Specific Aims (verbatim from original R03 application and status)

Aim 1 (Conduct a MLPCN Library Screen): We will complete a cell-based hCXCR6 HTS campaign to seek functional receptor antagonists.

• We have completed all primary and secondary assay development and conducted a LOPAC1280 pilot screen, with good Z′ and S/B values. We have defined a suitable critical path plan for Probe development. We will assist the Sanford-Burnham MLPCN Center in reconfiguring the cell-based critical path functional assay parameters (primary and secondary) from 384 to 1536 well format. All cell lines are available in the assay provider’s lab and will be transferred to the MLPCN center under the standard NIH material transfer agreement. This Aim has been completed.
• We will screen the MLSMR collection of 330,600 compounds according to the critical path workflow scheme and submit the assay protocol (AIDs) and screening data to PubChem in a timely manner. The HTS campaign was complete and data loaded into PubChem – This Aim has been completed.
• We will register our natural product extracts in PubChem and screen our library of 10,000 natural product extracts and peak fractions to expand novel chemical space outside of typical compounds available within the MLSMR. This will assure identification of suitable probe molecules after assay guided fractionation, dereplication, and isolation/structural identification of active components. Please note, we were not able to register extracts in PubChem given CID and SID limitations. Efforts in exploring the natural product lead extract are ongoing and are outside of the MLPCN effort.

Aim 2 (Develop Probe Characterization Tertiary Assays): We will develop and validate hCXCR6 probe characterization assays.

• We will develop new, and optimize existing, tertiary probe characterization assays including a membrane filtration radioligand binding assay (96 well), a Gαi coupled mechanism hCXCR6-based cAMP assay (96 well format), and in vitro chemotaxis and invasion assays using human and murine prostate cancer cells (LNCaP, LNCaP C4-2B, PC3, VCaP, RM1; 24 well format). Assays have been developed; however, the testing of compounds in these assays are post-probe characterization as listed in the planned future studies section on p 17.
• Demonstrate assay consistency and robustness using an available anti-CXCL16 antibody as a positive antagonist control. This has been completed.

Aim 3 (Characterize Probes): We will fully characterize resulting MLPCN hCXCR6 Probe(s) in the tertiary assays.

• Probe candidate and key SAR-related molecules identified in Aim 1 will be evaluated and characterized utilizing the mechanistic assays developed and optimized in Aim 2. All post-probe assay development has been completed as defined in the probe CPDP and related documents on file at the NIH. Given the fact that such assays are ‘post-probe’ characterization activities, this aspect of the program should be considered as ‘in progress’.

The approach described by these Specific Aims allowed for the successful screening and subsequent identification of novel small molecule hCXCR6 Probes. It is our expectation that validated and effective hCXCR6 chemokine receptor modulators will emerge, resulting in the downstream development of potential therapeutics for prostate cancer. Both orthosteric and allosteric receptor antagonists can be identified within this platform.

Background and Significance

We have recently generated experimental data providing validation that inhibition of the CXCR6/CXCL16 axis may have therapeutic benefit in modulating prostate cancer (PCa) cell proliferation and metastasis with attending bone invasiveness. In addition, a robust screening platform suitable for HTS and secondary characterization has been developed.

PCa is the second leading cause of cancer death in American men and its morbidity has increased globally in recent years. The high mortality rate is closely associated with the spread of malignant cells to various tissues including bone. Nearly 10% of patients whose conditions are diagnosed as PCa initially present with bone metastasis and almost all patients who die of prostate cancers have skeletal involvement.1–12 Identifying new mechanisms that control bone metastasis is of great consequence to facilitate the design of therapeutics aimed at decreasing metastatic risk and/or its complications. To address this unmet medical need, our team is actively engaged in exploring the chemical biology, medicinal chemistry, and therapeutic significance of modulating tumor cell trafficking and metastasis via chemokine receptor inhibition. The primary objective of this proposal is to use high throughput screening methods to identify small molecule antagonist probes that selectively inhibit CXCR6. Our team intends to address and explore a key hypothesis: The CXCR6/CXCL16 axis significantly contributes to PCa cell metastasis and subsequent bone invasion.13–15 A small molecule antagonist would block cancer cell trafficking; hence mediate a metastatic event and disease progression to bone. Thus, access to pharmacologically available small molecule antagonists will ultimately enable our studies in disease relevant models and allow for a more seamless translational advance to clinical applications.

