Citation
The antidiabetic drug metformin inhibits protein tyrosine phosphatase 1B activity in xenopus laevis oocytes

Material Information

Title:
The antidiabetic drug metformin inhibits protein tyrosine phosphatase 1B activity in xenopus laevis oocytes
Creator:
Holland, William Louis
Publication Date:
Language:
English
Physical Description:
ix, 64 leaves : illustrations ; 28 cm

Subjects

Subjects / Keywords:
Hypoglycemic agents ( lcsh )
Protein-tyrosine phosphatase ( lcsh )
Xenopus laevis ( lcsh )
Insulin resistance ( lcsh )
Non-insulin-dependent diabetes ( lcsh )
Hypoglycemic agents ( fast )
Insulin resistance ( fast )
Non-insulin-dependent diabetes ( fast )
Protein-tyrosine phosphatase ( fast )
Xenopus laevis ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 60-64).
Statement of Responsibility:
by Wliiliam Louis Holland.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
50741892 ( OCLC )
ocm50741892
Classification:
LD1190.L45 2002m .H64 ( lcc )

Downloads

This item has the following downloads:


Full Text
THE ANTIDIABETIC DRUG METFORMIN INHIBITS PROTEIN TYROSINE
PHOSPHATASE IB ACTIVITY IN XENOPUS LAEVJS OOCYTES
by
William Louis Holland
B.S., Biology, University of Colorado, 2000
B.S., Psychology, University of Colorado, 2000
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Arts
Biology
2002


This thesis for the Master of Arts
degree by
William Louis Holland
has been approved
by
David S. Albeck
4.
_______
Date


Holland, William Louis (M.A., Biology)
The Antidiabetic Drug Metformin Inhibits Tyrosine Phosphatase IB Activity in
Xenopus laevis Oocytes
Thesis directed by Professor Bradley J. Stith
ABSTRACT
For over 40 years, metformin has been used for the clinical treatment of
insulin resistance in type 2 diabetes. Despite the drugs widespread use, its cellular
mechanism of action has remained unclear. While previous research has indicated
that metformin stimulates the insulin receptor, the mechanism of metformins
stimulation of the receptor has not been described. In our current study, we report
that metformin stimulates the insulin receptor by inhibiting protein tyrosine
phosphatase IB (PTP-1B).
The binding of insulin to its receptor stimulates phosphorylation of a
regulatory domain in the receptor. Specifically, tyrosine residue numbers 1146,
1150 and 1151 constitute the regulatory phosphorylation sites of the receptor.
When all three of these sites have been phosphorylated, insulin receptor tyrosine
kinase (IRK) activity is fully activated. PTP-1B removes phosphate groups from
these regulatory sites, thus inhibiting IRK activity.


Using a new HPLC method, we report the ability to separate the full
spectrum of different phosphopeptides corresponding to the regulatory domain of
the insulin receptor. Metformin stimulated the ability of purified beta subunit of the
insulin receptor (PIRK) to phosphorylate a peptide containing the regulatory domain
of the insulin receptor. In another series of experiments using plasma membrane-
cortices (PMCs), metformin was found to increase tyrosine kinase activity and
decrease tyrosine phosphatase activity toward the regulatory phosphopeptide. The
PMCs contain both tyrosine kinase and phosphatase activities. Metformins ability
to decrease tyrosine phosphatase activity was negated by removal of soluble
proteins with a salt wash of the PMCs (although salt washes were not used with the
regulatory peptide).
In-vitro phosphatase assays of PTP-1B indicate that PTP-1B was not directly
inhibited by metformin, but metformin did inhibit PTP-1B in the presence of PMCs.
With experiments using a salt wash of the PMCs and the in vitro PTP-1B assay,
there are two lines of evidence that PTP-1B was indirectly inhibited by metformin.
This suggests that metformin inhibits tyrosine phosphatase activity through a
soluble intermediate.
This abstract accurately represents the content of the candidates thesis. I
recommend its publication.
Signed
IV


ACKNOWLEDGEMENT
I would like to thank Dr. Brad Stith for his support and friendship and for the
research opportunities he has provided me.
Additionally, I would like to thank Dr. Doug Petcoff, Dr. Martin Gonzalez and
Dr. Dave Albeck, for their endless hours of guidance, support and friendship.
Thank you to Kai Savi, Patricia Medina, Khulan Batbayar, Erinn Stauter, and
Timberley Roane for their support and friendship.


CONTENTS
Figures...............................................................ix
Tables.................................................................x
Chapter
1. Introduction.......................................................1
1.1 Diabetes Mellitus..................................................1
1.2 Metabolic Syndrome X...............................................3
1.3 Insulin Receptor Activation........................................5
1.4 Insulin Signaling..................................................8
1.5 Protein Tyrosine Phosphatase IB...................................11
1.6 Metformin.........................................................14
1.7 The Xenopus Cell Model............................................16
2. Materials and Methods.............................................17
2.1 Xenopus Maintenance...............................................17
2.2 Xenopus Oocytes...................................................17
2.3 Plasma Membrane-Cortices..........................................18
2.4 Analysis of Endogenous Tyrosine Phosphatase Activity.............19
2.5 Analysis of PTP-1B Activity In Vitro.............................22
2.6 Analysis of PTP-1B Activity with PMCs............................23


2.7 Development of a New HPLC Peptide Analysis for
Insulin Receptor Autophosphorylation...................................24
2.8 Analysis of the Phosphorylation State of Insulin
Receptor Regulatory Domain Peptides with (3IRK.........................29
2.9 Analysis of the Phosphorylation State of Insulin
Receptor Regulatory Domain Peptides with PMCs..........................30
3. Results................................................................31
3.1 Metformin Inhibited Tyrosine Phosphatase
Activity in PMCs.......................................................31
3.2 Metffomin Did Not Inhibit PTP-1B Activity In Vitro...................34
3.3 PMCs Allowed Metformin Inhibition of PTP-1B..........................36
3.4 Use of a Peptide Containing the Human Insulin
Receptor Regulatory Domain.............................................38
3.5 Metformin Stimulated piRK Activity In Vitro.........................42
3.6 The Human Insulin Receptor Kinase Preferentially
Added Phosphate to the 1150 Tyrosine...................................45
3.7 Metformin Stimulated Tyrosine Kinase Activity in PMCs...............47
3 .8 Metformin Inhibited Tyrosine Phosphatase Activity that
Removes Phosphate from the Insulin Regulatory Peptide..................49
4. Discussion.............................................................52
4.1 Does Metformin Directly or Indirectly Inhibit PTP-1B?..................54
4.2 Comparison with a Known Tyrosine Phosphatase
Inhibitor Vanadium...................................................55
4.3 Other Future Studies..................................................56
vii


5.0 Conclusion.........................................................59
References.............................................................60


FIGURES
Figure
1.3.1 Insulin Receptor Autophosphorylation...............................7
1.4.1 Insulin/Akt Pathway................................................10
1.5.1 PTP-1B Regulates Insulin Receptor Activity.........................13
2.4.1 Sample Preparation for Tyrosine Phosphatase Assay..................21
3.1.1 Metformin Inhibited Tyrosine Phosphatase Activity in PMCs..........33
3.2.1 Metformin Did Not Directly Inhibit PTP- IB In-Vitro................35
3.3.1 PMCs Rescued Metformin Inhibition of PTP-1B........................37
3.4.1 HPLC Separation of Regulatory Peptide..............................41
3.5.1 Metformin Stimulated piRK Activity In-Vitro........................44
3.6.1 piRK Preferentially Phosphorylated Tyrosine 1150
of the Regulatory Domain..........................................46
3.7.1 Metformin Stimulated Tyrosine Kinase Activity
in PMCs...........................................................48
3.8.1 Metformin Inhib ited Tyrosine Phosphatase Activity
Toward the Regulatory Peptide.....................................51
IX


TABLES
Table
2.7.1 A Summary of Regulatory Peptide Nomenclature and Elution Time.27
i
f
}
j
1
i
X


1. Introduction
LI Diabetes Mellitus
Diabetes mellitus is a disease characterized by insufficient synthesis, secretion
or response to the metabolic hormone insulin. Diabetes is clinically defined by the
American Diabetes Association by fasting glucose levels greater than 126 mg/dL
(greater than 20% above normal) (Zimmet et al., 1999). These high glucose levels, or
hyperglycemia, result in excessive urination and the development of painful and
debilitating complications.
There are two primary forms of diabetes. Type 1 diabetes, also known as
insulin-dependent diabetes mellitus (IDDM) generally strikes in childhood. This
juvenile form of diabetes is believed to be the result of autoimmune damage to the
insulin-producing (3 cells of the pancreas. Type 1 diabetics, who compose 5-10% of
the diabetic population, produce little to no insulin.
In type 2 diabetes or non insulin-dependent diabetes mellitus (NIDDM), the (3
cells are usually capable of producing some insulin; however target cells are insulin-
resistant and insulin production is insufficient to normalize plasma glucose
concentrations. Nearly 17 million Americans, or 3% of the population, currently
suffer from type 2 diabetes. Surprisingly, about 1 /3 of these cases are undiagnosed
(Margolis and Saudek, 2002).
1


