate immunity and the cGAS-STING-TBK1 pathway and the downstream signaling, which could be blocked using cGAS (1) or STING (2) inhibitors (LBD: ligand binding domain).

Recent studies by the laboratories of Li Gan and others working with both post-mortem human tissues from AD patients and transgenic (Tg) mouse models of tauopathy, including those partially or fully devoid of the cGAS gene or those carrying additional risk or protective genes, revealed numerous facts. First, it was demonstrated that cGAS is activated in both human AD brains and Tau P301S Tg (PS19) mouse brains.⁴ Second, both partial and full genetic ablation of cGAS and pharmacological inhibition using small-molecule inhibitors of cGAS, e.g., RU.521 or TDI-6570 for mouse (m-cGAS) reduce Aβ pathology, suppress neuroinflammation, and rescue memory deficits in PS19 mice.^4-6 Similarly, STING inhibition using H-151 reduces Aβ pathology, besides suppressing neuroinflammation and rescuing memory deficits in 5xFAD mice.^2,3 Third, Gan lab has shown that the risk factors, such as the APOE ε4 allele,⁹ the TREM2 R47H variant,^10-12 and female sex, promote microglial cGAS activation, leading to cellular senescence and neurodegeneration in tauopathy models, while the resilience allele, such as the Christchurch (CC) APOE3/R136S¹³ variant, provides resilience against tau pathology by suppressing cGAS–STING–IFN signaling.

The information described above and the promise that the pharmacological inhibition of cGAS could treat AD and other NDs prompted us to discover novel h-cGAS inhibitors (cGASI). Toward this goal, we optimized the previously described cell-free cGAS activity assay and performed high-throughput screening (HTS) of a chemical library containing 300,000 small molecules at the Rockefeller University (RU), followed by hit validation and optimization of one of the identified potent h-cGASI hits. Here, we report that the optimized screening method conveniently screened the 300 K small molecule library, identifying 38 hit inhibitors of h-cGAS in a relatively short period vis-a-vis the previous HTS campaign.¹⁴ Three distinct structural families (F), e.g., F1 (bicyclic cpds), F2 (sulfonamides), and F3 (thiadiazole cpds), showed low micromolar IC₅₀ (0.52 to 3.35 μM) in a cell-free assay. One of the sulfonamide derivatives, cpd 6, obtained by medicinal chemistry of the most potent hit, RU-0610270, which showed a slight improvement in its activity (IC₅₀ = 1.14 μM) against h-cGAS, and inhibited cGAMP production modestly in THP1 monocytes.

RESULTS AND DISCUSSION

Assay Development for cGAS High-Throughput Screening.

The cGAS enzyme catalyzes the reaction of GTP and ATP to produce cGAMP and two equivalents of pyrophosphate (Scheme 1), providing at least three different methods to assess the activity of a cGAS inhibitor under cell-free conditions.⁶ These can be characterized as (i) ATP-depletion assay, (ii) cGAMP production assay, and (iii) pyrophosphate (PPi) production assay. In the ATP depletion assay, cGAS-catalysis is determined by measuring the amount of unreacted ATP in the reaction using a luminescence-based Kinase-Glo kit.¹⁵ Measuring the ATP levels after approximately 50% conversion provides consistent results. On the other hand, in both the cGAMP and pyrophosphate production assays, cGAS catalysis is determined by measuring the levels of cGAMP and pyrophosphate, respectively. The cGAMP level can be determined using the Bell Brook cGAS/cGAMP Transcreener kits, cGAMP ELISA, or the RapidFire-MS,¹⁶ and the pyrophosphate levels can be measured using inorganic pyrophosphatase-coupled methods. All these methods have been previously used in HTS campaigns;⁶ we found that the ATP-depletion assay is simple and well-suited for large screens. Additionally, we used the cGAS/cGAMP Transcreener for performing a confirmatory assay in cell-free conditions and the cGAMP ELISA to determine cGAS activity in cell-based assays.^17,18

Scheme 1.

cGAS Catalyzes the Reaction of GTP and ATP to Produce cGAMP. Key: dsDNA, Double-Stranded DNA

Optimization of the ATP Depletion Assay for HTS.