Although an understanding regarding the role of chemokines that regulate cell movement in the spleen and lymphatic system has emerged, considerably less is understood regarding the role of these protein ligands and their corresponding cellular receptors particularly in cancer metastasis, proliferation and secondary tissue invasiveness. To date, evidence is available that supports the role of the hCXCR4/CXCL12 axis in cancer progression and there are currently clinical studies targeting antagonism of the hCXCR4 receptor with a selective small molecule compound for a variety of metastatic events including PCa.16–18 The CXCR4/CXCL12 axis only partially blocks metastatic behavior in vitro, and little is known about the relationship between PCa specific metastasis and other chemokine systems. Within the bone microenvironment hCXCR4 and CXCL12 appear to have conflicting roles. Genetic disruption of hCXCR4 enhances osteoclast activity and therefore stimulates tumor cell growth in bone yet data exists illustrating either growth stimulation or no effect of CXCL12 on osteoclasts thus indicating the role of this axis is not fully defined and understood. Recently, CXCL16 was identified as a ligand for hCXCR6 (STRL33), which is expressed by peripheral blood leukocytes.19 CXCL16 expression has been demonstrated in a variety of tissues and cells including activated endothelial cells. Additionally, CXCL16 has been shown to function as a potent and direct activator of NF-κB and induces κB-dependent pro-inflammatory gene transcription through interaction with heterotrimeric G-proteins triggering downstream PI3K, PDK-1, Akt, and IκB kinase (IKK) signal transduction events. Through a cytokine antibody array, Lu and coworkers15 reported that CXCL16 protein production was increased in aggressive PCa cells compared to the less aggressive PCa cells or benign prostate cells. We also found that both IL-1β and TNFα significantly induced CXCL16 production by LNCaP and PC3 cells, thereby indicating inflammatory cytokines may play a role in CXCL16 induction. These observations support the notion that hCXCL16/CXCR6 interactions may be important for PCa invasion and metastasis. As preliminary results, we have provided strong evidence that the hCXCR6/CXCL16 axis also plays a key role in PCa metastasis, tumor cell proliferation and invasiveness.

To date, no small molecule or antagonists are available to test our hypothesis. Given the knowledge gap, we propose that the development of pharmacologically available small molecule antagonists to CXCR6 will lay the foundation for a novel approach to mediate PCa and its deadly metastases. We have shown that the expression of CXCR6 1) is correlated with disease progression, 2) regulates migration and bone invasiveness, 3) is involved in proliferation of human PCa.14,15,20 These data are relevant as they set the foundation for target validation and for moving forward with our proposed aims.

Prior Art

A recent comprehensive search of the published scientific and patent literature on November 28, 2012 indicated that no small molecule CXCR6 antagonists have been reported. SciFinder was used to conduct the search. To follow up on the identification of MLS-0020556, a SciFinder substructure query not limited to CXCR6 relevance (Table 2) was conducted on December 4, 2012. This search identified 76 compounds and 12 associated references. No relevance to CXCR6 activity was noted. Although these references are not considered prior art for this program, they are discussed here for completeness. Briefly, one patent21 issued to colleagues at SBMRI have identified MLS-0020645 as a 4.67 μM inhibitor of HePTP relevant to leukemia. This was a patent claiming methods of treatment and did not disclose further SAR or examples related to the azabicyclo[3.3.1] scaffold. This compound was present in our MLSMR screening collection edition and was found to be not active. Four patents, written in Japanese, disclosed structures that the same cohort appears to have summarized in a manuscript.22 Compounds disclosed were of interest due to gastrointestinal prokinetic and dopamine D2 antagonist activity. The compounds (D2 compounds, see Table 1) contain the same azabicyclo[3.3.1] scaffold with bridgehead substitution involving an acid or ester linked via an alkyl chain. One additional reference was an analytical method development manuscript and the other four references disclosed different scaffolds and/or did not demonstrate utility of the azabicyclo[3.3.1] scaffold.

