Design, synthesis and biological evaluation of novel 5-hydroxy-2-methyl-4H-pyran-4-one derivatives as antiglioma agents

Two series of 5-hydroxy-2-methyl-4H-pyran-4-one derivatives were synthesized and their antiglioma activities were evaluated.

Abstract

d-2-Hydroxyglutarate (d-2HG) is frequently found in human brain cancers. Approximately 50–80% of grade II glioma patients have a high level of d-2HG production, which can lead to cancer initiation. In this study, a series of novel 5-hydroxy-2-methyl-4H-pyran-4-one derivatives were designed and synthesized as antiglioma agents, and their related structure–activity relationships are discussed. Among these novel compounds, 4a exhibited promising anti-proliferative activity against glioma HT1080 cells and U87 cells with an IC₅₀ of 1.43 μM and 4.6 μM, respectively. Further studies found that the most active compound (4a) shows an 86.3% inhibitory rate against the intracellular production of d-2HG at 1 μM, and dramatic inhibitory effects, even at 1 μM on the colony formation and migration of U87 and HT1080 cells.

1. Introduction

Excessive accumulation of intracellular d-2-hydroxyglutarate (d-2HG), which is a product of mutations of residue 132 (R132) in isocitrate dehydrogenase 1 (IDH1), results in hypermethylation of histones and DNA by competitive inhibition of several α-KG-dependent dioxygenases including histone demethylases and 5-methylcytosine hydroxylases in the TET family.1 The newly produced d-2HG has been proposed to act as an oncogenic metabolite inducing the dysregulation of methylation, and thus elevating the risk of malignant tumors.2

AGI-5198 (Fig. 1) has been reported to inhibit the production of cellular d-2HG and impair the growth of IDH1 mutant glioma both in vitro and in vivo. The potential clinical utility of such d-2HG scavengers has been identified, and much interest has been attracted to this area. To date, however, only four chemotypes of mutant IDH1 inhibitors with d-2HG scavenging effects, have been reported3 in addition to AGI-5198 as indicated in Fig. 1. Consequently, the exploration of structurally different types of d-2HG scavengers is necessary for the further confirmation of the feasibility of this antitumor strategy.

Fig. 1. Design strategy of 5-hydroxy-2-methyl-4H-pyran-4-one derivatives.

In the current study, several novel 5-hydroxy-2-methyl-4H-pyran-4-one derivatives that could be easily synthesized by an aldol reaction were designed by applying bioisosterism and a scaffold hopping strategy based on an analysis of the structures of the previously reported d-2HG inhibitors (Fig. 1), and evaluated as d-2HG scavengers. The compound (4a) was identified and showed a promising inhibitory effect on the d-2HG production as well as anti-proliferative activity against glioma cells, including cell growth, colony formation and migration inhibitory effects.

2. Chemistry

The compound SYC-435 shown in Fig. 1, was reported to be an excellent potent inhibitor of mutant IDH1 with IC₅₀ values as low as 190 nM.3f The X-ray structures of IDH1 (R132H) in a complex with NADPH and the inhibitor SYC-435 indicate that the 4-methyl-1-hydroxypyridin-2-one moiety of SYC-435 is the critical pharmacophore, for it forms several key H-bonds, and enjoys electrostatic and hydrophobic interactions with the protein. To enrich the type of inhibitor, compounds of series I were designed by replacing the 4-methyl-1-hydroxy-pyridin-2-one moiety (A ring) of SYC-435 with one of its bioisosteres, 5-hydroxy-2-methyl-4H-pyran-4-one, and the resulting compound was synthesized. Since the space of the active pocket of the IDH1 protein is bigger, and all the other reported inhibitors, including AGI-5198, VVS, GSK-321 and BRD-2879, appear to have an essentially triangular skeleton, compounds of series II were designed by introducing a second 5-hydroxy-2-methyl-4H-pyran-4-one moiety at the benzyl α-carbon position of the compounds in series I. The synthetic route of the designed compounds in Fig. 1 is outlined in Scheme 1 (series I: 3a–3e, series II: 4a–4f).

Scheme 1. Synthetic route to 5-hydroxy-2-methyl-4H-pyran-4-one derivatives. Reagents and conditions: (a) SOCl₂, r.t., 1 h; (b) Zn, HCl, H₂O, 70 °C, 4 h; (c) aldehyde, DABCO, dioxane : H₂O = 1 : 1, r.t., 12 h. (d) Triethylsilane, CF₃COOH, CH₂Cl₂, r.t., overnight; (e) aldehyde, DABCO, dioxane : H₂O = 1 : 1, 50 °C, 24 h.

