Discovery of Human PIM Kinase Inhibitors as a Class of Anthelmintic Drugs to Treat Intestinal Nematode Infections

Abstract

Soil-transmitted helminth (STH) infections affect one-fourth of the global population and pose a significant threat to human and animal health, with limited treatment options and emerging drug resistance. Trichuris trichiura (whipworm) stands out as a neglected disease, necessitating new drugs to address this unmet medical need. We discovered that several different chemical series of related human Provirus Integration sites for Moloney murine leukemia virus (PIM) family kinase inhibitors possess potent anthelmintic activity by using whole-worm motility assays. Systematic structure–activity relationship (SAR) studies based on the pan-PIM kinase inhibitor CX-6258 were conducted to identify compounds displaying improved in vitro motility inhibition of both adult hookworm (Ancylostoma ceylanicum) and adult whipworm (Trichuris muris) nematodes. A broad kinase selectivity screen of >450 human kinases confirms PIM1 kinase and others as potential targets for CX-6258 and analogues thereof. In addition, we demonstrated that CX-6258 significantly reduced worm burden and egg counts in the T. muris infection model of mice, establishing it as a new oral small molecule anthelmintic therapeutic.

Graphical Abstract

Approximately 24% of the global population is infected with soil-transmitted helminths (STH), mostly affecting communities with poor access to proper hygiene, sanitation, and clean water.¹ These infections reside in the intestinal tract, routinely causing multiple comorbidities and significant growth impairment of infected children.² Although the number of infections caused by STH remains high, there are only a limited number of currently available therapeutic options recommended by the World Health Organization (WHO), including albendazole and mebendazole as well as levamisole and pyrantel. These target only two therapeutic targets: β-tubulin (albendazole and mebendazole are inhibitors) and the nicotinic acetylcholine receptor (pyrantel and levamisole are agonists). However, treatments with these drugs suffer from low overall effectiveness in some cases due to (i) the emergence of drug resistance in veterinary medicine³ and the predicted high risk of resistance as a result of ongoing preventive chemotherapy in humans.⁴ Hence, the frequency of the β-tubulin resistance allele was significantly increased following Mass Drug Administration (MDA), indicating that drug pressure could result in a selection for benzimidazole resistance in human GINs,⁵ (ii) a lack of pan-species activity,⁶ which is required in light of the high rate of multispecies infection, thus establishing an urgent need for new anthelminthic agents and with new modes of action. Recent promising trials with the anthelmintic emodepside indicate that higher doses⁷ may be more efficacious and that multiple doses (used in veterinary treatments) have acceptable safety and tolerability profiles.⁸ However, some natural Caenorhabditis elegans strains demonstrated resistance to emodepside,⁹ suggesting the possibility of emodepside resistance in parasitic nematodes. Furthermore, emodepside is not yet an approved drug. Thus, additional new anthelmintics that are effective with a single dose are urgently needed to ensure that long-term WHO goals for STH elimination can be met.

The whipworm, Trichuris trichiura, the cause of human trichuriasis, is one of three well-documented STHs included in the WHO list for neglected tropical diseases (NTDs). To control STHs, the WHO recommends mass drug administration (MDA), which relies mainly on the two benzimidazole drugs for treating school-age children in STH-endemic countries. However, trichuriasis is not as well managed by MDA relative to the other two groups of STH, with only a 2% decline in prevalence between 2005 and 2015 versus a 10% decline for that same period in ascariasis (caused by another major STH species, Ascaris lumbricoides).¹⁰ Furthermore, single-dose treatment of T. trichiura infections is not efficacious with approved anthelmintics, so multiple doses over several days are necessary.¹¹ Taken together, there is a large unmet medical need for new anthelmintic drugs specifically targeting the intestinal worms T. trichiura (whipworm) but also hookworm (Ancylostoma ceylanicum).

We previously reported on the discovery of new classes of small molecule anthelmintics, where we first validated ten metabolic “chokepoint” enzymes that are taxonomically conserved across the phylum Nematoda.¹² Screening known inhibitors for each of the ten chokepoint targets, we identified hit compounds with pan-species anthelmintic activity in a whole-worm motility inhibition assay. One of the 10 enzyme targets of interest was the lipid kinase 1-phosphotidylinositol-4-kinase, or PI4K, where we established the mixed mTOR/PI3K/PI4K inhibitor Torin-1 as an effective treatment for A. ceylanicum hookworms, Trichuris muris whipworms, and Brugia pahangi filarial worms in vitro. Further testing of other known mTOR and PI3K inhibitors, such as Torin 2, led to unclear structure–activity relationship (SAR), so we expanded our evaluation of existing inhibitors to include a larger selection of diverse chemical series. Although we demonstrated good efficacy of Torin-1 in the treatment of hookworm infection in hamsters,¹² we could not identify improved inhibitors even after synthesizing multiple analogues. Using a systematic approach guided by the kinase selectivity profile of these PI3K and PI4K inhibitor scaffolds, we identified the mixed PI3Kγ/CK2 kinase inhibitor AS-252524, which reduced whipworm motility almost 90%@30 μM. Based on structural similarity to AS-252524, we identified multiple other compounds, which turned out to be potent but mixed PIM1 and PIM2 kinase inhibitors and were subsequently found to be very active in our whole-worm motility assays.

