Tải bản đầy đủ - 0 (trang)
4 Phlorizin, SGLTs, and Diabetes

4 Phlorizin, SGLTs, and Diabetes

Tải bản đầy đủ - 0trang

9.5 Phlorizin Analogs: O-Glucosides



Zemplen and Bognar in the 1940s (Scheme 9.2), the early analog programs focused

on O-glucoside analogs with superior characteristics to phlorizin: increased

potency versus SGLT2, increased selectivity over SGLT1, improved absorption in

the small intestine, increased resistance to b-glycosidase cleavage, and, for the

aglycone, reduced inhibition of GLUT transporters.

One of the first papers describing structure–activity relationships (SARs) around

phlorizin was published by investigators at Tanabe who profiled analogs in a rat in

vivo model of urinary glucose excretion (UGE) [25]. Activity in this model was

reduced by structural changes in the linker between aryl groups (A- and B-rings)

and required one, but not both, phenolic hydroxyl group on the A-ring (Figure 9.1).

A few sugar modifications were also explored, replacing glucose with galactose,

xylose (pyranoside), and the disaccharides maltose and lactose. In all these cases,

activity was attenuated. On the other hand, changes on the B-ring were often

tolerated and sometimes improved potency, for example, by replacement of the

phenolic hydroxyl group with a small alkyl substituent or fusion to form a 6,5-fused

heterocycle (e.g., benzofuran). From this work, which did not directly measure

potency versus SGLT1 and SGLT2, compound 7 (Figure 9.3) emerged as a

promising candidate based on its superior activity in the rat UGE model compared

to phlorizin. However, realizing that 7 would likely remain a substrate for

b-glycosidases, various ester and carbonate derivatives linked to the sugar moiety

were explored as potential prodrugs. Ideally, these would not be recognized by

intestinal b-glycosidases and would be cleaved by esterases to the parent drug (7).



Me



OH



O

H

N



RO



O



HO



O



7: R = H (T-1095A)

8: R = CO2Me (T-1095)



OH

OH

H

N



CF3



SMe



CF3



N



O



9



HO



SMe



OMe



N

HO



O



HO



O



RO



10



OH



O



O



OH



OH 1: R = H (sergliflozin)

13: R = CO2Et (sergliflozin etabonate)



HO



OH



OMe



N



S



HO



O



HO



O

OH



OH



O



N

RO



12 (AVE-2268)



O



O



OH



OH 11: R = H (remogliflozin)

14: R = CO2Et (remogliflozin etabonate)



HO



Figure 9.3 O-Glucoside SGLT inhibitor candidates.



307



308



9 From Natural Product to New Diabetes Therapy: Phlorizin and the Discovery



From among the derivatives of 7 evaluated as prodrugs, the methyl carbonate

(8, T-1095, Figure 9.3) was selected as a clinical candidate for T2DM,

becoming the first SGLT inhibitor to enter the clinic [26]. This compound

showed improved (1.3-fold) potency in the rat UGE model relative to 7, while

being 10-fold less active than 7 in rat brush border membrane vesicles, a

crude in vitro measure of SGLT activity.

A transition to structurally more distinct O-glucoside SGLT inhibitors was

suggested from a report that a series of 4-benzylpyrazolones represented by 9

(Figure 9.3) elicit a glucosuric response in mice [27]. This key observation led to the

suggestion that in mice, glucosidation occurs to form SGLT inhibitors such as 10

in vivo. Synthesis of 10 and analogs confirmed this hypothesis and, notably, high

potency versus SGLT2 and good selectivity against SGLT1 was achieved. Disclosures relating to 7 and 10 led to a flurry of new analog synthesis at a number of

companies, modifying or combining structural features of 7 and 10, and in some

cases, replacing the glycosidic or pyran oxygen with nitrogen or sulfur [28]. As a

result, several other O-glucoside SGLT2 inhibitors joined 8 in clinical development,

for example, sergliflozin (1) [20], remogliflozin (11) [29], and AVE-2268 (12) [30]. As

with 7, 1 and 11 were developed as carbonate derivatives of the C6 hydroxyl (13 and

14, respectively) to minimize glycosidase degradation and achieve systemic

exposure of the parent compound following oral administration (Figure 9.3).

