Dual-Target Compounds against Type 2 Diabetes Mellitus: Proof of Concept for Sodium Dependent Glucose Transporter (SGLT) and Glycogen Phosphorylase (GP) Inhibitors

A current trend in the quest for new therapies for complex, multifactorial diseases, such as diabetes mellitus (DM), is to find dual or even multi-target inhibitors. In DM, the sodium dependent glucose cotransporter 2 (SGLT2) in the kidneys and the glycogen phosphorylase (GP) in the liver are validated targets. Several (β-D-glucopyranosylaryl)methyl (het)arene type compounds, called gliflozins, are marketed drugs that target SGLT2. For GP, low nanomolar glucose analogue inhibitors exist. The purpose of this study was to identify dual acting compounds which inhibit both SGLTs and GP. To this end, we have extended the structure-activity relationships of SGLT2 and GP inhibitors to scarcely known (C-β-D-glucopyranosylhetaryl)methyl arene type compounds and studied several (C-β-D-glucopyranosylhetaryl)arene type GP inhibitors against SGLT. New compounds, such as 5-arylmethyl-3-(β-D-glucopyranosyl)-1,2,4-oxadiazoles, 5-arylmethyl-2-(β-D-glucopyranosyl)-1,3,4-oxadiazoles, 4-arylmethyl-2-(β-D-glucopyranosyl)pyrimidines and 4(5)-benzyl-2-(β-D-glucopyranosyl)imidazole were prepared by adapting our previous synthetic methods. None of the studied compounds exhibited cytotoxicity and all of them were assayed for their SGLT1 and 2 inhibitory potentials in a SGLT-overexpressing TSA201 cell system. GP inhibition was also determined by known methods. Several newly synthesized (C-β-D-glucopyranosylhetaryl)methyl arene derivatives had low micromolar SGLT2 inhibitory activity; however, none of these compounds inhibited GP. On the other hand, several (C-β-D-glucopyranosylhetaryl)arene type GP inhibitor compounds with low micromolar efficacy against SGLT2 were identified. The best dual inhibitor, 2-(β-D-glucopyranosyl)-4(5)-(2-naphthyl)-imidazole, had a Ki of 31 nM for GP and IC50 of 3.5 μM for SGLT2. This first example of an SGLT-GP dual inhibitor can prospectively be developed into even more efficient dual-target compounds with potential applications in future antidiabetic therapy.

As the constitution of the heterocycle may have an essential impact on the efficiency of the molecules against SGLTs, further replacements of the proximal benzene ring were envisaged (see target compounds C in Figure 1). The foreseen imidazole and 1,2,4-triazole derivatives may form H-bonds close to the position of the above phenolic OH. Furthermore, taking into account that the glucosyl biphenyl derivative B also inhibited the SGLTs [43], the study of compounds D was also envisaged. Since many of the D type derivatives proved to be potent inhibitors of glycogen phosphorylases (GP) [21,55,56], another validated target in combatting T2DM, these compounds may prove the dual-target concept for SGLT-GP which, to the best of our knowledge, has never yet been investigated.
As the constitution of the heterocycle may have an essential impact on the efficiency of the molecules against SGLTs, further replacements of the proximal benzene ring were envisaged (see target compounds C in Figure 1). The foreseen imidazole and 1,2,4-triazole derivatives may form H-bonds close to the position of the above phenolic OH. Furthermore, taking into account that the glucosyl biphenyl derivative B also inhibited the SGLTs [43], the study of compounds D was also envisaged. Since many of the D type derivatives proved to be potent inhibitors of glycogen phosphorylases (GP) [21,55,56], another validated target in combatting T2DM, these compounds may prove the dual-target concept for SGLT-GP which, to the best of our knowledge, has never yet been investigated.
As the constitution of the heterocycle may have an essential impact on the efficiency of the molecules against SGLTs, further replacements of the proximal benzene ring were envisaged (see target compounds C in Figure 1). The foreseen imidazole and 1,2,4-triazole derivatives may form H-bonds close to the position of the above phenolic OH. Furthermore, taking into account that the glucosyl biphenyl derivative B also inhibited the SGLTs [43], the study of compounds D was also envisaged. Since many of the D type derivatives proved to be potent inhibitors of glycogen phosphorylases (GP) [21,55,56], another validated target in combatting T2DM, these compounds may prove the dual-target concept for SGLT-GP which, to the best of our knowledge, has never yet been investigated.
As the constitution of the heterocycle may have an essential impact on the efficiency of the molecules against SGLTs, further replacements of the proximal benzene ring were envisaged (see target compounds C in Figure 1). The foreseen imidazole and 1,2,4-triazole derivatives may form H-bonds close to the position of the above phenolic OH. Furthermore, taking into account that the glucosyl biphenyl derivative B also inhibited the SGLTs [43], the study of compounds D was also envisaged. Since many of the D type derivatives proved to be potent inhibitors of glycogen phosphorylases (GP) [21,55,56], another validated target in combatting T2DM, these compounds may prove the dual-target concept for SGLT-GP which, to the best of our knowledge, has never yet been investigated.   [34] Chemically, the marketed gliflozins are (β-D-glucopyranosylaryl)methyl (het)arenes. This scaffold is the result of many rounds of iterations, including several O-and N-glycosidic compounds, to improve pharmacodynamic and pharmacokinetic properties of phlorizin to finally arrive at the C-glucosylic structure [19,20,[39][40][41][42]. After the fortuitous discovery of the highly beneficial meta arrangement of the glycosyl and the benzyl moieties on the proximal aromatic ring (Figure 1, A) [43], this skeleton has become the lead structure for further structure-activity relationship (SAR) studies to modify the sugar part and replace or substitute both aromatic rings and the methylene linker between them [19,20,[44][45][46].