We used the previously described HTS method, developed by Lama et al.,¹⁴ and optimized it further to increase the throughput and reduce the time required for screening (see the Supporting Information (SI) for the experimental details). The key difference between the two methods, as described in Table 1, includes the concentration of NaCl that greatly affected cGAS activity, measured by determining unconsumed ATP concentration in the cGAS-catalyzed conversion of ATP + GTP → cGAMP using the “Kinase-Glo Max Luminescent kinase assay kit” in 384-well plates. We evaluated three buffers: B1 contained 5 mM MgCl₂ and 1 μM ZnCl₂ in Tris–HCl buffer with 150 mM NaCl, B2 contained 5 mM MgCl₂ in Tris–HCl buffer with 50 mM NaCl, and B3 contained 2.5 mM MgCl₂ and 0.1 mM MnCl₂ in Tris–HCl buffer with 50 mM NaCl. B2 has recently been used by Skeldon et al.¹⁹ in the cGAS assay. We examined B3 containing MnCl₂ because the latter has been shown to activate cGAS. It was also described that MnCl₂ activated cGAS independently of ds-DNA.²⁰ In addition to the modified buffer, we used HIS-tagged human cGAS (h-cGAS) instead of untagged cGAS and herring testes DNA (ht-DNA) to substitute for ds-DNA. These changes facilitate cGAS generation and purification, as ht-DNA is commercially available at a low cost; otherwise, these were not expected to make much difference.

Table 1.

Comparison of Buffer Compositions for Optimizing cGAS Enzymatic Assays

reagents	buffer 114	buffer 2	buffer 3
Tris–HCl pH 7.4	20 mM	20 mM	20 mM
NaCl	150 mM	50 mM	50 mM
MgCl2	5 mM	5 mM	2.5 mM
ZnCl2	1 μM	-	-
MnCl2	-	-	0.1 mM
DTT	1 mM	1 mM	1 mM
Tween20	0.01%	0.01%	0.01%

We evaluated the activity of HIS-h-cGAS by incubating ATP and GTP (100 μM each) with the enzyme (100 nM) and ds-DNA (25 nM) or ht-DNA (5 μg/mL) in buffers B1–B3 at RT for 3, 5, or 7 h, keeping the ds-DNA (25 nM) and ht-DNA (5 μg/mL) concentration similar to that described in the original condition¹⁴ or from our study performed in 96-well plates. Identical reactions without cGAS or DNA were used as negative controls, and the progress of the reactions was measured using Kinase-Glo, which determined the unconsumed ATP. The results, shown in Figure 2A (in the table) and B (see the Supporting Information Plot for more detail), confirmed that no ATP conversion was observed in the absence of cGAS or DNA, as expected. Meanwhile, ATP levels decreased under all conditions over time. However, cGAS activity was significantly higher in B2 buffer compared to B1, and a 3-h incubation was sufficient to achieve greater than 50% ATP conversion, making it already suitable for the HTS campaign. The cGAS activity increased further in buffer B3 containing MnCl₂ at low concentration (0.1 mM) likely because of a known synergistic effect between dsDNA and MnCl₂ in cGAS activation and likely not independent of DNA as the latter require high MnCl₂ concentration (2 mM).^20,21 Subsequently, evaluation of the h-cGASI, G150,¹⁴ and TDI-8246²² showed that both possess low nanomolar IC₅₀ values as expected (Figure 2C). However, cGAS inhibition was found to be higher in B1, likely due to the low conversion. Based on these results, both B2 and B3 buffers together with a less costly and readily available ht-DNA instead of ds-DNA could be used for the HTS campaign. However, we used B2 buffer to minimize changes in the already established HTS method that had already enabled the identification of numerous low nanomolar potent cGASI hits.¹⁴

Figure 2.

