Pyridazinones and Structurally Related Derivatives with Anti-Inflammatory Activity

Persistent inflammation contributes to a number of diseases; therefore, control of the inflammatory response is an important therapeutic goal. In an effort to identify novel anti-inflammatory compounds, we screened a library of pyridazinones and structurally related derivatives that were used previously to identify N-formyl peptide receptor (FPR) agonists. Screening of the compounds for their ability to inhibit lipopolysaccharide (LPS)-induced nuclear factor κB (NF-κB) transcriptional activity in human THP1-Blue monocytic cells identified 48 compounds with anti-inflammatory activity. Interestingly, 34 compounds were FPR agonists, whereas 14 inhibitors of LPS-induced NF-κB activity were not FPR agonists, indicating that they inhibited different signaling pathways. Further analysis of the most potent inhibitors showed that they also inhibited LPS-induced production of interleukin 6 (IL-6) by human MonoMac-6 monocytic cells, again verifying their anti-inflammatory properties. Structure–activity relationship (SAR) classification models based on atom pair descriptors and physicochemical ADME parameters were developed to achieve better insight into the relationships between chemical structures of the compounds and their biological activities, and we found that there was little correlation between FPR agonist activity and inhibition of LPS-induced NF-κB activity. Indeed, Cmpd43, a well-known pyrazolone-based FPR agonist, as well as FPR1 and FPR2 peptide agonists had no effect on the LPS-induced NF-κB activity in THP1-Blue cells. Thus, some FPR agonists reported to have anti-inflammatory activity may actually mediate their effects through FPR-independent pathways, as it is suggested by our results with this series of compounds. This could explain how treatment with some agonists known to be inflammatory (i.e., FPR1 agonists) could result in anti-inflammatory effects. Further research is clearly needed to define the molecular targets of pyridazinones and structurally related compounds with anti-inflammatory activity and to define their relationships (if any) to FPR signaling events.


Introduction
Inflammation is a biological defensive mechanism in response to harmful stimuli, such as pathogens, damaged cells, or irritants, and it can last for a short period of time or it can result in a chronic condition [1,2]. Persistent inflammation can contribute to a variety of pathologies; therefore, adequate control of the inflammatory response is essential [3]. During the development of inflammation, phagocytes play a major role because they are involved in the host immune defense and are able to produce a variety of bioactive Table 1. Chemical structure and FPR1/FPR2 agonist activity of derivatives used for the evaluation of anti-inflammatory activity in human THP1-Blue cells.
Molecules 2022, 27, x FOR PEER REVIEW 3 of 26 In an effort to identify new anti-inflammatory compounds, we evaluated the antiinflammatory activity of a library of 177 compounds consisting mostly of pyridazinonelike derivatives that were previously evaluated for FPR1/FPR2 agonist activity in order to understand if FPR1/FPR2 agonist activity was related to anti-inflammatory activity. In initial screening assays, the effects of the compounds on lipopolysaccharide (LPS)-induced NF-κB activity in THP1 monocytes/macrophages was evaluated, as NF-κB activation is an important component of inflammatory responses [47,48]. Selected compounds were further evaluated for their effects on LPS-induced interleukin 6 (IL-6) production in monocytic MonoMac-6 cells. Finally, a structure-activity relationship (SAR) classification model based on atom pair descriptors and physicochemical ADME parameters was developed to achieve better insight into the relationships between the chemical structures of the investigated compounds and their biological activity. We identified a number of pyridazinone-like compounds with anti-inflammatory activity and concluded that the pyridazinone scaffold could be useful for the development of anti-inflammatory therapeutics.