As the constitution of the heterocycle may have an essential impact on the efficiency of the molecules against SGLTs, further replacements of the proximal benzene ring were envisaged (see target compounds C in Figure 1). The foreseen imidazole and 1,2,4-triazole derivatives may form H-bonds close to the position of the above phenolic OH. Furthermore, taking into account that the glucosyl biphenyl derivative B also inhibited the SGLTs [43], the study of compounds D was also envisaged. Since many of the D type derivatives proved to be potent inhibitors of glycogen phosphorylases (GP) [21,55,56], another validated target in combatting T2DM, these compounds may prove the dual-target concept for SGLT-GP which, to the best of our knowledge, has never yet been investigated.   [34] Chemically, the marketed gliflozins are (β-D-glucopyranosylaryl)methyl (het)arenes. This scaffold is the result of many rounds of iterations, including several O-and N-glycosidic compounds, to improve pharmacodynamic and pharmacokinetic properties of phlorizin to finally arrive at the C-glucosylic structure [19,20,[39][40][41][42]. After the fortuitous discovery of the highly beneficial meta arrangement of the glycosyl and the benzyl moieties on the proximal aromatic ring (Figure 1, A) [43], this skeleton has become the lead structure for further structure-activity relationship (SAR) studies to modify the sugar part and replace or substitute both aromatic rings and the methylene linker between them [19,20,[44][45][46].
As the constitution of the heterocycle may have an essential impact on the efficiency of the molecules against SGLTs, further replacements of the proximal benzene ring were envisaged (see target compounds C in Figure 1). The foreseen imidazole and 1,2,4-triazole derivatives may form H-bonds close to the position of the above phenolic OH. Furthermore, taking into account that the glucosyl biphenyl derivative B also inhibited the SGLTs [43], the study of compounds D was also envisaged. Since many of the D type derivatives proved to be potent inhibitors of glycogen phosphorylases (GP) [21,55,56], another validated target in combatting T2DM, these compounds may prove the dual-target concept for SGLT-GP which, to the best of our knowledge, has never yet been investigated.
6 Sotagliflozin (approved 2019: Zynquista ® ) 1.8 (20) [ 34] Chemically, the marketed gliflozins are (β-D-glucopyranosylaryl)methyl (het)arenes. This scaffold is the result of many rounds of iterations, including several Oand Nglycosidic compounds, to improve pharmacodynamic and pharmacokinetic properties of phlorizin to finally arrive at the C-glucosylic structure [19,20,[39][40][41][42]. After the fortuitous discovery of the highly beneficial meta arrangement of the glycosyl and the benzyl moieties on the proximal aromatic ring (Figure 1, A) [43], this skeleton has become the lead structure for further structure-activity relationship (SAR) studies to modify the sugar part and replace or substitute both aromatic rings and the methylene linker between them [19,20,[44][45][46]. Thus, the purpose of this study is to identify new SGLT2 inhibitors with dual activity at GP by characterizing their inhibitory properties towards both proteins. Beginning with the basic structure of clinically available SGLT2 inhibitors, the proximal aromatic group  While the exchange of the distal aromatic moiety was rather extensively investigated [19,20,[44][45][46], resulting in the marketed canagliflozin 4 and ipragliflozin 5, fewer studies targeted the replacement of the proximal benzene unit by hetarenes ( Figure 1) [19,21]. The known structures include thiophene [38,47,48], pyrrole [38], thiazole [49], pyridine [38], pyridazine [49], and pyrazine [38] rings, which show IC 50 values from the low nM to the low µM range. The IC 50 -s are strongly dependent on the heterocyclic moiety. The introduction of a phenolic OH into the ortho position of the proximal ring proved advantageous in several cases, indicating a possible role for a hydrogen bond-forming group in that region [38,[50][51][52][53][54].
As the constitution of the heterocycle may have an essential impact on the efficiency of the molecules against SGLTs, further replacements of the proximal benzene ring were envisaged (see target compounds C in Figure 1). The foreseen imidazole and 1,2,4-triazole derivatives may form H-bonds close to the position of the above phenolic OH. Furthermore, taking into account that the glucosyl biphenyl derivative B also inhibited the SGLTs [43], the study of compounds D was also envisaged. Since many of the D type derivatives proved to be potent inhibitors of glycogen phosphorylases (GP) [21,55,56], another validated target in combatting T2DM, these compounds may prove the dual-target concept for SGLT-GP which, to the best of our knowledge, has never yet been investigated.
Thus, the purpose of this study is to identify new SGLT2 inhibitors with dual activity at GP by characterizing their inhibitory properties towards both proteins. Beginning with the basic structure of clinically available SGLT2 inhibitors, the proximal aromatic group was exchanged with various heterocyclic groups (Figure 1, C). Cell lines stably transfected with SGLT1 and 2 were used to compare the inhibitory properties of these newly synthesized compounds with dapagliflozin and phlorizin, and to reveal selectivity between the two SGLTs. First, the activities of the type C compounds were examined against the SGLT and GP proteins, then the inhibitory effects of previously published GP inhibitors (Figure 1, D) were tested against SGLTs using the new functional assay system. Dual active compounds may serve as lead structures for further dual inhibitor design.