2.1. Synthesis of the target compounds in series I

The synthesis of the precursor allomaltol (1) was performed in two steps, starting with the reaction of kojic acid (5-hydroxy-2-(hydroxylmethyl)-4H-pyran-4-one) with thionylchloride, yielding the product, then, reduction with zinc powder under acidic conditions.4 Due to position-6 showing higher reactivity compared to position-3 on the pyrone ring of 1, appropriate functionalization at position-6 can easily be achieved by the aldol reaction. Thus, compound 3a–3e was synthesized by the condensation of allomaltol (1) (1.0 eq.) with different aldehydes (1.2 eq.) in the presence of a catalyst, triethylene diamine (DABCO, 1.2 eq.),5 then the α–hydroxyl group was reduced using triethylsilicane (5.0 eq.) and trifluoroacetic acid (5.0 eq.) in CH₂Cl₂.

2.2. Synthesis of the target compounds in series II

The aldol reaction of 5-hydroxy-2-methyl-4H-pyran-4-one, carried out with different aldehydes, leads to variously substituted double 5-hydroxy-2-methyl-4H-pyran-4-ones.6 It has been reported that substituted double kojic acid derivatives can be synthesized by microwave-assisted condensation of kojic acid with aldehydes.7 Here, we report a simple method with which to synthesize compounds 4a–4f by condensation of allomaltol (1) (1 eq.) with different aldehydes (0.5 eq.) in the presence of a catalyst, triethylene diamine (DABCO, 1.2 eq.).

3. Results and discussion

3.1. Inhibition of proliferation of human glioma cells

First, the growth inhibitory activities of the synthetic compounds in human brain glioma HT1080 cells and U87 cells were evaluated with an MTT assay. As shown in Table 1, compound 4a showed the best anti-proliferative activity, with an IC₅₀ of 1.43 ± 0.2 μM and 4.6 ± 0.3 μM against HT1080 cells and U87 cells, respectively. Based on these MTT results, some preliminary structure–activity relationships (SAR) could be summarized as follows: (a) a long R group containing a phenyl ring may be necessary for activity (4avs.4e and 4f), (b) substitution on the phenyl ring would reduce the activity (4avs.4b and 4c), and (c) when the benzene is connected through an alkyl chain, the activity will disappear (4avs.4d).

Table 1. Mutant IDH1 (R132H) inhibitory activities and cytotoxicity against the glioma HT1080 cell line and U87 cell line of 5-hydroxy-2-methyl-4H-pyran-4-one derivatives.

Compound	IC50 for MTT a (μM)		IC50 for IDH1 a (μM)
	HT1080	U87
AGI-5198 b	ND	20.1 ± 2.1	0.08 ± 0.01
3a	>50	>50	>50
3b	>50	>50	>50
3c	>50	>50	>50
3d	24.4 ± 0.4	>50	>50
3e	>50	>50	>50
4a	1.43 ± 0.2	4.6 ± 0.3	23.1 ± 2.8
4b	2.6 ± 0.3	5.2 ± 0.2	32.0 ± 2.9
4c	3.06 ± 0.2	>50	>50
4d	>50	27.2 ± 1.5	>50
4e	26.6 ± 0.4	29.5 ± 0.5	>50
4f	28.3 ± 0.6	30.9 ± 1.1	>50

^aAll data represent mean ± S.D. from different experiments performed in triplicate.

^b AGI-5198 acted as a positive control for the IDH1 (R132H) inhibitory activity assays and MTT assays.

3.2. Inhibition of d-2HG production

d-2HG is the specific product of the mutant IDH1 protein, which is often detected in IDH1-mutant cells including gliomas, and can be used as a biomarker for the activity of cancer cells.3a,b,d,8 So in order to ascertain if the cytotoxicity of the compounds resulted from the decline of intracellular d-2HG, the d-2HG level was tested using the human fibrosarcoma HT1080 cell line which harbors an IDH1 mutant gene and has relatively high cellular d-2HG accumulation in the absence or presence of 4a. The intracellular d-2HG concentrations were quantitatively determined by HPLC-MS (standard curve line is shown in Fig. S3,† R = 0.9966). As shown in Fig. 2, we found that the d-2HG level decreased significantly from 23.25 ± 4.03 μg mL^–1 in non-treated cells to 3.14 ± 0.77 μg mL^–1 after treatment of the cells with 1 μM of compound 4a. For the positive control, it decreased to 3.05 ± 0.70 μg mL^–1 after treatment with 1 μM AGI-5198. Here, compound 4a shows an excellent cellular d-2HG inhibitory activity.

Fig. 2. (A) d-2HG production inhibitory effects of 4a in HT1080 cells. The red peak represents the control, blue for compound 4a (1 μM), and green for the positive control (AGI-5198, 1 μM); (B) the quantitative analysis results according to the the peak and the standard curve line.