In the early stages of NIDDM, many patients produce normal to abnormally
high amounts of insulin, but the hormone fails to lower blood glucose levels. Thus,
NIDDM is associated with insulin resistance. Insulin resistance is defined as a need
for an additional 200 units of insulin daily to control hyperglycemia (Anderson,
1994). Although the mechanisms underlying insulin resistance haven't been
elucidated, human and animal models indicate that insulin properly binds to its
cellular receptor; however, a defect in the cells internal signaling results in
subnormal glucose transport (Sredy et al., 1995).
Insulin resistance is characterized by a lack of response to insulin by the
body's tissues. It appears to be caused by a defect in the insulin-signaling cascade
rather than by a defect in insulin binding to the receptor. In insulin resistance, insulin
binds its receptor, but glucose transporters are not recruited to the plasma membrane
(as they are in healthy tissue).
The glucose tolerance test is a way of quantifying insulin resistance. In this
test, patients fast for at least 12 hours and then consume a sugar solution. Blood
glucose levels are then measured. Glucose levels remain very high in the
bloodstream in diabetics, glucose intolerant or insulin resistant individuals. Current
American Diabetes Association guidelines define Type 2 diabetes by glucose levels
of greater than 126 mg/dl, whereas insulin resistant individuals have glucose levels of
2


111-126 mg/dl (Zimmet et al., 1999). Normal individuals have fasting glucose
concentrations below 110 mg/dL.
1.2 Metabolic Syndrome X
Insulin resistance and hyperglycemia are also symptoms of the newly defined
metabolic syndrome X (MSX) (Hansen, 1999). MSX is known by many names,
including insulin resistance syndrome and the cardiovascular disease risk factor
cluster. It is characterized by five major symptoms: obesity, hyperglycemia,
dyslipidemia, glucose intolerance, and insulin resistance. An estimated 75 million
Americans are believed to suffer from the syndrome; this is greater than one out of
every four people (Hansen, 1999)! It is a growing epidemic in the United States and
throughout the world.
The most evident major symptom of MSX is obesity central or visceral
obesity in particular. Central obesity is determined by a waist to hip ratio. Males
with a ratio of greater than 0.90 and females with greater than 0.85 are centrally obese
(Peeke and Chrousos, 1995). Central obesity, an excess of peritoneal fat, is a major
risk factor for heart disease and diabetes. Obese individuals with this pear-shaped
body type are at far greater disease risk than persons with peripheral obesity
(Bjomtorp, 2000). It is believed that the nature of central obesity, in this disorder, is
due to enlarged adipocytes, high insulin and cortisol levels, and increased
3


glucocorticoid receptor concentrations in the abdomen. Additionally, decreased leptin
sensitivity, often associated with MSX, has been linked to the buildup of adipose
tissue (Zimmet et al., 1999). Leptin is a hormone, released by fat cells, that reduces
hunger.
In MSX, abnormal lipid levels (dyslipidemia) are present in the bloodstream.
High triglyceride levels (triglyceridemia) and low high-density lipoprotein (HDL)
cholesterol (the good form) levels are examples of dyslipidemia. Triglyceridemia
is defined by fasting triglyceride levels above 200 mg/dl (Zimmet et al., 1999) and
low HDL cholesterol is defined as levels below 35 mg/dl. High total cholesterol and
low-density lipoprotein (bad form) levels are potent risk factors for heart disease
that are often present in the metabolic syndrome; however, they are not required.
Relatively little is known about the cause of the dyslipidemia in MSX, but it is
believed to be related to altered lipolysis induced by hormonal changes (Bjomtorp
and Rosmond, 1999).
Hypertension is the least consistently-associated symptom of the syndrome.
Blood pressure greater than 140/90 is generally considered sufficient for diagnosis of
high blood pressure (Zimmet et al., 1999). Although this symptom of MSX is seen
far less often than the five symptoms noted above, hypertension is also linked to
cardiovascular disease.
4


Each of the major symptoms for Metabolic Syndrome X is a risk factor for
cardiovascular disease and they are often expressed concurrently. A vast majority of
type 2 diabetics have MSX.
1.3 Insulin Receptor Activation
The insulin receptor is composed of two extracellular alpha subunits and two
transmembrane beta subunits all linked through disulfide bonds (Figure 1.3.1). Each
beta subunit possesses a tyrosine kinase domain and a regulatory domain. Each alpha
subunit contains a binding site for insulin.
When the peptide hormone insulin binds to this membrane receptor it induces
a conformation change, and this partially stimulates IRK activity. This weak tyrosine
kinase activity is sufficient to allow autophosphorylation of the receptor as the p
subunits phosphorylate themselves. In autophosphorylation, the kinase domains
place phosphate on the regulatory domains. The regulatory domain consists of
tyrosine residues 1146,1150, and 1151. All three tyrosine residues must be
phosphorylated to induce full activation of the insulin receptor.
Thus, when insulin binds to the receptor it causes a conformational change
that allows the Beta subunits to be autophosphorylated by their tyrosine kinase
domains. Upon complete phosphorylation of the regulatory domain of the insulin
receptor, the receptor becomes a fully active tyrosine kinase. Once fully activated,
5


the insulin receptor kinase (IRK) can phosphorylate other substrates and induce
numerous signaling pathways.
6


InsJin Fteceptor
1146,1150, and 1151 phosphorylated
Figure 1.3.1 Insulin Receptor Autophosphorylation. The insulin receptor
activates its tyrosine kinase activity via insulin induced autophosphorylation of the
regulatory domain (tyrosines 1146, 1150, 1151).
7


1.4 Insulin Signaling
Upon binding to an insulin receptor, insulin initiates a chain reaction of
biochemical events within numerous signaling pathways. Like a domino effect, each
reaction causes another to occur. The cascade of events in the Akt pathway leads to
the insulin induced uptake of glucose into cells and this lowers blood glucose
concentrations. Figure 1.4.1 illustrates the Akt signaling pathway.
Autophosphorylated insulin receptor tyrosine kinase will phosphorylate other
tyrosine-containing proteins. The primary substrate phosphorylated by the insulin
receptor is insulin receptor substrate 1 (IRS-1). Once phosphorylated, IRS-1 can
serve as a binding site for the p85 subunit of 3-phosphoinositide-dependent (PI3)
kinase. This binding stimulates the dimerization of this regulatory subunit (p85) with
the catalytic pi 10 subunit of PI3 kinase.
Once p85 binds pi 10, the activated PI3 kinase phosphorylates the membrane
lipid phosphatidylinositol 4,5 bisphosphate (PIP2), producing phosphatidylinositol
3,4,5 trisphosphate (PIP3). PIP3 serves as a binding site for proteins containing a
plextrin homology (PH) domain. PIP3 dependent kinase 1 (PDK1) and Akt are both
serine/threonine kinases that contain PH domains. As its name implies, PDK1 is
activated when bound to PIP3. Activated PDK1 will phosphorylate and activate Akt.
Activated Akt causes the movement of Glut 4 from internal stores to the plasma
membrane. Acting at the plasma membrane, glut4 allows the movement of glucose
8


into the cell. Recent evidence has suggested that Akt activation is necessary and
sufficient for insulin induced glucose transport to occur (Cohen, 1999).
9


p Kinase
Figure 1.4.1 Insulin/Akt Pathway. Insulin signaling stimulates glucose transport
via the Akt pathway. The autophosphorylated insulin receptor phosphorylates IRS-1.
IRS-1 activates PI3 kinase via binding. PI3 kinase produces PIP3, allowing for
PDK1 activation. PDK1 activates Akt, and Akt stimulates glut 4 translocation by an
unknown mechanism.
10


1.5 Protein Tyrosine Phosphatase IB
Protein tyrosine phosphatases (PTPases) are a family of enzymes that regulate
the insulin signaling pathway. By removing one or more regulatory phosphates from
the insulin receptor and IRS-1, they stop the domino effect within the cell (Figure
1.5.1); that is, PTPases turn off insulin's signaling cascade.
Protein tyrosine phosphatase IB (PTP-1B) has been implicated as the primary
PTPase that turns off the insulin receptor (Salmeen et al., 2000, Elchelby et al., 1999,
Cheung et al., 1999).
PTP-1B is known to have at least two regulatory phosphorylation sites. IRK
phosphorylates PTP-1B on a regulatory tyrosine, and inactivates phosphatase activity.
This will ensure continued IRK activity (Tao et al., 2001a).
In contrast, the cyclic AMP dependent kinase (PKA) phosphorylates PTP-1B
on a regulatory serine residue. When phosphorylated on this regulatory serine,
phosphatase activity is enhanced (Tao et al, 2001b). PKA also phosphorylates PTP-
1B on a threonine residue that appears to show no regulatory activity (Tao et al.,
2001b).
In addition to inhibiting the insulin signal, PTP-1B has also been found to
inhibit leptin signaling (Zabolotny et al., 2002). Leptin is a hormone that induces
satiety (feeling full). PTP-1B knock-out mice (where leptin signaling would be
activated to higher than normal levels) are lean (Zabolotny et al., 2002), while
11


overproduction of PTP-1B is associated with obesity.
12


ON AUTOPHOSPHORYLATED
OFF INSULIN
INSULIN RECEPTOR RECEPTOR
Figure 1.5.1 PTP-1B Regulates Insulin Receptor Activity. The deactivation of the
insulin receptor by a tyrosine phosphatase. PTP-1B removes phosphate from the
insulin receptor regulatory domain, thereby decreasing insulin receptor kinase
activity.
13