Biological Studies
To identify the inhibitory effects of the test compounds, new cell lines were created that stably overexpress SGLT1 and 2 using a lentiviral system. As shown in Figure 2A,C, SGLT1 and 2 are upregulated after lentiviral transfection (full blot pictures are shown in Figures S1 and S2 in the Supplementary Information). Figure 3 shows that glucose uptake is significantly increased in the TSA201 cells overexpressing SGLT1 and 2 compared to non-transfected cells. To determine the suitability of the new cell lines for assaying SGLT1 and 2 inhibitors, the known SGLT1 and 2 inhibitor, phlorizin 1, and the patented SGLT2 inhibitor, dapagliflozin 2, were tested. The data for phlorizin is shown in Figure 3. Phlorizin attenuated the increase in glucose uptake induced by overexpression of SGLT1 and 2, demonstrating the specificity of the assay. None of the compounds exhibited cytotoxicity, as shown in Figure 4. These data were utilized to generate dose-response curves for the calculation of the IC50s shown in Tables 2 and 3. Examples of dose-response curves for dapagliflozin and compounds 22b and 24b are shown in the Supplementary Figure  S3.

Biological Studies
To identify the inhibitory effects of the test compounds, new cell lines were created that stably overexpress SGLT1 and 2 using a lentiviral system. As shown in Figure 2A,C, SGLT1 and 2 are upregulated after lentiviral transfection (full blot pictures are shown in Figures S1 and S2 in the Supplementary Information). Figure 3 shows that glucose uptake is significantly increased in the TSA201 cells overexpressing SGLT1 and 2 compared to non-transfected cells. To determine the suitability of the new cell lines for assaying SGLT1 and 2 inhibitors, the known SGLT1 and 2 inhibitor, phlorizin 1, and the patented SGLT2 inhibitor, dapagliflozin 2, were tested. The data for phlorizin is shown in Figure 3. Phlorizin attenuated the increase in glucose uptake induced by overexpression of SGLT1 and 2, demonstrating the specificity of the assay. None of the compounds exhibited cytotoxicity, as shown in Figure 4. These data were utilized to generate dose-response curves for the calculation of the IC 50 s shown in Tables 2 and 3. Examples of dose-response curves for dapagliflozin and compounds 22b and 24b are shown in the Supplementary Figure S3.     Tests carried out with the newly synthesized and some previously prepared arylmethyl substituted glucosyl heterocycles (10, 13, 16, [18][19][20] revealed some new SGLT2 inhibitors with IC50-s in the micromolar range (Table 2). Although the best inhibitors of this set (10a, 13a, 18, and 19c) showed remarkably weaker inhibitory potencies than the known Cglucosyl arene type inhibitor 2, their activities exceeded that of phlorizin (1). Nevertheless, this study confirmed earlier observations for the C-glucosyl (het)arene type SGLT2 inhibitors demonstrating that replacement of the proximal aromatic ring with heterocycles leads to decreased inhibitory efficacy, which is strongly dependent on the structure of the heteroring [19,21].
The SGLT2 inhibitory effect of the benzyl substituted oxadiazoles indicated that the constitution of the oxadiazole ring did not significantly affect the inhibitory potential (Table 2, 10a vs. 13a). However, replacement of the heterocyclic oxygen by a H-bond donor NH group led to a 20-30-fold weakening in the inhibition by the corresponding 1,2,4-triazole (19a vs. 10a and 13a). The introduction of a phenyl ring to the N1-atom of the 1,2,4triazole resulted in significant improvement in the efficacy (19c vs. 19a), while attaching a hydroxyethyl group to the same position caused complete loss of the inhibition (19a vs. 19b). Compared to 1,2,4-triazole 19a imidazole 18, which also contained an NH group, proved to be ~20 times more effective. The role of the NH group seems contradictory, since compound 18 is stronger but 19a is a weaker inhibitor. We speculate that this might be the result of three possible tautomeric forms of the 1,2,4-triazole 19a, from which only one may bind to the protein, while for the imidazole 18 the formally existing tautomers are essentially the same due to a probable protonation in the biological medium [66].
Similar to other classes of SGLT2 inhibitors, the inhibition highly depends on substitution of the distal aromatic ring of the aglycon part. In the case of the 1,2,4-and 1,3,4oxadiazoles, only the unsubstituted derivatives ( Table 2, 10a and 13a) showed inhibitory effects, while the introduction of a methoxy or a chloro group into the para position of the benzene ring, such as in 10b and 13b and 10c and 13c, respectively, led to complete loss Tests carried out with the newly synthesized and some previously prepared arylmethyl substituted glucosyl heterocycles (10, 13, 16, [18][19][20] revealed some new SGLT2 inhibitors with IC 50 -s in the micromolar range (Table 2). Although the best inhibitors of this set (10a, 13a, 18, and 19c) showed remarkably weaker inhibitory potencies than the known C-glucosyl arene type inhibitor 2, their activities exceeded that of phlorizin (1). Nevertheless, this study confirmed earlier observations for the C-glucosyl (het)arene type SGLT2 inhibitors demonstrating that replacement of the proximal aromatic ring with heterocycles leads to decreased inhibitory efficacy, which is strongly dependent on the structure of the heteroring [19,21].
