Pancreatic Islet and Gut Hormones

Gut and pancreatic cells

Diabetes genetics and pharmacogenomics/ personalised treatments

Current treatments

Future treatments

Gut endocrine cells are responsible for >50% of insulin secretion

The particularly relevant ones are GLP-1 and GLP-1/PYY secreting L cells and GIP secreting K cells (See comment to right for more info)

Only 1% of crypt stem cells differentiate into enteroendocrine cells along the GI tract axis

At the top of the axis (near stomach) they secrete CCK, GIP and secretin

At the bottom they secrete GLP-1, PYY and INSL5

Most are open-type and have a lifespan of 5 days (preventing them from getting damaged directly, but stem cells can be), these secrete GLP-1, PYY, GIP and CCK

Some are closed-type, which is much more common in the stomach, these secrete somatostatin and ghrelin

Gut hormones amplify insulin secretion via the incretin effect

Discovered by Nauck (Diabetologica, 1986) using glycolycemic doses and shown to be impaired in D2M patients

Secretion of GLP-1 and GIP by gut cells is driving this, faulty GIP secretion in D2M patients is thought to be to blame for their lack of glucose response

CCK, GLP-1, PYY and potentially (based on new literature) GIP all reduce appetite and ghrelin increases it

Ghrelin increases somatostatin release in vitro (unclear physiological relevance), this decreases insulin and glucagon release. GLP-1 also increases somatostatin release, but only enough to decrease glucagon release. PYY may have an effect on pancreatic islets but there are very few receptors so this is somewhat doubtful

GLP-1 is produced by alternative processing of proglucagon

In pancreatic alpha cells only an inert form is produced. In L cells proglucagon is processed into GRPP, oxyntomodulin, (Gilcentin), GLP-1 (cutting at position 7 by prohormone convertase 1/3) and GLP-2

Production of GIP and GLP-1 is delayed from eating a meal

These were originally investigated using cell lines and dyes, before moving into transgenic mice models

More recently mouse organoids (Goldspink, 2018) and human organoids have come into use

Transcriptomic studies by Habib, Reimann, Gribble et al (2012) found that K and L cells ~produce the same chemicals under the right conditions and that they are effectively the same, it is their location that determines what they are secreting in vivo

Signalling in EECs involves Gq and Gs coupled receptors, nutrient transporters and ion channels

Each nutrient has a separate transporter/receptor

Experiments have shown that sugars must be absorbed by a cell for the incretin effect to be initiated (Na+ is also required) by using phloridzin to inhibit SGLT1 (and no incretins being secreted as a result)

Theories

1 - 'sweet taste' receptors localised on the basolateral side
There is no evidence for this though. Sglt1 KO mouse cells do not have a response to glucose even when it is added to the basolateral side (Nauck, Regul Peptides, 2004), additionally no evidence of them on primary L cells (Reimann, Cell Metab, 2008 - countered by Jang, PNAS, 2007)

2 - Sugar is sensed inside the cell
Blocking GLUT does not reduce GLP1 response, this is the main transporter of glucose so inhibiting it would reduce the amount of glucose that can be in the cell

3 - Sensing is directly linked to transport
Sodium entry alongside glucose causes a small membrane depolaristaion which then triggers CaVs to open and trigger secretion of incretin granules. It was shown that SGLT1 was responsible for glucose-stimulated incretin release by inhibiting it and showing that glutamine could still cause calcium influx (Parker, Diabetologica, 2012)

It has been theorised that K ATP channels may modulate responses to small SGLT1 current and aid propagation of the AP

It is not known why but SGLT1 KO impairs early GLP-1 secretion but increases late secretion. Was postulated by Powell to be the result of microbial glucose fermentation to short chain fatty acids which can then stimulate GLP-1 release through detection by FFA1

FFA1 signalling recruits TRPC3 currents, an unusual pathway for a Gq coupled receptor (Goldspink, 2018)

long chain FAs are transported through the endothelium by chylomicrons and then act on basolateral FFA1 to stimulate incretin secretion (Psichas, DIabetologica, 2017)
Such secretion also involves GPR119 (GsPCR - responds to OEA)

Peptones can also rigger incretin secretion, via the CASR. This is dependent on CaV action, suggesting a membrane depolarisation event occurs - shown by inhibiting CASR and CaV (Pais, peptides, 2016)

Microbes may play a part in gut cell response with indoline, SCFAs, LPS and secondary bile acid production

Glucose, amino acids and fatty acids stimulate this process

Pancreatic cells

Tests

Beta cell function

Original conflicting theories

IVGTT - pancreatic response (clamp is ~gold standard)
OGTT - combined gut and pancreatic response
Intraperitoneal tolerance test (animals) - bypasses gut so ~IVGTT