cGAS-catalysis is greatly enhanced at lower NaCl concentration and in the presence of MnCl₂. A summary of buffers and ht-DNA vs dsDNA showing effects on cGAS activity: (A) over 3-, 5-, and 7-h period depicted as a table and (B) at 3-h period depicted in a graph format. (C) IC₅₀ curves for known h-cGASI G150 and TDI-8246 in three buffers using ht-DNA (5 μg/mL conc.) vs dsDNA (25 nM conc.) (n = 3; mean ± S.D.). Buffer composition: B1: 20 mM Tris–HCl pH 7.4, 150 mM NaCl, 5 mM MgCl₂, 1 μM ZnCl₂, 1 mM DTT, and 0.01% Tween-20; B2: 20 mM Tris–HCl pH 7.4, 50 mM NaCl, 5 mM MgCl₂, 1 mM DTT, and 0.01% Tween-20; and B3: 20 mM Tris–HCl pH 7.4, 50 mM NaCl, 2.5 mM MgCl₂, 0.1 mM MnCl₂, 1 mM DTT, and 0.01% Tween-20.

After establishing buffer B2 and selecting ht-DNA as the stimulatory ligand to consistently and effectively activate HIS-h-cGAS in the assay, we determined the Michaelis Mentes parameters, establishing the optimal assay concentration of all the reagents, including cGAS and ht-DNA, required for cGAS catalysis of the ATP + GTP → cGAMP reaction (see the experimental methods are given in the Supporting Information and Figure 3 for details). Briefly, to determine optimal cGAS and ht-DNA concentration, we incubated ATP and GTP (100 μM each) in buffer B2 at RT using ht-DNA (0–20 μg/mL) and h-cGAS (0–400 nM), respectively, in two separate experiments and measured unconsumed ATP using Kinase-Glo at multiple time points. In these experiments, reactions without ht-DNA or cGAS were used as a negative control. We found that ht-DNA efficiently activated HIS-h-cGAS at a concentration of 0.625 μg/mL (Figure 3A) and a K_m of 0.25 μg/mL (Figure 3B), while 50 nM HIS-h-cGAS catalyzed a reaction greater than 50% as measured by ATP conversion (Figure 3C). Michaelis Menten kinetics showed that ht-DNA activated cGAS catalysis with high affinity, K_m, a parameter indicative of the substrate concentration at which the reaction velocity is half of the maximum velocity (V_max/2).²³ We selected a time point from our experiment between 2 and 4 h for the calculation and obtained a K_m = 0.25 μg/mL of ht-DNA titration. For the HTS, we used approximately 10 times the K_m value (2 μg/mL) to ensure the enzyme was nearly saturated with DNA, resulting in a robust assay and reduced sensitivity to substrate fluctuations. Additionally, we performed a DMSO tolerance experiment, which revealed that the luminescence signal was only slightly affected by the concentration of DMSO (Figure 3D). In all these experiments, the negative control conditions (“No DNA” or “No cGAS”) showed no activity, regardless of incubation time, providing us with the opportunity to use any of them as a negative control. We proceed with No DNA as a control besides using a lower DMSO concentration (0.2%) in our future experiments.

Figure 3.

(A) Titration of ht-DNA, normalized against “No DNA condition”. (B) Michaelis–Menten kinetics of ht-DNA. (C) Titration of cGAS, normalized to the “No cGAS condition”. The experiments were performed using ht-DNA (5 μg/mL conc.) and buffer B2. (D) DMSO tolerance assessment. Buffer composition: B2: 20 mM Tris–HCl pH 7.4, 50 mM NaCl, 5 mM MgCl₂, 1 mM DTT, and 0.01% Tween-20 (n = 4; mean ± S.D.).

Pilot Screening Using the LOPAC Library.

To determine and confirm that the screening method developed above is appropriate for a large HTS campaign (Supporting Information Figure S1A), we performed pilot experiments using the LOPAC library, comprising of 1266 diverse, active pharmaceutical compounds on two different days and determined the Z′ value of the results. The Z′ evaluates the statistical reliability and robustness of an experimental design by determining the standard deviations (σ) and means (μ) of both positive and negative control sets (p and n, respectively) and feeding into eq 1.²⁴ Ideally, Z′ values between 1 and 0.5 are preferred, with lower standard deviations indicating uniform controls and a Z′ value closer to one denoting optimal performance. We found that the hit rates from the pilot experiment were low, the hits identified each day were reproducible, and our screening yielded a Z′-prime value of >0.8, providing a high degree of confidence in our screening method (see Figure 4).

