Pharmacology of oral hypoglycaemic drugs

This chapter is related to Section U2(iii) from the 2017 CICM Primary Syllabus, which expect the exam candidate to "understand the pharmacology of oral hypoglycemic drugs". This could be an extremely difficult topic to cover if its relevance in the exam was greater, but fortunately only two past paper questions mentioned it, which means one may take a lackadaisical approach and gloss over a lot of the boring detail.

Those questions, for the record, were:

• Question 6 from the second paper of 2020
• Question 6 from the first paper of 2013

Examiner comments emphasised the need for "a strong and logical structure" and "more correct information", which seem like reasonable expectations. It also feels like it would suit the answers better to categorise these drugs by mechanism of action (as describing the mechanism of action was one of the main topics in the exam questions). This is the sub-grouping chosen for the summary below. 

Beyond what is included in this grey box, little else could be expected from the CICM exam candidate, as a detailed understanding of these agents is probably the province of an endocrinologist, and decisions about long-term diabetes management are usually not made at the bedside in the ICU. Following from this, the reader can safely regard the rest of this chapter as surplus to their requirements. In case it is ever required, a selection of these agents can be reviewed and compared against each other in the pharmacopoeia spreadsheet. Published peer-reviewed literature does exist, and the reader's tired eyes are directed to papers like Lorenzati et al (2010) for the pharmacodynamics, and Fowler (2007) or Kimmell & Inzucci (2005) for a class overview.

Chemical classes and chemical relatives

Any discussion of the chemical characteristics of oral hypoglycemics immediately becomes entangled in the massive abundance of available drug choices.  There is a huge amount of chemical diversity in this field, developing over decades of sustained industrial-scale research attention from pharmaceutical companies. Observe, a crude timeline of oral hypoglycemic agents, arranged by class and year of market availability:

Oral Hypoglycemic agents

Biguanides and diguanides
Galega officinalis	1652
Guanidine	1918
Decamethylene diguanide	1920
Buformin	1957
Phenformin	1957
Metformin	1957
First generation sulfonylureas
Tolbutamide	1956
Carbutamide	1956
Chlorpropamide	1957
Glycyclamide	1957
Tolazamide	1959
Acetohexamide	1964
Second generaltion sulfonylureas
Gliclazide	1972
Glibenclamide (glyburide)	1984
Glipizide	1984
Third generation sulfonylureas
Glimepiride	1995
α-Glucosidase inhibitors
Voglibose	1994
Acarbose	1995
Miglitol	1996
Meglitinides
Repaglinide	1997
Nateglinide	2000
Thiazolidinediones
Pioglitazone	1999
Rosiglitazone	1999
Dipeptidyl peptidase (DPP-4) inhibitors
Sitagliptin	2006
Vildagliptin	2007
Saxagliptin	2009
Linagliptin	2011
Amylin agonists
Pramlintide	2005
GLP-1 receptor agonists
Exenatide	2005
Liraglutide	2010
Dulaglutide	2014
Bile acid sequestrants
Colesevelam	2008
SGLT-2 inhibitors
Dapagliflozin	2012
Empagliflozin	2014

Probably the only logical way of covering this ground is to give some foundation and context to the chemistry of these substances, and then to discuss the pharmacokinetic and pharmacodynamic properties of some key representatives of each class, instead of trying to explore each member or each class individually. Ambassadors from each category were picked not by some objective scientific method, but by the modest mental commodities of the author, who chose those agents that he personally encounters in the field.

Biguanides 

This group was among the first drugs developed specifically to treat hyperglycemia, and their history has been one of gradual domestication and gentrification, moving from rather scary substances to progressively more tolerable ones.  The parent chemical guanidine is as toxic as a banned subreddit, and was found naturally in Galega officinalis (goat's rue or French lilac), a "medicinal" herb which was used by medieval peasants to poison each other randomly in the 17th century (as the conditions it was being used for, such as worms and plague, had nothing to do with its oral hypoglycemic properties). It certainly kills livestock very well, begging the question "why did those people decide to make tea from it", but that answer is lost in the mists of time. For humans raw pure guanidine is a directly corrosive alkali which can cause seizures and neuromuscular weakness by acting as a cholinergic toxin. These risks did nothing to deter people from listing herbal Galega extract in various pharmacopoeias all the way until the 1920s and 30s, when some slightly less toxic daughter molecules were finally developed.