3.3. Inhibition of migration of human glioma cells

Zhu Jian et al. have found that human glioma cells with high d-2HG levels have a faster migration rate compared to normal human glioma cells.9 Meanwhile, Luyuan Li et al. have reported that the first mutant IDH1 inhibitor AGI-5198 inhibits the migration of human glioma cells with the IDH1 mutation.8a Accordingly, the migration of human brain glioma U87 cells and the human fibrosarcoma HT1080 cell line was assessed with a scratch assay in the presence or absence of 4a. As depicted in Fig. 3A and B, the migration rate of U87 decreased in 12 h from 100% to 51.3% and in 24 h to 17.5% relative to the blank control. However, after treatment with 1 μM 4a, the migration rate decreased in 12 h from 100% to 82.8% and in 24 h to 69.5%, respectively. In the positive control (AGI-5198), the decreases were from 100% to 84.0% in 12 h and to 51.0% in 24 h. Because the migration rate of the HT1080 cells is faster, we shortened the interval time to 6 h. As shown in Fig. 3C and D, the scratch area in 6 h was 58.7%, 61.0%, and 70.0% for the control, 1 μM 4a, and the 10 μM AGI-5198 group, respectively, and in 12 h, it was 23.3%, 43.5%, and 56% for the control, 1 μM 4a, and the 10 μM AGI-5198 group, respectively. Our studies demonstrated that compound 4a is a potent d-2HG scavenger which also exhibits an obvious inhibitory effect on the migration of U87 and HT1080 cells.

Fig. 3. Cell migration inhibition assay. (A) Compound 4a inhibits U87 cell migration (photo), samples were photographed every 12 h after the addition of 1 μM 4a or 10 μM AGI-5198; (C) scratch area analysis of (A). (B) Compound 4a inhibits HT1080 cell migration (photo), samples were photographed every 6 h after the addition of 1 μM 4a or 10 μM AGI-5198; (D) scratch area analysis of (B).

3.4. Inhibition of colony formation of human glioma cells

Luyuan Li et al. have reported that the first mutant IDH1 inhibitor AGI-5198 with a d-2HG inhibitory effect can also inhibit the colony formation of human glioma cells with an IDH1 mutation.8a Thus, U87 and H1080 cells were further employed to examine the colony formation inhibitory activity of 4a. The results (Fig. 4) show that the colony formation of both U87 and H1080 cells were significantly inhibited after the addition of 4a in a concentration-dependent manner. The colony inhibitory rate against U87 cells was 11.3% (after the treatment of 0.5 μM 4a) and 96.8% (after the treatment of 1 μM 4a). For HT1080 cells, complete suppression with an inhibition rate of 99.5% was also observed at the higher concentration of 4a (1 μM). More importantly, compound 4a also showed a strong inhibitory effect (75.0%) on the colony formation of H1080 cells at a low concentration (0.5 μM). These more significant colony formation inhibitory effects of 4a against the cells that express mutant IDH1 highly suggest that these effects are likely to be attributed to the elimination of d-2HG.

Fig. 4. Colony formation assay of U87 and HT1080 cells. The gray column represents the U87 cell line, the other, HT1080 (black column). 0.5 and 1 μM 4a were added to the cells, and colonies were counted after 8 days. 5 and 20 μM AGI-5198 were the positive control drugs.

3.5. Inhibition of mutant IDH1 (R132H)

Considering that 4a shows potent d-2HG elimination activity, as well as obvious inhibition of antiproliferation, migration and colony formation, we investigated the effect of 4a on mutant IDH1, whose mutation is the critical cause of the high intracellular d-2HG levels. Recombinant IDH1 (R132H), the predominant (>90%) mutation found in IDH1 associated gliomas, was used for enzyme inhibitory activity screening. The purity of the recombinant IDH1 (R132H) protein is >98% (Fig. S1 in the ESI†), and the activity of the enzyme is good (Fig. S2 in the ESI†). The IDH1 (R132H) inhibitory activities (IC₅₀) of all the synthesized 5-hydroxy-2-methyl-4H-pyran-4-one derivatives are shown in Table 1. It was found that all of the compounds, including the active compound (4a) showed no obvious activity (IC₅₀ > 20 μM) against IDH1 (R132H). These compounds therefore are much weaker than the positive control AGI-5198 (0.08 ± 0.01 μM). Combined with the d-2HG production inhibition activity, we were puzzled because 4a exhibits good d-2HG production inhibitory activity but no obvious IDH1 direct inhibitory activity compared with AGI-5198. Therefore, 4a may employ a different inhibitory mechanism instead of combining with mutant IDH1 directly, and accordingly, we will, in the future, conduct an intensive study of the d-2HG inhibitory mechanism of this class of compounds.

4. Conclusion

In this study, two series of 5-hydroxy-2-methyl-4H-pyran-4-one derivatives were synthesized, and their bioactivities were evaluated. Among these compounds, one potent antiglioma agent (4a), which harbors excellent intracellular d-2HG scavenging activity was identified. Biological activity studies also indicated the significant inhibitory effects of 4a on the migration and colony formation of human glioma cells, suggesting that 4a may be a promising d-2HG scavenger and has potential to be developed as an antiglioma agent. An obvious d-2HG inhibition activity but no obvious mutant IDH1 inhibitory activity of 4a implied that our compounds may adopt a different inhibitory mechanism rather than combining with mutant IDH1 directly. Further studies of the mechanism of d-2HG elimination are in progress and it is hoped that we could successfully resolve this. These findings may be useful for the further development of antiglioma agents with novel mechanisms.

Conflicts of interest

The authors declare no competing interests.