1.6 Metformin
Since its April 1995 introduction to the United States, glucophage
(metformin) has become the most-prescribed diabetes pill in the nation. For a drug
that has effectively lowered the blood glucose concentrations of European and
Canadian diabetics since 1958, metformin's recent popularity should not be
surprising. Despite the drug's long, proven history, its biochemical mechanisms of
action still have not been elucidated.
Previous studies suggest that metformin doesnt increase insulin levels or
insulin binding to the insulin receptor (Prager and Schemthaner, 1983, Vigneri et al.,
1981), therefore it must be stimulating the insulin pathway via a post-binding site
modification.
Metformin increases insulin receptor activity (Stith et al., 1996, 1998). In
plasma membrane preparations isolated from whole cells, therapeutic concentrations
(10 |ig/ml) of the drug lead to a 300% increase in tyrosine kinase activity (Stith et al.,
1996). Later, Stith et al. (1998) reported that metformin induces a 25% increase in
purified insulin receptor kinase activity with a maximum stimulation at 1.0 pg/ml
metformin. This is a huge discrepancy! Metformins stimulation of tyrosine kinases
in the PMC is 12 fold greater than with isolated insulin receptor tyrosine kinase.
Stith et al. (1996) suggested that metformin must enter the cell and then acts at
or near the plasma membrane. This suggestion was based on the finding that
14


metformin acts more rapidly in isolated plasma membrane-cortices (PMCs) than with
whole cells. There was a time delay of approximately 90 minutes for maximal
metformin action in whole cells. This delay may be due to metformin entering the
cell and building up to effective levels. Since metformin can immediately stimulate
the insulin receptor in PMCs, metformin must enter the cell but act at or near the
membrane.
It has also been shown that metformin increases the phosphotyrosine content
of IRK (Stith et al., 1996) and IRS-1 (Stith et al., 1996, Pryor et al., 2000). The
increase in phosphotyrosine on the insulin receptor suggests that metformin may be
regulating the insulin receptor via altered phosphorylation of its regulatory domain.
Pryor et al. (2000) have shown that metformin stimulates the insulin signaling
pathway shown in Figure 1.4.1. Metformin stimulates PI3 kinase and Akt activities
and increases translocation of the glucose transporter Glut 4 to the plasma membrane
in insulin-resistant adipocytes. Insulin was used to induce an insulin-resistant state in
the adipocytes.
One major question remains about metformins ability to stimulate IRK
activity: How does metformin stimulate IRK activity? Our theory is that metformin
must enter the cell then inhibit PTP-1B which results in activation of the insulin
receptor.
15


1.7 The Xenopus Cell Model
Many factors make the Xenopus laevis oocytes an excellent cell model for our
studies. First, insulin signaling pathways have been well-characterized and well-
studied in Xenopus. Second, oocytes are easily maintained and attainable year-round.
Additionally, oocytes are very large cells (approximately 1.2 mm in diameter) that
can be manipulated very easily. For example, we can manually isolate the plasma
membrane-cortex (PMC). The ability to isolate a membrane is very unusual, but the
Xenopus oocyte has an unusual plasma membrane that is highly reinforced.
16


2. Materials and Methods
2.1 Xenopus Maintenance
Wild type Xenopus laevis, the African clawed frog, were obtained from
Xenopus Express (Ft. Lauderdale, Florida). Each female frog was fed 2 grams of
ground beef heart three every other day. To prevent infections, food was withheld for
two days after arrival at the University of Colorado at Denver. Frogs were kept in 30
gallon plastic tanks. The water was replaced with 15 gallons of fresh water about 4
hours after feeding. 10 ml of NovAqua (Novatek, Haywood, CA) was added to the
tank after refilling with fresh, cold tapwater. NovAqua removes CO2 and prevents
infection.
2.2 Xenopus Oocytes
Xenopus oocytes were acquired from frogs primed 3-10 days before
experiments with 50 IU Pregnant mare serum gonadotropin (PMSG) (Sigma, St.
Louis, MO). Growth of the ovarian follicles and the corpus luteum was stimulated by
priming with PMSG. This decreased the time required for oocytes to respond to
insulin (Stith et al., 1985). The animals were placed in ice water for 45 minutes and
then sacrificed with a rat guillotine. Ovaries were removed, blotted to remove excess
blood and placed in room temperature OR2 solution (83 mM NaCl, 0.5 mM CaC12, 1
mM MgC12, 10 mM HEPES). The ovaries were then sectioned into smaller segments
17


and placed in fresh 0R2 solution. Healthy, stage VI oocytes (1.1 mm or greater)
were manually isolated from follicles and maintained in OR2.
2.3 Plasma Membrane-Cortices
Plasma membrane-cortices (PMCs) are manually isolated from Xenopus
oocytes as previously described by Sadler and Mailer (1981). Each PMC is about 10-
50 nra thick and contains approximately 1 pg of total protein. PMCs contain the
plasma membrane, integral membrane proteins, peripheral membrane proteins, few
pigment granules, and the cortex (the outer 5 microns or more of the cytoplasm). All
proteins present in the PMC may retain normal function as insulin can stimulate the
insulin receptor and enzymes down-stream of the insulin receptor (Stith et al., 1996).
Before preparing PMCs, petri dishes (5 cm) were filled with Sadlers isolation
buffer (SIB) (10 mM NaCl, 10 mM HEPES, pH 7.2) and placed on a bed of crushed
iced for 30 minutes. Groups of 15 defolliculated oocytes were placed in the ice-cold
SIB. Oocytes were opened by piercing the cells with closed forceps and allowing the
forceps to open. The tom cells were then rotated, plasma membrane side up and
flattened against the bottom of the petri dish by gently depressing with forceps. The
PMC preparations were then stored on ice for 45 minutes to allow the yolk and
intracellular material to dissociate from the plasma membrane. PMCs were then
teased away from the cytosol and transferred to a 1.7 ml microfuge tube using a wide-
18


mouthed transfer pipette. The preparation can be washed more thoroughly. The more
extensively washed PMCs are clearer as more pigment granules and cytoplasm are
removed.
2.4 Analysis of Endogenous Tyrosine Phosphatase Activity
To examine the effect of metformin on phosphatase activity a phosphatase
assay in PMCs kit (kit 2, Upstate biotechnology, Lake Placid, NY) was used. These
experiments were conducted by a graduate student (Thomas Morrison) with my
assistance. The kit measures the release of phosphate from a tyrosine phosphopeptide
(T-S-T-E-P-Q-pY-Q-P-G-E-N-L). The phosphate release is corrected to determine
the moles of phosphate released. To measure phosphate the assay utilizes the fact
that the free phosphate binds to molecules of malachite green. When bound to
phosphate, malachite exhibits increased absorbance at 650 nm. Absorbance was
measured with a Milton Roy Spectronic 1001 spectrophotometer set at 650 nm.
After isolating PMCs in groups of 15, the microfuge tubes were flash spun
by holding the spin button on the Beckman microfuge for about 1 second to pellet
the PMCs. With one treatment group, the supernatant was removed and replaced
with lOOul SIB with 150 mM NaCl for 5 minutes. This is the salt-washed group. The
PMCs were then flash-spun again. The supernatant (containing peripheral membrane
proteins removed from the PMC) was removed from all groups and PMCs were
19


resuspended in 80 ul SIB. Phosphopeptide (20 ul of ImM) was added, samples were
gently vortexed and incubated at 15 C for 15 minutes. 30 seconds before the time-
point, samples were flash-spun and 20 pL of supernatant was removed and placed in
a separate 0.5 ml microfuge tube. The 20 pL of supernatant was analyzed for free
phosphate and the remaining sample was treated with either 2 pL of deionized water
(dH20) or 2 pL of 40 pg/ml metformin (10 pg/ml). The samples were vortexed to
resuspend the PMCs and incubated for another 15 minutes at 15 C. Then, the
samples were spun again and 20 pL of supernatant was removed for phosphate
analysis. Figure 2.4.1 summarizes the sample preparation.
20


Figure 2.4.1 Sample Preparation for Tyrosine Phosphatase Assay. 4 samples
were prepared. 1) samples of 15 untreated PMCs were incubated 15 minutes with
200 pM phosphopeptide. 2) samples of 15 untreated PMCs were incubated an
additional 15 minutes (30 minutes total) with PMCs. 3) samples of 15 salt washed
PMCs were treated with 10 pg/ml metformin after 15 minutes of phosphate release
and incubated for an additional 15 minutes (30 minutes total). 4) samples of 15 PMCs
(not salt washed) were treated with 10 pg/ml metformin after 15 minutes of
phosphate release and incubated for an additional 15 minutes (30 minutes total).
21