These molecular structures are included here only because they were already available in Bailey & Day (2004), and because they show off the double guanidine in the molecules of the "proper" biguanides. For the majority of them, side effects were still the main limitation on routine use, and many disappeared from the pharmacy shelves with the advent of modern 'formins in the 1950s. Of those, buformin and phenformin were far less popular than metformin because of their tendency to produce lactic acidosis, and now neither of them are available for commercial use (for example phenformin was booted out in 1977, the US authorities using stong language like "imminent hazard").  Metformin on the other hand has remained the mainstay of diabetes therapy and is apparently the fourth most commonly prescribed drug in the US. Metformin is therefore the sole representative of the biguanides in the pharmacology discussion that follows.

Sulfonylureas

This group was already available in the earliest days of the twentieth century, but we first came to know them as antibiotics, in the form of early sulfonamides. Already in the 1930s there was talk of their hypoglycemic properties, but not until 1942 did anybody document this effect in humans, and the mechanism of their action was not revealed until 1946. In the following ten years, the first sulfonylureas would be developed, to finally reach the market in the middle of the 1950s. 

The first generation of these drugs, in use from 1956 to 1966, will have names that are completely unrecognisable for the modern reader, and will never appear in anybody's exam papers, as most people writing the questions weren't even alive when these molecules slipped quietly out of the world. For a thorough review of their history the reader is invited to look up the 1980 work by Selizer or Srivastava et al (2019), with the usual caveat that it would achieve nothing useful. In short, these are all drugs with a central S-arylsulfonylurea structure from which some substituents hang, the exact nature of which tends to govern their pharmacodynamic effects and some of their pharmacokinetics:

The second generation of sulfonylureas were developed in ther 1970s and became available in the early 1980s in response to some concerns about the cardiovascular safety of the first-generation agents. The University Group Diabetes Program was responsible for generating these concerns. They were an early clinical trials group who ran studies starting as early as 1961, focusing on the safety and efficacy of sulfonylureas, and their controversial finding (that tolbutamide was probably doing more harm than good) resulted in so much drama and scandal that an entire paper has been dedicated to unravelling it (Blackburn & Jacobs, 2016). The pragmatic reader will ignore this historical detour and be simply satisfied with knowing that this group of  drugs remains in use today and includes gliclazide glibenclamide and glipizide.

Glimepiride is the only member of the third generation of sulfonylureas. Reading reviews of this drug (eg. Gale, 1999), one wonders whether there was really any point in creating a whole new generation just for this.  The main difference between glimepiride and the previous generation of sulfonylureas is the reduced incidence of adverse effects, specifically of hypoglycemia and of the cardiovascular risk (as glimepiride seems to have a lower affinity for myocardial ATP-sensitive potassium channels). Another difference was the once daily administration. At a molecular level, it does not cut a sufficiently distinct figure to make you think that it deserves its own podium (whereas there is a clear chemical distinction between the structures of the first and second generation), and in fact Lang & Light (2010) just grouped it with the second generation agents in this excellent diagram:

Thiazolidinediones

Understandably, most people reword this mellifluous class name as "glitazones". This category of oral hypoglycemic was discovered by accident when a Japanese pharma company were trying to create better fibrates to treat hyperlipidemia in diabetics (specifically, in a chronically voracious obesity model of the rat).  These guys were modifying clofibrate, a toxic discontinued antilipid agent, and the specifics of their biochemical manipulations are lovingly detailed by Yasmin & Jayaprakash (2017) for anybody who is into that sort of thing. In short the intended variable (triglycerides) were only modestly improved by these test substances, but alongside this effect the authors noted a reduction in the degree of insulin resistance. The first of these (ciglitazone) has subsequently been removed from the market because of hepatotoxicity; another (rosiglitazone) is no longer on sale in the US because of concerns regarding increased cardiovascular risk. 