The Provirus Integration site for Moloney murine leukemia virus (PIM) kinases is a family of highly conserved serine/threonine kinases that are known to promote cell survival and growth through transcriptional activation of cell proliferation genes.¹³ As there is a wide variety of human PIM kinase inhibitors and chemical scaffolds reported, it is an attractive target for potential drug repositioning.¹⁴ One inhibitor, the furanyl-indolinone CX-6258, was first reported by Ryckman et al. in 2006 as a pan-PIM kinase inhibitor that showed in vivo anticancer activity against PIM kinase-related xenograft models.¹⁵ The discovery of CX-6258 was accomplished first through high-throughput screening, which established the necessity of an oxindole group with the nitrogen in the lactam being free to participate in hydrogen binding, which was later confirmed via X-ray cocrystal structure showcasing interaction with the key PIM kinase hinge binding residue Asp186.¹⁶ Herein, we report on the repurposing and optimization of CX-6258 derived inhibitors as new anthelmintic drugs against parasitic intestinal whipworms and hookworms.

RESULTS AND DISCUSSION

Lead Identification and Optimization of PIM Kinase Inhibitors as Anthelmintics.

We previously published on the discovery of several new small molecule anthelmintic drugs using a whole-worm motility inhibition assay, using the “Worminator” system¹⁷ for testing compounds against 10 phylogenetically conserved “chokepoints” enzyme targets.¹² Two of the major enzyme classes discovered were several protein and lipid kinases.¹² We investigated several PI3K and PI4K inhibitor scaffolds and identified the inhibitor AS-252424 with excellent worm motility inhibition against T. muris (Figure 1A). AS-252424 is a potent dual inhibitor of CK2α/PI3Kγ with an IC₅₀ of 30 nM and 30-fold selectivity for PI3Kγ over PI3Kα, but only a weak inhibitor of PI3Kδ and β.¹⁸ However, the thiazolidinedione scaffold of AS-252424 is a known Pan-Assay Interference Compounds (PAINS)¹⁹ compound and is not attractive for further investigation. Therefore, we conducted a similarity search to find related compounds amenable to medicinal chemistry studies. Interestingly, we identified structurally similar compounds that were known PIM1 or PIM2 kinase inhibitors (Figure 1A), including SMI-4a, SMI-16a, and AZD1208 (Figure 1A). As shown in Figure 1B, SMI-4a and SMI-16a displayed moderate activity (71 and 57% motility inhibition, @48 h, respectively) while AZD1208 was inactive. Since AS-252424 also potently inhibits CK2α kinase, we also screened other known inhibitors which are structurally diverse, such as CX-4945 (Silmitasertib), TTP22, and IPI-549. However, when evaluated against adult T. muris, the CK2α inhibitors did not show measurable inhibition of worm motility (Figure 1). Upon further testing of other known CK2α inhibitors, we found multiple compounds that were potent inhibitors of worm motility in our whole-worm assay.

Figure 1.

Evaluation of known human PIM and CK2α kinase inhibitors against the adult whipworm T. muris. (A) Structures of AS-252424 and other kinase inhibitors returned from structure similarity and pharmacophore searches. (B) Motility inhibition (%, relative to dimethyl sulfoxide (DMSO) controls) testing at 30 μM showed that SGI-1776, CX-6258, and TCS-PIM had excellent inhibition of T. muris. ND = not determined. Error bars indicate standard error. The data represent the average of six adult whipworms. All compounds were tested for significance against matched DMSO-only controls with a one-way analysis of variance (ANOVA) test, followed by a Dunnet test for multiple comparisons (*P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001).

With this information in hand, we assessed other PIM kinase inhibitors derived from different chemical scaffolds, including TP-3654, SGI-1776, CX-6258, and Pim1/AKK1-IN-5 (Figure 1A), where we discovered SGI-1776, CX-6258, and TCS-PIM1 to be superior in their anthelmintic activity compared to the others tested with all three effectively blocking T. muris worm motility 100%@30 μM (Figure 1B). With these promising results, we engaged in a systematic medicinal chemistry lead optimization effort on CX-6258 to develop new broad-spectrum antihelmintic drugs targeting intestinal nematodes. In our studies, we are seeking significant improvements in the in vitro potency and pharmacokinetics (PK) of CX-6258 to identify new analogues with good efficacy in vivo. In parallel, we carried out the optimization of SGI-1776, which will be communicated in a following manuscript.

Design and Synthesis of CX-6258 Analogues and Derivatives.