2.5 Analysis of PTP-1B Activity In Vitro
Purified human recombinant PTP-1B was purchased from Upstate
Biotechnology (Lake Placid, NY). The effect of metformin on PTP-1B activity was
analyzed using a modification of the procedure suggested by the manufacturer of the
recombinant PTP-1B. This measurement does not require the difficult assay of
phosphate but records the amount of product (para-nitrophenol; pNP) by
spectrophotometric means. PTP-1B removes the phosphate from para-nitrophenol
phosphate (pNPP) (Upstate Biotechnology 20-106) in this non-specific phosphatase
assay, and this produces pNP. pNP, the product of the dephosphorylated pNPP,
absorbs strongly at 405 nm after phosphate is cleaved.
PTP-1B activity was assayed in a mixture of PTP-1B Assay Buffer (25 mM
HEPES, 50 mM NaCl, 5 mM dithiothreitol, 2.5 mM ethylene-diamine-tetraacetic acid
(EDTA), pH 7.2) and bovine serum albumin (BSA) solution (100 pg/ml in assay
buffer). Before preparing samples, 2 pL of PTP-1B stock was diluted with 198 pL of
BSA solution. To prepare each sample, 90 pL BSA solution, 60 pL Assay buffer, 10
pL dilute enzyme, 20pL 100 pg/ml metformin or deionized water, and 20 pL 50 mM
pNPP in PTP-1B assay buffer were combined. Samples were allowed to incubate
with metformin for 20 minutes prior to pNPP addition. This incubation was
performed to allow maximal interaction of metformin and PTP-1B. Each sample
contained 0.075 pg PTP-1B and a final concentration of 5 mM pNPP. After pNPP
22


addition, samples were then incubated for 10 minutes at 37 before stopping with 100
pL of 2M K2CO3. The final volume was raised to 1000 pL with dH20 and
absorbance was measured at 405 nm.
PTP-1B activity was calculated using the equation:
PTP-1B activity= [0.001L (A^s/e cm)]/ T/ pg.
In this equation, 0.001 L is the final volume (1 mL), A405 is the absorbance of
PTP-1B at 405 nm (absorbance is 0.015 for 0.05 pg of PTP-1B), e is the extinction
coefficient for pNP (1.78 X 10A4 /M cm), cm is the pathlength of light (1.0 cm), T
is the time for this assay (10 minutes), and pg PTP-1B per sample (0.75 pg).
2.6 Analysis of PTP-1B Activity with PMCs
PTP-1B activity was also assayed in the presence of PMCs. PTP-1B activity
measurement and PMC isolation were procedures described in sections 2.3 and 2.4
Calculations were performed to demonstrate that the phosphatase activity from PMCs
was negligible in comparison to the exogenous recombinant PTP-1B activity.
PTP-1B activity was assayed in a mixture of PTP-1B Assay Buffer (25 mM
HEPES, 50 mM NaCl, 5 mM dithiothreitol, 2.5 mM EDTA, pH 7.2) and bovine
serum albumin (BSA) solution (100 pg/ml in assay buffer). Samples were prepared
by aliquotting from prepared enzyme cocktails. This was done to help reduce the
23


between samples error due to pipetting. PTP-1B (4 pL) enzyme was diluted into
1596 pL of BSA solution and 960 pL of PTP-1B assay buffer was added to the
diluted enzyme. To prepare final enzyme cocktails, 640 pL of this dilute enzyme
mixture was then aliquoted into 2 microfuge tubes and treated with 80 pL of 100
pg/ml metformin or 80 pL of dH20 for control samples. Metformin was allowed to
incubate for 20 minutes with the enzyme solution (5 C). PMCs were flash-spun and
170pL of the enzyme cocktail was removed. Samples were then incubated for 20
minutes on ice and warmed to 37 for 5 minutes. Reactions were started by the
addition of 20 pL 50 mM pNPP. Each sample contained 0.1725 pg PTP-1B, 20
PMCs and a final concentration of 5 mM pNPP. Samples were incubated for 15
minutes at 37 before stopping with 100 pL of 2M K2CO3. The final volume was
raised to 1000 pL with dH20 and absorbance was measured at 405 nm. Enzyme
activity was calculated as before.
2.7 Development of a New HPLC Peptide Analysis for
Insulin Receptor Autophosphorylation
To examine the phosphorylation sites on the insulin receptor, many
'X'y
researchers use radioactive P to label various phosphotyrosines on the intact insulin
receptor. Then the intact insulin receptor is cut up into peptide fragments, and the
24


peptide fragments (some of which contain the regulatory phosphorylated amino acids)
are separated by high-performance liquid chromatography (HPLC). While we have
not yet used an intact insulin receptor for our studies, we have set up a peptide
separation method through modification of an HPLC method by Morris White (1990).
In other experiments, a peptide representing the insulin receptor regulatory
domain was added to PMCs. The use of this peptide would enable in-vitro analysis of
the phosphorylation content of the insulin receptor regulatory domain. Analysis of
insulin receptor phosphopeptides required that we set up a high performance liquid
chromatography (HPLC) procedure that would be able to separate the peptides.
High-pressure liquid chromatography (HPLC) separates molecules (in this
case phosphopeptides) according to subtle differences in their solubility
characteristics. All HPLC analysis was performed using a Rainin/Varian HPXL
model HPLC system consisting of two isocratic pumps, a high pressure mixer, UV
detector and Macintosh control interface. The UV detector was set to 270 nm (the
optimum wavelength for the detection of tyrosine residues).
The HPLC method uses a flow rate of 1 ml/minute. A Supelco LC-318
column provided efficient separation of all phosphopeptides of interest. A solvent
gradient consisting of water and acetonitrile, both acidified with 0.05%TFA, is
ramped from 5% acetonitrile to 19.1% acetonitrile over 40 minutes. The gradient
25


then ramps to 25% acetonitrile in 5 minutes then returns to 5% acetonitrile by the end
of the 50-minute run.
The HPLC method was characterized using insulin receptor phosphopeptides
(IRO, IR5, IR9, IR 10, IR 5,9,10) from Biomol (Plymouth Meeting, Pennsylvania).
Table 2.7.1 indicates the phosphorylation state of each peptide. The elution time for
each phosphorylated peptide was determined by running the standards individually
and in combination.
26


Table 2.7.1 A Summary of Regulatory Peptide Nomenclature and Elution Time.
The Biomol regulatory peptides consist of the insulin receptor residues 1142-1153.
This sequence of amino acids is known to regulate the insulin receptor based on its
phosphorylation state. The elution order for the 3 bisphosphorylated peaks has not
yet been determined, but they elute around 21-25 minutes.
Phosphorylated Tyrosine (amino acid number in the human insulin receptor) Biomol Name HPLC Elution Time (min)
None IRO 36
Monophosphorylated:
1146 IR5 32.7
1150 IR9 34.6
1151 IR10 33.7
Bisphosphorylated: 21-25
1146, 1150 IR 5,9
1146, 1151 IR 5, 10
1150,1151 IR 9, 10
Trisphosphorylated:
1146,1150, 1151 IR5, 9, 10 19
27


I would like to describe the HPLC methodology in more general terms. For
example, we note that the insulin receptor peptide with all three phosphates elutes at
19 min on our HPLC gradient. However, when the peptide is incubated with tyrosine
phosphatases (either from the Xenopus PMC or human PTP-1B itself),
dephosphorylation takes place (e.g., phosphate on sites 1146, or 1150 or 1151 is
removed from the peptide). Thus, we see new peaks appearing on the HPLC recorder
(occurring at 33-36 minutes); these new peaks include the peptide with only one
phosphate present. By comparing the size of peak areas versus time, we can
determine which phosphate is removed first, and also follow the rate at which
phosphate is removed from the specific regulatory sites on the insulin receptor
peptide. Perhaps metformin will inhibit the removal of phosphates located on a
certain tyrosine from the insulin receptor peptide.
We can also examine the action of metformin upon human PTP-1B. Our
model predicts that metformin, acting through an intermediate present in our plasma
membrane preparation, will inhibit PTP-1B. While the PMC contains endogenous
tyrosine phosphatases, the endogenous levels of tyrosine phosphatase activity are
relatively low, and we can add human PTP-1B to flood out the endogenous tyrosine
phosphatase activity. Thus, we can record the activity of a human tyrosine kinase
(PTP-1B) and measure the effects of metformin on both PTP-1B and the level of
phosphorylation of the human insulin receptor regulatory peptide.
28


Conversely, we can add the insulin receptor peptide without any phosphates
on it to PMCs, and record phosphate addition to the peptide. Perhaps metformin will
stimulate the addition of the phosphates to the insulin receptor peptide.
With the insulin receptor regulatory domains and PMCs, we can quantify
phosphatase and kinase activities simultaneously (as we only record the final sum
total of both kinase and phosphatase action).
2.8 Analysis of the Phosphorylation State of Insulin
Receptor Regulatory Domain Peptides with plRK
The effect of metformin on in-vitro (3 IRK (the commercially-available
purified beta subunit of the insulin receptor) activity was assayed using the new
HPLC methodology. Samples were prepared by diluting 1 pg of |3IRK into 68 pL of
(3IRK assay buffer with or without 10 pg/ml metformin. The reaction was started by
the addition of 32 pg IRO and carried out for 20 minutes at 37. Reactions were
stopped via the addition of 100 pL 10% trichloro-acetic acid (TCA). 150 pL (75%)
of each sample was injected onto the HPLC for analysis of kinase activity.
29