Biological data for representative O-glucoside SGLT inhibitors are summarized in

Table 9.1.

Despite the improvements made over phlorizin with respect to in vitro potency

and in vivo efficacy, the clinical doses explored for O-glucoside SGLT2 inhibitors

remained fairly high, typically ranging in the hundreds of milligrams to more than

a gram (Table 9.1). Furthermore, only modest levels of daily glucose output were

achieved. For example, once-daily administration of a 500 mg dose of sergliflozin

etabonate (13) to healthy volunteers produced urinary glucose output of 17 g/day,



Table 9.1



O-Glucoside SGLT inhibitors.

Selectivity

versus

SGLT1



SGLT2 Ki

(nM)



35

6.6

9



10

30

296–800



18.6

IC50 ¼ 50 nM

2.4



8

13



14



365–1100



12.4



14



Compound



Human

SGLT2

EC50 (nM)



Phlorizin (4)

T-1095A (7)

Sergliflozin

(1)

Remogliflozin

(11)

AVE2268 (12)



Adapted from Ref. [7].

a) Healthy normal volunteers.



Prodrug



None



Clinical

dose

(mg)a)



Glucose

output over

24 h (g)



200



12



500



17



1200

2000



14

21



9.6 Phlorizin Analogs: C-Glucosides



a small fraction of that expected from full inhibition of SGLT2 over 24 h. It now

appears that most, if not all, O-glucoside SGLT2 inhibitors have been discontinued

in development, being replaced by C-glucoside inhibitors, which possess higher

metabolic stability and therefore can achieve greater efficacy (over 60 g of glucose/

day) at much lower doses, even after once-daily (qd) administration.



9.6

Phlorizin Analogs: C-Glucosides



A logical alternative to prodrug derivatization as a strategy for improving

glycosidase stability was to explore analogs in which the glucosidic oxygen was

replaced with a methylene spacer (e.g., 15, Scheme 9.3). However, this approach

presented significant challenges starting with synthesis, which required CÀÀC

bond formation as opposed to the more familiar CÀÀO bond forming reactions

employed for obtaining O-glucosides. The first synthesis of a C-glucoside analog of

an SGLT inhibitor was reported in 2000 (Scheme 9.3) [31]. To prepare 15,

installment of the aglycone moiety was achieved via hydroboration of exo-glycal 16

and subsequent Suzuki cross-coupling with aryl triflate 17.



OMOM

MOMO

MOMO



O



1) 9-BBN,THF, reflux



OMOM

OMOM



2) Pd (dppf).CH2Cl2,

K3PO4, H2O, DMF



16



OMOM



OTf O



17



MOMO

MOMO

O



O



O



OMOM

OMOM



HCl,

aq. MeOH



HO



50%



OH



O



HO



O



O



O

OH



15



OH



Scheme 9.3 Synthesis of 15, a C-glucoside analog related to phlorizin (4) and T-1095A (7).



Another challenge in exploring C-glucoside analogs proved to be achieving

sufficient potency against the target, SGLT2. For example, compound 15 exhibited

activity at least 10-fold less than that of phlorizin [31]. Similarly, compound 18

(Figure 9.4), a C-glucoside related to sergliflozin (1), exhibited an SGLT2 EC50 of

500 nM, a reduction of more than 60-fold compared to the corresponding

O-glucoside (EC50 ¼ 8 nM) [32]. Attaching the aryl aglycone directly to the anomeric

carbon of the glucose ring also afforded disappointing results as exemplified by 19

(Figure 9.4), giving an SGLT2 EC50 >4000 nM. Ironically, the breakthrough for

increasing in vitro potency of C-glucosides came as a result of the attempted



309



310



9 From Natural Product to New Diabetes Therapy: Phlorizin and the Discovery

Me

A



B



O



HO



HO



HO



O



HO



OH



OH

OH



18



OH



19



Me



Me

HO



O



HO



HO



OH



O



HO

24



OH



Me

OH



OH



25



Figure 9.4 Early C-glucoside analogs.



