Apple Trees & SGLT2

“How long does it take to discovery a drug?” is a questions I am frequently asked by journalists, friends & family, BBQ goers and people generally interested in how medicines are discovered. My answer for the longest time was 10 to 20 years. I would then go on to describe the drug discovery process as a linear process starting with target identification. Before I even got to describe the intriguing ins and outs of assay development and hit identification the eyes of my originally curious conversation partner would glaze over. They would invariably need to urgently check on how the BBQ was coming along. I have recently changed my answer to “It can take over 100 years”, which seems to accelerate their departure. For those who are still with me, here is the story of SGLT2 Inhibitors.

the bark of the apple tree

In 1835, the French chemist Petersen isolated a molecule call phlorizin from the bark of apple trees. The bitter tasting compound was referred to back then as the ‘glycoside from the bark of the apple tree’. At the time, Phlorizin was thought to be a drug candidate for the treatment of fevers, infectious disease, and malaria. In 1903, Percy Stiles demonstrated that phlorizin normalised fasting and post-meal hyperglycaemia in animals: a potential drug to treat diabetes. Phlorizin has been an instrumental molecule in elaborating the biology of its target the sodium/glucose cotransporter (SGLT). However, it is not suitable as a medication for the treatment of diabetes in humans. It wasn’t until 2013 that Canagliflozin became the first approved diabetes medicine to target SGLT. 110 years after Stiles’ initial biological discovery and almost 180 years after the discovery of the molecule itself.

This begins to beg the question, when does a drug discovery project start. Cleary, some drug discovery stories begin with the discovery of the molecule. This leads to the elucidation of disease biology over time as occured with SGLT2. These stories though tend to be older ones. In the modern era, it is usually the initial biological insights that come first followed by a race to discover the molecule. But let’s delve a bit deeper into the SGLT story.

a diabetes medicine – but not for humans

It was clear in the early 1970s that phlorizin acted on transporters responsible for the resorption of glucose in the proximal tubule brush border of the kidney. From the late 1980s to the 1990s the actual glucose transporter proteins responsible were identified as the SGLT proteins. In fact, it was discovered that there were two such proteins. One which has a high affinity for glucose and one which has a low affinity. Although phlorizin normalized insulin levels in animal studies the compound is not suitable as a drug for humans. The reasons for this are three-fold. First, phlorizin is poorly absorbed from the gastrointestinal tract, this is referred to as oral bioavailability. Second, phlorizin inhibits both SGLT1 (primarily found in the gastrointestinal tract) and SGLT2 leading to intolerabilities. Third, phloretin (a metabolite of phlorizin) strongly inhibits the important glucose transporter 1 (GLUT1), which restricts glucose uptake in various tissues.

Oral bioavailability, selectivity (SGLT2 versus SGLT1 in this case) and metabolism represent so-called optimization parameters which medicinal chemists are required to improve through structural variation. Thus, the race to discovery an orally bioavailable, selective SGLT2 inhibitor began in the 1990s and continued until the end of the first decade of the 2000s.

The selectivity ratio for SGLT2 over SGLT1 started at 1:1 for phlorizin. This means that phlorizin inhibits both transporters equally. Inhibition of SGLT1, which mainly occurs in the small intestine, causes diarrhea, dehydration and malabsorption and hence needs to be avoided. Two early SGLT2 inhibitors, Sergliflozin and Remoglifozin, achieved SGLT2:SGLT1 selectivities of 300:1 and 400:1 respectively but these drugs still displayed poor pharmacokinetic stability despite being given as prodrugs. Even the non-prodrug compound AVE2268 with greater than 900-fold selectivity failed to have adequate pharmacokinetics.

organic synthesis to the rescue

Medicinal chemistry does not only encompass the design of molecules but also the synthesis of these molecules so that they can be tested in laboratory assays in order to assess whether a newly synthesized compound is improved over its predecessors. This is the field of organic chemistry. Because SGLTs are proteins that recognize and transport glucose it comes as no surprise that phlorizin and all SGLT2 inhibitors contain a sugar moiety in their structures. Importantly, this sugar group is attached to the rest of the molecule via a labile oxygen atom at carbon 1. This O-glucoside group was the Achilles heel of all early SGLT2 inhibitors. An alternative chemical structure was desperately needed.

The solution was the much more stable C-aryl glycosides for which a synthetic approach was discovered in 1994 and in 2000 the first C-glycosides relating to phlorizin appeared in the literature. This new synthetic methodology changed the field and the O-glycosides were surplanted by C-glycosides which were not plagued by the poor pharmacokinetic properties of their predecessors. In March 2013 Canagliflozin became the first SGLT2 Inhibitor to be approved by the FDA, closely followed by Dapagliflozin on January 2014 and Empagliflozin in August 2014.