Series G: Pyrazoles and Pyrazolones
Compd. Among the 177 compounds tested, 48 compounds inhibited the LPS-induced NF-κB activity in THP1-Blue cells ( Table 2), suggesting that they could potentially be novel anti-inflammatory molecules. The chemical structures of the most potent pyridazinones (compounds 71 and 84) are shown in Figure 1. Representative dose-response curves showing inhibition of the LPS-induced activation of NF-κB activity by these compounds are shown in Figure 2. Interestingly, 34 of these compounds were previously found to be FPR ligands with varying potencies and specificities (Table 1), whereas 14 of these NF-κB pathway inhibitors were shown previously to have no FPR activity, suggesting that they target FPR-independent anti-inflammatory pathways in THP1 cells. Nevertheless, there was not a good correlation between the FPR agonist activity and anti-inflammatory activity, as only 4 of the 17 most potent FPR agonists (EC 50 < 1 µM at FPR1 or FPR2) inhibited the LPS-induced NF-κB activity. Consistent with this conclusion, Cmpd43, a documented pyrazolone-based dual FPR1/2 agonist [40][41][42] and the high-affinity FPR1 and FPR2 peptide agonists (5 nM fMLF and 5 nM WKYMVM, respectively) also did not inhibit the LPS-induced NF-κB activity in THP1-Blue cells. Indeed, we previously found that peptide agonists of FPR1/FPR2 actually enhanced NF-κB activity in THP1-Blue cells by 2-2.5 fold in comparison with a solvent control [57]. Importantly, 46 of these compounds had no cytotoxicity, whereas 66 and 158 had some cytotoxicity (IC 50 = 22-26 µM) ( Table 2).
The agonist activity at FPR1/FPR2 is from Table 1. The reported agonist activity of the dual FPR1/FPR agonist Cmpd43 is from [42]. The FPR agonist activities of fMLF and WKYMVM are reported in [58]. N.T.: No cytotoxicity was observed at the highest concentration tested (50 µM [42]. The FPR agonist activities of fMLF and WKYMVM are reported in [58]. N.T.: No cytotoxicity was observed at the highest concentration tested (50 µM   The pyridazinone scaffold (1-120) was present in 70% of the active compounds in the anti-inflammatory assay, although this could be related to the greater number of molecules tested that had this specific structure. The chiral centers in 94/95, 98/99, and 102/103 affected the activity, even though it was not clear if there was a preferential enantiomer. Additionally, substitution of the chiral carbon led to an active compound only when small groups were present. The N-(4-bromophenyl)-acetamide group was present in almost all of the compounds tested, and all compounds with major modifications of this fragment were inactive, except for 109 and 113. On the other hand, some modifications in this group tended to improve the anti-inflammatory activity, as exemplified by compounds 2, 5, and 15, where either a shift from the para to the meta position or a replacement of the bromine with a nitro group or a fluorine atom improved activity compared to the inactive compound 1. Furthermore, changes in the 3-methoxybenzyl chain, which is present in many of the compounds, improved the anti-inflammatory activity, as seen for 30, where the methoxy group was removed, or for 64, where the same methoxy group was substituted with a fluorine atom. Interestingly, the pyridazinone derivative containing just a methyl group at position 3 (compound 67) was active. The pyridazinone scaffold (1-120) was present in 70% of the active compounds in the anti-inflammatory assay, although this could be related to the greater number of molecules tested that had this specific structure. The chiral centers in 94/95, 98/99, and 102/103 affected the activity, even though it was not clear if there was a preferential enantiomer. Additionally, substitution of the chiral carbon led to an active compound only when small groups were present. The N-(4-bromophenyl)-acetamide group was present in almost all of the compounds tested, and all compounds with major modifications of this fragment were inactive, except for 109 and 113. On the other hand, some modifications in this group tended to improve the anti-inflammatory activity, as exemplified by compounds 2, 5, and 15, where either a shift from the para to the meta position or a replacement of the bromine with a nitro group or a fluorine atom improved activity compared to the inactive compound 1. Furthermore, changes in the 3-methoxybenzyl chain, which is present in many of the compounds, improved the anti-inflammatory activity, as seen for 30, where the methoxy group was removed, or for 64, where the same methoxy group was substituted with a fluorine atom. Interestingly, the pyridazinone derivative containing just a methyl group at position 3 (compound 67) was active.
