Acute Effects of Canagliflozin, a Sodium Glucose Co-Transporter 2 (SGLT2) Inhibitor on Bone Metabolism in Healthy Volunteers

Background:

Canagliflozin is a new oral drug for the treatment of type 2 diabetes mellitus (T2DM), and is
one of four recently FDA approved sodium glucose co-transporter 2 (SGLT2) inhibitors, which
target renal glucose reabsorption and offer promising improvement in HbA1c. In the approval
process, the FDA Advisory Committee reviewed data suggesting that canagliflozin increased the
incidence of fractures. In addition, the drug induced changes in phosphate, bone resorption
markers, parathyroid hormone (PTH) and vitamin D metabolism which might mediate the adverse
changes in bone homeostasis. For a variety of reasons, the data on bone fracture risk are
relatively limited. First, the drug s development program was focused primarily on
demonstrating efficacy, and bone fractures were only one of many safety end-points which were
monitored. Second, only a minority of patients (approximately 1%) experienced bone fractures
in the course of the development program. Finally, there appears to be a lag time prior to
the time increased bone resorption translates into a significant increase in the rate of bone
fractures. We hypothesize that this class of drugs causes a cascade of hormonal changes
induced by increased phosphorus reabsorption that leads to significant changes in fibroblast
growth factor 23 (FGF23), PTH, and vitamin D metabolism which ultimately increase fracture
risk.

Aim:

The primary endpoint is to determine the effects of canagliflozin on bone health by
evaluating changes in the area under the curve (AUC) of FGF23 during the first 24-72 hours.

Secondary endpoints include the evaluation of canagliflozin on other biochemical parameters
in the early phase (1 week) of drug administration during which we hypothesize a new steady
state will be reached related to bone metabolism including PTH, 1,25 vitamin D, tubular
reabsorption of phosphate (TRP), and carboxy-terminal telopeptide (CTX).

Methods:

A randomized, blinded, placebo-controlled cross-over pilot study of healthy volunteers ages
18 years and older with a BMI of 20 30 kg/m(2).

Patients will be randomized to canagliflozin (300mg once daily) or placebo for 5 days and
will be studied as inpatients (NIH Clinical Center metabolic unit). Serial blood and urine
testing for 4 hours after daily drug administration and at 12, 16 and 24 hours thereafter
will be used to assess changes in the pre-specified endpoints. Each subject will be provided
a diet containing fixed contents of phosphate, sodium, and calcium throughout the study,
beginning 7 days prior to the administration of drug (or placebo).

Study Objectives:

The primary objective of this study is to determine the AUC change in FGF23 within 24-72
hours caused by the SGLT2 inhibitor canagliflozin. We will also examine the effects of
canagliflozin on various biochemical markers of bone metabolism (including phosphorus, PTH,
various hydroxylated vitamin D derivatives, calcium, and markers of bone turnover) during the
first 5 days of drug administration.

Background and Rationale:

Standards of Care in Diabetes Mellitus Type 2 (T2DM)

The epidemic of T2DM encompasses a spectrum of pathophysiological derangements that require a
focused and individualized therapeutic approach. When considering medication choices, the
American Diabetes Association (ADA) guidelines suggest inclusion of the duration of disease,
age, body weight, life expectancy, expense and variations in microvascular and macrovascular
complications. Therapy is usually initiated with metformin which is well known to decrease
hepatic gluconeogenesis and effectively lower hemoglobin A1c (HbA1c). This drug is usually
chosen as first line as it is well tolerated, has comparable efficacy, an attractive generic
price and low rate of toxicity. Over time, progressive deterioration of beta cell function
and insulin resistance usually require add-on therapies. Options are sulfonlyureas,
dipeptidyl peptidase-4 (DPP4) inhibitors and thiazoladinediones (TZDs). However, as Bennett
et al recently evaluated in a meta-analysis, most therapies have similar effectiveness in
reduction of HbA1c by 1 percent, but each class has different side effects.