synthesis of O-glucoside 20, which was not isolated from the reaction of the

bisphenol 21 with bromoglucoside 22 under typical glucosylation conditions

(Scheme 9.4). Rather, CÀÀC bond formation accompanied by p-methyoxybenzyl

migration occurred to provide C-glucoside 23 in low yield [33]. In light of the

emerging SAR in the O-glucoside series, the modest activity exhibited by 23

(EC50 ¼ 500 nM) was surprising given the number of polar substituents on the

A-ring. This prompted the chemists at Bristol-Myers Squibb to prepare b-Cglucoside 24 (Figure 9.4) lacking the polar phenolic hydroxyl and secondary amide

groups present in 23. Amazingly, 24 showed an SGLT2 EC50 of 22 nM with high

selectivity versus SGLT1 (>600-fold). As the first potent and selective C-glucoside

SGLT2 inhibitor to be characterized, 24 became the springboard for entry into a

new chemical series lacking any potential for instability due to glycosidases.

OH



OMe



N

O



AcO

AcO



Br



OH



OAc

OAc



22



OMe



HO



N

CdCO3, toluene

reflux



OH O



O



HO



O



O



OH



20



OH



21

HO

HO



O



HO



CONHEt

OH

A



OH



OMe

B



23



OH



Scheme 9.4 Serendipitous discovery of the C-aryl glucoside series at Bristol-Myers Squibb.



Early analog synthesis in the C-aryl glucoside series relied on chemistry

developed by Kishi and coworkers [34] and later extended to C-aryl glucosides by

Kraus and Molina [35] and others. This involved the addition of aryllithium



9.6 Phlorizin Analogs: C-Glucosides

O



BnO



O



BnO



ArLi, THF



BnO



OBn



O



OH

Ar



BnO



OBn



BF3∙OEt2

Et3SiH



OBn



BnO

BnO



MeCN



OBn



HO



H2, Pd(OH)2

or BCl3, CH2Cl2



O



HO



O



Ar

OBn



OBn



Ar

OH



OH



Scheme 9.5 General route to C-aryl glucoside analogs.



reagents to 2,3,4,6-tetra-O-benzyl-D-gluconolactone, followed by reduction of

the resulting lactol with triethylsilane in the presence of a Lewis acid

(BF3ÁOEt2) (Scheme 9.5). Formation of the desired b-C-aryl glucosides was

generally favored over the inactive a-isomers. However, the strongly basic

conditions used in the CÀÀC bond forming step, as well as the conditions

required for cleavage of the benzyl protecting groups (catalytic hydrogenolysis

or treatment with BCl3), limited the scope of aglycone pieces and substituents

that could be incorporated to develop SAR. Thus, initial analogs explored only

simple alkyl substituents on the B-ring (Figure 9.4) in the context of different

linkers between the aromatic rings (alkyl, O, and S). Nonetheless, this work

established key aspects of the SGLT2 SAR and identified compound 24 as

having the best profile among the early analogs synthesized. Changes in the

methylene linker between aromatic rings in compound 24 – deletion to form

a biaryl group and replacement with O, S, (CH2)2, or (CH2)3 – led to losses in

activity of at least 10-fold. Deletion of the 40 -methyl substituent (replacing it

with H) and placement of the methyl group at the 30 -position were also

detrimental changes, even in combination with changes in the linker. Linker

attachment at the para position of the A-ring as in 25 (Figure 9.4) was not

well tolerated, as had been observed with linker attachment at the ortho

position (e.g., 18).

Continued follow-up of the C-aryl glucoside series, exemplified by 24, demanded

the development of alternative routes or protecting group strategies to allow greater

scope for exploring the SAR around the aglycone. One approach was to attach the

linker and B-ring (Figure 9.4) at a late stage, employing the aryl bromide 26

(Scheme 9.6), which was prepared from 2,3,4,6-tetra-O-benzyl-D-gluconolactone



BnO



O



BnO



Br

OBn



OBn



Suzuki or Stille coupling,

then deprotection



HO



O



HO



R

OH



OH



26



Scheme 9.6 C-aryl glucosides: late-stage aglycone modification.