Figure 4.

Reproducibility of the LOPAC library analyzed over two different days, expressed as normalized percentage of inhibition (NPI) (R² = 0.97). Assay condition: 100 nM h-cGAS, 2 μg/mL ht-DNA, and 100 μM ATP/GTP in buffer B2, RT, 3 h.

$$Z^{\prime }=1-\frac{3({\sigma }_{\mathrm{p}}+{\sigma }_{\mathrm{n}})}{\mid {\mu }_{\mathrm{p}}-{\mu }_{\mathrm{n}}\mid }$$ (1)

Primary Screening and Secondary Confirmatory Assays.

Subsequently, we performed HTS of a new collection of chemical libraries acquired by DDRC at Rockefeller University. The library contains a collection of 302,000 compounds, obtained from various commercially available libraries (for more details, see the Supporting Information Figures S1-S3 and Tables S1-S5), and that had not been previously screened against human and mouse cGAS.^14,25 We initiated the primary screening using our modified HTS protocol, described above, and setting the concentration of the screening compounds at 10 μM and setting the minimum threshold to 40% inhibition of cGAS, otherwise following the general methods and the Kinase-Glo assay for detection as described previously.¹⁴ We anticipated that a lower concentration (10 μM vs 12.5 μM) and higher cGAS inhibition threshold would eliminate weak hit candidates in the primary screening.

From the initial screening campaign, all compounds exhibiting >40% inhibition of cGAS activity were considered for further evaluation. Several less active compounds bearing similar pharmacophores were also included to facilitate clustering into structural families and gather data for preliminary structure–activity relationship (SAR) analysis. Following visual inspection, scaffold diversity assessment, and prioritization of chemical novelty, a total of 51 hit compounds were selected for cherrypicking and follow-up studies. These hits were divided into 6 families, 5 of which present unique pharmacophores in full or part with 4–9 members, while family 6 contains 10 hits each with a unique pharmacophore (F1–F6, Supporting Information Figure S2 and Table S1). We re-evaluated all 51 hits using both Kinase Glo assay as well as BellBrooks’s Tanscreener cGAMP cGAS TR-FRET assay kits.²⁶ The latter determines cGAMP levels in the reaction mixture, unlike the Kinase Glo assay, which measures only the consumption of ATP, thereby providing secondary proof and confirmation for the identified hits. From this selection, 12 top hits were chosen as confirmed primary hits for further evaluation and analysis.

Representative Compounds from HTS.

Next, all 12 hits were assessed for their identities and purities by subjecting them to LC–MS analysis and determining the enzyme inhibitory constant (IC₅₀) using both the kinase Glo and the Transcreener assays, as described above. We found that all 12 hits have a purity of greater than 90% and possess the desired mass, and only 6 hits showed a measurable IC₅₀ below 16.67 μM, the maximum concentration used in this experiment. The results of the LC–MS analysis and IC₅₀ determination are provided in Supporting Information Table S2. Finally, we performed a DNA intercalation assay to identify and exclude hits that exhibited DNA intercalation (Supporting Information Table S3). Such hits may or may not target the enzyme’s catalytic site, but the DNA intercalation can cause toxicity. We found that all selected hits did not or minimally intercalate to DNA. We prioritized the top two hits RU-0350627 (JA-1, IC₅₀ = 0.52 μM) and RU-0610270 (JA-2, IC₅₀ = 1.88 μM) belonging to families F1 and F2 (Figure 5). Both hits possess drug-like properties and are readily available synthetically. Interestingly, we have also identified JA-1 as a potent h-cGAS inhibitor hit independently using the virtual HTS screening strategy and screening of the virtual primary hits, including numerous analogs of JA-1. A comprehensive study of JA-1 is currently underway and will be reported in a separate manuscript from our group. In this study, we focused on JA-2 as an h-cGAS inhibitor hit bearing a novel sulfonamide scaffold.

Figure 5.

Top h-cGAS inhibitor hits identified from HTS of 300 K compounds. Shown are the structure of the most potent hits from two families (F1 and F2) and analogs present in the HTS compound collection.