The SGLT2 inhibitory effect of the benzyl substituted oxadiazoles indicated that the constitution of the oxadiazole ring did not significantly affect the inhibitory potential ( Table 2, 10a vs. 13a). However, replacement of the heterocyclic oxygen by a H-bond donor NH group led to a 20-30-fold weakening in the inhibition by the corresponding 1,2, 4-triazole (19a vs . 10a and 13a). The introduction of a phenyl ring to the N1-atom of the 1,2,4-triazole resulted in significant improvement in the efficacy (19c vs. 19a), while attaching a hydroxyethyl group to the same position caused complete loss of the inhibition (19a vs. 19b). Compared to 1,2,4-triazole 19a imidazole 18, which also contained an NH group, proved to be~20 times more effective. The role of the NH group seems contradictory, since compound 18 is stronger but 19a is a weaker inhibitor. We speculate that this might be the result of three possible tautomeric forms of the 1,2,4-triazole 19a, from which only one may bind to the protein, while for the imidazole 18 the formally existing tautomers are essentially the same due to a probable protonation in the biological medium [66]. of the inhibition. In the pyrimidine series, not only the benzyl (16a), but also the 4-methoxybenzyl (16b) derivative, displayed inhibitory activity in the micromolar range. In contrast, compound 16c, which has a 4-chlorobenzyl residue, remained inactive. Replacement of the distal benzene ring by a 2-naphthyl moiety resulted in a non-inhibitory molecule 16d. None of the compounds listed in Table 2 had remarkable inhibitory effects against rabbit muscle GPb enzyme in the concentration tested. of the inhibition. In the pyrimidine series, not only the benzyl (16a), but also the 4-methoxybenzyl (16b) derivative, displayed inhibitory activity in the micromolar range. In contrast, compound 16c, which has a 4-chlorobenzyl residue, remained inactive. Replacement of the distal benzene ring by a 2-naphthyl moiety resulted in a non-inhibitory molecule 16d. None of the compounds listed in Table 2 had remarkable inhibitory effects against rabbit muscle GPb enzyme in the concentration tested. of the inhibition. In the pyrimidine series, not only the benzyl (16a), but also the 4-methoxybenzyl (16b) derivative, displayed inhibitory activity in the micromolar range. In contrast, compound 16c, which has a 4-chlorobenzyl residue, remained inactive. Replacement of the distal benzene ring by a 2-naphthyl moiety resulted in a non-inhibitory molecule 16d. None of the compounds listed in Table 2 had remarkable inhibitory effects against rabbit muscle GPb enzyme in the concentration tested. of the inhibition. In the pyrimidine series, not only the benzyl (16a), but also the 4-methoxybenzyl (16b) derivative, displayed inhibitory activity in the micromolar range. In contrast, compound 16c, which has a 4-chlorobenzyl residue, remained inactive. Replacement of the distal benzene ring by a 2-naphthyl moiety resulted in a non-inhibitory molecule 16d. None of the compounds listed in Table 2 had remarkable inhibitory effects against rabbit muscle GPb enzyme in the concentration tested. of the inhibition. In the pyrimidine series, not only the benzyl (16a), but also the 4-methoxybenzyl (16b) derivative, displayed inhibitory activity in the micromolar range. In contrast, compound 16c, which has a 4-chlorobenzyl residue, remained inactive. Replacement of the distal benzene ring by a 2-naphthyl moiety resulted in a non-inhibitory molecule 16d. None of the compounds listed in Table 2 had remarkable inhibitory effects against rabbit muscle GPb enzyme in the concentration tested. of the inhibition. In the pyrimidine series, not only the benzyl (16a), but also the 4-methoxybenzyl (16b) derivative, displayed inhibitory activity in the micromolar range. In contrast, compound 16c, which has a 4-chlorobenzyl residue, remained inactive. Replacement of the distal benzene ring by a 2-naphthyl moiety resulted in a non-inhibitory molecule 16d. None of the compounds listed in Table 2 had remarkable inhibitory effects against rabbit muscle GPb enzyme in the concentration tested. of the inhibition. In the pyrimidine series, not only the benzyl (16a), but also the 4-methoxybenzyl (16b) derivative, displayed inhibitory activity in the micromolar range. In contrast, compound 16c, which has a 4-chlorobenzyl residue, remained inactive. Replacement of the distal benzene ring by a 2-naphthyl moiety resulted in a non-inhibitory molecule 16d. None of the compounds listed in Table 2 had remarkable inhibitory effects against rabbit muscle GPb enzyme in the concentration tested.  Next, a series of known GP inhibitors 21-25 (Table 3) synthesized and investigated in our group [68][69][70][71][72][73] were assayed against SGLTs. The compounds were selected from non-inhibitory ones to low nanomolar GP inhibitors. Former GP inhibition studies revealed the essential role of the heterocyclic linker between the glucose unit and the distal aromatic moiety [21,55,56]. With respect to the latter aromatic group, the 2-naphthyl substituted derivatives 21b-25b surpassed the phenyl substituted ones 21a-25a by a factor of approx. 10 in most cases. The SGLT assay gratifyingly indicated low micromolar inhibitions for many of these compounds. Though thiazoles 23 proved inefficient towards both transporters in the studied concentration, the rest of the compounds had varying effects strongly indicating the substantial role of the heterocyclic linker in this biomolecular interaction. 23  vealed the essential role of the heterocyclic linker between the glucose unit and the distal aromatic moiety [21,55,56]. With respect to the latter aromatic group, the 2-naphthyl substituted derivatives 21b-25b surpassed the phenyl substituted ones 21a-25a by a factor of approx. 10 in most cases. The SGLT assay gratifyingly indicated low micromolar inhibitions for many of these compounds. Though thiazoles 23 proved inefficient towards both transporters in the studied concentration, the rest of the compounds had varying effects strongly indicating the substantial role of the heterocyclic linker in this biomolecular interaction. vealed the essential role of the heterocyclic linker between the glucose unit and the distal aromatic moiety [21,55,56]. With respect to the latter aromatic group, the 2-naphthyl substituted derivatives 21b-25b surpassed the phenyl substituted ones 21a-25a by a factor of approx. 10 in most cases. The SGLT assay gratifyingly indicated low micromolar inhibitions for many of these compounds. Though thiazoles 23 proved inefficient towards both transporters in the studied concentration, the rest of the compounds had varying effects strongly indicating the substantial role of the heterocyclic linker in this biomolecular interaction. vealed the essential role of the heterocyclic linker between the glucose unit and the distal aromatic moiety [21,55,56]. With respect to the latter aromatic group, the 2-naphthyl substituted derivatives 21b-25b surpassed the phenyl substituted ones 21a-25a by a factor of approx. 