311



312



9 From Natural Product to New Diabetes Therapy: Phlorizin and the Discovery



and 3-lithiobromobenzene in the usual way [32]. However, this method still required

benzyl group removal at the end of the synthesis, which could be problematic with

certain functionalities. Thus, a route employing trimethylsilyl (TMS) protection was

developed starting with 2,3,4,6-tetra-O-trimethylsilyl-D-gluconolactone (Scheme 9.7).

In this case, deprotection was achieved readily under the acidic conditions used for

formation of the intermediate O-methyl lactol (27) [36].



TMSO

TMSO



O



O



OTMS

OTMS

ArLi

THF



TMSO

TMSO



HO



O



OH

Ar



OTMS

OTMS



O



OMe

Ar



HO



OH

OH



MeSO3H

MeOH



AcO



HO



BF3∙OEt2

Et3SiH/MeCN



AcO

OAc



HO



27



Ac2O

pyridine

DMAP, CH2Cl2

OMe

O

Ar

BF3∙OEt2



Ac2O

pyridine

DMAP, CH2Cl2

AcO



Ar

OH



OH



OAc Et3SiH, MeCN

H2O



29



O



2: Ar = 4-Cl-3-(4ethoxybenzyl)



LiOH, THF

MeOH, H2O

O



AcO



Ar

OAc



OAc



28



Scheme 9.7 Optimization of C-aryl glucoside synthesis.



Employing these new routes for the synthesis of new C-aryl glucoside analogs,

additional important observations regarding SAR were made [7,33]. Most notably,

the addition of small lipophilic substituents (e.g., Cl) at the 4-position on the A-ring

was found to increase SGLT2 affinity further, with EC50 values close to or even less

than 1 nM. In place of methyl at the 4-position on the B-ring, other small- or

medium-sized lipophilic substituents (e.g., ethoxy) were often well tolerated or

improved SGLT2 affinity. As a result of this work, compound 2 emerged as a likely

candidate [36].

Characterization of 2 [36] revealed a promising biological profile: high potency

versus human SGLT2 (EC50 ¼ 1.1 nM) as well as high selectivity (>1300-fold)

versus human SGLT1 and the facilitative transporters GLUT1 and GLUT4. Besides

having a structure favoring oral bioavailability by eliminating any possibility for

glycosidase degradation, 2 exhibited physicochemical properties (molecular weight,

lipophilicity, solubility, etc.) well within the ranges favoring good oral exposure.

Together, these properties translated into good in vitro cell permeability, low

cytochrome P450-mediated metabolism in vitro (liver microsomes), and excellent

exposure (good absorption and elimination half-life) in test species.

As a result of the improved exposures and half-lives, 2 was efficacious at much

lower doses than compounds in the O-glucoside class. For example, a 0.1 mg/kg

dose of 2 produced a level of UGE in rats of about 550 mg glucose per 200 g body

weight over 24 h. In contrast, a 30 mg/kg dose of the prodrug sergliflozin etabonate

(13) elicited a glucosuric response in rats of only 400 mg glucose per 200 g body

weight over 24 h. This indicated a difference in potency of at least 300-fold, a much



9.6 Phlorizin Analogs: C-Glucosides



higher multiple than the 6-fold difference in affinity for SGLT2 between 2 and 1. In

further preclinical studies [37], the ability of 2 to cause glucosuria translated into

significant reductions in body weight, plasma glucose concentrations, and

postprandial glucose exposure following oral glucose administration.

Compound 2 progressed into clinical studies where it again demonstrated clear

differentiation from compounds in the O-glucoside class. For example, mirroring

the preclinical UGE results above, a 20 mg dose of 2 (assigned the generic name

“dapagliflozin”) elicited a 24 h glucose output of $64 g in healthy human

volunteers [7]. In stark contrast, a 500 mg dose of sergliflozin etabonate (13)

produced UGE of only 17 g over the same time interval, as noted earlier (Table 9.1).