TZDs, for example, offer comparative improvements in hyperglycemia, but have more side
effects than other agents. Additionally, evidence for serious adverse events emerged years
after this class of drugs had been on the market, thus delaying appropriate recognition.
Liver toxicity had been the first serious side effect to be discovered and lead to the
withdrawal of troglitazone, and restrictions were temporarily placed on rosiglitazone because
of concerns related to increased myocardial infarction risk. TZDs have also been shown to
increase the incidence of chronic heart failure (CHF) in select patients 8, bone fractures
especially in women, and bladder cancer especially in those receiving therapy for 5 years or
longer.

Risk assessments are important for all patients when considering initiation of a new drug. In
particular, vulnerable populations are at the highest risk, including older T2DM patients
with multiple co-morbidities who often suffer higher rates of adverse medication effect.
Likewise, patients diagnosed with T2DM at a younger age are likely to receive these drugs for
very long duration, which heightens concerns about long term drug toxicities. Therefore, it
is critical that selection of antidiabetic medications account for both short term and long
term consequences.

Rationale for SGLT-2 inhibitors as Novel Drugs for the Treatment of T2DM

Hyperglycemia is a complex disease state that has been described as an interplay among the
ominous octet 18 in which the pancreas, liver, small intestine, skeletal muscle, adipose
tissue, brain and kidney all contribute to glucose metabolism. Pharmacotherapy development
has effectively sought to target each of these mechanisms via various pathways. The renal
reabsorption of glucose has only recently become a focus for therapeutic intervention with
the introduction of sodium glucose co-transporter 2 (SGLT2) inhibitors in 2013.

SGLT2 is located in the S1 segment of proximal renal tubule and is a high-capacity,
low-affinity transporter that mediates the majority of the glucose reabsorption. Glucose
which escapes SGLT2 is reabsorbed by SGLT1, a low-capacity, high-affinity transporter located
in the S3 segment of the proximal renal tubule. Under conditions where SGLT2 is inactivated
by either a mutation or an inhibitor, urinary glucose excretion can range from 20-200 g/day
or more.

The evidence for SGLT2 as a major pathway for renal glucose reabsorption has come from
genetic studies of individuals with familial renal glucosuria (FRG), an autosomal recessive
disorder of SLGT2. It has been almost a century since the first case has been described which
causes glucosuria in the setting of a normal blood glucose concentration and the absence of
other signs of renal tubular dysfunction. Different phenotypes have been described based on
the extent of glucosuria. Individual and kindred studies have helped identify least 44
different mutations in the SLC5A2 gene. Long term follow up of these individuals has
demonstrated no major health problems. However, given inconsistency of follow up and a small
number of reported cases, it is difficult to infer the true extent of complications.

Another genetic defect of renal glucose transport is known to be due to mutations in the
GLUT2 transporter, which is responsible for Fanconi-Bickel syndrome (FBS). This is a rare
autosomal recessive disorder caused by homozygous or compound heterozygous mutations in the
GLUT2 (SLC2A2) gene. It is characterized by hepatorenal glycogen accumulation, proximal renal
tubular dysfunction (Fanconi nephropathy), and hypophosphatemic vitamin D dependent rickets
with severe growth retardation. The relation between glucose transport and phosphate
metabolism is not well understood. Mannstadt et al evaluated two families with varying
degrees of hypophosphatemic rickets and urinary phosphate wasting and found novel mutations
in GLUT2 as a causative factor for calcium and phosphate imbalances, thought to be due to
changes in the phosphate transporter Npt2c. To support this hypothesis, they studied
tgGlut2-/- mice who also demonstrated a decrease in Npt2c expression in the proximal renal
tubules. While the interplay between this genetic disease and hypophosphatemic rickets is
still being elucidated, changes in glucose homeostasis have been shown to have profound
effects on bone health and mineral metabolism.

Pharmacological support for the development of SGLT2 inhibitors comes from studies using
phlorizin, a natural product discovered in 1835 from root bark of apple trees. Von Mering
discovered that ingestion of phlorizin produced glucosuria in humans. Rossetti and DeFronzo
et al found that in partially pancreatectomized rats, phlorizin normalized blood glucose
levels. While this theoretical concept supported the idea that glucosuria could normalize
fasting and fed plasma glucose levels (and reverse insulin resistance in animals), it did not
immediately translate into pharmacotherapy due to phlorizin s multiple limitations: (1) low
selectivity for SGLT2 over SGLT1 causing frequent gastrointestinal side effects, (2)
degradation by the gut enzyme disaccharidase reducing the oral bioavailability, and (3) the
interaction of one of its metabolites, phloretin, which inhibits GLUT2 and GLUT1 glucose
reabsorption in the gut.