10 in most cases. The SGLT assay gratifyingly indicated low micromolar inhibitions for many of these compounds. Though thiazoles 23 proved inefficient towards both transporters in the studied concentration, the rest of the compounds had varying effects strongly indicating the substantial role of the heterocyclic linker in this biomolecular interaction. vealed the essential role of the heterocyclic linker between the glucose unit and the distal aromatic moiety [21,55,56]. With respect to the latter aromatic group, the 2-naphthyl substituted derivatives 21b-25b surpassed the phenyl substituted ones 21a-25a by a factor of approx. 10 in most cases. The SGLT assay gratifyingly indicated low micromolar inhibitions for many of these compounds. Though thiazoles 23 proved inefficient towards both transporters in the studied concentration, the rest of the compounds had varying effects strongly indicating the substantial role of the heterocyclic linker in this biomolecular interaction. vealed the essential role of the heterocyclic linker between the glucose unit and the distal aromatic moiety [21,55,56]. With respect to the latter aromatic group, the 2-naphthyl substituted derivatives 21b-25b surpassed the phenyl substituted ones 21a-25a by a factor of approx. 10 in most cases. The SGLT assay gratifyingly indicated low micromolar inhibitions for many of these compounds. Though thiazoles 23 proved inefficient towards both transporters in the studied concentration, the rest of the compounds had varying effects strongly indicating the substantial role of the heterocyclic linker in this biomolecular interaction. vealed the essential role of the heterocyclic linker between the glucose unit and the distal aromatic moiety [21,55,56]. With respect to the latter aromatic group, the 2-naphthyl substituted derivatives 21b-25b surpassed the phenyl substituted ones 21a-25a by a factor of approx. 10 in most cases. The SGLT assay gratifyingly indicated low micromolar inhibitions for many of these compounds. Though thiazoles 23 proved inefficient towards both transporters in the studied concentration, the rest of the compounds had varying effects strongly indicating the substantial role of the heterocyclic linker in this biomolecular interaction. vealed the essential role of the heterocyclic linker between the glucose unit and the distal aromatic moiety [21,55,56]. With respect to the latter aromatic group, the 2-naphthyl substituted derivatives 21b-25b surpassed the phenyl substituted ones 21a-25a by a factor of approx. 10 in most cases. The SGLT assay gratifyingly indicated low micromolar inhibitions for many of these compounds. Though thiazoles 23 proved inefficient towards both transporters in the studied concentration, the rest of the compounds had varying effects strongly indicating the substantial role of the heterocyclic linker in this biomolecular interaction.  [72,73] a The inhibition of SGLT2 and SGLT1 was measured in TSA-201 cells. b rmGPb: rabbit muscle glycogen phosphorylase b. The synthesis of the compounds is also described in the cited references. c NI: no inhibition at 1000 μM. d NI: no inhibition at 625 μM. Table 3 inhibited SGLT2 more than SGLT1 with a selectivity range of ~2-27. There were two exceptions to this trend: compounds 22b and 25b (both with a 2-naphthyl substituent) showed some preference for SGLT1 inhibition. In two comparisons, 21a-21b and 24a-24b the 2-naphthyl derivatives were stronger inhibitors of SGLT2 than SGLT1 by a factor of ~7 and ~26, respectively. The 2-(β-D-glucopyranosyl)-4(5)-(2-naphthyl)-imidazole 24b, with a 3.5 μM IC50 towards SGLT2, emerged as the best inhibitor. Taking into account its nanomolar inhibition of GP, 24b represents a dual-targeting compound, which can be a starting point for further antidiabetic studies in this direction. The SGLT inhibitors are highly sensitive to the substitution pattern of the aglycon as outlined in the introduction. On the other hand, the substituent effects in glucosyl heterocycles on GP inhibition are relatively unknown [56]. Thus, extensive computational and synthetic work will be necessary to optimize the dual-target efficiency. Our present efforts are extending in these directions.  [72,73] a The inhibition of SGLT2 and SGLT1 was measured in TSA-201 cells. b rmGPb: rabbit muscle glycogen phosphorylase b. The synthesis of the compounds is also described in the cited references. c NI: no inhibition at 1000 μM. d NI: no inhibition at 625 μM. Table 3 inhibited SGLT2 more than SGLT1 with a selectivity range of ~2-27. There were two exceptions to this trend: compounds 22b and 25b (both with a 2-naphthyl substituent) showed some preference for SGLT1 inhibition. In two comparisons, 21a-21b and 24a-24b the 2-naphthyl derivatives were stronger inhibitors of SGLT2 than SGLT1 by a factor of ~7 and ~26, respectively. The 2-(β-D-glucopyranosyl)-4(5)-(2-naphthyl)-imidazole 24b, with a 3.5 μM IC50 towards SGLT2, emerged as the best inhibitor. Taking into account its nanomolar inhibition of GP, 24b represents a dual-targeting compound, which can be a starting point for further antidiabetic studies in this direction. The SGLT inhibitors are highly sensitive to the substitution pattern of the aglycon as outlined in the introduction. On the other hand, the substituent effects in glucosyl heterocycles on GP inhibition are relatively unknown [56]. Thus, extensive computational and synthetic work will be necessary to optimize the dual-target efficiency. Our present efforts are extending in these directions.  [72,73] a The inhibition of SGLT2 and SGLT1 was measured in TSA-201 cells. b rmGPb: rabbit muscle glycogen phosphorylase b. The synthesis of the compounds is also described in the cited references. c NI: no inhibition at 1000 μM. d NI: no inhibition at 625 μM. Table 3 inhibited SGLT2 more than SGLT1 with a selectivity range of ~2-27. There were two exceptions to this trend: compounds 22b and 25b (both with a 2-naphthyl substituent) showed some preference for SGLT1 inhibition. In two comparisons, 21a-21b and 24a-24b the 2-naphthyl derivatives were stronger inhibitors of SGLT2 than SGLT1 by a factor of ~7 and ~26, respectively. The 2-(β-D-glucopyranosyl)-4(5)-(2-naphthyl)-imidazole 24b, with a 3.5 μM IC50 towards SGLT2, emerged as the best inhibitor. Taking into account its nanomolar inhibition of GP, 24b represents a dual-targeting compound, which can be a starting point for further antidiabetic studies in this direction. The SGLT inhibitors are highly sensitive to the substitution pattern of the aglycon as outlined in the introduction. On the other hand, the substituent effects in glucosyl heterocycles on GP inhibition are relatively unknown [56]. Thus, extensive computational and synthetic work will be necessary to optimize the dual-target efficiency. Our present efforts are extending in these directions.