In an initial strategy to develop structure–activity relationship (SAR) data and identify improved inhibitors with regard to anthelmintic activity, we divided the CX-6258 scaffold into three main portions (Figure 2) comprising: (A) the left-hand side (LHS) basic amine, (B) the central phenyl-furan ring system, and (C) the right-hand side (RHS) oxindole moiety. All compounds were obtained and purified, having a mixture of both E and Z alken isomers as determined by ¹H NMR and high-performance liquid chromatography (HPLC)/MS, which varied depending on the specific compound. For example, we determined the E/Z ratio of CX-6258 to be approximately 2:1 in both the whipworm and hookworm culturing media used for the motility Worminator assays (Figure S1). We found that it was not possible to isolate pure E or Z isomers due to rapid equilibrium. The synthesis of initial inhibitors 1, 2, 3a, and 3b is shown in Scheme 1A. The central phenyl furanyl ring system was constructed through a Suzuki coupling reaction with 3-carboxyphenylboronic acid and 5-bromofuran-carbaldehyde. A Knoevenagel type condensation of aldehyde 23 under basic conditions with 5-chlorooxindole gave intermediate 24. The addition of N-Boc-4-diazapane was then performed through amide formation via coupling with 24 using hexafluorophosphate benzotriazole tetramethyl uronium (HBTU) to give Bocprotected compound 1. The removal of the Boc group using standard 4 N HCl in dioxane returned compound 2, which was diversified into compound 3a by acetylation with acetic anhydride and 3b by reductive amination with isovaleraldehyde. Further exploration of the scaffold for determining SAR was accomplished via amide coupling of intermediate 24 with various basic diamines facilitated by hexafluorophosphate azabenzotriazole tetramethyl uronium (HATU) or HBTU, which generated analogues 4–6 (Scheme 1B). Additional analogues 7–22 were synthesized by reaction of intermediate 23 with various substituted oxindoles to yield the corresponding carboxylic acids 25a–g, which were converted to the target amide analogues 7–22 (Scheme 1C).

Figure 2.

Chemistry strategy for developing SAR based on (A) the left-hand side (LHS) basic amine, (B) the central phenyl-furan ring system, and (C) the right-hand side (RHS) oxindole moiety of CX-6258.

Scheme 1.

Synthesis of CX-6258 Analogues^a

^a(A) Synthesis of substituted homopiperidine analogues 3a and 3b; (B) synthesis of methylamino piperidine analogues 4–6; (C) synthesis of oxindole modified derivatives 7–22.

Biological Evaluation of New Compounds as Anthelmintics.

There are significant challenges to new anthelmintic drug discovery, including (1) the assay utilized is a whole organism assay and is based on worm motility, so there is significant variability upon repeat assays, (2) the limitation in the number of adult worms harvested from the intestine of infected animal models (mice for whipworm and hamsters for hookworm), and each adult worm is manually picked into the test well. Thus, we were compelled to adopt single-concentration assay as a primary screening method for the newly synthesized analogues, followed by IC₅₀ determination only for active analogues, therefore keeping utilizing the adult worms and infected animals held to a minimum, (3) the worm cuticle is extremely difficult to penetrate^20–22 and thus increased or decreased inhibition can be due to either changes in permeability or target inhibition. These challenges complicate SAR analysis and necessitate the adoption of medicinal chemistry strategies more suited to address these limitations when directly utilizing adult parasitic nematodes. Another challenge for the rational design of improved inhibitors is that the target of these inhibitors is not established, although we are operating on the premise that the compounds target one or more PIM kinases for which there are orthologues in both whipworm and hookworm (discussed later). Furthermore, the exposure time of compounds to worms in the intestinal tract is limited due to many factors, including metabolism and/or being cleared from the gastrointestinal tract by bowel movement or entering the systemic circulation. Thus, compounds that induce irreversible damage to the worms are favored.

To ensure that CX-6258 was causing irreversible damage to whipworms, we incubated whipworms in the presence of the drug for 24 h prior to replacing with fresh media. We recorded motility 4 and 24 h after removing CX-6252 and saw no recovery in whipworm motility after removing CX-6258 (Figure S2), indicating irreversible damage to whipworms. Such an effect could be explained, for example, by targeting pathways essential for survival or by a lack of whipworms to metabolize or excrete CX-6258; as a result, the drug is accumulated within the worm body in enough molarity to disrupt the target. With these key results now established, we tested all new synthetic analogues for their ability to inhibit worm motility using CX-6258 as a positive control.

All compounds were first tested in a primary assay against the whipworm target species, T. muris, at a single concentration (30 or 50 μM, motility measured @24 and @48 h). Of the 61 compounds tested, 10 analogues had greater than 50% motility inhibition at 48 h. To determine if these inhibitors encompass pan-intestinal anthelmintic activity or selective whipworm (clade I of a V clade phylogeny²³) activity, we next tested them against A. ceylanicum hookworms (clade V of Nematoda phylogeny). Of the compounds tested against adult A. ceylanicum (30 μM, motility measured @24 and @48 h), 8 showed better than 50% motility inhibition, with two analogues being hookworm selective (8 and 19; Table S1). We found that changing the N-methyl substitution on the 1,4-diazapane ring to an acetyl (3a), a longer alkyl chain (3b), or a carbamate (1) negatively impacted compound activity in the STH motility assay (Figure 3). Remarkably, we also discovered that removing the terminal diazapane N-methyl group of CX-6258 as in 7 resulted in a complete loss of activity against T. muris, but not against A. ceylanicum. In contrast, removing the piperidine N-methyl of compound 4 as in 6 resulted in significantly increased inhibition of both T. muris, and A. ceylanicum motility, almost equivalent to the CX-6258 potency (Figure 3).