When a para fluorine was present on the phenyl of the acetamide chain in position 1, a substitution of the methyl group with a para-phenyl-substituted group at position 5 (compounds 42, 46, and 47) led to active compounds. The same result occurred even when there was an N-(4-bromophenyl)-acetamide moiety, as in compound 49 compared to the inactive compound 19. Position 4 of the pyridazinone scaffold was mostly unsubstituted, When a para fluorine was present on the phenyl of the acetamide chain in position 1, a substitution of the methyl group with a para-phenyl-substituted group at position 5 (compounds 42, 46, and 47) led to active compounds. The same result occurred even when there was an N-(4-bromophenyl)-acetamide moiety, as in compound 49 compared to the inactive compound 19. Position 4 of the pyridazinone scaffold was mostly unsubstituted, but the insertion of an increasingly bulkier group led to an improved activity, as shown by compounds 69, 71, 78, 79, and 80, which contain a methyl, ethyl, n-propyl, and n-butyl, respectively. Reduction of the double bond on C3-C4 was not favorable for activity (dihydro derivatives 121-132), nor was the presence of a condensed ring on the pyridazinone scaffold (compounds 174, 175, 176, and 177). Likewise, compounds having an indole (133 and 134) or a pyridazine core (135-138) were inactive, even though the latter might be due to the lack of an N-1 acetamide chain or, moreover, a pyridinone core, with the only exception being compound 141. Considering the small number of pyrimidin-2,6-dione derivatives analyzed (149-154, 6 compounds) and that 4 out of 6 compounds were active, this might also be the most promising scaffold to further investigate for anti-inflammatory compound development. Analysis of the 5-member ring compounds (155-173) showed that a number of these compounds were active and could be potential leads for further development.
Although the molecule in this group having the best IC 50 (compound 158) was cytotoxic, it could also be further investigated for potential modifications that may reduce/eliminate cytotoxic effects.
The most potent anti-inflammatory compounds (based on the inhibition of LPSinduced NF-κB activity in THP1-Blue cells) were also evaluated for their effect on LPSinduced IL-6 production in MonoMac-6 cells, which is also considered to be an inflammatory response ( Table 3). All of these compounds, except 46, were also found to be active in this cell-based assay system and inhibited IL-6 production. Compound 71 again had the highest activity (IC 50 = 2.0 µM), whereas compound 84 had reduced activity (IC 50 = 30.7 µM) compared to its effectiveness in THP1-Blue cells. Representative dosedependent response curves showing inhibition of the LPS-induced activation of IL-6 production by the pyridazinones 71 and 84 are shown in Figure 3. The pyrimidine-2,6-dione derivatives (150 and 153) had an IC 50 < 10 µM and were also potent IL-6 inhibitors, as was compound 169, which has a 5-member ring scaffold. The only compound exhibiting cytotoxicity in MonoMac-6 cells was compound 9. It should be noted that among the 24 non-cytotoxic inhibitors of IL-6 production, 6 structurally related pyridazinones (2, 38, 42, 46, 47, and 83) were not FPR1 or FPR2 agonists (Table 1), again supporting our previous suggestion that FPR agonist activity does not correlate well with anti-inflammatory activity. Indeed, the potent peptide agonists of FPR1 and FPR2 (fMLF and WKYMVM, respectively) also did not affect the LPS-induced IL-6 production by MonoMac-6 cells. Table 3. The effects of selected compounds and FPR peptide agonists on the LPS-induced production of IL-6 in human MonoMac-6 cells, cytotoxicity, and Ca 2+ mobilization in FPR1/FPR2-transfected HL-60 cells.

Classification Tree Analysis
To further evaluate the relationships between the chemical structures of the investigated compounds and their anti-inflammatory activity (i.e., the inhibition of LPS-induced NF-κB activity), we performed analysis based on a binary classification tree approach. The classification tree was developed using ADME parameters and atomic pairs as 2D descriptors. A total of 147 compounds were included in the analysis, as the racemic mixtures along with the individual enantiomers corresponding to these racemates were excluded. The ADME parameters were calculated using SwissADME [59], and the selected parameters found are shown in Supplementary Table S1. The total number of different atom pairs present in at least one compound was 1029, with topological distances between atoms ranging from 1 to 22 chemical bonds. The number of occurrences of an atom pair in a molecule was considered to be the value of the corresponding 2D descriptor [60]. Atom pairs have an important advantage in SAR analysis over many other types of descriptors, since they can be easily related to certain substructures in a molecule [60][61][62]. Therefore, the results can be interpreted in terms of molecular fragments and/or functional groups. The classification tree is based on five descriptors [four atom pairs and one ADME parameter, known as topological polar surface area (tPSA)]. The misclassification matrix in graphical and tabular forms are shown in Figure 4 and Table 4, respectively.