2.9 Analysis of the Phosphorylation State of Insulin
Receptor Regulatory Domain Peptides with PMCs
The effect of metformin on kinase and phosphatase activity contained in the
PMCs was assayed using the new HPLC method. Samples were prepared by adding
84 pL of pIRK assay buffer to 15 PMCs. Reactions were started by the addition of
16 pL (16 pg) IR9 phosphopeptide. The reaction proceeded for 40 minutes at room
temperature before stopping it with 100 pL 10% TCA. 75% (150pL) of each sample
was injected onto the HPLC for analysis.
30


3. Results
3.1 Metformin Inhibited Tyrosine Phosphatase
Activity in PMCs
With the malachite green assay for tyrosine phosphatase and our PMC
preparations, metformin (10 pg/ml) significantly decreased phosphatase activity
(Figure 3.1.1). Metformin was added after a 15 minute control period. As opposed to
a control rate of removal of phosphate from tyrosine located on a phosphopeptide (see
line labeled C), metformin strongly inhibited the rate of removal of phosphate. The
use of PMCs enables us to reduce the amount of background noise (error caused by
interference from cytoplasmic kinases and phosphatases) by emphasizing
dephosphorylation taking place at the membrane. This is where the insulin receptor is
located (as opposed to dephosphorylation events taking place deeper in the
cytoplasm).
For these tyrosine phosphatase assays, all groups of 15 PMCs were incubated
with a phosphopeptide substrate. After 15 minutes, a sample was taken from each
sample and the amount of free phosphate was measured. The experimental samples
were allowed to incubate another 15 minutes with or without metformin addition. In
Figure 3.1.1, phosphate release (a measure of tyrosine phosphatase activity) is
strongly inhibited by metformin: compare control release (C) with that in the
31


presence of metformin (M). These data lend support to our idea that metformin
acts by inhibiting a tyrosine phosphatase.
Salt washing of the PMCs prior to treatment with metformin (M+W)
nullified the inhibiting effects of metformin. The tyrosine phosphatase activity in
washed samples did not significantly differ from unwashed controls, indicating that
tyrosine phosphatases were not removed by the salt wash. The results of the salt
wash indicate that an intermediate, necessary for the effects of metformin, was
removed by the salt wash. So, metformin may be inhibiting PTP-1B indirectly.
32


Figure 3.1.1 Metformin Inhibited Tyrosine Phosphatase Activity in PMCs.
Metformin (M) inhibited phosphatase activity in plasma membrane-cortex
preparations. Metformin did not inhibit PTPase activity if the PMC was first washed
with 150 mM NaCl first (W+M). Compare the release of phosphate from a tyrosine
phosphatase peptide substrate in the control group C with a 10 pg/ml metformin
group M. When PMCs were washed with high salt to remove peripheral proteins,
metformin was unable to inhibit the phosphatase activity (W+M). For each group,
N=9. An asterisk represents P less than 0.05.
>-
I-
>
h-
O
<
UJ
%
<
CL
if)
o
X
0.
UJ
>-
I-
Q
HI
if)
<
UJ
UJ

UJ
H-
<
CL
if)
O
CL
33


3.2 Metformin Did Not Inhibit PTP-1B Activity In Vitro
We next wanted to see if metformin would directly inhibit human PTP-1B (a
commercially-available tyrosine phosphatase known to inhibit the insulin receptor).
Our prior work (Figure 3.1.1) suggested that metformin acts indirectly. We combined
various concentrations of metformin with purified PTP-1B. We found that metformin
did not inhibit the in vitro PTP-1B activity (Fig. 3.2.1).
These data suggest that metformin does not act upon PTP-1B directly, and this
supports the idea that metformin acts through an intermediate to inhibit PTP-1B. The
intermediate would only be present in our plasma membrane-cortex preparation
(unwashed PMCs) or the intact cell.
34


(ug/ml)
Figure 3.2.1 Metformin Did Not Directly Inhibit PTP-1B In Vitro. Metformin
was unable to inhibit the activity of human recombinant PTP-1B in an in-vitro assay.
Phosphate removal from the enzyme substrate pNPP was measured. Human PTP-1B
activity was measured with pNPP (a phosphatase substrate). The number of
determinations for each sample is 12. Groups were not significantly different.
35


3.3 PMCs Allowed Metformin Inhibition of PTP-1B
We then determined whether metformin would inhibit human PTP-1B when
plasma membrane-cortex was present. The PMC may provide the unknown
intermediate that would enable metformin to inhibit PTP-1B.
Whereas metformin did not directly inhibit PTP-1B (Figure 3.2.1), after
PMCs were added (Figure 3.3.1) metformin inhibited PTP-1B by 50% (P less than
0.01). Metformin was effective at concentrations ranging from 1 to 100 pg/ml (10
pg/ml is the therapeutic concentration) The conclusion of these two experiments is
that metformin acts through an intermediate in order to inhibit PTP-1B.
36


30 -w
25
&
:e m
oa S
' O
t s
£ 3
20
10 -
5 -
Pi
lCs';^S- :
0
' /i*
BIS
0.1
Denotes p < 0.005
* *
X.
1.0
10
T ili? X

' *** **T*I . MW- ^ V
Metformin Concentration
(ug/ml)
Figure 3.3.1 PMCs Rescued Metformin Inhibition of PTP-1B. The effect of
metformin on Human PTP1-B was assayed in the presence of PMCs. Fifteen PMCs
were added to human recombinant PTP-1B, and the phosphatase activity was
measured using pNPP. Compare to the lack of effect of metformin on PTP-1B shown
in 3.2.1. For each sample N=12. An asterisk denotes significance P less than 0.005.
37


3.4 Use of a Peptide Containing the Human Insulin
Receptor Regulatory Domain
One additional line of evidence that would support our theory that metformin
acts through inhibition of a tyrosine phosphatase would be to describe any metformin-
induced changes in individual phosphorylation sites on the regulatory domain of the
insulin receptor. Our model predicts that, in the presence of metformin, certain
regulatory phosphorylation sites on the insulin receptor would show higher levels of
phosphorylation (and this would result in the measured higher insulin receptor
activity). Thus, we wanted to switch from artificial substrates, such as pNPP which
was used in Figure 3.1.1, to a peptide that contains the insulin receptor regulatory
domain.
The insulin receptor is regulated through phosphorylation of tyrosines located
in a domain of the receptor from about amino acid 1142 to amino acid 1153. In the
experiments described below, we use a peptide derived from the insulin receptor that
contains these residues (amino acid sequence: 1142-1153: TRDIYETDYYRK).
Under conditions where tyrosines in the insulin receptor regulatory domain
are phosphorylated, the insulin receptor would be stimulated. We have purchased this
peptide with and without various phosphates already attached to tyrosines. One
peptide lacks phosphate on any tyrosine on the peptide, and we purchased other
peptides with one phosphate on the peptide, and another with three phosphates on the
38


peptide. The phosphorylation can occur on tyrosine 1146, or tyrosine 1150 or on
tyrosine 1151 (or on all three) of the peptide. Experiments with these peptides a new
high performance liquid chromatography (HPLC) procedure that separates the
different phosphopeptides.
To examine the phosphorylation sites on the insulin receptor, many
researchers use radioactive 32P to label various tyrosines on certain amino acids of the
insulin receptor. Then the insulin receptor is cut up into peptide fragments, and the
peptide fragments (some of which contain the regulatory amino acids) are separated
by HPLC. While we have not yet used an intact insulin receptor for our studies, we
have set up just such a peptide separation method through modification of an HPLC
method by Morris White (1990).
Figure 3.4.1 illustrates a typical UV trace from our insulin receptor peptide
separation method. With our HPLC method, we can separate the peptides that have
no phosphate from those that have one or more phosphates; we can even separate the
peptides based on which tyrosine has the phosphate. Thus, we can examine whether
metformin increases or decreases the level of the phosphate on the insulin receptor
peptide. Our model would predict that metformin would increase the level of
phosphorylation on certain regulatory sites known to stimulate the insulin receptor.
39


Thus, we have developed a new method that should be of great interest to
diabetes researchers. The advantages of this system are that we can examine:
the human insulin receptor kinase activity
the phosphorylation of specific regulatory sites on a human insulin receptor
peptide or on the proteolyzed insulin receptor
the action of metformin on receptor tyrosine kinase and tyrosine phosphatase
activity
metformin and insulin action at the plasma membrane (through use of the
plasma membrane-cortex or PMC preparation)
combined action of both tyrosine kinase and tyrosine phosphatase regulation
on the insulin receptor peptide (through the total phosphorylation changes of
receptor peptide)
human tyrosine phosphatase PTP-1B, human insulin receptor (the
intracellular portion of the human insulin receptor or BIRK), and
metformin action.
40


Figure 3.4.1 HPLC Separation of Regulatory Peptide. Unphosphorylated
regulatory domain peptide (32 pg) was incubated with piRK for 40 minutes. The
reaction was stopped with trichloro-acetic acid and the resulting phosphopeptides
were separated via HPLC using our new method. Four groups of regulatory domains
(circled below) were separated by HPLC and quantified by UV detection of tyrosine.
(See also Table 2.7.1)
1. The trisphosphorylated regulatory domain peptide
2. The bisphosphorylated regulatory domain peptides
3. The monophosphorylated regulatory domain peptides
4. The nonphosphorylated regulatory domain peptide
3 4
ELUTION TIME
41