Figure 3.

SAR of selected LHS analogues, specifically investigating the relationship between changing the amine moiety and worm motility at 48 h for both whipworm (T. muris, tested at 50 μM) and hookworm (A. ceylanicum, tested at 30 μM). Values represent motility inhibition relative to the DMSO controls. Error bars indicate the standard error. The data represent the average of six adult whipworms. All compounds were tested for significance against matched DMSO-only controls with a one-way ANOVA test followed by a Dunnet test for multiple comparisons (*P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001).

Compared with the LHS modifications, changes to the RHS oxindole moiety were generally better tolerated. For example, the removal of the chlorine substituent (11) or its movement to the 7′ position from the 5′ position (8) provided neither an increase nor decrease to activity. The replacement of the chlorine with an amino group at the 5′ position (7) or the 6′ position (9) completely rendered the compounds inactive against both species. Exchanging the 5′ position chlorine with a methyl group (10) also inactivated the inhibition of both species, but changing the methyl group in the 7′ position (12) increased activity against A. ceylanicum while retaining activity against T. muris (Figure 4).

Figure 4.

SAR of selected RHS analogues, specifically showing the relationship between changing the substituents of the oxindole and worm motility at 48 h for both whipworm (T. muris, tested at 50 μM) and hookworm (A. ceylanicum, tested at 30 μM). Results indicate that substituents in the 7′ position are generally well tolerated but not the addition of an amino group. Values represent motility inhibition relative to DMSO controls. Error bars indicate standard error. The data represent the average of six adult whipworms. All compounds were tested for significance against matched DMSO-only controls with a one-way ANOVA test followed by a Dunnet test for multiple comparisons (*P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001).

Armed with this knowledge, we designed “combination” analogues that leveraged the LHS and RHS moieties from the compounds that were most active. As the best compounds contained either 7-chlorooxindole or 7-methyloxindole from the right-hand side of the SAR analysis, we synthesized a series of compounds with the most active diamines (Figure 5). Interestingly, the addition of either 7-chloroxindole or 7-methyloxindole failed to improve the activity of compounds with N-methyl-4-piperidinylmethanamine (15, 17). The addition of the more active oxindole moieties generally increased the activity of previously poor performing diamines 1,4-diazepane and 4-(aminomethyl)piperidine in T. muris (16, 21, and 22). Depending on the compound, it has either a detrimental or no effect on the activity against A. ceylanicum.

Figure 5.

“Combination” analogues against whipworm (T. muris) and hookworm (A. ceylanicum). Compound IDs are indicated in bold numbers in each box, and the box colors indicate the motility inhibition of each compound after 48 h (50 μM in whipworm, 30 μM in hookworm). LHS substitutions are indicated for each column, and RHS substitutions are indicated for each row.

Cytotoxicity evaluation using a HepG2 MTT viability assay (Table 1; see the Experimental Section) revealed clear SAR with respect to cellular toxicity. For example, when the chlorine substituent on the oxindole ring is moved from the 5′ position (CX-6258) to the 7′ position (8), we see a 2.4-fold increase in cytotoxicity (measured by CC₅₀) from 37.5 to 15.5 μM. When a methyl group replaces the 7′ position chlorine (12), cytotoxicity is eliminated for concentrations up to 50 μM. Changes to the N-methyl homopiperazine moiety by replacement of the terminal tertiary amine with a secondary amine in a homopiperazine (22) or 4-piperidinemethanamine (6) increased cytotoxicity (to 13.0 and 24.5 μM, respectively). The increase in cytotoxicity remained even when changing the subsistent on the oxindole ring to a 7′ position chlorine (16) or 7′ position methyl (21).

Table 1.

IC₅₀ Values against T. muris Whipworms and A. ceylanicum Hookworms and Human HepG2 CC₅₀ Cytotoxicity Values for CX-6258 and Top Analogues^a

