Conversion of oxadiazolo[3,4-b]pyrazines to imidazo[4,5-b]pyrazines via a tandem reduction-cyclization sequence generates new mitochondrial uncouplers

Abstract

We report new mitochondrial uncouplers derived from the conversion of [1,2,5]oxadiazolo[3,4-b]pyrazines to 1H-imidazo[4,5-b]pyrazines. The in situ Fe-mediated reduction of the oxadiazole fragment followed by cyclization gave access to imidazopyrazines in moderate to good yields. A selection of orthoesters also allowed functionalization on the 2-position of the imidazole ring. This method afforded a variety of imidazopyrazine derivatives with varying substitution on the 2, 5 and 6 positions. Our studies suggest that both a 2-trifluoromethyl group and N-methylation are crucial for mitochondrial uncoupling capacity.

Modulation of cellular respiration by mitochondrial uncouplers is an attractive target for the development of therapeutics for cancer,^1,2 neurodegenerative diseases,³ and metabolic disorders.^4,5 In the mitochondria, cellular respiration occurs by the breakdown of nutrients to produce adenosine triphosphate (ATP), an energy rich molecule that powers cellular function and signaling pathways.^6,7 This process is driven by a proton motive force (pmf) generated by the efflux of protons from the matrix to the mitochondrial intermembrane space. Protons returning to the matrix pass through ATP synthase, enabling the enzyme to phosphorylate adenosine diphosphate (ADP) into ATP. Mitochondrial uncoupling occurs when protons are shuttled from the intermembrane space to the matrix independent of ATP synthase; hence nutrient oxidation is uncoupled from ATP synthesis.^8,9 This proton movement may occur naturally through basal and protein-induced leaks.^10,11 Alternatively, uncoupling via small molecules is possible, particularly by the activity of protonophores.

Protonophores are most commonly lipophilic weak acids that shuttle protons across the mitochondrial innermembrane.^12,13 In the literature, the identified scaffolds for small molecule mitochondrial uncouplers are largely based on existing drugs and natural products, such biarylsulfonamides, biarylureas, hydroxyl benzamides, salicylates, and phenols (Fig. 1).^12–14 Development of these particular mitochondrial uncouplers is complicated by their established mechanisms of action on other biological targets. Other small molecules such as 2,4-dinitrophenol (DNP) and FCCP have been identified, but they are not selective for the mitochondrial membrane and result in cytotoxicity that limits their utility. Thus, the chemical diversity among reported mitochondrial uncouplers remains narrow and underexplored. The preclinical success of BAM15 and OPC-163493^15,16 showcase the appeal of developing novel chemical entities. Our laboratories have a long-standing interest in the development of mitochondrial uncouplers, particularly derivatization of BAM15, a protonophore selective for the mitochondrial membrane.^17–22.

Fig. 1.

Chemical structures of common small molecule mitochondrial uncouplers.

BAM15, containing two ortho-fluoroanilines on a [1,2,5]oxadiazolo [3,4-b]pyrazine core, is efficacious in obesity and cancer mouse models (Fig. 2).^17,23 An unsymmetrical analog of BAM15, SHC517, is a more potent uncoupler in vitro and efficacious in obesity and fatty liver disease mouse models.^18,24 Replacing an aniline with a hydroxyl moiety gave SHS4121705 with improved aqueous solubility and pharmacokinetic properties over BAM15.²² SHS4121705 resulted in a 2-point reduction in nonalcoholic fatty liver activity score in a STAM mouse study. Removal of furazan produced pyrazine analogs that were generally less potent and efficacious as uncouplers.^19,21 Furazan, which is recently more utilized in medicinal chemistry,²⁵ is strongly electron withdrawing and contributes to both the acidity and lipophilic properties of these compounds. We envisioned the possibility of replacing furazan with a different heterocycle to increase chemical diversity and improve physicochemical properties. Replacing the oxygen atom in furazan with a carbon yields an imidazo[4,5-b]pyrazine derivative (Fig. 2). Imidazo [4,5-b]pyridines and imidazo[4,5-b]pyrazines represent a privileged scaffold in medicinal chemistry, in part because they are isosteres of the 9H-purine scaffold.²⁶ These compounds have shown biological activity as kinase inhibitors, cardiotonic agents, antiviral, antifungal, antimalarial, anticancer, anticonvulsant, and anxiolytic amongst others.^27–34 A common strategy for the synthesis of imidazo[4,5-b]pyrazines includes amination of aryl halides, typically catalyzed by palladium, followed by cyclization in either a one pot or step-wise approach (Fig. 3).^35–39 For example, You and co-workers developed a palladium-catalyzed bis amination of 1,2-dihalopyrazine with a variety of amidines to afford imidazo[4,5-b]pyrazine derivatives, which had opto-electronic properties (Fig. 3A).³⁵ More recently, Clark reported the imidazo ring formation through a palladium-catalyzed amidation-cyclization sequence using 3-amino-2-chloropyrazine analogs with formamide (Fig. 3B).^37,38 An alternative approach is the reaction of 1,2-diaminopyrazine with carbon electrophiles such as orthoesters or acid chlorides.^40–46 However, compounds bearing amine functional groups on the 5- and 6-position as well as other moieties on the 2 position remains challenging. To streamline our medicinal chemistry efforts, we envisaged accessing imidazo[4,5-b]pyrazine-5,6-diamines from our library of [1,2,5]oxadiazolo[3,4-b]pyrazine-5,6-diamines that are readily synthesized and scalable.¹⁸ Ninkovic previously employed Fe in the reduction of nitroaromatics followed by cyclization with orthoester to yield imidazo[4,5-b]pyrazine (Fig. 3C).⁴⁷ We postulated that the furazan is analogous to a nitro group and may provide access to the requisite diamine. Herein, we disclose a strategy for the synthesis of imidazo[4,5-b]pyrazine-5,6-diamines from [1,2,5]oxadiazolo[3,4-b]pyrazine-5,6-diamines via a reductive ring opening of the [1,2,5]oxadiazole moiety to form an unstable 1,2-diaminopyrazine in situ that is captured by a carbon electrophile amenable to cyclization (Fig. 3D).