With the hit compound JA-2 selected for further evaluation, we first synthesized it using commercially available starting materials and confirmed its activity in a cell-free assay. As the sulfonamide scaffold in JA-2 presents a new opportunity to develop cGAS inhibitors and explore their fit within the catalytic binding site of the enzyme, we decided to gather primary SAR information by evaluating the commercially available analogs. For this, we obtained 149 analogs from the ENAMINE collection and screened in vitro using the Kinase Glo assay. Surprisingly, none of the commercially available analogs, which differ considerably from JA-2, demonstrated better activity than the original hit (data not shown). Therefore, we designed and prepared 49 new JA-2 analogs that retain the sulfonamide linker and possess new substituents in one or both phenyl rings, 1 and 2 (Scheme 2). Some compounds also lacked the ─CH₂─ spacer between ring 2 and the ─CO– function. All 49 analogs of JA-2 were synthesized using the commercially available sulfonamide (I) and acid (II) derivatives as starting materials, as outlined in Scheme 2.

Scheme 2.

Sulfonamide Derivatives Design and Retrosynthesis

We assessed the in vitro activity of the resulting JA-2 analogs using Kinase glo assay, as described previously. Initially, we evaluated all 49 compounds by incubating with 100 μM ATP and GTP, 100 nM h-cGAS, and 2 μg/mL ht-DNA at RT, as used in the HTS campaign. The IC₅₀ for synthetic JA-2 (cpd 1) and top analogs is provided in Figure 6 and Table 2. Thus, we identified five active compounds with IC₅₀ values between 1.1 and 5.8 μM and two with IC₅₀ values of 24.5 and 48.6 μM. We also tested the activity of these seven analogs by performing experiments using a lower cGAS concentration (25 nM) at physiological temperature, 37 °C. We found no significant difference in the IC₅₀ of these compounds.

Figure 6.

IC₅₀ plots selected sulfonamide analogs. Kinase Glo (left, 37 °C and 25 nM enzyme) and (right, 25 °C and 100 nM enzyme, n = 3; mean ± S.D.). Corresponding IC₅₀ values are listed in Table 2.

Table 2.

IC₅₀ Data for Selected Sulfonamide Derivatives in Kinase Glo Under Different Experimental Conditions

Cpd	structure 1R1	kinase Glo IC50 (μM)a	kinase Glo IC50 (μM)b
1	4-Me	4.298	3.417
2	4-CF3	24.50	31.10
4	4-Cl	2.919	3.285
5	4-Br	3.724	4.423
6	4-I	1.143	0.666
9	4-OPh	48.60	58.41
10	4-SMe	5.819	5.593

^aExperimental condition: 25 nM enzyme, 37 °C.

^bExperimental condition: 100 nM enzyme, 25 °C.

SAR Analysis of the Compounds.

Initially, we modified the substituents on ring 2, keeping ring 1 unchanged in compounds (cpds) 1–24 and assessed their effects on activity (Supporting Information Table S5). We found that an ortho –CH₃ group, as present in cpd 13, exhibited reduced activity, while substituting halogens in the para position revealed the following activity trend: I (cpd 6) > Br (cpd 5) > Cl (cpd 4) > F (cpd 3) (see Table 2). Several factors can explain this trend: (i) Iodine’s large size and high polarizability enhance van der Waals interactions, increasing activity, whereas fluorine, with low polarizability, contributes weakly. (ii) Although fluorine’s high electronegativity reduces phenyl ring electron density, polarizability and interaction potential often outweigh this effect. (iii) Iodine’s higher hydrophobicity improves binding within hydrophobic pockets. (iv) A larger size of iodine is likely needed to create a stronger van der Waals interaction of the molecule in the protein active site. (v) Substituent position is critical, as evidenced by reduced activity for ortho-iodine (cpd 16) compared to para-iodine (cpd 6), highlighting the importance of steric compatibility with the binding pocket.^27-31