Table 2

Summary of Prior Art Search.

2.1. Assays

Table 3 summarizes details for the assays that enabled this probe discovery project. A detailed description of the Primary assay can be found in the Appendix at end of the probe report and in the PubChem AIDs listed in Table 3.

Table 3

Summary of Assays and AIDs.

2.2. Probe Chemical Characterization

b. Probe chemical structure including stereochemistry if known

The probe ML339 is the exo-positional isomer. Although it possesses two stereogenic centers, it is an achiral molecule due to a plane of symmetry.

Figure 1Structure of ML339

c. Synthesis and Structural Verification Information of probe SID 150864418 corresponding to CID 60202254 (See Scheme 1)

Scheme 1

Synthesis of ML339, conditions. i) 3,4,5-trimethoxybenzoyl chloride, Et₃N, CH₂Cl₂, 2 h, r.t.; ii) trifluoroacetic acid, CH₂Cl₂, 1 h, r.t.; iii) chloroacetyl chloride, DMAP (cat.), Et₃N, CH₂Cl₂, 2 h, r.t., 99%; iv) (B), Et₃N, CH₂Cl₂, 18 h, r.t., 24 % overall. (more...)

The exo-positional isomer of an analog (7) was confirmed by x-ray crystallography.23 This analog was synthesized from the same exo-building block as the probe, and there was no stereochemical ambiguity associated with the synthesis.

Figure 2X-ray Structure of 7

2.3. Probe Preparation

3.1. Summary of Screening Results

Results of Screening and Hit Validation

The screening funnel results are summarized in Figure 3. We completed a full-deck cell-based screen of the available MLSMR library of ~360,000 compounds at 20 μM final library compound concentration at 1% (v/v) final DMSO concentration for inhibitors of the human chemokine (C-X-C motif) receptor 6 (CXCR6) in CHO-K1 cells using a β-Arrestin-2 readout. Cells were pre-treated with compound before stimulation with of the endogenous CXCL16 ligand or addition of DMSO only for 90 minutes. Inhibition of β-Arrestin recruitment was measured using the β-galactosidase PathHunter Enzyme Complementation Assay Platform (DiscoveRx). The assay was validated using anti-hCXCR16 as an antagonist positive control tested in the presence of an EC₈₀ concentration of hCXCL16 (IC₅₀ = 2.0 nM; % response = 100%). Figure 3 summarizes the hit triage and prosecution. An initial 1305 hits were obtained at hit threshold of >40% inhibitory activity. Cheminformatic filtering was applied to eliminate known PAINS24 (Pan Assay Interference Compounds) and promiscuous compounds. Additionally, our experienced project chemists visually inspected the chemical structures of the filtered hits and eliminated any compounds they judged were a priori undesirable, intractable or were previously problematic in other screens and projects. The final number of hits thus selected for reorder was 929. We requested all of these as “cherry-picks” of fresh DMSO stock solutions from the NIH MLSMR (Molecular Libraries Small Molecule Repository) managed by Evotec (South San Francisco, CA) and received 854 of them for a 91.9% order fulfillment.

Figure 3

Summary of Screening Triage.