Fig. 2.

Derivatization of BAM15 to novel and potent mitochondrial uncouplers.

Fig. 3.

Strategies towards imidazo[4,5-b]pyrazines.

Our investigation began with Fe (mesh) as the reducing agent and ytterbium triflate as the catalyst for cyclization with acetic acid as the solvent.⁴⁷ Based on our previous studies, the strong electron withdrawing effects of the furazan were important for activity.^19,21 We sought to incorporate a trifluoromethyl group on the 2 position of the imidazopyrazine to maintain a strong electron pull. A variety of fluorinated carbon electrophiles were investigated (Table 1). Using BAM15 in the presence of 1,1,1,5,5,5-hexafluoropenta-2,4-dione at 95 °C afforded 2-trifluoroimidazolopyrazine 1a in 28% yield (entry 1). Trifluoroacetic anhydride and trifluoroacetic acid were significantly less efficient providing only a trace amount of product (entries 2–3). Gratifyingly, the use of ethyl trifluoroacetate provided 1a in 61% yield (entry 4). A survey of temperature and reaction times in entries 5–10 suggested that the reaction proceeded most efficiently at 95 °C within 4 h (entry 9). In the absence of the ytterbium triflate under more concentrated conditions, 1a was afforded in slightly lower yields (entries 11–12). While ytterbium triflate is not required for this reaction, we found that the catalyst can improve the yields for certain substrates, particularly on small scale. Thus, the catalyst was incorporated into our substrate synthesis.

Table 1.

Optimization of reaction conditions.^a

Entry	Electrophile	T (°C)	Time (h)	Yieldb
1	1,1,1,5,5,5-Hexafluoropenta-2,4-dione	95	3	28
2	(CF3CO)2O	95	3	5
3	CF3CO2H	95	3	Trace
4	CF3CO2Et	95	3	61
5	CF3CO2Et	75	3	5
6	CF3CO2Et	85	3	45
7	CF3CO2Et	105	3	60
8	CF3CO2Et	95	2	40
9	CF3CO2Et	95	4	79
10	CF3CO2Et	95	6	75
11c,d	CF3CO2Et	95	4	73
12c,e	CF3CO2Et	95	4	55

^aGeneral procedure: BAM15 (0.1 mmol), Fe (1.0 mmol), Yb(OTf)₃ (0.01 mmol), HOAc (0.2 M), electrophile (1 mmol).

^bIsolated yield.

^cWithout Yb(OTf)₃.

^dHOAc (0.4 M).

^eHOAc (0.6 M).

With the optimum reaction conditions in hand, we proceeded to explore the mitochondrial uncoupling activity of the 5,6-bisaniline 2-trifluoromethyl-imidazo[4,5-b]pyrazines. Mitochondrial uncoupling activity is measured as a function of the oxygen consumption rate (OCR) in rat L6 myoblast cells, using the Agilent Seahorse XF analyzer.⁴⁸ Efficacious compounds will cause an increase in OCR relative to basal cellular respiration, whereas toxic compounds will cause a decrease in OCR.^8,9,12–14 Inactive compounds will not impact OCR values while efficacious compounds will have a dose–response activity. First, symmetrical compounds where both anilines are identical were tested (Fig. 4). Compound 1a, synthesized from BAM15 in 79% yield, did not increase oxygen consumption at 1 μM, but was moderately active at 10 μM with a ca. 50% increase in OCR over the baseline and 150% at 50 μM. Aniline 1b and p-methylaniline 1c had moderate to low activity at 50 μM. Synthetically, reaction yields were lowest for compounds with electron donating groups. Electron withdrawing groups with multiple fluorine groups were generally higher yielding and more active, especially at 10 μM with >150% increase in OCR for 1d, 1g, and 1h. However, cellular toxicity became apparent as the concentration was increased to 50 μM where a corresponding decrease in OCR was evident in 1d–1l.