Materials and Methods
17.72 ± 2. 32 12.98 ± 2.88 0.41 [72,73] a The inhibition of SGLT2 and SGLT1 was measured in TSA-201 cells. b rmGPb: rabbit muscle glycogen phosphorylase b. The synthesis of the compounds is also described in the cited references. c NI: no inhibition at 1000 µM. d NI: no inhibition at 625 µM.
Similar to other classes of SGLT2 inhibitors, the inhibition highly depends on substitution of the distal aromatic ring of the aglycon part. In the case of the 1,2,4-and 1,3,4-oxadiazoles, only the unsubstituted derivatives (Table 2, 10a and 13a) showed inhibitory effects, while the introduction of a methoxy or a chloro group into the para position of the benzene ring, such as in 10b and 13b and 10c and 13c, respectively, led to complete loss of the inhibition. In the pyrimidine series, not only the benzyl (16a), but also the 4-methoxybenzyl (16b) derivative, displayed inhibitory activity in the micromolar range. In contrast, compound 16c, which has a 4-chlorobenzyl residue, remained inactive. Replacement of the distal benzene ring by a 2-naphthyl moiety resulted in a non-inhibitory  Table 2 had remarkable inhibitory effects against rabbit muscle GPb enzyme in the concentration tested.
Next, a series of known GP inhibitors 21-25 (Table 3) synthesized and investigated in our group [68][69][70][71][72][73] were assayed against SGLTs. The compounds were selected from non-inhibitory ones to low nanomolar GP inhibitors. Former GP inhibition studies revealed the essential role of the heterocyclic linker between the glucose unit and the distal aromatic moiety [21,55,56]. With respect to the latter aromatic group, the 2-naphthyl substituted derivatives 21b-25b surpassed the phenyl substituted ones 21a-25a by a factor of approx. 10 in most cases. The SGLT assay gratifyingly indicated low micromolar inhibitions for many of these compounds. Though thiazoles 23 proved inefficient towards both transporters in the studied concentration, the rest of the compounds had varying effects strongly indicating the substantial role of the heterocyclic linker in this biomolecular interaction.
Most of the compounds in Table 3 inhibited SGLT2 more than SGLT1 with a selectivity range of~2-27. There were two exceptions to this trend: compounds 22b and 25b (both with a 2-naphthyl substituent) showed some preference for SGLT1 inhibition. In two comparisons, 21a-21b and 24a-24b the 2-naphthyl derivatives were stronger inhibitors of SGLT2 than SGLT1 by a factor of~7 and~26, respectively. The 2-(β-D-glucopyranosyl)-4(5)-(2-naphthyl)-imidazole 24b, with a 3.5 µM IC 50 towards SGLT2, emerged as the best inhibitor. Taking into account its nanomolar inhibition of GP, 24b represents a dualtargeting compound, which can be a starting point for further antidiabetic studies in this direction. The SGLT inhibitors are highly sensitive to the substitution pattern of the aglycon as outlined in the introduction. On the other hand, the substituent effects in glucosyl heterocycles on GP inhibition are relatively unknown [56]. Thus, extensive computational and synthetic work will be necessary to optimize the dual-target efficiency. Our present efforts are extending in these directions.