Insulin sensitivity

HOMA - prediction from fasting plasma glucose and insulin
Hyperinsulinaemic euglycaemic clamp
Insulin tolerance test (in animals)

T2DM results from failure of beta cells or resistance to glucose
Is known now to be due to beta cell failure

Does glucose stimulate insulin secretion by being sensed or metabolised
Now known that it is the result of metabolism - non-metabolisable sugars are ineffective and secretion is prevented by metabolism inhibitors (the reasons this was resisted is that the cell is sensing changes in [ATP], which is ordinarily kept constant

Consensus model of glucose sensing

GLUT2 draws glucose in, which is metabolised to produce ATP. This blocks K+ATP channels, causing a small depolarising current, which stimulates CaV channels and subsequent insulin release

K+ATP channel modulators have different affinities for SUR1&2, most used clinically seem to have some action on SUR2 but avoid CV side effects

KO of Kir6.2 abolishes insulin secretion, SUR1 KO effectively does the same, although there is a slight response (Miki, 1998, PNAS & Seghers, 2000, J Biol Chem)

Insluin secretion is potentiated by glucose, resulting in a longer duration of secretion for the same membrane depolarisation
Appears to be via a two phase process and glucose (or a metabolite) determines the length of the second phase - it requires energy though as granules must be continuously replenished

Regulation

GPCRs act through Gq proteins to activate IP3Rs, PKC and (somehow)TRPC3 and through Gs proteins to increase [cAMP]

Insulin release is potentiated by GqCPR pathways and GsPCR pathways, the latter pathway is stimulated by incretins (Rosselin, 1998, Endocrine)

GLP-1 triggered insulin secretion is dependent on glucose because without it there is no signal to potentiate. Hence GLP-1 therapies carry low risks of hypoglycaemia

Sulphonylureas can trigger insulin secretion at any glucose secretion, potentially causing hypoglycaemia (particularly in older thinner people)

Sulphonylureas are used for treatment of patients with open K+ATP channels and show a good response to oral but not i.v. glucose (Pearson, 2006, New Eng J Med) as it allows beta cells to respond to gut hormones

Action in the body

Unclear how physiologically relevant the action of incretins on glucagon is

High doses of GLP-1 but not GIP stimulate insulin release and suppress glucagon levels in D2M patients (Nauck, 1993, J Clin Invest)

GLP-1

Beneficial effects have been seen in patients with D2M (Zander, 2002, Lancet) by looking at C-peptide, and glucose concentrations in the blood

However GLP-1 is a poor therapy as it is rapidly degraded by DPP4 in vivo

DPP4 inhibitors have been developed to prolong GLP-1 half-life (Vidagliptin, alogliptin etc) but these can have off-target effects as DPP4 has multiple targets

GLP-1 analogues and mimetics have also been developed to prevent the issue of DPP4 degradation (occurs between 8th and 9th residue)

Liraglutide binds to albumin, protecting it from DPP4. This is also now used in the treatment of obesity

Semaglutide is also available orally (one of few as these are peptides)

Dulaglutide has a much longer half life and means that treatment is only needed once per week

These therapies have shown reasonable efficacy as measured by absolute reduction in HbA1c and body weight (Drucker & Nauck, 2006, Lancet)

Exenatide has shown weight loss effects on D2M patients as a long acting molecule and as a short acting molecule (Drucker, 2008, Lancet)

Is also thought that this may be the result of GLP-1 stimulation of somatostatin

How GLP1 suppresses food intake is currently unclear. GLP1Rs are expressed in afferent vagus (could detect in the intestines), brainstem, hypothalamus (but can GLP-1 penetrate these nuclei?) and mesolimbic reward areas (again, can GLP-1 penetrate this deep?)

Mouse model studies seem to suggest that the BBB is leaky to GLP-1, which suggests that GLP-1 could be acting in the brainstem or elsewhere (Cork, 2015, Mol Metab)

Seems to slow gastric emptying, but this is thought to be more important for post-prandial glucose control than beta cell stimulation (Meier, 2012, Nat Rev, Endocrin)

There has been some evidence given that incretins can cause damage to the pancrease (pancreatitis, pancreatic cancer and thyroid C-cell cancer) (Butler, 2013, Diabetes) but the data was seen as poor and other evidence refuted it (Egan, 2014, NEJM)

Little advancement on the front of personalised treatments but good progress in terms of pharmacogenomics

Type 1 diabetes is a serious medical emergency, it only kicks in after 6 months (once the immune system is up and running)