Canagliflozin was the first SGLT2 inhibitor to be approved by the FDA, and its approval took
place in January, 2013. However, data on dapagliflozin had been presented to the FDA for
review as early as July, 2011, but its approval was delayed until January 2014, mostly
because of concerns about increased risks for bladder and breast cancer.

The SGLT2 inhibitor which we will study in this protocol, decreases the renal threshold for
glucose excretion (occurring on average > 70-90mg/dL in the presence of canagliflozin) and
increases glucosuria to approximately 80-100 grams/day, thereby reducing plasma glucose. At
the highest FDA approved dose of canagliflozin (300mg/day) given as monotherapy, HbA1c
improves by 1.2%. In addition, canagliflozin therapy leads to several other pharmacological
effects: evidence of 2-3% weight loss, decreases in systolic blood pressure by about 4.5mmHg
from baseline, and an increase in insulin sensitivity all of which contribute to a beneficial
clinical profile for patients with T2DM.

SGLT2 Inhibitor Changes in Bone Health: A Summary of the Data

The first indication that SGLT2 inhibitors could possibly influence bone health came from
animal studies 38. Rats exposed to high doses of dapagliflozin (the first developed drug of
this class) were observed to have disturbances in calcium homeostasis, tissue mineralization,
increased trabecular bone and renal medullary tubular degeneration. The proposed mechanism
was thought to be due to the high dose of dapagliflozin causing non-selective inhibition of
SGLT1 in the gut, thereby altering the pH balance in the intestinal lumen and increasing
calcium reabsorption. These findings are not thought to be relevant to human pharmacology
when SGLT2 inhibitors are administered at approved doses. Nevertheless, given these findings,
special attention was paid to bone mineralization and changes during clinical development.

Specifically, data presented at the FDA advisory committee meeting on dapagliflozin in July
2011 demonstrated an increase in fracture rate after 104 weeks in a dedicated study of
patients with moderate renal impairment (eGFR 30-59mL/min/1.73m(2)) with 9.4% fractures on
the 10 mg dose and 6% fractures on the 5 mg dose, versus zero fractures on placebo. In
patients with normal renal function, they also noted a 2-fold increase in fractures. However,
the sponsor argued that the imbalance was non-significant because of the associated small
laboratory changes, possible influence of weight loss on fracture rate, minimal effects on
bone mineral density (BMD) and inconsistencies between various short term and long term
studies. Ultimately, the drug was not approved for other safety related reasons, but further
studies dedicated to bone health were recommended. Later, Ljuggren et al reported no
significant changes in bone markers or changes in DEXA at 50 weeks, but mean increases in
phosphate were observed and trends towards increased CTX and procollagen type 1 N-terminal
propeptide (P1NP) were also evident.

Canagliflozin, the second drug reviewed by the FDA in January 2013, also underwent a
dedicated analysis on bone safety due to concerning findings in animal studies similar to
those of dapagliflozin. Specific parameters that were evaluated included calcium, phosphorus,
bone turnover markers and fracture rates. A phase II study found a 23-37% rise in the bone
resorption marker CTX by week 3 which persisted until week 12. Likewise, an increase in PTH
was observed by week 3, which returned toward baseline by week 6-12. At high doses, this
study also demonstrated slight decreases in 25-hydroxyvitamin D as well as 1,25
hydroxyvitamin D.

In a pooled analysis of placebo controlled studies (n=2,313) for 26 weeks, a 5.1% mean change
in serum phosphate was observed among the highest dose of 300mg compared to placebo. In
another 26-week analysis (n= 269) of renal impairment patients (eGFR greater than or equal to
30 and < 50 ml/min/1.73 m(2)), serum phosphate increased 7.8% above baseline and
1,25-hydroxyvitamin D decreased by 8.1% from baseline 44. Interestingly, the most acute
changes in serum phosphate (0.5 mg/dL mean change from baseline) was reported within three
weeks of study initiation.