According to the classification tree model, 123 of the 147 compounds evaluated (83.7%) were classified correctly, whereas 11 of 36 active compounds were erroneously classified as inactive (Supplementary Table S2). This may be due to the uneven distribution of compounds within the class, i.e., there were relatively many inactive and few active compounds. According to the logical rules defined by the classification tree, a compound was classified as active if it had at least one of the C3_7_F or C4_7_NO atom pairs or more than eight C4_7_CA atom pairs (terminal nodes 5, 7, and 3, respectively). Otherwise, the compound was recognized as active if it had a tPSA value of not greater than 84.685 Å and, at the same time, contained more than four CA_8_O1 atom pairs (terminal node 11). Examples of compounds containing the mentioned 2D descriptors are shown in Figure 5.  Table 1). N.T.: No cytotoxicity was observed at the highest concentration tested (50 µM). N.A.: No activity was observed at the highest concentration tested (50 µM). b fMLF and WKYMVM were tested at 5 nM.   The atom pairs C4_7_CA were present in molecules containing simultaneously alkyl groups and aromatic fragments at a distance of several chemical bonds. The 2D descriptor C3_7_F corresponds to the presence of a fluorine-containing substituent on the phenyl or benzyl group attached to the non-aromatic pyridazinone heterocycle, while the C4_7_NO atom pair corresponds to the presence of a nitrophenyl fragment in the acetamide derivative. The condition on the last split of the tree, using the CA_8_O1 atom pair, corresponds to an arylacetamide derivative with a pyridazinone heterocycle and an aryl substituent in this heterocycle in the position opposite of the carbonyl group. This results in at least five CA_8_O1 atom pairs in the molecule. a molecule was considered to be the value of the corresponding 2D descriptor [60]. Atom pairs have an important advantage in SAR analysis over many other types of descriptors, since they can be easily related to certain substructures in a molecule [60][61][62]. Therefore, the results can be interpreted in terms of molecular fragments and/or functional groups. The classification tree is based on five descriptors [four atom pairs and one ADME parameter, known as topological polar surface area (tPSA)]. The misclassification matrix in graphical and tabular forms are shown in Figure 4 and Table 4, respectively.  The importance of the C3_7_F and C4_7_NO atom pairs for the NF-κB inhibitory activity correlates well with the SAR analysis, described above. Indeed, these 2D descriptors are due to fluorine and nitro substituents at certain topological distances from the acetamide and/or heterocyclic moiety (see examples in Figure 5). The topological polar surface area (tPSA) is also an important descriptor in the binary classification tree (Figure 4). Thus, 53 compounds with tPSA > 84.685 Å 2 were recognized as inactive (terminal node 9), which agrees well with the known fact that a high number of polar groups is unfavorable for cell permeability [63,64].
The classification rules encoded in the binary tree can be generally translated into "chemical" language applied to the series of compounds that we investigated. According to the classification model, a compound is active if at least one of the following three conditions is satisfied: Condition 1 (corresponds to node 3 of the tree)-a compound contains a pyridazinone heterocycle which has two or more substituents with sp 3 carbon atoms (e.g., phenylacetamide fragment or benzyl group) and an aliphatic moiety with more than one sp 3 carbon atom (e.g., isopropyl, 10; cyclohexyl, 18; propyl, 79; etc.). Condition 2 (corresponds to nodes 5 and 7 of the tree)-a compound contains a pyridazinone heterocycle that has a phenyl or benzyl substituent containing a strong acceptor (i.e., a fluorine atom, 47, 49, 64; a trifluoromethyl group, 66; or a p-nitro group, 169), or a compound contains a pnitroacetamide moiety (5). Condition 3 (corresponds to node 11 of the tree)-a compound contains at least two C=O groups and two or more benzene rings (e.g., pyridazinones with the phenyl group in position 3 and the acetamide moiety in position 1 meet these structural criteria), and the compound has a relatively low tPSA value (≤84.685 Å 2 ).

Predicted
Active 25 13 Non-active 11 98 According to the classification tree model, 123 of the 147 compounds evaluated (83.7%) were classified correctly, whereas 11 of 36 active compounds were erroneously classified as inactive (Supplementary Table S2). This may be due to the uneven distribution of compounds within the class, i.e., there were relatively many inactive and few active compounds. According to the logical rules defined by the classification tree, a compound was classified as active if it had at least one of the C3_7_F or C4_7_NO atom pairs or more than eight C4_7_CA atom pairs (terminal nodes 5, 7, and 3, respectively). Otherwise, the compound was recognized as active if it had a tPSA value of not greater than 84.685 Å and, at the same time, contained more than four CA_8_O1 atom pairs (terminal node 11). Examples of compounds containing the mentioned 2D descriptors are shown in Figure 5.