	T. muris		A. ceylanicum		HEPG2 cells
	IC50 μM (95% CI)		IC50 μM (95% CI)		CC50 μM (95% CI)
compound	24 h	48 h	24 h	48 h	24 h
CX-6258	17.8 (11–27)	14.7 (9–23)	31.6 (11–97)	5.5 (2.5–12)	37.5 (26–49)
6	41.5 (28–62)	57.6 (32–111)	6.0 (3.7–9.5)	5.6 (3.8–8.3)	24.5 (18–31)
8	42.0 (30–58)	>90b	40.0 (28–56)	21.0 (15–30)	15.5 (12–19)
11	>90b	>90b	25.2 (16–37)	8.8 (5–12)	22.2 (17–27)
12	17.2 (10–28)	14.7 (8–25)	7.7 (3–19)	2 (1.1–19))	>50
16	64.9 (50–84)	27.7 (15–49)	15.9 (11–22)	16.9 (10–25)	14.5 (12–17)
21	50.0 (30–84)	28.7 (18–27)	32.2 (20–50)	23 (15–33)	11.4 (9–14)
22	53.3 (32–91)	61.3 (37–106)	73.3 (34–181)	20.2 (11–38)	13.0 (9–17)
albendazole	>300b	>300b
ivermectin	>300b	89.6
emodepside	0.0143 (0.007–0.022)	0.0345 (0.01–0.06)	0.0033 (0.002–0.005)	0.0011 (0.0004–0.002)

^aValues for three currently used anthelmintic drugs are presented at the bottom of the table.

^bSome of the IC₅₀s were outside the range of the tested dose response, and to avoid extrapolation, we report the highest dose tested in the dose response.

To rank compounds based on the concentrations associated with 50% motility inhibition (IC₅₀), we tested top analogues for their concentration response against T. muris whipworms and A. ceylanicum hookworms. The 48 h IC₅₀ values against T. muris whipworms ranged from 14.7 μM in CX-6258 and 12 to >90 μM in 8 and 11. On the other hand, the IC₅₀ values against A. ceylanicum hookworms ranged from 2 μM in 12 and 5.5 μM in CX-6258 to 23 μM in 21 (Table 1).

It was established that many of our most potent CX-6258 analogues had comparable IC₅₀ values to CX-6258 itself in whipworm, and substantially improved IC₅₀ values in hookworm, although the values were substantially higher than for emodepside (Table 1). All compounds had better IC₅₀ values for hookworm compared to whipworm, with 12 and 22 being the best performing analogue in both species.

Differential Activity Is Potentially Related to Worm Cuticle Membrane Penetration and Uptake.

In previous studies, we reported that the differential activity of some compounds for different species may be a result of poor permeability of the compounds through the worm’s cuticle.^20–22 To determine the permeability of CX-6258, we utilized a liquid chromatography–mass spectrometry (LCMS)-based bioaccumulation assay to determine the compound’s relative concentration inside the worms. Freshly harvested worms were incubated in a culture medium with 100 μM from each compound for six h and then washed with phosphate-buffered saline to remove excess drugs attached to the cuticle. The highest concentration found inside T. muris was 50 μM for 12 compared to CX-6258 with a concentration of 12 μM (Table 2).

Table 2.

Bioaccumulation of Compounds Inside Whipworm and Hookworm Following 100 μM Incubation for 6 h (n = 3 for All Comparisons)^a

	T. muris		A. ceylanicum
compound	Avg conc (μM)	Std error	Avg conc (μM)	Std error
CX-6258	12.4	1.4	0.0	0.0
12	50.5	3.0	3.0	1.1
22	28.6	4.2	0.0	0.0

^aCompound 12 was the only compound detected in the A. ceylanicum samples, while all three compounds were present in T. muris.

One possible explanation for the increase in relative concentration in 12 may be that the change from a 5′-chlorine on the oxindole moiety to a 7′-methyl improves uptake through the cuticle. Interestingly, CX-6258 was not detected in the A. ceylanicum samples, most likely suggesting that the 6 h bioaccumulated concentration is below the detection level as CX-6258 was found to be active against the hookworm when scored at 24 h of exposure. However, analogue 12 was detected in A. ceylanicum at a low concentration (3.0 μM). The low detection of CX-6258 may also explain the reduced 24 h in vitro activity of CX-6258 against A. ceylanicum with IC₅₀ of 31.6 μM compared to 17.8 μM in T. muris (Table 2).

Evolutionary Conservation and Gene Expression of PIM Kinases across the Life Cycles.

PIM kinases are evolutionarily conserved in eukaryotes, including the nematode C. elegans,²⁴ indicating that existing mammalian PIM kinase inhibitors can have anthelmintic activity. This was validated in this study, as we demonstrated the anthelmintic activity of a subset of the tested mammalian PIM inhibitors (Figure 1). This concept has been further supported by Brehm and co-workers,²⁵ who reported PIM kinase inhibitors SGI-1776 and CX-6258 had a detrimental effect on both the larvae and stem cells of the tapeworm Echinococcus multilocularis.