α-Glucosidase inhibitors

Why interfere with the complicated inner clockwork of the carbohydrate metabolic apparatus, when you can just prevent carbohydrate absorption? That is the main objective of inhibiting α-glucosidase, a brush border enzyme of the small intestine which is mainly involved in the hydrolysis of terminal starch residues which releases glucose. To disable this enzyme means to slow the metabolism and absorption of carbohydrates, which achieves the laudable goal of reducing postprandial hyperglycemia. Chemically, this is achieved by presenting the enzyme with some kind of decoy to chew on, something that resembles the original substrate but which the enzyme can't do anything interesting with. Thus, voglibose and acarbose are both weird aminosugars that vaguely resemble D-glucose:  acarbose is a maltose molecule with an acarviosin moiety attached to it, and voglibose is an N-substituted derivative of valiolamine. You don't have any enzymes interested in absorbing or metabolising these alien chemicals, and they are mainly broken down by gut bacteria. Miglitol is a more recent addition, a sugar alcohol  (structurally very similar to glucose but derived from deoxynojirimycin) which can be absorbed systemically (though it simply passes through politely without touching anything and exits via the kidneys).

Meglitinides

Otherwise abbreviated as "glinides", these are benzamides which share their mechanism of action with sulfonylureas. They are structurally dissimilar, even though they might have similar names - for example repaglinide is a piperidine derivative, and is recognisably different from nateglinide, which is made from phenylalanine. Their structure has some vague similarities to the structure of the latter generations of sulfonylureas (for example their molecules curl up into the same sort of U-shaped conformation), which is thought to be the explanation for their sulfonylurea-like behaviour (i.e. the bind the same target sites but with weaker affinity and for shorter periods).

Dipeptidyl peptidase 4 (DPP)-4 inhibitors

Unsurprisingly, many people tend to refer to these substances as "gliptins". Baetta & Corsini (2011) describe their activity in some detail, and the reader is referred there for molecular minutae. In passing, it is interesting to note that they are all structurally rather different, and all do weirdly different things to the DPP-4 enzyme - for example, sitagliptin pretends to be a substrate, linagliptin interacts with the catalytic site in a way that prevents it from accepting proper substrates, and vildagliptin and saxagliptin engage in some kind of strange reversible covalent bond with the enzyme, deactivating it for a much longer period than what one might expect from their short plasma half lives. 

Glucagon-Like Peptide-1 (GLP-1) receptor agonists

Glucagon-Like Peptide-1 (GLP-1) is also occasionally known as "incretin". All of the available agents that mimic the work of this endogenous small bowel hormone are small peptides that resemble human GLP-1 to a lesser or greater extent. Exenatide, the first of these, was developed from exendin-4, which is a minor constituent of the venom of Heloderma suspectum, the Gila monster. The Gila monster of course has not been evolving oral hypoglycemic agents in the expectation of one day becoming pharmacologically useful to diabetics, and this peptide's resemblance to GLP-1 is purely accidental (it only has a 57% homology with the human peptide), so in general hypoglycemia is not one of the features of helodermid envenomation.

From this weird herpetological origins, we now have a remarkable array of highly specific GLP-1 agonists, listed and explored by Hinnen (2017). Of the other agents, the modern Western audience will probably be most familiar with liraglutide and dulaglutide.

The reader outraged at the inclusion of these drugs in the chapter supposed to be about oral hypoglycemic agents has every right to be disappointed. To be sure, these are all rather large molecules, and none of them could never be administered enterally. For example,  dulaglutide is attached to an FC fragment of IgG, and weighs about 60 kDa, whereas liraglutide is decorated with a fatty acid molecule that causes it to self-associate into large oligomers. However, they do exist, and ICU trainees are probably expected to be at least vaguely aware of them. Though the author may have used a more inclusive term for the chapter title ("non-insulin agents used in the control of diabetes" for example),  it seemed better to follow the wording of the CICM syllabus, as that is a more natural search string for an exam candidate.