C4_7_CA in compound 89 (12) C3_7_F in compound 47 (1) C4_7_NO in compound 169 (1) CA_8_O1 in compound 38 (7) Figure 5. Examples of molecular structures with atom pairs corresponding to C4_7_CA, C3_7_F, C4_7_NO, and CA_8_O1 that were used for predicting the inhibition of the LPS-induced NF-κB activity. One atom pair for each structure is highlighted in red and the total occurrence of the atom pair of a given type is indicated in parentheses. The notation of the atom types is as follows: CA, aromatic carbon; C3, olefin-type or imino carbon; C4, tetrahedral sp3-hybridized carbon; CO, carbonyl carbon; NA, pyridine nitrogen; N1, nitrogen in nitro group; N2, amino-, amido-, or iminonitrogen; O1, carbonyl or nitro oxygen; O2, two-coordinated alcohol or ether-type oxygen; BR, bromine.
The atom pairs C4_7_CA were present in molecules containing simultaneously alkyl groups and aromatic fragments at a distance of several chemical bonds. The 2D descriptor Figure 5. Examples of molecular structures with atom pairs corresponding to C4_7_CA, C3_7_F, C4_7_NO, and CA_8_O1 that were used for predicting the inhibition of the LPS-induced NF-κB activity. One atom pair for each structure is highlighted in red and the total occurrence of the atom pair of a given type is indicated in parentheses. The notation of the atom types is as follows: CA, aromatic carbon; C3, olefin-type or imino carbon; C4, tetrahedral sp3-hybridized carbon; CO, carbonyl carbon; NA, pyridine nitrogen; N1, nitrogen in nitro group; N2, amino-, amido-, or imino-nitrogen; O1, carbonyl or nitro oxygen; O2, two-coordinated alcohol or ether-type oxygen; BR, bromine.
Thus, taking into account the 2D descriptors (atom pairs) in combination with the ADME physicochemical parameters made it possible to develop SAR models based on binary trees containing simple classification rules, with the possibility of interpreting the rules in terms of chemical substructures in the compounds under study.

Comparative Analysis
To evaluate the pairwise relationship between anti-inflammatory activity of these compounds in the cell-based assay in THP1-Blue cells and their FPR1/2 agonist activity in the Ca 2+ mobilization assay using transfected HL-60 cells, we used the Jaccard index. This is defined in mathematical terms as the ratio of the intersection over the union for two subsets, A and B, with different activities within the investigated series of compounds [65] and as follows: where the numerator corresponds to the number of compounds exhibiting both biological activities, while the denominator is the number of compounds exhibiting at least one of the activities. As shown in Table 5, the relationship between FPR1 and FPR2 agonist activities is tight, with a Jaccard index of 0.831, while the inhibition of the LPS-induced NF-κB activity is only weakly related to the compound's FPR1/FPR2 agonist activity. No correlation was found when plotting the logarithms of IC 50 and EC 50 values for the NF-κB inhibitory and FPR1/FPR2 activation assays, respectively, and for compounds active in both types of assays (r = 0.086 for NF-κB vs. FPR1 and r=0.061 for NF-κB vs. FPR2). However, a relatively strong linear correlation (r = 0.742, n = 80) was obtained for agonists of both FPR1 and FPR2 ( Figure 6).
Thus, the calculated Jaccard indices and Pearson correlation coefficients indicate that there is not a significant relationship between NF-κB inhibitory activity (i.e., antiinflammatory activity) and FPR1/FPR2 agonist activity. Indeed, this result is also consistent with the lack of anti-inflammatory activity observed for the FPR1 and FPR2 peptide agonists.
Among the pyridazinones and related derivatives evaluated for anti-inflammatory activity were 98 FPR1/FPR2 agonists and 79 compounds with no FPR agonist activity. Our analysis revealed that 46 of these compounds were not cytotoxic and exhibited antiinflammatory activity, as determined by their ability to inhibit the LPS-induced NF-κB activity in human THP1-Blue monocytic cells. The most potent of these compounds also inhibited monocyte/macrophage IL-6 production, further supporting their anti-inflammatory potential. While the anti-inflammatory compounds identified included both FPR agonists and non-agonists, the calculated Jaccard indices and Pearson correlation coefficients indicated that there was not a significant relationship between the anti-inflammatory activity and FPR1/FPR2 agonist activity. This is an interesting finding, as several groups have reported that FPR agonists can exhibit anti-inflammatory activity in in vivo inflammatory models. For example, compound 84 (a.k.a. FPR1 agonist Cmpd17b) was reported to induce endothelium-independent relaxation in mouse aortas and had a cardioprotection effect in myocardial infarction injury in mice [22,37,38]. We also found that this compound had antiinflammatory activity using in vitro assays. In addition, an evaluation of the effect of an FPR1 antagonist (Boc-MLF) on the ability of compounds 84 and 71 (both are FPR1-specific agonists) showed that their anti-inflammatory effects on LPS-induced NF-κB activation were independent of FPR1, since blocking FPR1 had no effect on their inhibitor activity in LPS-treated THP1 cells (Figure 7).