In order to identify and classify kinases in the target nematode species, OrthoFinder²⁶ was used to define orthologous protein families (OPFs) across whipworms (T. muris and T. trichiura), hookworms (A. ceylanicum and Necator americanus), the free-living model nematode species C. elegans, and mammalian host species (human and mouse), using the complete deduced protein sets from each species. We identified the orthologues of the human PIM kinases in both A. ceylanicum and T. muris in one OPF (OG0000556), which included human PIM-1, PIM-2, and PIM-3, as well as one orthologue in T. muris (TMUE_3000011049) and three orthologues in A. ceylanicum (ACEY_02787, ACEY_06843 and ACEY_07950). Helminth members of this OPF also had the top BLAST sequence similarity for the E. multilocularis putative CX-6258 target EmuJ_000197100²⁵ (TMUE_3000011049, E = 8.06 × 10⁻⁵⁸ and ACEY_07950, E = 1.01 × 10⁻³⁹). All of the OPF membership data are available in Table S2.

Evaluation of Human PIM and CK2α Kinase Activity for CX-6258 and Selected Analogues.

To determine whether our compounds had selective activity against worm targets, we tested CX-6258 and selected analogues against human CK2α and PIM 1, 2, and 3. In summary, the data (Table 3) show that 6, 12, 16, and 22 have lowered inhibition of all human PIM kinases relative to parent CX-6258. Since the whipworm motility IC₅₀ of 12 is the same as CX-6258, the 100-fold loss of human PIM1 and PIM3 kinase inhibition from 12 might suggest we are demonstrating selectivity for the worm kinase(s) but not necessarily human PIM orthologues. Compound 6 also returns an IC₅₀ in human PIM1 similar to that of CX-6258 and retains activity against hookworms similar to that of CX-6258 but has reduced activity against the whipworm. It is noteworthy that 16 only has a 10-fold decrease in kinase IC₅₀ but has higher worm IC₅₀, and 22 is >100-fold weaker for human PIM kinase inhibition, which is also correlated with weaker worm motility than 16 and 12.

Table 3.

Human Kinase Inhibition Compounds

	human kinase IC50a (nM)
compound	CK2α	PIM1	PIM2	PIM3
CX-6258	3901	0.69	41.4	>0.508
6	1431	1.00	429.2	1.80
12	>10,000	69.5	4996	9.88
16	6009	9.46	>10,000	7.54
22	>10,000	93.4	>10,000	20.3
GW5074	128.7
staurosporine		1.93	18.6	0.113

^aCompounds were tested for their inhibition using 10 different doses with 3-fold serial dilution starting at 10 μM.

Evaluation of Kinase Selectivity for CX-6258 in a Broad Panel of All Human Kinase Families.

To elucidate the full range of potential kinase targets for inhibitors, we tested CX-6258 for inhibition against a broad panel of 485 human kinases (SelectScreen Kinase Profiling Services, Thermo Fisher Scientific, Madison, WI). We determined the specific kinase activity of CX-6258 at a concentration of 500 nM, where it was found to inhibit several human kinases by over 80%, including PIM-1, PIM-3, AAK1, MYLK4, STK17A, GSG2 (haspin), and MAPK15 (ERK7), and an additional 26 kinases inhibited between 50 and 80% (Table S3). The identification of PIM kinases as targets was supportive of the known “pan-PIM kinase” activity of CX-6258¹⁵ (PIM-2 was not included in the assay of available kinases), with previous activity against FLT-3¹⁶ and Haspin²⁷ having also been reported in the literature. This study establishes additional human kinase targets of CX-6258 such as AAK1, CLK2, and others, providing evidence for broad kinase selectivity and a higher potential for targeting hookworm and whipworm kinase targets. All human kinase assay results are provided in Table S3.

Compound Solubility, Permeability, and Pharmacokinetics.

T. muris and A. ceylanicum are intestinal parasites, where adult worms occupy and propagate in the large intestine (T. muris) or the small intestine (A. ceylanicum). To predict whether CX-6258 and the best analogues, 12 and 22, remain in the intestine instead of being metabolized or eliminated immediately from lack of solubility, we pursued solubility testing in simulated gastric fluid (SGF) and simulated intestinal fluid (SIF). As expected, all three compounds were more soluble in SGF (pH = 1.2) than in SIF (pH = 6.8) since they all contain an ionizable basic amine that would be protonated at low pH. We determined that CX-6258 was approximately 30 times more soluble in SGF than in SIF, 12 was ~16 times more soluble, and 22 was ~8 times more soluble. While all of the inhibitors were more soluble in SGF, changing the 5-chloro to a 7-methyl in 12 more than doubled the solubility in SIF while also increasing the solubility in SGF. Removing the N-methyl group further improved the solubility in 22 over that of CX-6258 in SIF, though the solubility in SGF remained constant (Table 4). Increased solubility in the gastric fluid predicts better or more rapid transport of the compound into the gastrointestinal tract, where these nematodes reside. While better solubility in the intestinal fluid could indicate a longer residence time versus elimination in the bile and feces, it could also suggest increased potential for oral penetration, which is undesired as a therapeutic for intestinal worms.

Table 4.

Kinetic Solubility Calculated in Simulated Gastric Fluid (SGF) and Simulated Intestinal Fluid (SIF) for CX-6258, 12, and 22^a

		kinetic solubility
compound	media	μM	μg/mL
CX-6258	SGF	151	69.7
	SIF	5.05	2.33
12	SGF	190	83.9
	SIF	11.9	5.26
22	SGF	149	69
	SIF	18.7	8.67

^aAll compounds were more soluble in SGF than SIF.