Again, the improved potency of dapagliflozin relative to sergliflozin etabonate (13)

and other O-glucosides was attributed largely to an improved half-life. In healthy

volunteers, the half-life of dapagliflozin was $15 h, significantly longer than halflives of compounds in the O-glucoside class such as sergliflozin (1) with a half-life

in humans of less than 1 h [38].

As the most advanced SGLT2 inhibitor in development, dapagliflozin has now

completed extensive evaluation in multiple clinical studies in more than 5000

patients with T2DM. These studies have solidly established the ability of the

compound to produce significant reductions in body weight, as well as levels of

glycated hemoglobin (HbA1c), the gold standard indicator of average glucose

exposure over long periods of time. Importantly, acting via a glucose-dependent

and insulin-independent mechanism (SGLT2 inhibition), dapagliflozin treatment

minimizes the risk of hypoglycemia and could potentially help preserve pancreatic

b-cell mass [39].

The discovery of dapagliflozin stimulated further refinement in C-aryl glucoside

synthesis, especially in connection with its progression into toxicology studies and

subsequent human clinical trials. In the earlier synthesis of dapagliflozin

(Scheme 9.7) [36], peracetylation of crude 2 was carried out to provide the

crystalline tetraacetate derivative 28. Recrystallization of this material, followed by

hydrolysis, allowed removal of small amounts of the undesired a-anomer formed

during the reductive removal of the anomeric methoxy group. In contrast, the

preferred method for scale-up of dapagliflozin involves peracetylation prior to the

reduction step [40]. Despite concerns that anchimeric assistance by the C2 acetoxy

group would direct delivery of hydride from the b-face of the oxocarbenium ion

intermediate, reduction of the tetraacetate 29 (Scheme 9.7) using triethylsilane,

BF3ÁOEt2, and 1 equiv of water provides almost exclusively the desired b-anomer 28

in high yield (b:a ratio >20 : 1). The high b-selectivity of the process is attributed to

overriding steric and electronic influences from the stabilizing C1 aryl group on the

oxocarbenium ion intermediate. It was proposed that the addition of water

promotes generation of F3BOỵH2, a strong Brứnsted acid, which accelerates

oxocarbenium ion formation.

Whereas O-glucosides have long been prepared via glucosyl bromide derivatives

(as in Scheme 9.2), C-glucosides have generally not been accessible via this

approach, which circumvents a lactol reduction step. Recently, however, building



313



314



9 From Natural Product to New Diabetes Therapy: Phlorizin and the Discovery

OPiv

PivO



O



PivO

OPiv



30



Br



Ar2Zn



OPiv



toluene

n-Bu2O

90oC



PivO

PivO



O

O

O



31



PivO



O



PivO



Ar

OPiv



OPiv

NaOMe

MeOH



2, etc.

Scheme 9.8 “Direct” C-aryl glucoside synthesis.



upon the work of Gagne [41], Knochel and coworkers developed a high-yielding

stereoselective method for preparing dapagliflozin and other C-aryl glucosides via

the “direct” reaction of diaryl zinc reagents (generated in situ from aryllithium

or aryl Grignard reagents and ZnBr2ÁLiBr complex) with glucosyl bromide 30

(Scheme 9.8) [42]. The key to the success of the method was the use of a

solvent system such as dibutyl ether/toluene. In this system, decomposition of

30 is relatively slow compared to that seen using tetrahydrofuran (THF).

Unlike the method developed by Gagne, a transition metal catalyst is not

required. The high degree of b-stereoselectivity presumably arises from attack

of the organometallic reagent on the oxocarbenium ion 31 formed via

anchimeric participation of the pivalate group at the 2-position. The choice of

pivalate as hydroxyl protecting group was also critical in that it is less reactive

toward attack by organometallic reagents than other esters (e.g., acetate), and

yet is still capable of neighboring group participation in forming an

oxocarbenium ion.