HLA, INS, PTPN22 and IL2RA are the primary genetic biomarkers of D1M (Pociot, 2010, Diabetes)

Often polygenic in nature

Neonatal and maturity onset diabetes of the young are monogenic diseases

Type 2 diabetes (D2M) is highly polygenic with large environemental factors influencing severity and speed of onset

MODY (Maturity Onset Diabetes of the Young) is caused by one of 7 genes but MODY3 accounts for ~65% of cases. Many of these are TF mutations that regulate differentiation to beta cells and beta cell function (Servitja, 2004, Diabetologica)

Neonatal diabetes is diagnosed within six months of birth, is still technically Type II
Insulin independent, causes by K+ATP channel subunit mutations (~50%), glucokinase null mutations or mutations in the insulin gene (Stoey, 2010, Endocr Metab Disord)

Mutation examples

MODY2 = glucokinase mutations
Can be heterozygous or homozygous, the former causes MODY, the latter causes neonatal diabetes (Njolstad, 2001, NEJM)
Activating mutations rarely cause dominant persistent hyperinsulinism

Insulin mutations are typically either:
Processing mutations - misfolded protein accumulates and causes ER stress and apoptosis, hence these need only be heterozygous for an effect
Biosynthesis mutations - decreased [insulin] due to faulty transcription or mRNA processing, need to be homozygous for an effect
MODY7 = insulin mutations

M210K causes MODY if heterozygous and neonatal diabetes if homozygous

Glucokinase activators have been theorised as able to treat MODY and in clinical trials have benefits but these are very short lived and there is increasing evidence of liver steatosis. They did work in vitro and in mice

HNF1alpha mutations alter metabolism in mature beta-cells by reducing glycolysis and ATP generation. This has been shown to reduce insulin, GLUT2, pyruvate kinase etc expression and increase UCP2 mRNA and protein (Wang, 2000, EMBO J)

Sulphonylureas

They can bypass issues caused by mutations in metabolic enzymes/TFs or K+ATP channels but cannot be used to mutations that result in impaired insulin synthesis

Particularly effective in patients with HNF1alpha mutations - shown by comparison of gliclazide and metformin (Lancet, 2003)

Persistent hyperinsulinaemic hypoglycaemia of infancy/ congenital hyperinsulinism (PHHI/CHI)

Mostly caused by autosomal recessive or sporadic mutations, some are in Kir6.2, most are in SUR1 (Kane, Nat Med, 1996)

Treated by glucose in fusion, high sugar diet + drugs to inhibit insulin release or subtotal pancreatectomy

SUR1 mutations cause truncated or misfolded protein, reduced surface expression or partially functional channels
Some mutations (K67N in Kir6.2) have no effect on the channel in vitro

Neonatal diabetes may be associated with DEND syndrome

Activating mutations in Kir6.2 and SUR1 subunits are common

Disease severity is determined by ATP sensitivity of the mutant channels (ND alone, iDEND or DEND)

Theorised to work in neonatal diabetes as they could still bind to K+ATP channels in Xenopus models

Initial trials were successful but the genotype can still influence the benefit received from these drugs (Pearson, 2006, NEJM)

Benefit is largely derived from sensitivity of K+ATP channels

K+ATP channels in other areas of the body are that to cause DEND syndrome (neurons = SUR1 + Kir6.2, skeletal & cardiac muscle = SUR2A & Kir6.2, smooth muscle = SUR2B + some Kir6.1)

Mutations in muscle results in weakness and in neurons it results in decreased muscle stimulation, shown by looking at V59M mutation in mice (Clark, 2010, Science)

E23K mutation in Kir6.2 increases D2M risk by ~15% and is linked to S1369A mutation in SUR1 (Hamming, 2009, Diabetes)

Strong environmental effect of obesity , making it difficult to draw trends from genetics and delineate between increased risk of diabetes and obesity, or a third factor that causes both

Beta cell mass decrease in D2M patients, but this appears to be correlated to time since diagnosis than being a causative factor

TCF7L2 is thought to be especially predictive of D2M, but may be more predictive of obesity. PPARG and KCNJ11 were the first two genetic biomarkers identified (Prasad, 2015, Genes)

These findings have mainly informed pharmacoepignomics rather than triggered a wave of new therapies

Bariatric Surgery

Types include:
Gastric band
Roux-en-Y Gastric Bypass
Sleeve Gastrectomy
Endoluminal sleeve

For morbid obesity has shown significant improvements in weight loss - Swedish NEJM 2007 study showed 25% weight reduction with gastric bypass.
Arterburn, JAMA, 2015 shows 1/3 drop in mortality after 12 years

Roux-en-Y GB seems to be the most effective for diabetes, followed by sleeve gastroectomy

