Structure – activity relationships in Toll-like receptor 7 agonistic 1 H -imidazo[4,5- c ]pyridines †

Engagement of TLR7 in plasmacytoid dendritic cells leads to the induction of IFN- α / β which plays essential functions in the control of adaptive immunity. We had previously examined structure – activity relationships (SAR) in TLR7/8-agonistic imidazoquinolines with a focus on substituents at the N 1 , C 2 , N 3 and N 4 positions, and we now report SAR on 1 H -imidazo[4,5- c ]pyridines. 1-Benzyl-2-butyl-1 H -imidazo[4,5- c ]-pyridin-4-amine was found to be a pure TLR7-agonist with negligible activity on TLR8. Increase in potency was observed in N 6 -substituted analogues, especially in those compounds with electron-rich substituents. Direct aryl – aryl connections at C6 abrogated activity, but TLR7 agonism was reinstated in 6-benzyl and 6-phenethyl analogues. Consistent with the pure TLR7-agonistic behavior, prominent IFN- α induction in human PBMCs was observed with minimal proin ﬂ ammatory cytokine induction. A benzologue of imidazoquinoline was also synthesized which showed substantial improvements in potency over the parent imidazopyridine. Distinct di ﬀ erences in N 6 -substituted analogues were observed with respect to IFN- α induction in human PBMCs on the one hand, and CD69 upregulation in lymphocytic subsets, on the other.