3.5 Metformin Stimulated plRK Activity In Vitro
We have noted that metformin directly stimulates in-vitro piRK activity (Stith
et al., 1998). In these earlier metformin experiments, radioactive ATP was added to
purified piRK and radioactivity on a peptide substrate was recorded. We wanted to
confirm these results with the insulin receptor regulatory domain peptide and our new
HPLC method. We have now confirmed that metformin does have a small (11%)
stimulatory effect on the insulin receptor (Figure 3.5.1). We combined:
the intracellular portion of the human insulin receptor (BIRK)
metformin
IRO insulin receptor peptide without phosphate
We added all constituents under optimal conditions (it took us weeks to
determine these conditions; we altered enzyme amounts, substrate, time course, etc.).
We then determined that BIRK was able to phosphorylate the insulin receptor peptide
without phosphate (refer to section 3.6). That is, piRK catalyzes IRO
(unphosphorylated peptide) phosphorylation to IR 9 (one phosphate at the 1150
position).
Metformin was able to stimulate the in vitro phosphorylation of IRO (Figure
3.5.1) however, like the prior in vitro studies of Stith et al. (1998), this stimulation
was not large (only 11 %). With our PMCs or whole cells, we achieved a much larger
42


stimulation (300%) of the tyrosine kinase activity (Stith et al., 1996). As is discussed
elsewhere, we believe that this larger stimulation may be through indirect inhibition
of PTP-1B, and the PMC supplies an inhibitory intermediate between metformin and
the tyrosine phosphatase. This difference in stimulation may be due to the use of
different substrates that would reflect different tyrosine kinases being measured.
43


8.0
a>
3
a
V
a
o
d
a
w
0
JB
a
ON
a
&B
3
7.5 -
7.0 -
6.5
6.0
5.5
5.0
T
2
3
0
u
*
ff
a
Figure 3.5.1 Metformin Stimulated plRK Activity In Vitro. 0.5 |ig piRK was
incubated with 200 [iM IRO for 20 minutes at 37 with or without 10 fig/ml
metformin. The asterisk denotes P less than 0.05. N=6
44


3.6 The Human Insulin Receptor Kinase Preferentially
Added Phosphate to the 1150 Tyrosine
We next wanted to examine the purified beta subunit of the human insulin
receptor in our new assay. As noted in Figure 3.6.1, we determined that [3IRK prefers
to add phosphate to tyrosine 1150. That is, the purified beta subunit of the insulin
receptor preferred to add phosphate to the 1150 (not the 1146 or 1151) position. Next
PIRK preferred to add phosphate to 1151 then to 1146.
In these experiments, the regulatory peptide without any phosphates on it
(IRO) was added to the human insulin receptor (BIRK) and this produced a new peak
on our HPLC gradient corresponding to IR9 (the peptide with tyrosine 1150
phosphorylated). Relative to phosphorylation of tyrosine 1150, the insulin receptor
adds less phosphate to tyrosine 1151 and then, less yet to tyrosine 1146. It is well
known (White, 1990) that the human insulin receptor autophosphorylates tyrosine
1150 (and other sites) and that this stimulates the receptor kinase activity.
Does metformin change the ability of the human insulin receptor to
phosphorylate tyrosine 1150? In other words, does metformin change the preference
of the human insulin receptor to phosphorylate this 1150 to 1146? Our data (Figure
3.6.1) demonstrate that metformin does not change the preference of the insulin
receptor. Thus, we do not suggest that metformin alters the substrate specificity for
the human insulin receptor.
45


Figure 3.6.1 |1IRK Preferentially Phosphorylated Tyrosine 1150 of the
Regulatory Domain. Metformin had no significant effect on the site specificity of
(3IRK. For each sample, N=4.
Control
Metformin
46


3.7 Metformin Stimulated Tyrosine Kinase Activity in PMCs
We also examined whether metformin could stimulate the phosphorylation of
the insulin receptor peptide in a system containing both tyrosine phosphatase and
tyrosine kinase activities (the PMC preparation). Thus, we switched from piRK to
PMCs for these experiments. As previously noted, increased levels of
phosphorylation of the regulatory peptide would be associated with increased insulin
receptor activity in an intact cell.
To be able to simultaneously record both tyrosine kinase and phosphatase
activities, we used the peptide with one phosphate at the 1150 position (IR9). After
incubation with 15 PMCs, we examined the appearance of peaks representing
peptides with two phosphates (kinase action; this section) or without phosphate
(phosphatase action; next section).
In the absence of metformin, we noted that phosphate was added to the IR9
peptide at a rate of 0.18 pg/40 min/15 PMCs (see control bar in Figure 3.7.1).
Metformin significantly increased the rate of a second phosphorylation of this insulin
receptor peptide by 31%. So, metformin stimulates tyrosine kinase activity that
phosphorylates the first tyrosine (1150) by 11% (Figure 3.5.1) and metformin
stimulates the addition of a second phosphate to the regulatory peptide by 31%
(Figure 3.7.1).
47


Figure 3.7.1 Metformin Stimulated Tyrosine Kinase Activity in PMCs. Fifteen
PMCs were incubated 40 minutes with 200 jiM IR9 phosphopeptide. Peak areas of
all three bisphosphorylated peptides were combined. N=6. An asterisk indicates P
less than 0.05.
48


3.8 Metformin Inhibited Tyrosine Phosphatase Activity that
Removes Phosphate from the Insulin Regulatory Peptide
Our sample analysis of the last set of experiments (Figure 3.7.1) also recorded
the amount of IR9 was dephosphorylated to IRO. PMCs and insulin receptor
phosphopeptide (IR9) were added with or without 10 pg/ml metformin. With time,
the insulin receptor peptide without any phosphate (IRO) increased in amount (Figure
3.8.1).
Metformin inhibited total tyrosine phosphatase activity in PMCs by 100% in
our previous experiment (Figure 3.1.1). Total tyrosine phosphatase activity would
include those phosphatases that act upon the insulin receptor and others presumably
not related to insulin signaling. As noted in Figure 3.3.1, metformin in the presence
of PMCs inhibited human PTP1B (a known regulator of the insulin receptor) by
approximately 50%. In this new experiment, metformin inhibited the appearance of
the dephosphorylated form of the peptide that would be associated with an inactive
insulin receptor (see bar on the right side of the Figure 3.8.1). In Figure 3.8.1 (next
page), metformin inhibited (by 50%) a tyrosine phosphatase, located in the PMC, that
specifically removes phosphate from the human insulin receptor regulatory domain.
This inhibition of the removal of phosphate from the receptor regulatory domain
would stimulate the insulin receptor kinase activity (and thus, fight diabetes).
49


These last results are strong support for our model of action for metformin
since it utilizes the human insulin receptor peptide.
50


Figure 3.8.1 Metformin Inhibited Tyrosine Phosphatase Activity Toward the
Regulatory Peptide. In these experiments, the monophosphorylated IR 9 was added
to 15 PMCs (containing endogenous tyrosine kinases and tyrosine phosphatases).
After 40 minutes, the peptide (IRO) in the supernatant was measured. N=6, asterisk
denotes P less than 0.05.
51


4. Discussion
The work of Stith et al. (1996) reported that metformin may fight diabetes
via its ability to stimulate insulin receptor tyrosine kinase activity. It was also
suggested that metformin stimulates receptor tyrosine kinase activity in response to
an increase in phosphorylation on the regulatory domain of the receptor (Stith et al.,
1996).
In this present work, we have shown that metformin inhibits a tyrosine
phosphatase (Figure 3.8.1) that removes phosphates from a peptide corresponding to
the regulatory domain of the human insulin receptor. Thus, we have evidence that
metformin fights diabetes (and stimulates insulin receptor tyrosine kinase) through
the inhibition of tyrosine phosphatase activity.
Although very different methods were used, the in-vitro stimulation of piRK
by metformin (Figure 3.5.1) (15%) was similar to that previously published (25%)
(Stith et al., 1998). However, the increase in tyrosine kinase activity (31%) in PMCs
(Figure 3.7.1) was vastly lower them the 300% increase in PMC tyrosine kinase
activity reported by Stith et al. (1996). This difference is not yet explained; however,
the fact that different substrates (regulatory peptide IRO used here: RR-Src used in
Stith et al., 1996) were used may account for the discrepancy.
The insulin receptor peptide should be largely phosphorylated by the insulin
receptor, however many tyrosine kinases will phosphorylate by RR-Src. Since
52