Thus, the question is whether some of the FPR agonists reported to have anti-inflammatory activity actually mediate their effects through FPR-independent pathways, as is suggested by our results with these series of compounds. This could explain how treatment with some agonists known to be inflammatory (i.e., FPR1 agonists) could result in anti-inflammatory effects. At the very least, it should be considered in FPR studies that some FPR agonists may mediate their effects in vivo through both FPR-dependent and FPR-independent pathways, possibly to include homologous down-regulation of FPR in a process called functional antagonism [66]. As another example, FPR agonists 69, 71, and 73 were all reported previously to reduce pain hypersensitivity in rats with CFA-induced inflammatory arthritis [26]. Interestingly, compounds 69 and 71 inhibited LPS-induced NF-κB activity, and compound 71 inhibited LPS-induced IL-6 production, whereas 73 was inactive. Thus, FPR-dependent and FPR-independent anti-inflammatory effects should be considered when making conclusions regarding the therapeutic potential of these compounds. In comparison, the anti-inflammatory effects of the pyrazolone-based FPR1/FPR2 dual ago-nist (i.e., Cmpd43) have been reported in several murine models of inflammation [40][41][42][43][44]; however, this compound did not inhibit an LPS-induced NF-kB activity, suggesting that it is likely acting in vivo through signaling pathways not involving NF-kB. Nevertheless, additional targets should also be considered when evaluating the anti-inflammatory effects of this class of compounds. For example, several pyridazinones structurally related to compounds 1-120 were reported to be positive allosteric modulators of the α7 nicotinic acetylcholine receptor (nAChR) [67]. It is known that such positive allosteric modulators can act on α7 nAChRs producing anti-nociceptive activity [68,69]. The nAChRs are expressed not only by neurons but also by macrophages and other immune cells involved in inflammation [70,71]. In these cells, stimulation of the nAChR by specific agonists suppresses the release of pro-inflammatory cytokines and could have a therapeutic potential for the alleviation of rheumatoid arthritis [72]. Another potential target of pyridazine compounds could be monoamine oxidase (MAO), and pyridazinones structurally related to the compounds evaluated here were reported as MAO-A inhibitors [73,74]. MAO inhibitors have anti-inflammatory effects in the CNS and a variety of non-CNS tissues [75]. Moreover, MAO-A inhibitors could be used for the reprogramming of tumor-associated macrophages to improve cancer immunotherapy [76]. No correlation was found when plotting the logarithms of IC50 and EC50 values for the NF-κB inhibitory and FPR1/FPR2 activation assays, respectively, and for compounds active in both types of assays (r = 0.086 for NF-κB vs. FPR1 and r=0.061 for NF-κB vs. FPR2). However, a relatively strong linear correlation (r = 0.742, n = 80) was obtained for agonists of both FPR1 and FPR2 ( Figure 6).