Of the 854 received cherry-picks, 563 confirmed in duplicate at the original 20 μM test concentration and 366 compounds showed IC₅₀ < 20 μM inhibition of CXCR6 receptor β-Arrestin recruitment. 176 compounds showed activity in a β-galactosidase enzyme assay using the same PathHunter reagents and were eliminated as interfering with the protein complementation system. The remaining 190 compounds were tested against two additional GPCRs using the same β-Arrestin platform from DiscoveRx, the human CCR6 chemokine receptor and the human non-related Apelin (APJ) receptor to determine overall GPCR selectivity. A single compound showed initial 6 μM IC₅₀ activity in the primary CXCR6 assay and was devoid of activity at inhibiting either the CCR6 or APJ receptors. This compound represented a chemically tractable scaffold and was advanced through multiple rounds of medicinal chemistry and structure-activity relationship (SAR) studies. These studies resulted in one first in class small molecule that was potent in antagonizing β-arrestin recruitment of the CXCR6 receptor and was inactive in at the CCR6, CXCR5, CXCR4 and APJ receptors. The synthetic strategies for this scaffold are described further.

3.2. Dose Response Curves for Probe

The multiple dose response titrations in Figure 4 of the probe ML339 against an EC₈₀ concentration of the natural ligand for hCXCR6, hCXCR4, hCXCR5 and hCCR6 receptors highlight the high degree of selectivity the compound has for the human CXCR6 receptor.

Figure 4

Specificity of ML339 Inhibition for the hCXCR6 receptor vs. human hCXCR4, CXCR5 and CCR6 receptors.

3.3. Scaffold/Moiety Chemical Liabilities

ML339 contains three methoxy groups that may compromise metabolic stability. This is consistent with the low metabolic stability in human and mouse liver microsomes see during in vitro ADME/T profiling (see Table 10).

Table 10

Summary of in vitro ADME Properties of probe ML339.

3.4. SAR Analysis

The screening program identified MLS-0020556 as a singleton hit. Given the ambiguous database structure, the exo/endo positional amide isomer question required resolution. This defined several Key Questions which our SAR program has addressed successfully. These included:

1. Are there commercially available analogs with similar activity?
2. Is the [3.3.1]azabicyclononane required for activity?
3. In the hit MLS-0020556, is the reverse amide substituent in the exo or endo position?
4. If the [3.3.1] system is required or preferred, is the exo or endo position required or preferred?
5. Is the 3,4,5-trimethoxy aryl motif required?
6. Does substitution of the benzamide group appended to the bridgehead nitrogen influence activity?

Tables 6, 7, and 8 summarize SAR, which addresses these questions.

Table 6

SAR of Purchased Analogs.

Table 7

SAR of Alternative Scaffolds.

Table 8

SAR of exo-[3.3.1] Scaffold Compounds.

Table 6 addresses Key Question #1. Briefly, this table showed that MLS-0020556 (Entry 1) was, in fact, a singleton hit. Five commercially available structural analogs (Entries 2–6) were identified and purchased along with a new batch of Entry 1 powder. The newly purchased analogs included the unsubstituted case (Entry 2) as well as cases with either fewer methoxy groups (Entries 4 and 6) or with substitution on the phenyl ring linked to the bridgehead (Entries 3–6). Test results showed no activity for Entries 2–6, but reconfirmed the activity of the parent hit (Entry 1). None of the compounds in Table 6 were active in the CCR6 or CXCR5 counterscreens. These data prompted us to explore the importance of the [3.3.1] system and address the exo vs. endo question.

Key Questions #2, #3 and #4 are addressed in Table 7. Overall, this table shows that the exo-[3.3.1] scaffold is required. Entries 7 and 8 were initially synthesized from a building block which was a 1:1 exo / endo mixture, and the two positional isomers were isolated by flash chromatography. As only the exo building block was commercially available, this separation was necessary to obtain both positional isomers from the same synthetic sequence. Subsequently, another batch of MLS-0020556 (Entry 13, Table 8) was synthesized from the pure exo building block. As this sample was identical to the Entry 7 compound, Entry 7, as well as the parent singleton hit (Table 6, Entry 1) was known to be exo and the Entry 8 compound, by default, was known to be endo. Entry 7 was further characterized by x-ray crystallography to unambiguously establish the exo positional isomer. Entries 9 and 10 were synthesized from the commercially available pure exo and pure endo building blocks. Comparison of Entries 7, 9, and 11 (n = 3, 2, 0, respectively, with no other changes) revealed about 4 times more potency for the [3.3.1] scaffold than for the corresponding [3.2.1] variant. The monocyclic analog (Entry 11) was inactive. Comparison of Entries 7 and 8 showed that, for the [3.3.1] scaffold, the exo positional isomer was required for activity. The same trend was noted for the less active [3.2.1] scaffold (Entries 9 and 10). None of the compounds in Table 7 were active in the APJ counterscreen. These data prompted us to explore the SAR at positions R₁ and R₂.