We further investigated modifications in ring 2 by focusing on para substituents ─CF₃ (cpd 2), ─CH₃ (cpd 1), and ─SCH₃ (cpd 10). Notably, the activity followed the trend ─CH₃ > ─SCH₃ > ─CF₃ (see Supporting Information Table S5), which could be influenced by steric and electronic effects as well as their lipophilicity. Presumably, the small ─CH₃ substituent in cpd 1 minimizes its steric hindrance and fits well in the binding pocket. In contrast, the bulkier ─CF₃ and ─SCH₃ substituents in cpds 2 and 10 could cause potential steric clashes and the substituent-induced conformational changes, thereby disrupting ligand–target interactions and reducing their activities.^32,33 Alternatively, a moderate lipophilicity and electron-donating property of the ─CH₃ group could be optimal for supporting both hydrophobic and electronic interactions, unlike the most lipophilic and electronegative substituent ─CF₃ among the three substituents present in cpds 1, 2, and 10.^32,33

Subsequently, we modified ring 1 by altering its substituents, positions, or aromatic ring type while keeping the rest of the molecule constant. Surprisingly, none of the compounds exhibited higher activity than cpd 1 (see Supporting Information Table S5). For instance, replacing ─Br with hydrogen (cpd 26) led to a loss of activity, as did changing ─OCH₂CH₃ to ─OCH₃ (cpd 27) or removing all substituents on ring 1 (cpd 25). Similarly, para substituents ─OCF₃ in cpd 28, ─CCH₃ in cpd 29, and ─CF₃ in cpd 30 showed negligible activity compared to cpd 1, which features ─Br and ─OCH₂CH₃. Substituting phenyl ring 1 with benzo[d]thiazole (cpd 47) or quinoline (cpd 48) or phenyl ring 2 with furan (cpd 41) also resulted in inactive analogs. These findings highlight the importance of the ─Br and ─OCH₂CH₃ substituents in ring 1 as well as the phenyl scaffold for maintaining activity. Attempts to modify the substituents, alter their positions, or replace the aromatic ring consistently resulted in reduced or abolished activity compared to that of cpd 1.

Cell-Based Assay.

Based on the cell-free activity of all compounds, we chose and assessed the inhibitory effects of cpd 6 on h-cGAS activity in a cell-based assay using THP1 Dual cells, a human monocytic reporter cell line engineered to simultaneously monitor NF-κB and IFN pathway activation via secreted luciferase and alkaline phosphatase reporters, respectively, as described previously.¹⁴ Briefly, cells plated in 96-well plates were incubated with cpd 6 at 10, 30, and 100 μM concentration and ht-DNA/lipofectamine for 16 h, and their effects on IFN and NF-kB expression were determined by measuring the luciferase and phosphatase activity in supernatants and on cell viability by measuring the ATP production from cell pellets. We found that cpd 6 reduced IFN levels by 76%, but it did not affect NF-kB levels at 100 μM and showed some (~20%) toxicity (Figure 7A-C) at this concentration. Furthermore, the compound reduced cGAMP levels by 72% at a 100 μM concentration and 27% at a 30 μM concentration in THP1 cell lysates (Figure 7D), as well as the downstream signals for the IP10 (CXCL10) level by 71% at 100 μM as measured in cell supernatants (Figure 7E).

Figure 7.

Effects of cpd 6 in THP1-dual cells. Cells incubated with cpd 6 in presence of ht-DNA inhibits IFN (A), cGAMP (D) and IP10 (E), but no effects on NF-kB (B) and cell viability (C). DMSO alone and ds-DNA treated cells were included as controls. (n = 3; mean ± SEM).

Finally, we evaluated the permeability of cpd 6 using the MCDKII-MDR1 cells, which not only estimate the cell and brain permeability of a compound but also measure efflux due to P-glycoprotein (Pgp). The results, summarized in Table 3, indicate that cpd 6 exhibits a high cell permeability and low efflux, supporting its suitability for further optimization. This data also confirms that the low potency in cellular assays is not due to issues with cellular permeability.

Table 3.