A dedicated bone study in older adults age 55-80 (n=715), the highest dose of 300mg of
canagliflozin increased serum CTX over baseline by 24.9% at 26 weeks and by 22% at 52 weeks,
with P1NP decreasing by 6.9% in the highest dose group at 26 weeks. Also in the highest dose
group, DEXA scanning found a decrease of 0.7% mean percent change in the lumbar spine and
total hip, and quantative-CT demonstrated a decrease of 1.9% and 1.6% in the lumbar spine and
total hip, respectively.

In terms of fractures, a prospectively adjudicated analysis across all phase III studies
demonstrated a 0.6% difference in the fracture rate between canagliflozin and
non-canagliflozin groups. Analysis by fracture location and trauma classification
demonstrated an imbalance in upper limb fractures not favoring canagliflozin. This imbalance
persisted in low trauma upper limb fractures and spine (Figure 1). In a pooled analysis of
all treatment arms, females had a higher incidence of upper limb fractures (1.2% [31/2608])
compared to placebo (0.4% [5/1338]). The location of distal extremities, including the hand,
distal forearm and wrist are similar to the wide variety of fracture sites found in other
diseases such a primary hyperparathyroidism.

- INCLUSION CRITERIA:

We are targeting healthy greater than or equal to 18 years old, inclusive of all races and
ethnicity within a BMI of 20 30 kg/m(2). Specifically, we have defined healthy to mean:
normal fasting glucose and hemoglobin A1c less than or equal to 6%, normal Hb, no
glucosuria, normal renal function (based on normal serum creatinine + Cystatin C), urine
albumin:creatinine ratio, protein:creatinine ratio, and GFR > 80 as calculated by the
CKD-Epi equation and normal lab urinalysis.

EXCLUSION CRITERIA:

If you have any of the following health issues, you cannot participate in the study:

- Presence of heart disease, untreated high blood pressure (>140/90 mm Hg), orthostatic
hypotension or symptomatic hypotension, cancer, diabetes, recurrent symptomatic
hypoglycemia and /or history of recurrent genital or urinary tract infection, thyroid
disease, or any other condition that affects bone health

- Past history of eating disorder or psychiatric disorders, including severe depression,
anxiety, or psychosis or presently on treatment with medications for any of these
conditions

- Taking certain medications, especially those that affect bone metabolism (e.g., high
dose vitamin D [>1000 units daily] or calcium supplements [>800mg daily], high dose
vitamin A [>20,000 units daily], phosphate binding antacids, calcitonin, calcitriol,
growth hormone, or any anti-seizure medications for any reason including valproic
acid, lamotrigine), certain medications for high blood pressure (diuretics), steroids
including inhalers, diet/weight loss medications, or any other medications at the
discretion of the principal investigator and/or study team

- Have started, increased or decreased calcium [>400mg daily] or vitamin D [>1000 units
daily] supplements within 2 weeks of the study

- Dependence or regular use of alcohol (>2 drinks per day), tobacco (smoking or
chewing), amphetamines, cocaine, heroin or marijuana over the past 6 months

- Volunteers will be excluded if they have abnormal blood concentrations of

- inorganic phosphate level (less than or equal to 2.5 mg/dl or greater than or
equal to 4.8 mg/dl),

- parathyroid hormone (PTH) (less than or equal to 60 pg/ml),

- creatinine (less than or equal to 1.5 mg/dl) or eGFR (< 80 ml/min/1.73sq.m),

- fasting glucose (greater than or equal to 100 mg/dl),

- hemoglobin (less than or equal to 11 g/dl),

- liver function tests (more than twice normal),

- testosterone (less than or equal to 260 ng/dl)

- Participation in a vigorous exercise program (>3h/day of vigorous activity)

- Consume more than 300 mg/day of caffeine (about two to three 8 fluid ounce servings)

- Have strict dietary concerns (e.g., vegan or kosher diet, multiple food allergies)

- Cannot commit to the research experience at the Clinical Research Center as required
by the study timeline

- Have previous hypersensitivity reaction to canagliflozin (including but not limited to
rash, raised red patches on your skin (hives), swelling of the face, lips, tongue, and
throat that may cause difficulty in breathing or swallowing).

- Positive urine pregnancy test and/or planning to become pregnant during the course of
the study.

- You are unwilling to use effective contraceptive methods for duration of study
(hormonal or barrier.

- Irregular menstrual cycles