SGLT-2 inhibitors

One could just as easily classify these drugs alongside osmotic diuretics, as their main site of action is the kidney, and the main activity is to prevent the reabsorption of a solute. They are all vaguely based on phlorizin, a naturally occurring nonselective SGLT-1 and SGLT-2 inhibitor which is too unstable to be available orally, and tends to block intestinal SGLT-1 transporters, with predictably humorous effects on the gut microbiome (Kalra, 2014). The structure and function relationship of these effects is not completely clear but the 'flozin molecules all share a glucose moiety as a feature, suggesting some kind of substrate mimicry is probably playing a role.  The main modifications which spawned dapagliflozin and empaglflozin from phlorizin have been directed at making them more selective for SGLT-2 and increasing their chemical stability to help their oral bioavailability.

Exotic hypoglycemic agents

The road to safe effective management of diabetes is littered with abandoned wreckages of failed and forgotten pharmacological agents, the wind rustling through their fading promotional literature. As a condition that affects a large proportion of the affluent West, the market is large, rich, white and old, and so one might expect a lot of pharma company research attention to be directed to finding the hot new thing, which by necessity means multiple safety-related retractions, false starts, failures, and needless duplications.  Into this category fall amylin agonists like pramlintide and bile acid sequestrants like colesevelam, names which the reader will immediately forget with no adverse professional effects. Apart from these, a lot of drugs can cause hypoglycemia as a side effect (eg. bromocryptine and THAM), but that does not mean that we should be using them routinely to lower the blood glucose. They are mentioned here as asides, for no reason other than the worship of completeness. 

Route of administration

Surely they are called oral hypoglycemics for a reason. For the selection of drugs discussed here, oral administration is the only accepted method. Only the GLP-1 receptor agonists disagree with this trend and require subcutaneous administration.

Absorption and oral bioavailability of oral hypoglycemics

Overall, the oral bioavailability and enteric absorption of all of these drugs is excellent:

	Enteric absorption	Bioavailability
Metformin	Completely absorbed	40-60%
Gliclazide	Completely absorbed	97%
Glibenclamide	Variably and poorly absorbed	If absorbed completely, 95% oral bioavailability
Glimepiride	Completely absorbed	97%
Acarbose	Minimally absorbed (less than 2%)	Less than 1%
Repaglinide	Completely absorbed	62%
Rosiglitazone	Completely absorbed	99%
Sitagliptin	Completely absorbed	87%
Dulaglutide	Not absorbed	100% (s/c injection)
Empagliflozin	Completely absorbed	78%

Glibenclamide and acarbose are the only drugs of this group which are not especially well absorbed via the oral route. Acarbose obviously is non-absorbable as you have no mechanisms for its uptake, but it does all of its work in the gut anyway, so it can safely remain there. Glibenclamide is the black sheep of the sulfonylurea group because its oral absorption is frustrated by extremely poor water solubility, and many articles are dedicated to creating and testing various excipients to make it dissolve better. The bioavailability of the drug therefore depends on the proprietary formulation of each preparation, and most studies tend to compare AUCs of concentration/time curves instead of giving you a specific percentage of bioavailability. If it were to somehow become absorbed completely, it would have excellent oral bioavailability because the first pass metabolism only takes out about 5% of the administered dose. 