metformin stimulates other tyrosine kinases besides the insulin receptor (Stith et al.,
1998), this may explain why the use of RR-Src recorded a larger metformin
stimulation. That is, the insulin receptor may have a relatively higher affinity for the
regulatory peptide: thus use of the regulatory peptide may quantify insulin receptor
kinase activity as opposed to other tyrosine kinases. Thus, Metformin may stimulate
the insulin receptor by about 30%, but the drug may stimulate another tyrosine kinase
by about 250%. In addition, RR-Src should have many more phosphorylation sites as
compared with regulatory peptides.
Previous work by Chavanieu et al (1992) indicated that the regulatory peptide
IRO can act as a competitive inhibitor of IRK and PTP-1B. High concentrations (2
mM; compared to 200 pM in our assays) of the nonphosphorylated peptide inhibited
insulin receptor autophosphorylation some 70%. The trisphosphorylated peptide
inhibited phosphatase activity toward the receptor. Although the
monophosphorylated form IR9 has not previously been indicated to have effects as a
competitive inhibitor, it is possible (if not likely) that the phosphopeptide is effecting
rates of autophosphorylation of the IRK. Thus, the substrate we used may inhibit the
tyrosine kinase activity and negate some of metformins effects. We will search for
this possible inhibition in future experiments.
Assays of tyrosine phosphatase activity in PMCs revealed a strong inhibition
of phosphatase activity (100%) by metformin (Figure 3.1.1); however this result may
53


be slightly misleading because there may be an ATP regeneration system present in
the PMCs that would remove phosphate. Thus, free phosphate (measured in this
assay) that is produced by the action of a phosphatase may then be recycled by
enzymes present in the PMC and artificially lower the free phosphatase measured.
Thus, inhibition of tyrosine phosphatase by metformin is probably less than 100%.
Our other phosphatase assays did not measure free phosphate, thus free phosphate
removal would not artificially alter the results of the other assays. Two more direct
assays of phosphatase activity using HPLC/UV and pNPP both indicated that
metformin inhibited PTP-1B phosphatase activity by about 50% (Figure 3.3.1).
4.1 Does Metformin Act Directly or Indirectly on PTP-1B?
We have two lines of evidence indicating that metformin inhibits tyrosine
phosphatase activity through an intermediate. First, salt washed PMCs (where the
putative intermediate, but not tyrosine phosphatase, would be removed) are
unresponsive to metformin (Figure 3.1.1). Second, there is a lack of a direct action
by metformin on PTP-1B (Figure 3.2.1), but the addition of PMCs (not salt washed,
thus containing the supposed intermediate) rescues this inhibition (Figure 3.3.1).
Additional, future studies will attempt to identify the intermediate. We will
determine whether PTP-1B is inhibited via dephosphorylation and if metformin can
stimulate this dephosphorylation. We have not yet examined whether metformin can
54


stimulate a phosphatase that removes phosphate from PTP-1B (this would stimulate
PTP-1B activity).
Another set of experiments that we propose is to see if metformin stimulates a
protein that then binds to PTP-1B and this binding would inhibit the tyrosine
phosphatase.
A third possibility would be if the intermediate within the PMCs is a kinase.
That is, metformin would stimulate a protein kinase that would then phosphorylate
and inhibit PTP-1B. However, when we reproduced the conditions shown in Figure
3.3.1 in the presence of radioactive P-ATP, we did not detect any metformin-
'X'J
induced increase in radioactive P on PTP-1B (results not shown). That is,
autoradiographs of polyacrylamide gel-purified PTP-1B indicate that PTP-1B is not
phosphorylated in the presence of metformin. However, PTP-1B was poorly
separated from the bovine serum albumin in the assay buffer (both have molecular
weights of 65 kDa), and these experiments should be repeated without BSA to
decrease the background radiation at the 65 kDa bands.
4.2 Comparison with a Known Tyrosine Phosphatase
Inhibitor Vanadium
We also briefly examined whether metformin inhibits tyrosine phosphatases
in a manner similar to vanadium. Our results (not shown) indicate that Xenopus
55


phosphatases were also inhibited by vanadium compounds like human phosphatases.
However, since our evidence suggests that metformin acts through a mechanism
dependent upon an unknown intermediate found in the PMCs, we believe that the
mechanism used by metformin must be different from that used by vanadium. That
is, as opposed to metformin, vanadium acts directly upon the phosphatase, can inhibit
in-vitro PTP-1B activity and does not require an intermediate for this inhibition.
4.3 Other Future Studies
There are several studies that need to be performed. We would like to analyze
the effect of metformin on (3IRK activity with the monophosphorylated (IR9) peptide,
and on PMC tyrosine kinase activity using IRO. These experiments would serve to
compliment the PMC experiments with IR9 (Figure 3.8.1) and the (3IRK experiments
with IRO (Figure 3.7.1). Addition of (3IRK and IR9 would examine the rate of second
site phosphorylation as a function of metformin.
The future experiment with PMCs and IRO may lend insight. As previously
mentioned, the incubation of PMCs with IR9 resulted in low level metformin
stimulation of tyrosine kinase activity as compared with prior PMC experiments
(using RR-Src as the substrate) (31% versus 300%) (Stith et al., 1996). Depending
upon whether IRO acts as an inhibitor or a better (more phosphorylation possible with
56


IRO than with IR9) substrate, assaying the tyrosine kinase activity in the PMC with
IRO might result in a lesser or greater stimulation by metformin.
In addition, to determine possible inhibitory effect of the receptor peptides on
PMC kinase activity, we would also like to assay insulin stimulated PMCs in the
presence of IR9 or IRO. In these experiments, RR-Src would be used as a substrate
for tyrosine kinase activity, or we might measure insulin receptor phosphorylation
(via Western blot), while IRO and IR9 would serve as potential inhibitors. We will
also microinject IRO or IR9 into whole oocytes to see if they can inhibit insulin-
induced meiosis.
Additionally, we would like to determine the effect of metformin on PTP-1B
activity using the trisphosphorylated form of the receptor peptide. This experiment
would allow us to determine if metformin affects PTP-1B specificity (which sites are
preferred?) as well as effects on overall activity.
We still need to show that metformin increases the phosphorylation state of
the insulin receptor regulatory domain in an intact insulin receptor in a whole cell.
This can be done via two techniques. The first step is to allow autophosphorylation to
occur in the presence of 32P ATP, thus allowing the regulatory domain to be
phosphorylated with radiolabeled phosphate. We would then purify the insulin
receptor via wheat germ affinity, immunoprecipitation or similar techniques. Then,
the receptor must be cleaved with trypsin to form the same regulatory peptides used
57


in our HPLC analyses. The final step is to separate by HPLC and quantify (via liquid
scintillation counting) the amount of phosphate on the insulin receptor regulatory
peptides. Instead of peptide separation by HPLC, we may use two-dimensional thin-
layer electrophoresis. Unfortunately both methods have failed (so far) to yield good
results because of poor separation.
58


5. Conclusion
In summary, we have evidence that metformin fights diabetes by acting
through an intermediate to inhibit PTP-1B. Inhibition of PTP-1B would block
phosphate removal from the insulin receptor and stimulate receptor tyrosine kinase
activity.
59


REFERENCES
Anderson, Kenneth N. Mosbys Medical, Nursing, and Allied Health Dictionary. St.
Louis: Mosby-Year Book, 1994.
Bevan, P., (2001). Insulin signaling. Journal of Cell Science 114(8): 1429-30.
Bjomtorp P, (2000). Metabolic implications of body fat distribution. Diabetes Care
14: 1132-1143.
Bjomtorp P, Rosmond R, (1999). Hypothalamic origin of the metabolic syndrome X.
Ann NY Acad Sci 892: 297-307.
Chavanieu, A. Calas B., Vaglio, P., Grigorescu, F., (1992). Kinase inhibition by a
phosphorylated peptide corresponding to the major insulin receptor
autophosphorylation domain. Eur J Biochem 208(2): 367-373.
60


Cheung, A., Kusari, J., Jansen, D., Bandyopadhyay, D., Kusari, A., Bryer-Ash, M.,
(1999). Marked impairment of protein tyrosine phosphatase IB activity in
adipose tissue of obese subjects with and without type 2 diabetes mellitus. J Lab
ClinMed 134(2): 115-23.
Cohen, P., (1999). The croonian lecture 1998. Identification of a protein kinase
cascade of major importance in insulin signal transduction. Philos Trans R Soc
Lond B Biol Sci 354(1382): 485-95.
Elchebly, M, et al., (1999): Increased insulin sensitivity and obesity resistance in mice
lacking the protein tyrosine phosphatase-IB gene. Science 283(5407): 1544-8.
Hansen PC, (1999). The metabolic syndrome X. Ann NY Acad Sci 892:1-24.
Margolis, Simeon, and Christopher Saudek. The Johns Hopkins White Papers:
Diabetes Mellitus. Baltimore: The Johns Hopkins Medical Institutions, 2002.
Meir, R., Hemmings, R. A., (1999). Regulation of protein kinase B. Receptor Signal
Transduction Research 19(1-4): 121-8.
61


Peeke PM, Chrousos GP, (1995) Hypercortisolism and obesity. Ann NY Acad Sci
774: 665-76.
Pellegrini, M. C., Liang H., et al., (1998). Mapping the subsite preferences of protein
tyrosine phosphatase PTP-1B using combinatorial chemistry approaches.
Biochemistry 37 (45): 15598-606.
Prager, R., and Schemthaner, J., (1983). Insulin receptor binding to monocytes,
insulin secretion, and glucose tolerance following metformin treatment. Results
of a double-blind cross-over study in type II diabetics. Diabetes 32(12): 1083-6.
Sadler, S.E., and Mailer, J.L., (1981). Progesterone inhibits adenylate cyclase in
Xenopus oocytes: action on the guanine nucleotide regulatory protein. Journal of
Biological Chemistry 256: 6368-6373.
Salmeen, A., Andersen, J., Myers, M.P., Tonks, N.K., Barford, D., (2000). Molecular
basis for the dephosphorylation of the activation segment of the insulin receptor
by protein tyrosine phosphatase IB. Molecular Cell 6: 1401-1412.
62