Introduction
Host responses to pathogens are mediated via highly coordinated mechanisms involving both the innate and adaptive limbs of the immune system. The innate immune system utilizes germline-encoded pattern recognition receptors (PRRs) to detect pathogen-associated molecular patterns (PAMPs) that are distinct and unique to the pathogen. [1][2][3] PRRs encompass a broad range of molecules 4 that are secreted into the extracellular environment (such as the collectins, 5 ficolins, 6 pentraxins, 7 alarmins 8 ), exist in the cytosol (examples of which include the retinoic acid-inducible gene I-like receptors, 9 and the nucleotide-binding domain and leucine-rich repeat-containing receptors 10 ), or are present on membranes.
Important among the transmembrane PRRs include the Toll-like receptors 11 (TLRs) which are either expressed on the plasma membrane or in the endolysosomal compartments. 1,12 At least 10 functional TLRs are encoded in the human genome, each with an extracellular domain having leucine-rich repeats (LRR) and a cytosolic domain called the Toll/IL-1 receptor (TIR) domain. 13 The ligands for these receptors are highly conserved microbial molecules such as lipopolysaccharides (LPS) (recognized by TLR4), lipopeptides (TLR2 in combination with TLR1 or TLR6), flagellin (TLR5), single stranded RNA (TLR7 and TLR8), double stranded RNA (TLR3), CpG motif-containing DNA (recognized by TLR9), and profilin present on uropathogenic bacteria (TLR11). 14,15 TLR1, -2, -4, -5, and -6 recognize extracellular stimuli, while TLR3, -7, -8 and -9 function within the endolysosomal compartment. 13 The activation of TLRs by their cognate ligands leads to production of inflammatory cytokines, and up-regulation of major histocompatibility complex (MHC) molecules and co-stimulatory signals in antigen-presenting cells as well as activating natural killer (NK) cells (innate immune response), which leads to the priming and amplification of T-, and B-cell effector functions (adaptive immune responses). [16][17][18][19] The Type I interferon (IFN) family in humans include approximately 20 IFN-α subtype genes in addition to individual genes encoding IFN-β, -κ, -ε and -ω; these monomeric secreted proteins bind to a single IFN-α/β receptor, which is constitutively expressed in virtually all cell types. 20 Occupancy of TLR7 [21][22][23] or TLR9 24,25 in professional antigen-presenting cells (APCs), particularly plasmacytoid dendritic cells ( pDCs), leads to the induction of IFN-α/β. Although the Type I IFNs are best known historically for their antiviral activities, 26 recent studies show that they have many essential functions in the control of adaptive immunity. 2 First, Type I IFNs promote cross-priming through direct stimulation of DCs, leading to specific CD8 + lymphocytic responses to soluble antigens. 27 Second, Type I IFNs potently enhance the primary antibody responses to soluble antigens, inducing sustained and durable humoral responses with appropriate isotype switching, as well as the induction of immunological memory. 28 B lymphocytes can differentiate into two distinct types of functionally polarized effectors: B-effector-1-cells (Be-1), producing a Th-1-like cytokine pattern, or Be-2, characterized by a Th-2-like profile. 29,30 It is of particular interest that recent reports suggest that IFN-α may serve as an initial trigger for Be-1-biased differentiation pattern. 31 Third, Type I interferons secondarily induce Type II IFN (IFN-γ) secretion, also driving Th-1-biased adaptive immune responses. 32 Type I IFN-inducing TLR ligands may therefore hold promise as vaccine adjuvants.
In an effort to identify optimal immunostimulatory chemotypes, we have screened representative members of virtually the entire compendium of known TLR agonists in a series of hierarchical assays including primary TLR-reporter assays, secondary indices of immune activation such as IFN-α/β/γ and cytokine induction, activation of lymphocytic subsets in whole human blood, and tertiary screens characterizing transcriptomal activation patterns. 33 In these assays, small-molecule agonists of TLR7 were uniquely immunostimulatory; they were potent inducers of Type I IFN and, unlike TLR-4, -5, or -8 agonists, 33 did not evoke dominant proinflammatory cytokine responses, suggesting that they may be effective, yet safe vaccine adjuvants, a premise that we have been actively exploring. 34 Small molecule TLR7 agonists are also being investigated as orally bioavailable, endogenous Type I IFN inducers for the management of chronic viral diseases, 35 especially hepatitis C and hepatitis B. Current therapeutic regimens for the therapy of hepatitis C and hepatitis B include parenteral IFN-α. 36 Clinical trials of TLR7 agonists for hematological malignancies are also currently underway. 37 The currently known small molecule agonists of TLR7 occupy a very small chemical space, and are represented by the 1H-imidazo [4,5-c]quinolines, 38 8-hydroxy- [39][40][41] or 8-oxoadenines, 42 and guanine nucleoside analogues. 43,44 We had previously reported structure-activity relationships (SAR) in the imidazoquinolines with a focus on substituents at the N 1 , C 2 , N 3 and N 4 positions, 45,46 and we had observed that relatively minor structural modifications at these positions yielded compounds with widely differing immunomodulatory properties. 34,[47][48][49] It was of interest, therefore, to extend our SAR studies to the quinoline ring system. We asked if a partstructure (imidazopyridine) or a benzologue (benzoimidazoquinoline) would alter the biological properties of the parent imidazoquinoline compound. Examination of the structures of 3M-003 50 and the 8-hydroxy-and 8-oxoadenines (Fig. 1) suggested that the quinoline system may be dispensable, and activity would be retained in imidazopyridines. Indeed, imidazopyridine derivatives with alkyl groups at C 6 and C 7 positions, 51 hydroxyalkyl, 52 oxime and hydroxylamine-bearing substituents 53 at C 2 , and alkylsulfonamide substituents at the N 1 position 54 have been reported in the patent literature.
Detailed activity profiles of these compounds, however, are not available, perhaps owing to the fact that the investigations of such compounds precede the discovery of the TLRs.
Incorporating substituents that we had previously determined to be optimal in the imidazoquinolines (N 1 -benzyl and C 2 -butyl; IMDQ, Fig. 1), we embarked on the syntheses and biological evaluation of novel 1H-imidazo [4,5-c]pyridine analogues with modifications at the N 4 -and C 6 positions. The parent imidazopyridine compound, 1-benzyl-2-butyl-1Himidazo [4,5-c]pyridin-4-amine, exhibited moderate TLR7-agonistic activity. N 4 -Acyl or -alkyl substitutions abrogated activity. The majority of C 6 derivatives bearing aryl groups were also inactive, but analogues with N 6 -benzyl substituents gained TLR7-specific activity. Particular N 6 substituents were found to augment TLR7-specific agonistic potency without compromising specificity at TLR7; consistent with their pure TLR7 activity (and undetectable TLR8 agonism), these compounds potently induced IFN-α in human peripheral blood mononuclear cells (PBMCs), upregulated CD69 in lymphocytic subsets, and yet showed very weak proinflammatory cytokine-inducing activities. Strong Type I IFN inducers, especially in conjunction with attenuated proinflammatory profiles are expected to be potently adjuvantic without inducing prominent local or systemic inflammation.
Modest gains in potency were obtained in analogues with short aliphatic substituents with N 6 -butyl (19b) and N 6 -cyclohexylmethyl (19d), but potency diminished in the N 6 -heptyl analogue (19c). The N 6 -phenyl-substituted compound 19e was marginally weaker than 5; however, the potency of the N 6 -benzyl analogue 19f was ∼7.6 times that of 5 (Table 1, Fig. 2), triggering a detailed SAR investigation on various aryl substituents at N 6 . Both steric and electronic effects appear to play a role in governing TLR7-agonistic potency, since the biphenylmethyl-substituted compound 19o was active, whereas the naphthylmethyl analogue 19n was quiescent; to a first approximation, electron-rich N 6 substituents appear to be preferred, with the methoxybenzyl derivatives (19g and 19h) and the pyridinylmethyl compounds (19l and 19m) being marginally more active than the trifluoromethyl-(19i) or chloro-(19j) substituted analogues. Compounds 19p and 19q were also active in primary screens, with EC 50 values of 0.26 and 0.37 μM, respectively (Table 1). In the C 6 -alkyl or -aryl analogues (Scheme 6), the SAR appeared more stringent. Whereas the C 6 -butyl compound 23a was more active than 5, direct aryl-aryl connections at C6 (23b-f ) abrogated activity, but TLR7 agonistic properties were restored in the 6-benzyl (23g) and 6-phenethyl analogues (23j). Taken together with activity data of compounds of the 19 series, we surmised that rotational constraints about the C 6aryl groups may hinder TLR7 occupancy. Unlike TLR2, TLR3, TLR4, 57 and TLR5 58 for which crystal structures are available as complexes with their cognate ligands, a detailed structural characterization of the mode of binding of TLR7 ligands is not yet available to guide structure-based design of small molecule agonists of TLR7, necessitating classical SAR approaches to refine successive iterations of ligand design.
The benzologue 30 was synthesized as shown in Scheme 7. It showed substantial improvements in potency over the parent imidazopyridine 5 (Fig. 3, Table 1), but the two most potent compounds in the entire series as adjudged by primary screens were the N 6 -(4-methoxybenzyl) and N 6 -(furan-2ylmethyl) analogues (19g and 19k, respectively), both of which were approximately twenty-fold more potent than 5 (Fig. 2, Table 1).
We had previously shown that of all the various classes of innate immune stimuli, TLR7 agonists were extraordinarily immunostimulatory, stimulating virtually all subsets of lymphocytes (assessed by quantifying CD69 expression), and yet without inducing dominant proinflammatory cytokine responses, 33 and we wished to confirm the rank-order potency observed in IFN-α induction assays described above. We observed considerable dissociation between Type I IFN induction on the one hand (Fig. 4), and CD69 upregulation in lymphocytic subsets on the other (Fig. 5). Whereas the subset of active compounds induced IFN-α with similar potencies (EC 50 values between 0.3-2 μM; Fig. 4), pronounced differences were observed in CD69 expression in natural killer, cytokineinduced killer and B lymphocytic subsets with 19p being as active as the reference TLR7 agonist IMDQ, and 19d showing virtually no activity (Fig. 5). Possible mechanisms underlying the differential activity in these two compounds are being investigated.