To predict the intestinal oral bioavailability and efflux of the CX-6258, 12 and 22 were tested in a cellular permeability assay using MDR1-MDCK II cells. The results indicated that CX-6258 and both analogues have low permeability (Table 5), with the P_app (A–B) being significantly less than 1, especially CX-6258. In the case of intestinal nematodes, low predicted permeability is preferred and suggests that the compounds will have reduced oral absorption to the host intestinal cells, which also may result in compound accumulation in the lumen of the intestinal tract where gastrointestinal nematodes reside. In contrast to CX-6258 and 12, compound 22 exhibited a high efflux ratio/P_app (B to A) of 10.3, signifying that 22 is not predicted to lack oral bioavailability. Taken together, these results predicted that CX-6258 and 12 have absorption, distribution, metabolism, and excretion (ADME) properties that are suited for inhibiting STH parasites with limited oral absorption and extended half-life in the intestines.

Table 5.

Cellular Permeability and Efflux of Selected Compounds in MDR1-MDCK II Cells^a

	mean Papp (10−6 cm/s)			mean solution recovery %		Rank
compound	A to B	B to A	efflux ratio	A to B	B to A	P app
nadolol	0.050	NDb	ND	83.1	ND	low
metoprolol	25.6	ND	ND	97.4	ND	high
digoxin	0.151	3.43	22.8	77.5	92.7	low
CX-6258	0.049	0.0897	1.84	5.4	20.2	low
12	0.132	0.212	1.61	9.1	33.6	low
22	0.305	3.15	10.3	27.9	36.9	low

^aCX-6258 and analogues 12 and 22 exhibited low permeability.

^bND = not determined.

Prior to efficacy studies, we evaluated the new lead compound 12 for its pharmacokinetics (PK) in mice. The mouse PK for CX-6258 has already been published elsewhere.¹⁵ As shown in Table 6, compound 12 is rapidly cleared from the mouse plasma after an intravenous (IV) dose (3 mg/kg) having compound exposure out to 8 h and a high volume of distribution (V_d = 11) but with a very short 0.9 h terminal half-life (T_1/2) (Table S4). However, the T_1/2 after oral dosing is much higher at 4.9 h, and promisingly, the oral bioavailability is very low, which is desired for an anthelmintic with an F = 6.7% (Table 6) and is detected in the feces (Table S5). Like compound 12, CX-6258 has significantly lower plasma clearance but increased oral bioavailability relative to 12 (F = 23%). It is notable that CX-6258 has an incredibly high V_d of 85 L, which explains the high t_1/2 of 27 h and suggests the drug is highly absorbed in tissues. Taken together, these results suggest that both compounds 12 and CX-6258 could achieve good exposure in the intestines and were predicted to show in vivo efficacy in treating mice infected with whipworms.

Table 6.

Pharmacokinetics (PK) of CX-6258 and 12 in Mice^a

	oral (PO) dosing	IV dosing
compound	%F	Cl (L/h/kg)	Vd (L/kg)	T1/2 (h)
CX-6258	23	2.1	85	27
12	6.7	14	11	0.9

^aAnalogue 12 showed decreased bioavailability and a shorted half-life than CX-6258 (previously published¹⁵ PK profile of CX-6258).

In Vivo Activity of CX-6258 in Whipworm-Infected Mice.

To determine the efficacy of CX-6258 and 12 as new anthelmintic drugs, we performed in vivo studies (repeated two times) against T. muris whipworm infection in mice and A. ceylanicum infection in hamsters: both well-established models for human whipworm²⁸ and hookworm²⁹ infections. Mice were treated via oral gavage with water suspension containing 100 mg/kg/day for three consecutive days with CX-6258, 12, or untreated water as a placebo control. Treatment with CX-6258 did not affect the visually observed health and vitality of treated animals, nor were any changes in body weight (Figure S3). Nonetheless, treatment with CX-6258 significantly reduced both the fecal egg count (EPG) by 86% (P = 0.0234) and the total worm burden by 61% (P = 0.0201) as compared with the placebo control (Figure 6A,B). In an independent repeat of the in vivo challenge, CX-6258 showed consistent activity against T. muris with a significant reduction in EPG by 57% (0.0374) and in total worm burden by 40% (P = 0.024) compared to the placebo control (Figure 6C,D). Treatment with 12, however, did not induce a significant reduction in egg counts and induced a marginally significant (P = 0.045) reduction in worm burden (Figure S4). Treating hamsters infected with A. ceylanicum with the same dose and regimen did not reduce egg number or worm burden, indicating that CX-6258 and 12 are suitable for selective anthelmintic activity toward whipworm parasites (Figure S5).

Figure 6.