Thus, the calculated Jaccard indices and Pearson correlation coefficients indicate that there is not a significant relationship between NF-κB inhibitory activity (i.e., anti-inflammatory activity) and FPR1/FPR2 agonist activity. Indeed, this result is also consistent with the lack of anti-inflammatory activity observed for the FPR1 and FPR2 peptide agonists. Among the pyridazinones and related derivatives evaluated for anti-inflammatory activity were 98 FPR1/FPR2 agonists and 79 compounds with no FPR agonist activity. Our analysis revealed that 46 of these compounds were not cytotoxic and exhibited anti-inflammatory activity, as determined by their ability to inhibit the LPS-induced NF-κB activity in human THP1-Blue monocytic cells. The most potent of these compounds also oprotection effect in myocardial infarction injury in mice [22,37,38]. We also found that this compound had anti-inflammatory activity using in vitro assays. In addition, an evaluation of the effect of an FPR1 antagonist (Boc-MLF) on the ability of compounds 84 and 71 (both are FPR1-specific agonists) showed that their anti-inflammatory effects on LPSinduced NF-κB activation were independent of FPR1, since blocking FPR1 had no effect on their inhibitor activity in LPS-treated THP1 cells (Figure 7). Thus, the question is whether some of the FPR agonists reported to have anti-inflammatory activity actually mediate their effects through FPR-independent pathways, as is suggested by our results with these series of compounds. This could explain how treatment with some agonists known to be inflammatory (i.e., FPR1 agonists) could result in anti-inflammatory effects. At the very least, it should be considered in FPR studies that some FPR agonists may mediate their effects in vivo through both FPR-dependent and FPR-independent pathways, possibly to include homologous down-regulation of FPR in a process called functional antagonism [66]. As another example, FPR agonists 69, 71, and

Compounds
Our studies included 175 compounds that were previously synthesized and characterized at NEUROFARBA, Pharmaceutical and Nutraceutical Section, University of Florence, Italy. We previously evaluated the FPR1/FPR2 agonist activity of these compounds by evaluating their ability to induce Ca 2+ mobilization in FPR-transfected HL60 cells and human neutrophils [22][23][24][25][26][27][28][29]53,54]. Compounds 61 and 139 were synthesized as described (see Supplementary Materials). All test compounds were dissolved in DMSO at stock concentrations of 10 mM and stored at −20 • C. For primary screening assays (analysis of NF-κB activation, IL-6 production, and cytotoxicity), the following final concentrations of the compounds were used: 3.125, 6.25, 12.5, 25, and 50 µM; the final concentration of DMSO was 1% in all samples. The most potent compounds were retested at final concentrations in the submicromolar and low micromolar range (0.1-5.0 µM).

Analysis of NF-κB Activation
The activation of NF-κB was measured using an alkaline phosphatase reporter assay in human monocytic THP1-Blue cells stably transfected with a secreted embryonic alkaline phosphatase gene that is under the control of a promoter inducible by the NF-κB. The THP1-Blue cells (2 × 10 5 cells/well) were pretreated with the test compound or DMSO for 30 min, followed by an addition of 250 ng/mL LPS for 24 h. Alkaline phosphatase activity was determined in cell supernatants using a QUANTI-Blue mix (InvivoGen) and measured at 655 nm and compared with positive control samples (LPS). For active compounds, the concentrations of the inhibitors that caused a 50% inhibition of the NF-κB reporter activity (IC 50 ) were calculated.

IL-6 Analysis
A human IL-6 ELISA kit (BD Biosciences, San Jose, CA, USA) was used to measure IL-6 production. MonoMac-6 cells were plated in 96-well plates at a density of 2 × 10 5 cells/well in a culture medium supplemented with 3% (v/v) endotoxin-free FBS. The cells were pretreated with the test compound or DMSO for 30 min, followed by addition of 250 ng/mL LPS for 24 h. The IC 50 values for IL-6 production were calculated by plotting the percentage inhibition against the logarithm of the inhibitor concentration (at least five points).

Cytotoxicity Assay
The cytotoxicity was analyzed with a CellTiter-Glo Luminescent Cell Viability Assay Kit from Promega (Madison, WI, USA), according to the manufacturer's protocol. The luminescent ATP assays are based on a luciferase enzymatic reaction which uses ATP from viable cells to generate photons. For this assay, the cells were incubated with the compounds under investigation for 24 h. After treatment, the cells were equilibrated to room temperature for 30 min, luciferase + luciferin substrate was added, and the samples were analyzed with a Fluoroscan Ascent FL (Thermo Fisher Scientific, Waltham, MA, USA). The IC 50 values were calculated by plotting the percentage inhibition against the logarithm of the inhibitor concentration (at least five points).