Key Questions #5 and #6 are addressed in Table 8. Overall, this table shows good opportunity for improving potency via modification of R₁, suggests requirement of the 3,4,5-trimethoxy substitution pattern at R₂, and identifies the new probe ML339. All compounds in Table 8 were prepared from the commercially available pure exo building block. A synthesized batch of the unsubstituted R₁ = R₂ = H case (Entry 12) reconfirmed the inactivity of the purchased batch (Table 6, Entry 2), and proved retroactively that the purchased Entry 2 material was, in fact, the exo positional isomer. The SAR at the R₂ position is clearly seen by comparing the set of five R₂ = 3,4,5-trimethoxy cases (Entries 13–17) with the corresponding five 3,4-dimethoxy (Entries 20–24), 3,5-dimethoxy (Entries 25–29), 3-methoxy (Entries 30–34), 4-methoxy (Entries 35–39), and 3,4-methyleneoxy (Entries 40–44) cases. For these analogous sets, activity was only observed for the trimethoxy cases. At the R₂ position, 4-Cl was about two-fold better than the 4-H case (Entries 15 and 13), whereas 4-F, 4-Me, and 4-OMe were about two to five fold worse than no substitution (Entries 14, 16, 17, and 13). Substitution at the 2-position was markedly better than at the 4-position, with the 2-OMe and 2-Cl cases both being 10-fold better than the corresponding 4-OMe and 4-Cl cases (Entries 19, 18, 17, and 15). The 2-Cl case (Entry 18) was clearly the best of all, and was therefore scaled-up to be nominated as ML339. The data for the submitted scaled-up 2^nd batch (IC₅₀ = 0.14 μM; % response 100%) are also shown in Table 1. None of the compounds in Table 8 were active in the CCR6, CXCR5 or APJ counterscreens when tested.

3.5. Cellular Activity

The murine assay was validated using anti-mCXCR16 as an antagonist positive control tested in the presence of an EC₈₀ concentration of mCXCL16 (IC₅₀ = 1.5 nM; % response = 100%). Probe ML339 was shown to be 100-fold less active at the murine CXCR6 receptor using the DiscoveRx β-arrestin cell-based assay format (Table 9). This is an interesting observation because we previously demonstrated that the human and murine ligands have activity within 2-fold at each respective species-based receptor in this assay format (as agonist). Within the CPDP, we elected to state that a suitable chemical probe would have equipotent to 3-fold species variance based on comparing assay development EC₅₀ values between the human and murine CXCR6 receptors and ligands (agonist mode). The probe ML339 will now be able to assist in determining whether activity at hCXCR6 is based on a species specific allosteric vs. orthosteric functional inactivation of the receptor. While the need for activity at mCXCR6 and stated 3-fold stringency is not critical for our human cell in vitro PCa chemotaxis and related post-probe studies, it would assure the final probe series will be useful to future Investigators interested in the role of CXCR6 in murine lymphatic cells for immunology-based studies. Although this observation limits the current probe utility outside of immunology, it fully is adequate for planned post probe PCa-based human cell optimization and for use in associated murine xenograft studies. Furthermore, the set of compounds shown in Table 9 gave no response when screened against CXCR5 (Table 8) and CXCR4, which underscore their selectivity. Thus, a suitable ‘first in class’ pharmacological tool, useful for exploring the role of CXCR6 related PCa, has been discovered.

Table 9

SAR of Selected Compounds as Mouse CXCR6 Antagonists.