General Methods
Melting points were measured on a Kofler hot-stage and the values are uncorrected. Optical rotations were determined on a P-2000 polarimeter (Jasco, Easton, MD, USA) at rt. NMR spectra were recorded with DRX360 (360/90 MHz for 1 H/ 13 C) and DRX400 (400/100 MHz for 1 H/ 13 C) spectrometers (Bruker, Karlsruhe, Germany). Chemical shifts are referenced to internal Me 4 Si ( 1 H) or the residual solvent signals ( 13 C). To avoid ambiguities in assignment of the NMR signals sugar protons and carbons are marked by primed numbers. HRMS spectra were obtained by a Bruker maXis II spectrometer using the electrospray ionization technique. TLC was performed on DC Alurolle Kieselgel 60 F 254 (Merck, Darmstadt, Germany) and the plates were visualized by gentle heating. Column chromatography was carried out by applying Kieselgel 60 (particle size  µm; Molar Chemicals, Halásztelek, Hungary) silica gel. Anhydrous solvents were obtained by standard distillation methods. CH 2 Cl 2 , CHCl 3 , CH 3 CN (4 Å molecular sieves), toluene, and m-xylene (sodium wires) were distilled from P 4 O 10 and stored over the indicated drying agents. MeOH was distilled over Mg turnings and iodine. Distillation of 1,4-dioxane was performed from sodium benzophenone ketyl and the solvent was stored over sodium wires. To a solution of C-(2,3,4,6-tetra-O-benzoyl-β-D-glucopyranosyl)formamidoxime (7) in anhydrous 1,4-dioxane (3 mL/100 mg substrate), the corresponding acid chloride (1.1 equiv.) was added and the reaction mixture was stirred at rt. After the disappearance of the starting amidoxime 7, judged by TLC (EtOAc-hexane 1:2), the solvent was removed by diminished pressure and the crude product was purified by column chromatography. To a solution of the corresponding O-acyl-C-(2,3,4,6-tetra-O-benzoyl-β-D-glucopyranosyl)formamidoxime (8) in anhydrous toluene (3 mL/100 mg substrate), a catalytic amount of tetrabutylammonium fluoride (0.1 equiv., 1 M solution in THF) was added. The reaction mixture was heated at boiling temperature until TLC (EtOAc-hexane 1:2) showed that the starting material was completely consumed. When the reaction was complete the solvent was removed under reduced pressure and the resulting crude product was purified by column chromatography. The 5-(2,3,4,6-tetra-O-benzoyl-β-D-glucopyranosyl)tetrazole (11) and the corresponding acid chloride (3 equiv.) were suspended in anhydrous m-xylene (3 mL/100 mg tetrazole) and the reaction mixture was heated at 140 • C. After completion of the reaction monitored by TLC (EtOAc-hexane 2:3), the mixture was evaporated under reduced pressure and the resulting crude product was purified by column chromatography.

General Procedure 4 for Cleavage of the O-Benzoyl Protecting Groups by the Zemplén Method to Get Test Compounds 10 and 13
An O-benzoyl protected compound was dissolved in an anhydrous MeOH-CHCl 3 solvent mixture (5 mL:1 mL/100 mg substrate) and a catalytic amount of~1 M solution of NaOMe in MeOH was added. The reaction mixture was left to stand at rt until the TLC (CHCl 3  To a solution of C-(β-D-glucopyranosyl)formamidine hydrochloride (14) in anhydrous CH 3 CN (2 mL/100 mg substrate,) the corresponding 4-(trimethylsilyl)but-3-in- 2-one (1 equiv.), Na 2 CO 3 (2 equiv.), and one drop of water were added. The reaction mixture was heated under reflux temperature for 8 h, then treated with activated charcoal and filtered. The solution was evaporated in vacuo and the residue was purified by column chromatography.

General Procedure 6 for Removal of O-Benzyl Protecting Groups by Using BCl 3 to Get Test Compounds 16 and 18
The corresponding O-perbenzylated C-glucosyl heterocycle (15 or 17) was dissolved in anhydrous CH 2 Cl 2 (5 mL/100 mg substrate). The stirred reaction mixture was cooled to -78 • C and BCl 3 (5 equiv.,~1 M solution in CH 2 Cl 2 ) was added. The stirring was continued at this temperature and the reaction was monitored by TLC (EtOAc-hexane 1:2 and CHCl 3 -MeOH 3:1). After the complete disappearance of the starting material (4 h), MeOH (5 mL) was added to the reaction mixture and the mixture was warmed to rt. Solvents were removed under diminished pressure and the residue was purified by column chromatography (CHCl 3 -MeOH = 19:1 → 9:1 gradient).

Lentiviral Transfection
Cells (100,000 cells, TSA201) were plated into each well of a 24-well plate and incubated overnight at 37 • C. Six wells were transfected with lentiviral particles (NM_000343 for SGLT1, and NM-003041 for SGLT2) containing the mammalian expression vectors with SGLT1 or SGLT2 and the puromycin resistance gene, using 8 µg/mL hexadimetrine bromide (MOI = 3.5 for the best transfection). Two wells were treated only with hexadimetrine bromide as a control. After 24 h of transfection, the medium was changed to fresh medium without antibiotics. Cells were collected and transferred into T-25 flasks after 3 days. On the fifth day, 10 µg/mL puromycin was added to select only the transfected cells. The nontransfected cells cannot survive this high concentration of antibiotic. To keep up the selection, the medium including 10 µg/mL puromycin was replaced every two days. The cells lived under puromycin pressure until the day of the experiments.

Western Blot
Cells were grown in 60-mm  USA) containing 10% heat-inactivated fetal bovine serum, and 1.5% L-glutamine at 37 • C in humidified atmosphere of 5% CO 2 .
The cytotoxicities of all earlier and newly synthesized inhibitors were measured by fluorimetric CyQUANT assay. Cells (25,000/well) were plated into 96-well plates. The next day, the media was removed, and replaced with DMEM containing 10% FBS, 1.5% L-glutamine, and synthesized the compounds (1000 µM). Controls were treated with vehicle (DMSO). After 4 h, the supernatant was removed and cells were washed with PBS (Phosphate Buffer Solution pH 7.4). Then, the plates were frozen. The next day the plates were thawed and 200 mL CyQUANT reagent was added to each well. The plates were incubated for 5 min at room temperature, protected from light. The fluorescent signal was measured using a fluorimeter (excitation wavelength: 480 nm, emission wavelength: 520 nm). The obtained values were plotted against control.