Conclusions
These findings raise the possibility of utilizing these compounds in selectively targeting Type I IFN induction versus lymphocytic activation, and are being explored in greater detail.
The potential advantages of strong Type I IFN inducers as candidate vaccine adjuvants have been discussed earlier. Such compounds, especially in conjunction with attenuated proinflammatory cytokines, are expected to be potently adjuvantic without inducing prominent local or systemic inflammation. As mentioned earlier, the prominent Type I IFN inducing abilities of the imidazopyridines may also find utility as an alternative therapeutic strategy to address disease states wherein systemic IFN-α is of proven benefit. A clear delineation of structural features that confer TLR specificity not only charts a rational course for the development of effective, yet (v) a. C 4 H 9 COCl, NEt 3 , THF; b. NaOH, EtOH; (vi) mCPBA, CH 2 Cl 3 , CHCl 3 , MeOH; (vii) a. Benzoyl isocyanate, CH 2 Cl 2 ; b. NaOMe, MeOH.

Materials and equipment
All of the solvents and reagents used were obtained commercially and used as such unless noted otherwise. Moisture-or   Mass spectrometer (mass accuracy of 5 ppm) operating in the positive ion (or negative ion, as appropriate) acquisition mode.

Synthesis of compound 23a
To a solution of compound 14 (120 mg, 0.31 mmol) in 1 mL of dioxane were added cesium carbonate (303 mg, 0.93 mmol) in H 2 O (0.5 mL), [1,1′-bis(diphenylphosphino)ferrocene]dichloropalladium(II) (Pd(dppf )Cl 2 ) (15 mg, 0.019 mmol) and n-butylboronic acid (98 μL, 0.46 mmol) under N 2 . The reaction mixture was then heated at 90°C in a sealed vial for 18 h. It was cooled to room temperature and filtered through celite and washed with MeOH. The solvent was removed and the crude residue was purified using column chromatography (15% EtOAc-hexane) to obtain the compound 20a (97 mg, 76%). To a solution of compound 20a (94 mg, 0.23 mmol) in 10 mL of MeOH were added zinc dust (149 mg, 2.30 mmol) and ammonium formate (145 mg, 2.30 mmol). The reaction mixture was stirred at room temperature for 10 min and filtered through celite. Then the solvent was evaporated and the residue was dissolved in water. This was extracted with EtOAc (3 × 20 mL), washed with water and dried over sodium sulfate. The solvent was removed under vacuum to obtain compound 21a (45 mg, 51%). To a solution of compound 21a (42 mg, 0.11 mmol) in 7 mL of anhydrous THF were added triethylamine (16 μL, 0.12 mmol) and valeryl chloride (13 μL, 0.11 mmol). The reaction mixture was refluxed for 1 h. The solvent was then removed under vacuum, and the residue was dissolved in 5 mL of EtOH and NaOH (10 mg, 0.22 mmol) in 1 mL of H 2 O was added. The reaction mixture was refluxed for 18 h. The solvent was then removed under vacuum, and the residue was dissolved in EtOAc and washed with water. The EtOAc fraction was dried using sodium sulfate and evaporated and purified using column chromatography (20% EtOAchexane) to obtain the compound 22a (25 mg, 51%). Compound 22a (22 mg, 0.049 mmol) was dissolved in 1 mL of HCl (4 M in dioxane) and stirred at room temperature for 30 min. Then the solvent was removed under vacuum to obtain compound 23a (11 mg, 69%).  13   Phenylboronic acid was used as a reagent. Yellow solid (28 mg, 57%). 1