Speaking of solubility: as one can clearly see, only a handful of these drugs are well-dissolved in water, namely metformin sitaglipin and acarbose. For the rest, water solubility is described as "sparing" or "minimal".

	pKa	Solubility
Metformin	12.4	Good water solubility
Gliclazide	5.8	Basically insoluble in water; but at least slightly soluble in lipid and ethanol
Glibenclamide	5.5	Terrible solubility in either water or lipid
Glimepiride	8.07	Basically insoluble in water; but at least slightly soluble in lipid and ethanol
Acarbose	5.1	Excellent solubility in water
Repaglinide	4.8	Poorly water-soluble, but good lipid solubility
Rosiglitazone	6.8	Poorly water-soluble, but good lipid solubility
Sitagliptin	8.7	Good water solubility
Dulaglutide	n/a	Made soluble with the addition of excipients
Empagliflozin	12.57	Very slightly water-soluble

These chemical properties are listed here not for some kind of deeper educational purpose, but because they can somewhat explain the distribution kinetics and protein binding in the following section.

Distribution and protein binding of oral hypoglycemics

The vast majority of oral hypoglycemic agents have a very small volume of distribution, mainly because they are almost completely protein-bound. This refers to all the agents with extremely poor water solubility. The exceptions are agents which are at least modestly water-soluble, like metformin and sitagliptin - their volume of distribution is larger, and their protein binding is lower. 

	VOD	Protein binding
Metformin	1-4 L/kg	Minimally protein bound
Gliclazide	0.2-0.4 L/kg	85-97% protein bound
Glibenclamide	0.2-0.45 L/kg	99.9% protein bound
Glimepiride	0.11 L/kg	99.5% protein bound
Acarbose	n/a	not really absorbed, so...
Repaglinide	0.4L/kg	97% protein bound
Rosiglitazone	0.24L/kg	99.8% protein bound
Sitagliptin	3-4 L/kg	38% protein bound
Dulaglutide	0.05 L/kg	is an actual protein, so...
Empagliflozin	1L/kg	86% protein-bound

Metabolism and elimination of oral hypoglycemics

Unsurprisingly, of this selection of heterogeneous chemicals, those that have good water solubility and minimal protein binding are also those that undergo renal elimination and little hepatic metabolism, like metformin and sitagliptin. Among the half-lives, the standouts are acarbose (which is basically candy) and repaglinide which is broken down over 60 minutes.

	Metabolism	Elimination	Half life
Metformin	Minimal metabolism	90% of the dose is cleared renally	6 hrs
Gliclazide	Extensive hepatic metabolism	Only 4% of the drug is renally eliminated	11 hrs
Glibenclamide	Extensive hepatic metabolism; two major metabolites are also active	Active metabolites are renally eliminated	15 hrs
Glimepiride	Extensive hepatic metabolism; one intermediate metabolite is about one third as active as the parent drug	Active and inactive metabolites are renally eliminated	24 hrs
Acarbose	Metabolised by gut bacteria	Breakdown products eliminated in the faeces	2 hrs
Repaglinide	Extensive hepatic metabolism	90% of the inactive metabolites are eliminated in the faeces	60 minutes
Rosiglitazone	Extensive hepatic metabolism	Inactive metabolites are eliminated mainly in the urine (65%)	7 hrs
Sitagliptin	Hepatic metabolism plays a minor role in the elimination	79% excreted renally as unchanged drug	8-14 hrs
Dulaglutide	Degraded into amino acids and peptides by the reticuloendothelial system	Most of the amino acids and peptide breakdown products are reclaimed	1 week
Empagliflozin	Mainly hepatic metabolism into inactive metabolites	10-20% of the dose is eliminated renally as unchanged drug	6-13 hours

For the rest, hepatic metabolism plays a major role, which makes them slightly safer in renal failure. A note of caution must be left on the sulfonylureas, which undergo mainly hepatic metabolism, but among which many have active metabolites that are dependent on renal elimination. 

Mechanisms of action of oral hypoglycemics

The pharmacodynamics of oral hypoglycaemic agents is a tour of some dangerous metabolic back alleys. There must be some way to explain these mechanisms without the reader becoming lost and mugged by gangs of biochemists. What follows is a crude reductionist attempt to simplify and stereotype these drug effects into something that could be quickly and easily digested by a revising exam candidate. Whatever detail is lost in that process can be recovered by any reader who can click and open a link, as references are offered pointing to detailed review articles (of which there is a glorious abundance).