Results from two in vivo efficacy experiments for CX-6258 activity in T. muris-infected mice (100 mg/kg, three times). Panels (A) and (C) show the eggs per gram of feces (EPG) in fecal samples collected 7 days after the third treatment and the placebo control. Panels (B) and (D) show the total number of T. muris adults (worm burden) harvested after 7 days from the last treatment from the large intestines of treated animals. CX-6258 showed a significant reduction in both replicate final EPG and worm burden compared to the control group. Numbers represent P values using one-tailed unpaired t tests.

CONCLUSIONS

We have demonstrated through both in vitro and in vivo studies that existing chemical series of human PIM kinase inhibitors have high therapeutic potential as a new class of kinase targeted anthelmintic drugs. Several improved synthetic analogues of CX-6258 that we designed had increased inhibition of worm motility in both whipworms and hookworms. To help guide our studies, we utilized aqueous solubility and cellular permeability data of the compounds in addition to a bioaccumulation assay to determine penetrability through the cuticle to predict bioavailability in the worm and to use in rationally designing improved compounds with increased exposure in the intestinal tract. We hypothesize that the observed differential activity of the compounds between different species of intestinal nematodes maybe due to specific binding to a target or may be the result of an innate inability to permeate the cuticle. While we have not yet confirmed the protein target of CX-6258 and analogues in either species of worm, we screened CX-6258 against 485 human kinases to determine other target potential kinase targets besides PIM1 and PIM3. We used the results of this kinase screening to mine for orthologues in both T. muris and A. ceylanicum, revealing additional enzymes that may have potential in future anthelmintic drug discovery ventures. We have shown that CX-6258 has efficacy as an oral anthelmintic that significantly reduced both the whipworm burden and the number of eggs per gram of feces in vivo for T. muris. Future studies will report on lead compound optimization of this and other series of putative PIM kinase inhibitors in addition to investigations into the mechanism of action of CX-6258, other kinase inhibitors, and their best analogues.

EXPERIMENTAL SECTION

Brief methods are provided below, but detailed methods for all experiments, including the synthesis of compounds and ¹H NMR and HPLC/MS spectra, are provided for all final compounds in Supporting Information Methods.

Ethics Statement.

All animal experiments were carried out under protocols approved by the UMass Chan Medical School (PROTO202000044 and PROTO202000071) Institutional Animal Care and Use Committees (IACUC). All housing and care of laboratory animals conformed to the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals in Research (see 18-F22) and all requirements and regulations issued by the United States Department of Agriculture (USDA), including regulations implementing the Animal Welfare Act (P.L. 89–544), as amended (see 18-F23).

Parasite Maintenance.

Following standard protocols, A. ceylanicum parasites were maintained in golden Syrian hamsters (R), and T. muris whipworm parasites were maintained in stat6−/− mice (R).

Compound Screening with Adult-Stage of Intestinal Nematodes.

A. ceylanicum: adult hookworm parasites were harvested 18 days post inoculation from A. ceylanicum-infected hamsters. In vitro assays were carried out as previously described.³⁰ Adult worms were manually sorted into each well of a 48-well plate with one worm/well containing hookworm culture medium (Roswell Park Memorial Institute (RPMI) 1640 of pH 7.2 supplemented with 10% fetal bovine serum, 100U penicillin, 100 μg/mL streptomycin, 2.5 μg/mL amphotericin) and the test drug (6 wells/test drug). Chokepoint inhibitors were tested against hookworm at either 30 or 50 μM in 1% DMSO. Each assay plate contained six wells of positive control (30 μM ivermectin) and six wells of negative control (1% DMSO). Worms were scored for motility using either standard motility index or automated imaging system Worminator and WormAssay-6 software after 24 and 48 h of incubation at 37 °C and 5% CO₂.³¹

T. muris: In vitro assays were carried out as described.³⁰ Adult T. muris parasites were harvested 35 days post inoculation from Stat6−/− mice. Adult T. muris worms were manually sorted into each well of a 48-well plate with one worm/well containing whipworm culture medium (RPMI 1640 of pH 7.2 supplemented with 5% fetal bovine serum, 100U penicillin, 100 μg/mL streptomycin, 2.5 μg/mL amphotericin) and the test drug (6 wells/test drug). Inhibitors were initially tested against whipworms at 50 μM in 1% DMSO. Each assay plate contained six wells of positive control (50 μM levamisole) and six negative control wells (1% DMSO). Worms were scored for motility using either standard motility index (R) or automated imaging system Worminator and WormAssay-6 software (R) after 24 and 48 h of incubation at 37°C and 5% CO₂.

To generate the IC₅₀ values for potential inhibitors, adult worms (A. ceylanicum or T. muris) were tested against dose responses of 0.3, 1, 3, 10, 30, and 90 μM with six wells/dose. Each dose–response plate contained negative and positive controls. Worm motility was scored after 24 and 48 h of incubation at 37 °C and 5% CO₂, using the Worminator platform and WormAssay-6 software (R). IC₅₀ values were established using the nonlinear regression of inhibitor doses versus normalized motility using Prism 9 (GraphPad, La Jolla, CA).