Paper Organic & Biomolecular Chemistry
The induction of NF-κB was quantified using HEK-Blue-7 (hTLR7-specific) and HEK-Blue-8 (hTLR8-specific) cells as previously described by us. 33,45,46 HEK293 cells stably co-transfected with human TLR7 or human TLR8, MD2, and secreted alkaline phosphatase (sAP), were maintained in HEK-Blue™ Selection medium containing zeocin and normocin. Stable expression of secreted alkaline phosphatase (sAP) under control of NF-κB/AP-1 promoters is inducible by appropriate TLR agonists, and extracellular sAP in the supernatant is proportional to NF-κB induction. HEK-Blue cells were incubated at a density of ∼10 5 cells per mL in a volume of 80 µL per well, in 384-well, flat-bottomed, cell culture-treated microtiter plates until confluency was achieved, and subsequently stimulated with graded concentrations of stimuli. sAP was assayed spectrophotometrically using an alkaline phosphatase-specific chromogen ( present in HEK-detection medium as supplied by the vendor) at 620 nm.

Immunoassays for interferon (IFN)-α, and cytokines
Fresh human peripheral blood mononuclear cells (hPBMC) were isolated from human blood obtained by venipuncture with informed consent and as per institutional guidelines on Ficoll-Hypaque gradients as described elsewhere. 64 Aliquots of PBMCs (10 5 cells in 100 μL per well) were stimulated for 12 h with graded concentrations of test compounds. Supernatants were isolated by centrifugation, and were assayed in triplicates using either high-sensitivity multi-subtype IFN-α ELISA kits (PBL Interferon Source, Piscataway, NJ and R&D Systems, Inc., Minneapolis, MN), or analyte-specific multiplexed cytokine/ chemokine bead array assays as reported by us previously. 65 Flow-cytometric immunostimulation experiments CD69 upregulation was determined by flow cytometry using protocols published by us previously, 33 and modified for rapid-throughput. Briefly, heparin-anticoagulated whole blood samples were obtained by venipuncture from healthy human volunteers with informed consent and as per guidelines approved by the University of Kansas Human Subjects Experimentation Committee. Serial dilutions of selected imidazopyridine compounds (and imiquimod, used as a reference compound) were performed using a Bio-Tek Precision 2000 XS liquid handler in sterile 96-well polypropylene plates, to which Organic & Biomolecular Chemistry Paper were added 100 μL aliquots of anticoagulated whole human blood. The plates were incubated at 37°C for 16.5 h. Negative (endotoxin free water) controls were included in each experiment. Following incubation, fluorochrome-conjugated antibodies (CD3-PE, CD56-APC, CD69-PE-Cy7, 10 μL of each antibody, Becton-Dickinson Biosciences, San Jose, CA) were added to each well with a liquid handler, and incubated at 37°C in the dark for 30 min. Following staining, erythrocytes were lysed and leukocytes fixed by mixing 200 mL of the samples in 2 mL pre-warmed Whole Blood Lyse/Fix Buffer (Becton-Dickinson Biosciences, San Jose, CA) in 96 deep-well plates. After washing the cells twice at 200 g for 8 minutes in saline, the cells were transferred to a 96-well plate. Flow cytometry was performed using a BD FACSArray instrument in the tricolor mode (tri-color flow experiment) and two-color mode (two-color flow experiment) for acquisition on 100 000 gated events. Compensation for spillover was computed for each experiment on singly-stained samples. CD69 activation in the major lymphocytic populations, viz., natural killer lymphocytes (NK cells: CD3 − CD56 + ), cytokine-induced killer phenotype (CIK cells: CD3 + CD56 + ), nominal B lymphocytes (CD3 − CD56 − ), and nominal T lymphocytes (CD3 + CD56 − ) were quantified using FlowJo v 7.0 software (Treestar, Ashland, OR).