Biguanides decrease blood glucose mainly by decreasing hepatic glucose production through their actions on AMP-activated protein kinase (AMPK), though there are probably also multiple other mechanisms involved (Rena et al, 2017). AMPK is a ubiquitous fuel-sensing enzyme present in basically all mammalian cells, and its main role is to coordinate a switch from energy consumption to energy generation, for example when exercising skeletal muscle needs to take more glucose from the bloodstream. Hepatic AMPK also does something like this, and to activate this enzyme has a largely catabolic effect, stimulating fatty acid oxidation and suppressing protein synthesis and glucose release - mainly by AMPK phosphorylating all kinds of key enzymes in those pathways. 

Metformin does not do anything to AMPK directly. Instead, that enzyme becomes activated as the reaction to an act of mitochondrial terrorism. Metformin is a highly positively charged molecule, and becomes concentrated inside mitochondria as a result, with the intramitochondrial concentrations several orders of magnitude higher than the extracellular fluid. Once inside, it sabotages ATP synthesis by disabling Complex I of the respiratory chain. The result is a decreased ATP:ADP ratio, which is a potent stimulus for AMPK activation (as it would normally be viewed as a signal that the cell is starving and requires immediate metabolic substrate support). AMPK then dutifully activates the catabolic machinery of the hepatocytes, and abolishes all forms of charitable export behaviours, among them the production and systemic delivery of glucose by glycogenolysis and gluconeogenesis. The onset of this effect is said to be about three hours following administration.

The reader is reminded that metformin has whole PhDs of different mechanisms, but the only one the ICU trainee really needs to know about is this mitochondrial toxin aspect, mainly because it explains a common toxicological presentation. By disabling the mitochondrial metabolism of oxygen, metformin produces lactic acidosis, which can be rather impressive in magnitude, and which comes up quite often in exam papers as a differential. 

Sulfonylureas act by stimulating insulin secretion, an activity which relies on the existence of residual pancreatic β-cells (because otherwise where would it come from). A secondary effect is the decrease of insulin clearance by the liver, which seems to occur over some weeks with sustained treatment (Sola et al, 2015). Sulfonylureas achieve these effects by binding to a specific receptor on pancreatic β-cells which has come to be known as sulfonylureas receptor (SUR1), as we would not have found it otherwise. This thing is a transmembrane protein which - together with several others - forms the ATP-sensitive potassium channels that mediate insulin release. The binding of sulfonylureas to this complex tends to have the same effect as ATP and raised blood glucose, i.e. to block the outward flow of potassium, which results in the depolarisation of the β-cell and the release of insulin. This effect is fairly rapid in onset, i.e. as soon as the drug is systemically absorbed. 

Meglitinides, like sulfonylureas, are "insulinotropic" or "secretagogue" molecules that stimulate the release of insulin from pancreatic β-cells. They also bind to the SUR1 receptor, albeit with less affinity, and produce the same β-cells-depolarising effect. The main difference from sulfonylureas is the duration of effect, which is much shorter, and therefore much less likely to produce hypoglycaemia (Guardado-Mendoza et al, 2013).

α-Glucosidase inhibitors act as pseudocarbohydrates, i.e. they perform as impostor substrates for intestinal blush border enzymes, thereby inhibiting the digestion of real carbohydrates (Derosa & Maffioli, 2012). They are usually taken along with the first bites of a main meal. There are actually several α-glucosidase enzymes at the brush border, such as sucrase, maltase, dextranase and glucoamylase, and acarbose interferes with all of them. Their normal role is to digest more complex carbohydrates until they turn into the sort of monosaccharides that can be easily absorbed through the intestinal mucosa. Theoretically, this means to block them all would result in the complete failure of all carbohydrate digestion, and the delivery of undigested carbohydrate directly to the colon, which is in fact what happened when Puls et al (1980) overdosed some rats with acarbose. 