Ca 2+ Mobilization Assay
Changes in the intracellular Ca 2+ [Ca 2+ ] i were measured with a FlexStation II scanning fluorometer (Molecular Devices). The cells, which were suspended in Hank's balanced salt solution without Ca 2+ and Mg 2+ but with 10 mM HEPES (HBSS − ), were loaded with 1.25 µg/mL Fluo-4 AM dye and incubated for 30 min in the dark at 37 • C. After the dye loading, the cells were washed with HBSS − containing 10 mM HEPES, resuspended in HBSS + containing Ca 2+ , Mg 2+ , and 10 mM HEPES (HBSS + ), and aliquoted into the wells of black microtiter plates (2 × 10 5 cells/well). For the evaluation of the direct agonist activity, the compounds of interest were added from a source plate containing dilutions of the test compounds in HBSS + and changes in the fluorescence were monitored (λ ex = 485 nm, λ em = 538 nm) every 5 s for 240 s at room temperature after the automated addition of the compounds. The maximum change in the fluorescence during the first 3 min, expressed in arbitrary units over the baseline, was used to determine a response. The responses for the FPR1 agonists were normalized to the response induced by 5 nM fMLF for FPR1-HL60 cells or 5 nM WKYMVM for FPR2-HL60 cells, which were assigned a value of 100%. A curve fitting (5-6 points) and a calculation of median effective inhibitory concentrations (IC50) were performed by nonlinear regression analysis of the dose-response curves generated using Prism 8 (GraphPad Software, Inc., San Diego, CA, USA).

Molecular Modeling
For SAR classification analysis, we used an atom pair representation of molecular structures with each atom pair denoted as T1_D_T2, where T1 and T2 are the types of atoms in the pair and D represents the topological distance or number of bonds in the shortest path between these atoms in a structural formula. In our investigation, T1 and T2 were defined with symbolic codes used in HyperChem, Version 7 (Hypercube, Inc., Waterloo, ON, Canada) for an atom type representation within an MM+ force field [60][61][62]. For example, CA, CO, and C3 codes were used for sp 2 -hybridized aromatic, carbonyl, and pyrazole carbon atoms, respectively. This approach allows easy generation of atom pairs directly from the output file containing the molecular structure (HIN file) built by HyperChem.
A total of 147 compounds were included in the analysis. The racemic mixtures along with individual enantiomers corresponding to these racemates were excluded, as they are indistinguishable in the atom pair 2D representation of the molecular structures. All 1029 unique atom pairs possible for non-hydrogen atoms in the 147 investigated compounds were generated. This 147×1029 data matrix was automatically built by our CHAIN program based on HIN files created in HyperChem. By convention, a matrix element at the intersection of the i th row and j th column was equal to the j th atom pair occurrence in the i th molecule. The data matrix obtained in this way for the 147 compounds was supplemented with columns corresponding to the physicochemical ADME parameters calculated with the use of a SwissADME online tool (http://www.swissadme.ch, accessed on 1 August 2021) [59].
The total matrix containing both atom pair and ADME descriptors was used for the construction of SAR rules with the use of a binary classification tree methodology [77] with STATISTICA 6.0 using discriminant-based univariate splits with estimated prior probabilities and equal misclassification costs for classes.

Conclusions
Previously, we identified a number of FPR1/FPR2 agonists and found that many of the active compounds had a pyridazinone core structure [22][23][24][25][26][27][28][29]. In an effort to identify novel anti-inflammatory compounds, we evaluated a library of 177 pyridazinone-like compounds and related scaffolds. Although the pyridazinone scaffold was present iñ 70% of the compounds that inhibited the LPS-induced NF-κB activity in a cell-based anti-inflammatory assay, there was not a good correlation between the FPR1/FPR2 agonist activity and the anti-inflammatory activity. We suggest that some of these compounds could mediate their anti-inflammatory effects through FPR-independent pathways, which may explain why treatment with some agonists that are known to be inflammatory (i.e., FPR1 agonists) results in anti-inflammatory effects. We developed a SAR classification model based on atom pair descriptors and physicochemical ADME parameters and found that the C4_7_CA, C3_7_F, CA_8_O1, and C4_7_NO atom pairs, together with the topological polar surface area (tPSA), a physicochemical descriptor, were important. This classification model could provide a relatively simple approach for de novo design of anti-inflammatory compounds with pyridazinone-like scaffolds. Further research is clearly needed to define the molecular targets of pyridazinones and structurally related compounds with antiinflammatory activity and to define their relationships (if any) to FPR signaling events.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/molecules27123749/s1. Synthesis of compounds 61 and 139. Table S1: Selected ADME parameters of the investigated compounds calculated with SwissADME web tool. Table S2: Experimentally observed and predicted classes for inhibition of NF-κB activity for the pyridazinones and related derivatives in according to classification tree analysis [78].

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.