9.7

C-Glucosides: Aglycone Modifications



As would be expected, the disclosures related to dapagliflozin and its advantages

over O-aryl glucoside-based SGLT2 inhibitors stimulated another wave of activity in

SGLT2 pharmaceutical research to capitalize on this important breakthrough.

During the early part of this wave, efforts were focused largely on modifications to

diarylmethylene (“aglycone”) side chain with the goal of establishing a proprietary

foothold with regard to the composition of matter. Fueling these efforts were the

advancements in synthesis described earlier for the synthesis of dapagliflozin and

analogs. Some representative structures emerging from this aglycone modification

strategy are shown in Figure 9.5; canagliflozin (32) [43], ipragliflozin (33) [44], and

empagliflozin (34) [45] are in clinical development. Key data on these compounds

and other SGLT2 inhibitors in development are shown in Table 9.2.

In the course of exploring heterocyclic variants of the aglycone side chain, it was

found that the aryl lithium/gluconolactone addition step generalized in Scheme 9.7

does not work in certain cases [46]. Although the reasons for failure are not



9.7 C-Glucosides: Aglycone Modifications

F



F



S

HO



O



HO



HO



OH



HO

canagliflozin (32)



OH



Cl

HO



ipragliflozin (33)



S

O

HO



OH



N



O



HO

empagliflozin (34)



OH



OH

OH



O



O



HO



S



O



OH

35



OH



F

HO

HO



N



O



HO



HO



OH

OH



O



HO

36



OH

OH



37



Figure 9.5 SGLT2 inhibitor candidates bearing aglycone modifications.



discussed, an alternative route avoiding the problem was developed by workers at

Tanabe (Scheme 9.9). Here, the aglycone side chain is introduced as an aryl

bromide that is coupled to a glucal boronate [47]. This approach was applied to the

synthesis of benzisothiazole analog 35 (Figure 9.5) [46]. As exemplified by 36

(SGLT2 IC50 ¼ 1.9 nM) [48], aglycone modification of the C-aryl glucoside series has

included heterocyclic variants wherein the aryl side chain is linked to the sugar via

nitrogen (Figure 9.5).



Table 9.2



C-Glucoside and sugar-modified SGLT inhibitor clinical candidates.



Compound



Human SGLT2 IC50

(nM)



Selectivity versus

SGLT1



Dapagliflozin (2)

Canagliflozin (32)



1.1

2.2



1200

414



Ipragliflozin (33)

Empagliflozin (34)

Ertugliflozin (3)

LX4211 (38)

Tofogliflozin (43)

Luseogliflozin (47)



7.4

3.1

0.9

1.8

2.9

2.3



255

>2500

2235

20

2930

1765



Adapted from Ref. [6b].



Dose range qd

(mg)

2.5–10 (Phase III)

100–300

(Phase III)

50–300 (Phase II)

10–25 (Phase III)

1–25 (Phase II)

150–300 (Phase II)

2.5–40 (Phase II)

0.5–5 (Phase II)



315



316



O

t-Bu Si

O

t-Bu



9 From Natural Product to New Diabetes Therapy: Phlorizin and the Discovery



O



O

B



OTIPS



HF, Et3N, 60°C

or TBAF, 23°C



O



ArX, Pd(PPh3)2Cl2

DME, Na2CO3

85°C



HO



O



HO



O

t-Bu Si

O

t-Bu



O



Ar



OTIPS



BH3∙THF, 0 to 23°C

then H2O2, NaOH



O

t-Bu Si

O

t-Bu



O



Ar



OH

OTIPS



Ar

OH



OH



Scheme 9.9 Alternative approach to C-aryl glucosides.



A clear downside to the aglycone modification strategy was that changes

required to secure proprietary chemical space often came at the expense of

optimizing drug-like physicochemical properties such as molecular weight and

lipophilicity. In the clinic, this could result in poor absorption, suboptimal

pharmacokinetic (PK) properties (e.g., high clearance and short half-life), and

increased dose. In some cases, structural features associated with reactive

metabolite formation (e.g., thiophene) or ultraviolet light absorption (extended

unsaturated ring systems) were introduced, thereby posing increased risk of

idiosyncratic toxicity or phototoxicity [49]. The highly unusual azulene analog

37 (Figure 9.5), an advanced compound patented by Astellas [50], is a good

example of the structural extremes taken in the quest for proprietary aglycone

side chains.