Thiazoledinediones operate in the shadowy world of peroxisome-proliferator–activated receptors (PPARs), which are another group of nuclear receptors that regulate gene expression much like the receptors for corticosteroids. Under normal circumstances, the natural ligands for these receptors are all sorts of fatty acids and bile acids 

• PPARα receptors activate fatty acid oxidation and regulate lipoprotein synthesis. They are the targets of fibrates such as gemfibrosil.
• PPARβ receptors are expressed in the adipose tissue CNS and skin, appear to be involved in myelination, and we currently have no drugs to target them
• PPARγ receptors are mainly found in adipose tissue and pancreas, and are the target of thiazoledinediones

 Cheatham (2010) and Yki-Järvinen (2004) explain the function much better, but if "better" is not as good as "shorter" for the reader, the function of thiazoledinediones can be simplified as "increased insulin sensitivity". The exact mechanisms of how they do this are still being determined, but it appears that the presence of activated PPARγ receptors enhances the transcription of all the proteins involved in the machinery of glucose uptake and processing, especially in adipose tissue. In this fashion, the response to any insulin binding to that cell is enhanced. By this mechanism it appears the glitazones redistribute the deposition of fat into the fatty tissue (and away from the liver), increase the sensitivity of the liver and fatty tissue to insulin, and increase insulin secretory responses from the pancreas. Because of their indirect gene-transcription-modifying function these drugs also have a host of nonglycemic effects, some antiinflammatory and some antiatherogenic. 

Dipeptidyl peptidase (DPP-4) inhibitors are well described by Thornberry & Gallwitz (2009) or Baetta & Corsini (2011). Their target, DPP-4, is a member of a large family of transmembrane proteins, and is also known as CD-26 (when it is observed on the surface of lymphocytes). It is found in many tissues, particularly the vascular endothelium, and some part of it seems to be able to break off and sail the bloodstream as a soluble enzyme, retaining full activity. That activity is to basically break down hormones - theoretically anything with a proline or an alanine in the penultimate position on the N-terminal will get cleaved, but practically in humans the only known substrates for this thing are glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide 1(GLP-1). 

GIP and GLP-1 are released in response to food ingestion, and their most interesting effect is to increase the secretion of insulin. GLP-1 binds to its G-protein-coupled receptors on pancreatic β-cells, where it increases intracellular cAMP, and therefore the availability of intracellular calcium that drives insulin exocytosis. Over longer timeframes an increased exposure to GLP-1 also leads to the increased synthesis of insulin and β-cell hypertrophy. GLP-1 also suppressess the secretion of glucagon, which possibly has an equally great importance in controlling postprandial sugar levels. Additionally, GLP-1 receptors are expressed in the CNS where it appears to affect satiety, with all sorts of positive flow-on effects on behaviour modification and weight loss. In short, if you are an obese diabetic, you want more GLP-1 in you, and that is what DPP-4 inhibitors produce. 

GLP-1 receptor agonists bypass the need to reactive DPP-4 and go straight to the source. In fact GLP-1 agonist drugs out-GLP the native enzyme by having higher affinity for its receptors and a more stable pharmacokinetic profile, remaining active for longer (Cornell, 2020). Their pancreatic and extrapancreatic effects are otherwise the same as those of GLP-1.

If the outcome is the same, what would help you choose between DPP-4 inhibitors and GLP-1 agonsist, apart from the convenient oral availability of the former? To discriminate between the two classes, Brunton (2014) reviewed the available clinical trials and concluded that GLP-1 agonists were actually more effective at achieving all kinds of meaningful targets (eg. HbA1c reduction).

SGLT-2 inhibitors block the Type 2 sodium-glucose cotransporter that you tend to find in the kidney. SGLT-1, on the other hand, is mainly intestinal, and to block this would have an acarbose-like malabsorption effect (Dardi et al, 2016).  That was the mechanism of action of phlorizin, the original precursor of this drug class. The blockade of SGLT-1 would unfortunately do nothing about the release of glucose from the liver. Instead, the blockade of SGLT-2 prevents the reabsorption of already circulating glucose. SGLT-2 is responsible for the reabsorption of about 90% of filtered glucose, so theoretically their blockade could result in some impressive glycosuria (140g/day!) but realistically in clinical use normal doses of these drugs only tend to block about 30-50%, resulting in the loss of 50-90g of the total daily filtered glucose.