Permeability Results of cpd 6 and Control Compounds in the MDCKII-MDR1 Cell Line

compound ID	Papp (A-B) (10−6, cm/s)	Papp (B-A) (10−6, cm/s)	efflux ratio	recovery (%) A-B	recovery (%) B-A
metoprolol	28.56	27.96	0.98	109.14	97.41
digoxin	0.46	11.12	24.89	95.47	101.32
Cpd 6	17.30	23.38	1.35	95.95	97.54

CONCLUSION

In summary, we found sulfonamide as a new scaffold in our HTS targeting human cGAS. The selected family allowed us to explore new modifications and their effects on observed activity. We successfully obtained compounds with submicromolar to low micromolar activity. Evaluation using THP1 dual cells showed that the top analog, cpd 6, inhibited both the cGAMP production as well as IP10 and IFN levels, albeit at a high concentration. Further analysis of the SAR generated in this study is in progress and will be used with other ongoing studies to develop more potent cGAS inhibitors for measuring the effects on other downstream signals and the proteome-wide reactivity studies.

EXPERIMENTAL SECTION

Reagents and Materials.

All reagents are of the highest purity commercially available. ATP, GTP, and ht-DNA were purchased from Sigma-Aldrich. The cGAS enzyme was custom-made by Genscript, human cGAS (type: Recombinant enzyme; lot number: U6746IA020-3/P3IB001; expression host: E. coli; purification: Protein was obtained from supernatant of cell lysate, Ni column + superdex 200; concentration: 1.11 mg·mL⁻¹ (by Bradford); purity: ≥ 90% (SDS-PAGE under reducing condition); mass: Confirmed by LC–MS; storage Buffer: 50 mM Tris–HCl, 500 mM NaCl, 10% Glycerol, pH 8.0).

cGAS Protein Length = 528 amino acids Sequence: MHHHHHHQPWHGKAMQRASEAGATAPKASARNARGAPMDPTESPAAPEAALPKAGKFGPARKSGSRQKKSAPDTQERPPVRATGARAKKAPQRAQDTQPSDATSAPGAEGLEPPAAREPALSRAGSCRQRGARCSTKPRPPPGPWDVPSPGLPVSAPILVRRDAAPGASKLRAVLEKLKLSRDDISTAAGMVKGVVDHLLLRLKCDSAFRGVGLLNTGSYYEHVKISAPNEFDVMFKLEVPRIQLEEYSNTRAYYFVKFKRNPKENPLSQFLEGEILSASKMLSKFRKIIKEEINDIKDTDVIMKRKRGGSPAVTLLISEKISVDITLALESKSSWPASTQEGLRIQNWLSAKVRKQLRLKPFYLVPKHAKEGNGFQEETWRLSFSHIEKEILNNHGKSKTCCENKEEKCCRKDCLKLMKYLLEQLKERFKDKKHLDKFSSYHVK-TAFFHVCTQNPQDSQWDRKDLGLCFDNCVTYFLQCLRTEKLENYFIPEFNLFSSNLIDKRSKEFLTKQIEYERNNEFPVFDEF.

Buffer Compositions.

B1: 20 mM Tris–HCl pH 7.4, 150 mM NaCl, 5 mM MgCl₂, 1 μM ZnCl₂, 1 mM DTT and 0.01% Tween-20.

B2: 20 mM Tris–HCl pH 7.4, 50 mM NaCl, 5 mM MgCl₂, 1 mM DTT and 0.01% Tween-20.

B3: 20 mM Tris–HCl pH 7.4, 50 mM NaCl, 2.5 mM MgCl₂, 0.1 mM MnCl₂, 1 mM DTT and 0.01% Tween-20.

Kinase-Glo Assay.

The experimental conditions followed Lama et al.¹⁴ with modifications. Briefly, reactions were performed in 384-well plates using the cGAS enzyme, ht-DNA, ATP, and GTP at the indicated concentrations (100 nM, 2 μg/mL, 100 μM, respectively) with compounds tested at 10 μM in B2. After a 3-h incubation, the reactions were stopped with Kinase-Glo Max, and luminescence was measured. Compounds showing greater than 40% inhibition were further tested in dose–response studies. Full details are provided in the Supporting Information.

TR-FRET Assay.