Fig. 4.

Changes in relative oxygen consumption rate, compared to baseline respiration, in rat L6 myoblasts treated with symmetrical bisaniline analogs. Standard reaction conditions: [1,2,5]oxadiazolo[3,4-b]pyrazine-5,6-diamines (0.1 mmol), ethyl trifluoroacetate (10 equiv.), Fe (10 equiv.), Yb(OTf)₃ (0.1 equiv.), AcOH (0.2 M), 95 °C, 4 h; Isolated yields are reported in parenthesis.

Next, unsymmetrical analogs, where a 2-fluoroaniline was maintained, were investigated akin to SHC517 (Fig. 5). Similarly, reaction yields were highest for fluorinated anilines (1q–1t). The most notable compound is 1q, synthesized from SHC517 in 73% yield, which resulted in a 100% OCR increase at 10 μM and 150% at 50 μM without any sign of toxicity. Aniline 1n and p-methylaniline 1o resembled their symmetrical counterparts (1b and 1c respectively) with only modest increase in OCR of ca. 50% at 50 μM. Electron-donating methoxy in 1p was not tolerated. Similar to the symmetrical compounds, compounds 1r–1t became toxic at 50 μM. Interestingly, cyclohexyl substituted 1u was active at 50 μM with 125% increase in OCR. Other aliphatic groups in 1v and 1w were not tolerated.

Fig. 5.

Changes in relative oxygen consumption rate, compared to baseline respiration, in rat L6 myoblasts treated with unsymmetrical bisamine analogs. Standard reaction conditions: [1,2,5]oxadiazolo[3,4-b]pyrazine-5,6-diamines (0.1 mmol), ethyl trifluoroacetate (10 equiv.), Fe (10 equiv.), Yb(OTf)₃ (0.1 equiv.), AcOH (0.2 M), 95 °C, 4 h; Isolated yields are reported in parenthesis.

The symmetrical compounds were then methylated to probe the impact of the imidazole N—H (Fig. 6). Interestingly, methylation alleviated the toxicity observed in the N—H analogs as none of the active methylated analogs were toxic at elevated concentrations. This is apparent in comparing 1j and 2j. Compound 1j with the N—H increased OCR by 100% at 10 μM but was toxic at 50 μM (Fig. 4). After methylation to 2j, the compound’s OCR activity improved in a dose dependent manner (>200% at 50 μM, Fig. 6). The most notable analog is 2m which elevated the OCR by nearly 100% at 10 μM and >300% at 50 μM (Fig. 6). This is in stark contrast to the analogous N—H derivative 1m which had a maximum activity at 50 μM of only a 100% increase in OCR (Fig. 4).

Fig. 6.

Changes in relative oxygen consumption rate, compared to baseline respiration, in rat L6 myoblasts treated with N-methylated analogs. Standard reaction conditions: compound 1, CH₃I, Et₃N, CH₂Cl₂, 40 °C, 16 h; Isolated yields are reported in parenthesis.

Finally, to investigate the role of the trifluoromethyl group, derivatives of 1a where the 2-trifluoromethyl was replaced with hydrogen (3a), methyl (3b), and phenyl (3c) were tested (Scheme 1). For the synthesis, the ethyl trifluoroacetate was replaced with the requisite orthoester. These reactions had comparable outcomes with compound 3a having a 72% yield. The three compounds were inactive thus showing the importance of the electron withdrawing effects of the trifluoromethyl group.

Scheme 1.

Scope of orthoesters for installation on the imidazole 2-position. Standard reaction conditions: BAM15 (0.1 mmol), orthoester (10 equiv.), Fe (10 equiv.), Yb(OTf)₃ (0.1 equiv.), AcOH (0.2 M), 95 °C, 4 h; Isolated yields reported.

In summary, we developed a method for the synthesis of 2-substituted-imidazo[4,5-b]pyrazines from their [1,2,5]oxadiazolo[3,4-b]pyrazine counterparts in moderate to good yields. The process involves an in situ reduction of the oxadiazolo moiety to the corresponding diaminopyrazine, which is trapped and cyclized using the requisite carbon electrophile. The capacity of these compounds to increase cellular respiration was assayed by measuring oxygen consumption rates in cells, revealing potent mitochondrial uncouplers. Our studies demonstrated that the 2-trifluoromethyl group and N-methylation of the imidazole ring is crucial for activity yielding 2m as the best performing uncoupler in this series. Future work will involve characterization of these compounds in vivo.

Data availability

All necessary data was included in the supplemental information.