Adverse effects of oral hypoglycemics

Probably one of the main reasons that we have such a vast plethora of different antidiabetic medications (apart from the inherent attractiveness of marketing something for elderly Westerners to take every day for many years) is that most of them have some fairly serious side-effects, some of which are merely embarrassing, whereas others may be life-threatening. Probably the single best paper on these is a review by Lorenzati et al (2010), which also happens to cover the pharmacodynamics of each class in just enough detail for the tired exam candidate.

Biguanides can, by the direct extension of their mechanism of action, cause severe lactic acidosis. This is not very common (at least not as common as it was with phenformin) but it still happens, particularly where metformin accumulates due to renal failure. Its other side effects consist of gastro-abdominal stuff, for example a metallic taste in your mouth, diarrhoea, abdominal discomfort, anorexia, etc.

Sulfonylureas have a distinct tendency to cause undesirable hypoglycaemia, which is a direct extension of their therapeutic effect. They stimulate insulin secretion no matter the glucose concentration, and therefore remove the normal regulatory safeguards that prevent insulin release during periods of normal and low blood glucose. The result is essentially an insulin overdose.  Other side effects can include hypokalemia, weight gain,  skin eruptions and photosensitivity.

α-glucosidase inhibitors are essentially osmotic laxatives, as their mechanism of action is centered on creating malabsorption. They leave a lot of unfinished carbohydrate scraps in the lumen of the bowel, and as you might imagine, there are plenty of microbial scavengers down there who are ready to pounce on these molecules, gladly metabolising them into clouds of flatus and torrents of diarrhoea. These are potentially dinner-party-ending consequences, as these drugs are taken immediately at the commencement of a meal.  Overall, such features have not charmed the population of users, with only 30% or so continuing to take these drugs in the long term.

Meglitinides, like sulfonylureas, directly stimulate the release of insulin, but because they are very short-acting and taken immediately before a meal, the possibility of severe hypoglycaemia is somewhat diminished. Other (weird, unexpected) adverse effects in studied populations were headaches, and an increased risk of upper respiratory tract infections - for some reason specifically sinusitis.

Thiazolidinediones are generally not likely to cause life-threatening hypoglycaemia, even though they increase insulin sensitivity. Unfortunately, by making adipose tissue more interested in glucose, they make it grow in size, and noticeable weight gain results. They also effect PPAR receptors in bone, and are associated with an increase in the risk of fracture. PPAR activation in renal tubular cells, on the other hand, leads to fluid retention, and oedema results. Additionally, individual agents have distinct and unpleasant risk profiles: rosiglitazone is associated with myocardial ischaemia and pioglitazone slightly increases the risk of bladder cancer. 

Dipeptidyl peptidase (DPP-4) inhibitors have a fairly benign side effect profile. They are generally said to be "weight-neutral", i.e. the population of patients starting sitagliptin therapy remain as obese as they were before treatment. The incidence of serious hypoglycaemia with this class is also fairly low. The most commonly reported adverse effects were sinus infections and headache. 

GLP-1 receptor agonists are also rather free from serious life-threatening side effects, with the exception of some idiosyncratic pancreatitis episodes reported with exenatide. Nausea and anorexia are also reported. Weight loss does tend to occur with these drugs, resulting in as much as 5% body weight loss, but for the majority of patients this is a desirable feature. 

SGLT-2 inhibitors are basically osmotic diuretics, and have a tendency to produce polyuria and volume depletion. At the same time the urine becomes sweeter, and the incidence of urinary tract infection increases. The most interesting side effect is probably euglycaemic ketoacidosis, where a sustained low BSL produces a downregulation of insulin release and a concomitant increase in ketogenic hormones like glucagon adrenaline and cortisol.