Causes a reduction in weight loss, typically mediated by reduced appetite (only banding doesn't do this)
Can also resolve diabetes - some patients can cease taking therapies within a week. Calorie restriction improves insulin sensitivity even before surgery
There is a potential role for gut hormones in both of these

78.1% of patients in the Buchwald 2009 study (Am J Med) had complete resolution of D2M, which increased to 80.3% for those with RYGB

Is thought that the ideal therapy is one which mimics bariatric surgery without surgery as medication is easier to distribute

RYGB has shown clear improvements in glucose tolerance and a reemergence of the incretin effect after 1 week (Joergensen, Am J Phys, 2012)

Stomach exclusion seems to be key to this, duodenal exclusion has a role but only occurs in RYGB so cannot be the major factor

Calorie restriction

Caloric restriction reduces fat build up at ectopic sites (lipotoxicity)

Even w/o weight loss it can have a large impact

600kcal diet has been shown to restore glucose levels to normal in D2M patients (Taylor, Diabetologica, 2011). Experiment showed reemergence of two peaks in insulin conc after a meal

Experiments have shown that gastrectomy improvements in insulin secretion are mediated by improvements in GLP-1 secretion by inhibiting GLP1R with exendin-9, the result is an abolition of improvements seen after RYGB (Larraufie, Cell Reports, 2019)

GLP-1 drives hypoglycaemia when patients lose a lot of weight post-surgery and can be treated with exendin-9

Mechanism of increased GLP-1 and PYY release

There are no transcriptional or hormone biosynthetic changes in each cell following bariatric surgery. However the make up of cells has changed as many have been removed, meaning overall peptide production has shifted (Larraufie, Cell Reports, 2019)

Case report found in Dirksen, 2010, Diabetes Care

The changes reflect altered nutrient flow, there is now a higher [glucose] at the site of L cells (Roberts, 2019, Diabetes)

In mouse models GLP-1 has a less clear role in improving glucose tolerance (Garibay, 2016, Endocrinology). Things to watch out for are mice with different initial and final body weights and the method of glucose administration

Current hot topics are whether the gut produces glucagon after bariatric surgery and whether islets produce GLP-1, and the importance if they do

Also the pathway of post-surgical weight loss is not understood. GLP1R and PYY do not seem to be relevant (Boland, Mol Metab, 2019) but FXR may be (Seeley, Nature, 2014) although mice were lighter to start with which may have prevented weight loss

PYY's role

Is at low levels in obesity and PYY (3-36) is related to food intake (Batterham, 2003, NEJM)

Is known to contribute to decreased food intake after RYGB. Inhibiting GLP1R and Y2R with exendin and sitagliptin reduces food intake showing that they both have a role in the reduction in food intake post-bariatric surgery

Stimulation of L cells may do this (mimicking high [glucose] at them) by increasing PYY and GLP-1 secretion

Injectable peptides (GLP-1 already exists, PYY and oxyntomodulin are under evaluation) but these are diffiult to work with

Dual GLP-1 & glucagon/GIP agonist peptides (Tschoep, Cell Metab, 2016)

Triple/multiple agonistss are currnently in development (Tan, 2017, Clin End Metab)

Glucagon reduces food intake so on balance seems to favour weight loss

Cotadutide (GLP-1/glucagon agonist) recently underwent PhIIa trial and showed positive results (Parker, 2019, J Clin Endocrin Metab)

Tirzepatide (GIP/GLP-1 agonist) is now in PhIII trials after passing PhII (Frias, 2018, Lancet)

May be achieved by reducing proximal absorption using SGLT1 inhibitors, slowing digestion with orlistat or increasing dietary fibre to trap nutrients in complex structures

Small molecules targeting GPCRs that ordinarily respond to aas, LCFAs, bile acids and SCFAs are in the clinic

Metformin

Activates AMPK, is a complex I inhibitor and causes a degree of lactic acidosis in patients. Is recognised that it is concentrated in the intestinal epithelium because apical transporters take it up but there are no suitable transporters basolaterally. Additionally new formulations with low systemic bioavailability are equally as effective as the standard

It increases fasting and postprandial GLP-1 and PYY concentrations (DeFonzo, 2016, Diabetologica). There are conflicting reports on it stimulating L cells. Effects may be related to gut motility

May mediate its action through GDF15 as levels of it are raised in obese individuals treated with metformin. In mice metformin only triggered weight loss in GDF15 mice (GDF15 KO mice received no benefit from metformin) (Coll, 2019, BioRxiv)

Action may also be related to increased glucose uptake from the circulation, which can be shown using radiography 18F-FDG PET