9.8

C-Glucosides: Sugar Modifications



Despite potentially greater synthesis challenges, sugar modification in the

C-glucoside series emerged as an attractive strategy to avoid the potential pitfalls

associated with aglycone modification. Among the first disclosures pertaining to

this approach were patent applications from Boehringer Ingleheim outlining the

preparation of a number of C-glucoside analogs bearing changes at the C4 or C6

positions on the pyranose ring [51]. The syntheses (Scheme 9.10) employed wellestablished carbohydrate protecting group strategies starting with an unmodified

C-glucoside analog [52]. Unfortunately, the patent applications revealed very little

information relating to SGLT2 SAR around the sugar modifications exemplified.

Furthermore, it appears that none of the disclosed compounds advanced beyond

preclinical evaluation.

In 2010, Pfizer published the results of a systematic study to explore the SAR

associated with hydroxyl group deletion and methylation at each of the positions

around the pyranose ring [53]. This effort clearly established that structural change

is well tolerated at the C5 and C6 positions. For example, in the case of where the

C-glucoside bears a 4-chloro-3-(4-methoxybenzyl)phenyl side chain, hydroxyl

deletion (replacing OH for H) at C6 results in only a 2.4-fold increase in SGLT2



9.8 C-Glucosides: Sugar Modifications



317



further functionalization at C6



O



HO

MeO



Ar



O



O



HO



OH



Ar

SEMCl



HO



O



O

O

MeO



OMe



OMe



SEMO



O



HO



i-Pr2NEt



Ar

O



O

MeO



OMe



further

functionalization

at C4



diacetyl, BF3OEt2

HC(OMe)3, MeOH

6



HO



O

5

4



HO



3



Ar



1) TrCl, pyridine



OH



2) NaH, BnBr



1

2



HO

BnO



3) acid



OH



O



Ar

OBn



OBn



further

functionalization

at C6



1) PhCH(OMe)2, H+

2) NaH, BnBr

O



O

Ph



O



Ar

OBn



OBn



AlCl3, LiAlH4



BnO

HCl, NaBH3CN



O



HO



Ar

OBn



OBn



further

functionalization

at C4



Scheme 9.10 C-Aryl glucosides: general approaches to sugar modification.



IC50 (1–2.4 nM). In contrast, hydroxyl deletion at C2 and C3 leads to greater than

2000-fold loss in potency. The C4 hydroxyl deletion was tolerated somewhat

(20-fold loss in activity).

The tolerance of SGLT2 inhibition for structural modification at C5 and C6

was independently discovered and exploited by multiple research groups,

including those at Green Cross [54] and Lexicon [55]. A clinical candidate

emerged from the effort at Lexicon, the C5 methylthio analog LX4211 (38)

[56], a dual inhibitor of SGLT1 and SGLT2 (Figure 9.6 and Table 9.2). The

synthesis of this compound was reported recently in a patent application

(Scheme 9.11) [57].

Nucleophilic addition of the organometallic species derived from aryl iodide

39 to amide intermediate 40 (synthesized in four steps and 43% overall yield

starting from L-xylose) gave aryl ketone 41. Diastereoselective reduction of 41

using sodium borohydride in the presence of cerium trichloride, followed by

pyranose ring formation under acidic conditions, then provided 42 after

peracetylation. Sequential trapping by thiourea of the putative oxocarbenium

ion formed upon reaction of 42 with TMSOTf, treatment with iodomethane,

and deprotection produced LX4211 (38) in 45% overall yield over six steps

from 41.

A different approach to glucoside modification was undertaken by medicinal

chemists from Chugai who explored spirocyclization from the aglycone onto



Tài liệu bạn tìm kiếm đã sẵn sàng tải về

4 Phlorizin, SGLTs, and Diabetes

Tải bản đầy đủ ngay(0 tr)

×