SGLT2 inhibitors have been proven to be effective in several preclinical rodent models of non-alcoholic fatty liver disease (NAFLD). Using a choline deficient diet to recapitulate some of the histological features of human non-alcoholic steatohepatitis (NASH), it was found that 5 weeks of SGLT2 inhibition led to significant reductions in hepatic triglyceride content and improved markers of liver fibrosis. Similarly, 4 weeks of treatment in obese mice led to improved glucose tolerance, reduced hepatic steatosis and reduced markers of liver oxidative stress in a dose dependent manner. These findings corresponded with an improvement in traditional liver function tests including the aminotransferases (ALT and AST). The widely used antidiabetic agent metformin has been shown in rodent models to increase hepatic insulin sensitivity and lower liver fat content which is in contrast to the findings in humans where metformin increases hepatic insulin sensitivity, reduces body weight but does not decrease liver fat content. The reason for the discrepancy between the animal and human studies, with regards to liver fat content remains unclear.

The investigators hypothesise the following:

• SGLT2 Inhibitors have the potential to decrease lipid accumulation in the liver through reduced de novo lipogenesis (DNL)
• There will be no decrease in endogenous lipid synthesis (DNL) with metformin and thus no change in liver fat content.

There are two arms to this study.

• Arm 1: x10 participants with poorly controlled type 2 Diabetes (T2DM) who have been recommended to start an SGLT2 inhibitor called dapagliflozin will be recruited.
• Arm 2: x13 participants with insulin resistance who have not yet started any diabetic medication will be recruited and will be prescribed metformin at standard clinical doses.

The two arms will run in parallel and all participants will undergo identical investigations before and after 3 months of treatment with either dapagliflozin or metformin. Investigations will include liver magnetic resonance imaging/spectroscopy, fat biopsy, fat microdialysis sampling, two-step hyperinsulinaemic euglycaemic clamp, breath sampling and stable glucose and palmitate isotope infusions.

The investigators aim to show that SGLT2 inhibition decreases liver fat whereas we aim to demonstrate why liver fat remains unchanged in humans, treated with metformin. These data will provide the first evidence for the use SGLT2 inhibitors in NAFLD, and will be highly informative for the design of future clinical studies. Moreover, the data gained from the metformin arm of the study will provide the first mechanistic evidence in humans of the effects of metformin on hepatic fatty acid metabolism.

Inclusion Criteria:

Arm 1: SGLT2 inhibitors

• Volunteers with diagnosis of Type 2 Diabetes on oral anti-diabetic therapy at a stable dose for ≥3 months including one of the following:

  i. Metformin monotherapy ii. Sulphonylurea monotherapy iii. Metformin and Sulphonylurea dual therapy

• All volunteers will be due to start SGLT2 inhibitor therapy for inadequate glycaemic control and it will be prescribed according to licensed indications.

Arm 2: metformin

• Insulin resistant treatment naive individuals as defined by fasting insulin and / or glucose in top 10th percentile

Both arms:

• Participant is willing and able to give informed consent for participation in the study.
• Male or Female, aged between 18 years and 70 years.
• BMI: 25-45 kg/m2
• HbA1C: 42-86mmol/mol
• Normal renal function

Exclusion Criteria:

Arm 1: SGLT2 inhibitors Volunteers taking insulin, glucagon-like peptide 1 analogues, thiazolidinediones, or dipeptidyl peptidase IV inhibitors

Arm 2: metformin Volunteers taking insulin, glucagon-like peptide 1 analogues, thiazolidinediones, SGLT2 inhibitors, metformin or dipeptidyl peptidase IV inhibitors

Both arms

• Age <18 or >70 years
• Body mass index <25 or >45kg/m2
• A blood haemoglobin <120mg/dL
• History of alcoholism or a greater than recommended alcohol intake (Recommendations > 21 drinks on average per week in men and > 14 drinks on average per week in women)
• Pregnant or nursing mothers
• History of severe claustrophobia
• Presence of metallic implants, pacemaker
• Haemorrhagic disorders
• Anticoagulant treatment
• History of albumin allergy