Modulation of Janus kinase/signal transducer and activator of transcription (JAK/STAT) signaling is a promising method of treating autoimmune diseases, and the profound potency of clinical compounds makes this mode of action particularly attractive. Other questions that remain unanswered also include: What is the ideal selectivity between JAK1 and JAK3? Which cells are most relevant to JAK blockade? And what is the ideal tissue distribution pattern for addressing specific autoimmune conditions? We hypothesized that JAK3 selectivity is most relevant to low-dose clinical effects and interleukin-10 (IL-10) stimulation in particular, that immune cells are the most important compartment, and that distribution to inflamed tissue is the most important pharmacokinetic characteristic for in vivo disease modification. To test these hypotheses, we prepared modified derivatives of JAK3 specific inhibitors that target C909 near the ATP binding site based on FM-381, first reported in 2016; a compound class that was hitherto limited in uptake and exposure in vivo. These limits appear to be due to metabolic instability of side groups binding in the selectivity pocket. We identified derivatives with improved stability and tissue exposure. Conjugation to macrolide scaffolds with medium chain linkers was sufficient to stabilize the compounds and improve transport to organs while maintaining JAK3 affinity. These conjugates are inflammation targeted JAK3 inhibitors with long tissue half-lives and high exposure to activated immune cells.

Short-chain fatty acids (SCFAs) have a range of effects in metabolism and immune regulation. We have observed that delivery of SCFAs to lysosomes has potent immune regulatory effects, possibly as a surrogate signal for the presence of anaerobic organisms. To better understand the pharmacology of lysosomal SCFA donors, we investigated the distribution and metabolism of propionate and butyrate donors. Each analog (1 a and 2 a) can donate three SCFA equivalents via ester hydrolysis through six intermediate metabolites. The compounds are stabilized by low pH, and stability in cells is usually higher than in medium, but is cell-type specific. Butyrate derivatives were found to be more stable than propionates. Triesters were more stable than di-or mono-esters. The donors were surprisingly stable in vivo, and hydrolysis of each position was organ specific. Jejunum and liver caused rapid loss of 4'' esters. The gut metabolite pattern by i. v. differed from that of p.o. application, suggesting luminal and apical enzyme effects in the gut epithelium. Central organs could de-esterify the 11position. Levels in lung relative to other organs were higher by p.o. than via i. v., suggesting that delivery route can influence the observed pharmacology and that gut metabolites distribute differently. The donors were largely eliminated by 24 h, following near linear decline in organs. The observed levels and distribution were found to be consistent with pharmacodynamic effects, particularly in the gut.

Topical imiquimod based creams are indicated as immune stimulants for papillomas and various skin neoplasms. Imiquimod is considered a TLR7 ligand. These creams are also used in research to induce skin inflammation in mice as a model for psoriasis. We observed that this inflammatory response was not strictly imiquimod dependent and we set out to establish which components drive the proinflammatory effects. To this end, we examined the induction response in a BALB/cJRj mouse model, in which 50 mg of cream is applied to 2 cm 2 of skin (125 mg/kg imiquimod -5% W/V, and/or 625 mg/kg isostearic acid -25% W/V). Comparing cream formulations containing isostearic acid, imiquimod and the combination, we observed that isostearic acid causes skin inflammation within 2 days, whereas imiquimod requires up to 5 days for initial signs. Isostearic acid activated an inflammasome response, stimulated release of proinflammatory cytokines and upregulated the IL-23/17 axis. Animals treated with isostearic acid had enlarged livers (+40% weight), which was not observed with imiquimod alone. Imiquimod was readily metabolized and cleared from plasma and liver, but was maintained at high levels in the skin throughout the body (200 mM at area of application; 200 µM in untreated skin). Imiquimod application was associated with splenomegaly, cytokine induction/release and initial body weight loss over 3 days. Despite high imiquimod skin levels throughout the animal, inflammation was only apparent in the treated areas and was less severe than in isostearic acid groups. As the concentrations in these areas are well above the 10 µM required for TLR7 responses in vitro, there is an implication that skin inflammation following imiquimod is due to effects other than TLR7 agonism (e.g., adenosine receptor agonism). In brain, isostearic caused no major changes in cytokine expression while imiquimod alone sightly stimulated expression of IL-1β and CCL9. However, the combination of both caused brain induction of CCL3, -9, CXCL10, -13, IL-1β and TNFα. The implication of these data is that isostearic acid facilitates the entry of imiquimod or peripherally secreted cytokines into the brain. Our data suggest that psoriaform skin responses in mice are more driven by isostearic acid, than generally reported and that the dose and route used in the model, leads to profound systemic effects, which may complicate the interpretation of drug effects in this model.