A brief description of the TR-FRET assay is provided here. In summary, the assay detects cGAMP production by displacing a fluorescent tracer, resulting in a decrease in the intensity of the TR-FRET signal. Reactions were set up in 384-well plates with the cGAS enzyme (25 nM), ht-DNA (2 μg/mL), or dsDNA (60 nM), ATP (100 μM), and GTP (100 μM) under specified conditions. Compounds were added at the desired concentrations and incubated; the reactions were then stopped before the TR-FRET signal was read. Calibration curves were performed with cGAMP standards. Complete details, including buffer composition, controls, and modifications, are provided in the Supporting Information.

DNA Intercalator Assay.

We performed a DNA intercalator assay using fluorescence polarization (FP) to evaluate HTS hits and exclude those with intercalation effects following the method described by Lama et al.¹⁴ Briefly, reactions were set up in 384-well plates using dsDNA (200 nM) or ht-DNA (40 μg/mL), Acridine Orange (50 nM), and Mitoxantrone (20 μM, positive control). Compounds were tested at 33.3 μM. After 30 min of incubation at RT, FP was measured with excitation at 485 nm and emission at 530 nm. Full experimental details are provided in the Supporting Information.

Cell Culture.

THP1-Dual cells (InvivoGen) were cultured in RPMI 1640 (Gibco) supplemented with 10% FBS and 1% penicillin/streptomycin. To maintain luciferase expression, 100 μg/mL of zeocin and 10 μg/mL of blasticidin were added to the growth medium every other passage.

Cell-Based IRF-Lucia Luciferase and Cell Viability.

THP1-Dual cells were seeded in 96-well plates and treated with compounds or DMSO control for 4 h. Cells were transfected with ht-DNA/Lipofectamine 2000 overnight. Luciferase activity in supernatants was measured using QUANTI-Luc reagent. % Inhibition was calculated relative to controls. Cell viability was assessed with CellTiter-Glo 2.0. Full details are provided in the Supporting Information.

cGAMP ELISA Assay.

0.25 million THP1-dual cells were treated with serial dilutions of compounds for 4 h, then transfected with ht-DNA/Lipofectamine 2000, and incubated overnight. Cells were lysed in RIPA buffer containing protease and phosphatase inhibitors. cGAMP levels in cell lysates were measured using a 2,3-cGAMP ELISA kit (Arbor) following the manufacturer’s protocol. Full details are provided in the Supporting Information.

Preparation of Final Compounds.

All new analogs were synthesized using the general methods described below and outlined in Scheme 3. The ¹H NMR and LC–MS data of all compounds can be found in the Supporting Information.

Scheme 3.

Synthesis of Structures 1–4 Sulfonamide Analogs

Cpds 1–24 and 39–46.

A mixture of 3-bromo-4-ethoxybenze-nesulfonamide S1 (50 mg, 0.178 mmol), substituted carboxylic acids S2–25 or S40–47 (0.196 mmol), EDC·HCl (0.357 mmol), and DMAP (0.375 mmol) in dichloromethane (2 mL) was stirred at RT for 12 h. The reaction was monitored by TLC and LC–MS. After completion of the reaction, the reaction mixture was diluted with water and acidified to a pH of 1 with the addition of 10% aq. HCl and followed by extracted with dichloromethane. The organic layer was dried over Na₂SO₄, and the organic solvent was evaporated on a rotary evaporator. The crude residue was purified by column chromatography on silica gel to afford the desired products cpds 1–24 and 39–46.

Cpds 25–38.

A mixture of 2-(p-tolyl)acetic acid S2 (30 mg, 0.196 mmol), substituted sulfonamides S26–39 or S48–50 (0.178 mmol), EDC·HCl (0.357 mmol), and DMAP (0.375 mmol) was similarly treated in dichloromethane (2 mL), as above, and worked-up and purified by column chromatography to afford the desired products, cpds 25–38 and 47–49.

Supplementary Material

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acschembio.5c00649.

• Experimental methods, buffer optimization, screening workflows, representative hit structures, cGAMP ELISA, and cytotoxicity assay; activity data: Kinase-Glo, TR-FRET, LC–MS, and DNA intercalation assays; screening library information; and ¹H NMR/LC–MS characterization for sulfonamide derivatives (PDF)