3.2.6. 2-NBDG Uptake Measurement TSA201 cells (25,000) with and without overexpressed SGLT1 or 2 were put into poly-Llysine (P5999, Sigma-Aldrich, St. Luis, MO, USA) treated wells of a 96-well plate. The next day, the cells were starved for 1 h with 90 µL glucose-free medium (D5036-10X1L Sigma-Aldrich, St. Luis, MO, USA) (with or without inhibitor). After starving, cells were treated with 100 µM 2-NBDG and incubated for 30 min (again with or without inhibitor). The incubation time of cells with possible inhibitors was 90 min altogether. After incubation, the medium was removed, the wells were washed with PBS twice, lysed with 100 µL lysis buffer (0.1 M KH 2 PO 4 + 1% Triton-x, pH = 11), and stored at -20 • C. After thawing, the cells were resuspended and the fluorescent signal was measured using a fluorimeter (excitation wavelength: 485 nm, emission wavelength: 538 nm). Subsequently, the protein concentration was measured and NBDG uptake was normalized to protein amount.

Calculation of IC 50 for SGLT1 and 2
We selected the optimal concentration to identify the IC 50 of the compounds using dose-response measurements, then we chose the optimal concentrations between 20-80% of inhibition. After selecting optimal concentrations, each compound was tested at four different concentrations. Concentrations of 0.1, 1, 10, 100, and 1000 µM were used and the lowest and highest concentrations were removed, depended on the dose-response measurements. Examples of dose-response curves for dapagliflozin and compounds 22b and 24b are shown in the Supplementary Figure S3. To determine the specific effects of the compounds on SGLT1 or 2, the glucose uptake values in TSA201 cells without SGLT overexpression were subtracted from the values in TSA201 cell overexpressing SGLT. Overexpressed control minus original control was 100%. All calculations were prepared using the GraphPad Prism software (GraphPad Software, Inc., SanDiego, CA, USA). The data were plotted on a logarithmic scale as a function of inhibitor concentration. The IC 50 values were calculated by non-linear regression analysis from sigmoidal doseresponse curves. The results of kinetic experiments were calculated from at least three independent experiments. Each experiment contained four parallel samples that were averaged. The standard deviations were calculated based on the average values for the independent experiments.

Determination of Inhibitory Constants (K i ) for Glycogen Phosphorylase
Enzyme activity in the direction of glycogen synthesis was assayed. Kinetic data were collected using the muscle phosphorylase b (dephosphorylated, GPb) isoform. Kinetic data for the inhibition of phosphorylase by glucose analogues were obtained in the presence of 10 µg/mL enzyme, varying concentrations of α-D-glucose-1-phosphate (4-40 mM), constant concentration (1%) of glycogen, and AMP (1 mM). Enzymatic activities were presented in the form of a double-reciprocal plot (Lineweaver-Burk). The plots were analysed by a non-linear data analysis program. The inhibitor constants (K i ) were determined by secondary plots, replotting the slopes from the Lineweaver-Burk plot against the inhibitor concentrations. The means of standard errors for all calculated kinetic parameters averaged to less than 10% [78,79].

Statistical Analyses
Protein expression measurements, cytotoxicity assays, and glucose uptake were repeated in at least 3 independent experiments. The data for these measurements are presented as means ± SE. Protein expression levels, cytotoxicity assay, and glucose uptake levels were compared using ANOVA with Fisher's posthoc analysis. Statistical comparisons were performed using Statistica software (v.13.6.0, TIBCO Software, Palo Alto, CA, USA).
All IC 50 and K i calculations were prepared using the GraphPad Prism software. The data were plotted on a logarithmic scale as a function of inhibitor concentration. The IC 50 values were calculated by non-linear regression analysis from sigmoidal doseresponse curves. The results of kinetic experiments were calculated from at least three independent experiments. Each experiment contained four parallel samples that were averaged. The standard deviations were calculated based on the average values for the independent experiments.

Conclusions
New series of (C-β-D-glucopyranosylhetaryl)methyl benzene derivatives with 3glucosyl-1,2,4-and 2-glucosyl-1,3,4-oxadiazole, 2-glucosylpyrimidine, and 2-glucosylimidazole skeletons were prepared. The deprotected compounds, together with some previously obtained 1-glucosyl-1,2,3-triazole and 3-glucosyl-1,2,4-triazole derivatives and known β-Dglucopyranosyl heterocyclic glycogen phosphorylase inhibitors, were assayed as SGLT1 and 2 inhibitors in a transfected TSA201 cell system. The (C-β-D-glucopyranosylhetaryl)methyl benzene type compounds showed low micromolar inhibition of SGLT2 in some cases, demonstrating that the structure of the proximal heterocyclic ring has a strong bearing on the inhibitory effect. On the other hand, none of these derivatives inhibited GP. Among the GP inhibitors, some low micromolar inhibitors of SGLT2 were found. The most efficient compound, 2-(β-D-glucopyranosyl)-4(5)-(2-naphthyl)-imidazole, is also the best glucose analogue inhibitor of GP known to date. This finding expands and proves the dual-target concept to the SGLT2-GP proteins as a potential antidiabetic strategy for the first time. Current computational and synthetic efforts are directed at fine-tuning the glucosyl-heterocyclic skeleton and its substitution pattern to find more efficient dual-target compounds.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/ph14040364/s1, Figure S1: Western blot analysis of SGLT1 and GAPDH protein in TSA, transfected TSA201 cells, and mouse kidney, Figure S2: Western blot analysis of SGLT2 and GAPDH protein in TSA, transfected TSA201 cells, and mouse kidney, Figure S3: Representative dose-response curves for calculating IC 50 values.
Author Contributions: Á.S. performed the biochemical assays, E.S., N.É.H., S.K., and K.E.S. prepared the compounds, K.U. contributed to the description of the biological experiments and manuscript editing, L.S. conceived the research and wrote the paper, T.D. evaluated the biochemical measurements and wrote the paper, É.B. conceived the research, designed the synthetic work, and wrote the paper. All authors have read and agreed to the published version of the manuscript.