TLR Agonists have promising activity in preclinical models of viral infection and cancer. However, clinical use is only in topical application. Systemic uses of TLR-ligands such as Resiquimod, have failed due to adverse effects that limited dose and thus, efficacy. This issue could be related to pharmacokinetic properties that include fast elimination leading to low AUC with simultaneously high cmax at relevant doses. The high cmax is associated with a sharp, poorly tolerated cytokine pulse, suggesting that a compound with a higher AUC/cmax-ratio could provide a more sustained and tolerable immune activation. Our approach was to design TLR7/8-agonist Imidazoquinolines intended to partition to endosomes via acid trapping using a macrolide-carrier. This can potentially extend pharmacokinetics and simultaneously direct the compounds to the target compartment. The compounds have hTLR7/8-agonist activity (EC50 of the most active compound in cellular assays: 75-120 nM hTLR7, 2.8-3.1 µM hTLR8) and maximal hTLR7 activation between 40 and 80% of Resiquimod. The lead candidates induce secretion of IFNα from human Leukocytes in the same range as Resiquimod but induce at least 10-fold less TNFα in this system, consistent with a higher specificity for human TLR7. This pattern was reproduced in vivo in a murine system, where small molecules are thought not to activate TLR8. We found that Imidazoquinolines conjugated to a macrolide or, substances carrying an unlinked terminal secondary amine, had longer exposure compared with Resiquimod. The kinetics of pro-inflammatory cytokine release for these substances in vivo were slower and more extended (for comparable AUCs, approximately half-maximal plasma concentrations). Maximal IFNα plasma levels were reached 4 h post application. Resiquimod-treated groups had by then returned to baseline from a peak at 1 h. We propose that the characteristic cytokine profile is likely a consequence of altered pharmacokinetics and, potentially, enhanced endosomal tropism of the novel substances. In particular, our substances are designed to partition to cellular compartments where the target receptor and a distinct combination of signaling molecules relevant to IFNα-release are located. These properties could address the tolerability issues of TLR7/8 ligands and provide insight into approaches to fine-tune the outcomes of TLR7/8 activation by small molecules.

Topical imiquimod based creams are indicated as immune stimulants for papillomas and various 13 skin neoplasms. Imiquimod is considered a TLR7 ligand. These creams are also used in research to 14 induce skin inflammation in mice as a model for psoriasis. We observed that this inflammatory 15 response was not strictly imiquimod dependent and we set out to establish which components 16 drive the proinflammatory effects. To this end, we examined the induction response in a 17 BALB/cJRj mouse model, in which 50 mg of cream is applied to 2 cm2 of skin (125 mg/kg imiquimod 18 – 5% W/V, and/or 625 mg/kg isostearic acid – 25% W/V). Comparing cream formulations contain- 19 ing isostearic acid, imiquimod and the combination, we observed that isostearic acid causes skin 20 inflammation within 2 days, whereas imiquimod requires up to 5 days for initial signs. Isostearic 21 acid activated an inflammasome response, stimulated release of proinflammatory cytokines and 22 upregulated the IL-23/17 axis. Animals treated with isostearic acid had enlarged livers (+40% 23 weight), which was not observed with imiquimod alone. Imiquimod was readily metabolized and 24 cleared from plasma and liver, but was maintained at high levels in the skin throughout the body 25 (200 mM at area of application; 200 µM in untreated skin). Imiquimod application was associated 26 with splenomegaly, cytokine induction/release and initial body weight loss over 3 days. Despite 27 high imiquimod skin levels throughout the animal, inflammation was only apparent in the treated 28 areas and was less severe than in isostearic acid groups. As the concentrations in these areas are 29 well above the 10 µM required for TLR7 responses in vitro, there is an implication that skin in- 30 flammation following imiquimod is due to effects other than TLR7 agonism (e.g., adenosine re- 31 ceptor agonism). In brain, isostearic caused no major changes in cytokine expression while 32 imiquimod alone sightly stimulated expression of IL-1β and CCL9. However, the combination of 33 both caused brain induction of CCL3, -9, CXCL10, -13, IL-1β and TNFα. The implication of these 34 data is that isostearic acid facilitates the entry of imiquimod or peripherally secreted cytokines 35 into the brain. Our data suggest that psoriaform skin responses in mice are more driven by isos- 36 tearic acid, than generally reported and that the dose and route used in the model, leads to pro- 37 found systemic effects, which may complicate the interpretation of drug effects in this model.