3.6. Profiling Assays

As a pro forma activity, the SBCCG is committed to profiling all final probe(s) compound(s) and in certain cases key informative analogs in the PanLabs full panel as negotiated by the MLPCN network. Additional commercial profiling services will be considered for funding by SBCCG as deemed appropriate and informative. ML339 was evaluated in a detailed in vitro pharmacology screen as shown in Table 10:

ML339 can achieve concentrations >10× IC₅₀ in aqueous buffer between a pH range of 5.0–7.4.

The PAMPA (Parallel Artificial Membrane Permeability Assay) assay is used as an in vitro model of passive, transcellular permeability. An artificial membrane immobilized on a filter is placed between a donor and acceptor compartment. At the start of the test, drug is introduced in the donor compartment. Following the permeation period, the concentration of drug in the donor and acceptor compartments is measured using UV spectroscopy. Consistent with its solubility data, ML339 exhibits good permeability with increased pH of the donor compartment.

Plasma protein binding is a measure of a drug’s efficiency to bind to the proteins within blood plasma. The less bound a drug is, the more efficiently it can traverse cell membranes or diffuse. Highly plasma protein bound drugs are confined to the vascular space, thereby having a relatively low volume of distribution. In contrast, drugs that remain largely unbound in plasma are generally available for distribution to other organs and tissues. ML339 was highly plasma protein bound.

Plasma stability is a measure of the stability of small molecules and peptides in plasma and is an important parameter, which can strongly influence the in vivo efficacy of a test compound. Drug candidates are exposed to enzymatic processes (proteinases, esterases) in plasma, and they can undergo intramolecular rearrangement or bind irreversibly (covalently) to proteins. ML339 showed good stability in human plasma, and moderate stability in mouse plasma.

The microsomal stability assay is commonly used to rank compounds according to their metabolic stability. This assay addresses the pharmacologic question of how long the parent compound will remain circulating in plasma within the body. ML339 is almost completely metabolized in human and mouse liver microsomes within 1 hour.

ML339 shows no toxicity (>50 μM) towards immortalized Fa2-N4 human hepatocytes.

Profiling against other GPCRs

The probe, ML339 (CID 60202254), was submitted to the Psychoactive Drug Screening Program (PDSP) at the University of North Carolina (Bryan Roth, PI) and the data against a GPCR binding assay panel is shown in Figure 5. Preliminary results indicate ML339 shows a low level of promiscuity across a range of GPCRs, with the only activity appearing against 5-HT2B and DAT. It is not known whether these activities in binding assays are translated into functional modification of the activities of these receptors.

Figure 5

Profile of ML339 against the Psychoactive Drug Screening Program (PDSP) % inhibition at 10 μM.

4.1. Comparison to existing art and how the new probe is an improvement

ML339 is the first reported selective small molecule CXCR6 antagonist. It represents a meaningful improvement in potency over the HTS hit compound, maintains good GPCR selectivity, and offers an additional starting point for lead optimization efforts and eventual in vivo studies.

4.3. Planned Future Studies

Representative samples and probe ML339 are under evaluation in a radioligand binding assay format along with chemotaxis in PC3 cells in the Taichman laboratory. Future, post-probe research may include additional SAR, functional assay, pharmacology, and in vivo experiments. Future rounds of SAR may include further exploration of the R₂ substituent, alternatives to the trimethoxy benzamide at R₁, and modification of the glycine-like linking unit between the bridgehead and the R₂ amide group. Consistent with the goal of achieving proof of concept in an in vivo model, structural perturbations which may enhance metabolic stability and activity against mouse CXCR6 will be emphasized. Profiling assays may be expanded to include radioligand binding, calcium mobilization, cAMP, and chemotaxis. Pharmacological studies may be extended to include multipoint time-based profiling of chemical and metabolic stability, including metabolite identification.

6.1. Assay Details

6.2. ¹H NMR and LC-MS spectra of ML339

¹H NMR Spectrum of ML339 (500 MHz, DMSO-d₆)

¹³C NMR Spectrum of ML339 (125 MHz, DMSO-d₆)

LC/MS for ML339.

MS spectrum for M339