Sredy, J., Sawicki, D.R., Flam, B.R., Sullivan, D., (1995). Insulin resistance is
associated with abnormal dephosphorylation of a synthetic phosphopeptide
corresponding to the major autophosphorylation sites of the insulin receptor.
Metabolism 44(8): 1074-81.
Stith, B. J., Woronoff, K., and Wiemsperger, N., (1998). Stimulation of the
intracellular portion of the human insulin receptor by the antidiabetic drug
metformin. Biochemical Pharmacology 55: 533-536.
Stith B. J., Goalstone, M. L., Espinoza, R., Mossel, C., Roberts, D. and Wiemsperger,
N., (1996). The antidiabetic drug metformin elevates receptor tyrosine kinase
activity and inosital 1,4,5-trisphosphate mass in Xenopus oocytes. Endocrinology
137: 2990-2999.
Tao, J., Malbon, C.C., Hsien-yu, W., (2001a). Insulin stimulates tyrosine
phosphorylation and inactivation of protein-tyrosine phosphatase IB in Vivo.
The Journal of Biological Chemistry 276(31): 29520-25.
63


Tao, J., Malbon, C.C., Wang, H., (2001b). Gai2 enhances insulin signaling via
suppression of protein-tyrosine phosphatase IB. The Journal of Biological
Chemistry 276(43): 39705-12.
Vigneri, R., Gullo, D., Pezzino, V., (1981). Metformin and insulin receptors.
Diabetes Care 7 Suppl 1:113-7.
White M.F., (1990). The insulin receptor tyrosine kinase. Peptide Hormone Action
223-250.
Zabolotny JM, Bence-Hanulec KK, Stricker-Krongrad A, Haj F, Wang Y, Minokoshi
Y, Kim YB, Elmquist JK, Tartaglia LA, Kahn BB, Neel BG (2002). PTP1B
regulates leptin signal transduction in vivo. Developmental Cell 2(4): 489-95.
Zimmet P, et al., (1999). Etiology of the metabolic syndrome X: potential role of
insulin resistance, leptin resistance and other players. Ann NY Acad Sci 892: 25-
44.
64


Full Text

PAGE 1

THE ANTIDIABETIC DRUG METFORMIN INHIBITS PROTEIN TYROSINE PHOSPHATASE IB ACTIVITY IN XENOPUS lAEVIS OOCYTES by William Louis Holland B.S., Biology University of Colorado, 2000 B.S Psychology, University of Colorado, 2000 A thesis submitted to the University of Colorado at Denver in partial fulfillment of the requirements for the degree of Master of Arts Biology 2002 r'--, A I L.! '. -..:

PAGE 2

This thesis for the Master of Arts degree by William Louis Holland has been approved by David S. Albeck

PAGE 3

Holland, William Louis (M.A., Biology) The Antidiabetic Drug Metformin Inhibits Tyrosine Phosphatase 1 B Activity in Xenopus laevis Oocytes Thesis directed by Professor Bradley 1. Stith ABSTRACT For over 40 years, metformin has been used for the clinical treatment of insulin resistance in type 2 diabetes Despite the drug's widespread use, its cellular mechanism of action has remained unclear. While previous research has indicated that metformin stimulates the insulin receptor, the mechanism of met form in's stimulation of the receptor has not been described. In our current study, we report that metformin stimulates the insulin receptor by inhibiting protein tyrosine phosphatase IB (PTP-1B) The binding of insulin to its receptor stimulates phosphorylation of a regulatory domain in the receptor. Specifically, tyrosine residue numbers 1146, 1150 and 1151 constitute the regulatory phosphorylation sites of the receptor When all three of these sites have been phosphorylated, insulin receptor tyrosine kinase (IRK) activity is fully activated. PTP-1B removes phosphate groups from these regulatory sites, thus inhibiting IRK activity III

PAGE 4

Using a new HPLC method, we report the ability to separate the full spectrum of different phosphopeptides corresponding to the regulatory domain of the insulin receptor. Metfonrun stimulated the ability of purified beta subunit of the insulin receptor to phosphorylate a peptide containing the regulatory domain of the insulin receptor. In another series of experiments using plasma membranecortices (PMCs), metformin was found to increase tyrosine kinase activity and decrease tyrosine phosphatase activity toward the regulatory phosphopeptide The PMCs contain both tyrosine kinase and phosphatase activities. Metformin's ability to decrease tyrosine phosphatase activity was negated by removal of soluble proteins with a salt wash of the PMCs (although salt washes were not used with the regulatory peptide). In-vitro phosphatase assays ofPTP-lB indicate that PTP-IB was not directly inhibited by metformin, but metformin did inhibit PTP-IB in the presence ofPMCs. With experiments using a salt wash of the PMCs and the in vitro PTP-IB assay, there are two lines of evidence that PTP-IB was indirectly inhibited by metformin This suggests that metformin inhibits tyrosine phosphatase activity through a soluble intermediate This abstract accurately represents the content of the candidate's thesis I recommend its publication. Signed IV

PAGE 5

ACKNOWLEDGEMENT I would like to thank Dr. Brad Stith for his support and friendship and for the research opportunities he has provided me. Additionally, I would like to thank Dr. Doug Petcoff, Dr. Martin Gonzalez and Dr. Dave Albeck, for their endless hours of guidance, support and friendship. Thank you to Kai Savi, Patricia Medina, Khulan Batbayar, Erinn Stauter, and Timberley Roane for their support and friendship.

PAGE 6

CONTENTS Figures ............. .. ... ....... ... .. .. .. ................... ............................ .ix Tables .................. .. '" ........ '" ............ .. .............................. ... ... x Chapter 1 Introduction ....................... ............... .................... .. .................. 1 1.1 Diabetes Mellitus .. ............... ... ......... ........ ... .. .. ........ ........... 1 l.2 Metabolic Syndrome X ...... ... ..................................... ... ... ... ....... 3 1 3 Insulin Receptor Activation ... ... ... .. .. ......................... . .. ... .. .... 5 l.4 Insulin Signaling .. ......... ...... ............. ................... ................. .... 8 l.5 Protein Tyrosine Phosphatase lB ..... .. .................. ....................... 11 1 6 Metformin ...... .. ..... ..................... .. .. .............................. .. ... 14 1.7 The Xenopus Cell Model. .......... ..... ... .. ........ .......................... .. 16 2. Materials and Methods ..................... .. '" ... ... ......... .............. ... .. ... 17 2 1 Xenopus Maintenance ... '" ....................... ..................... ........ .. 17 2 2 Xenopus Oocytes ...... ... ...................... .................... ......... ........ 17 2.3 Plasma Membrane-Cortices ......................................................... 18 2.4 Analysis of Endogenous Tyrosine Phosphatase Activity .. ......... .......... ... 19 2 5 Analysis ofPTP-IB Activity In Vitro .............................................. 22 2 6 Analysis ofPTP-IB Activity with PMCs ......................................... 23 Vi

PAGE 7

2 7 Development of a New HPLC Peptide Analysis for Insulin Receptor Autophosphorylation .. .... ..... . .. ........ . .. .. .. 24 2.8 Analysis of the Phosphorylation State of Insulin Receptor Regulatory Domain Peptides with f3IRK ...... ............ ........... 29 2 9 Analysis of the Phosphorylation State ofInsulin Receptor Regulatory Domain Peptides with PMCs .. .. ..... ... .. . ... .. .. 30 3 Results . ... ...... ..... ........... ... .. .. ..... .. ........... .. ....... ... . ..... 31 3 1 Metformin Inhibited Tyrosine Phosphatase Activity in PMCs ..... ... ...... .......... ........ .. ..... .............. .. .... .. 31 3 2 Metfromin Did Not Inhibit PTP-IB Activity In Vitro .. .. .. ...... .. ... . 34 3 3 PMCs Allowed Metformin Inhibition ofPTP-IB ..................... .... .. 36 3 4 Use ofa Peptide Containing the Human Insulin Receptor Regulatory Domain .. .. .. .......... ........ ... .. .. ........ .. .... 38 3 5 Metformin Stimulated f3IRK Activity In Vitro ... .......... .. .. ... ... ... ... 42 3.6 The Human Insulin Receptor Kinase Preferentially Added Phosphate to the 1150 Tyrosine .............. ....... .. ..... .. .. ... 45 3 7 Metformin Stimulated Tyrosine Kinase Activity in PMCs .. .. .. .. . ... ... 47 3 8 Metformin Inhibited Tyrosine Phosphatase Activity that Removes Phosphate from the Insulin Regulatory Peptide ...................... 49 4 Discussion ..... . ..... .. ... ... ....... ... .... .. .. ........... ... .. ... .... 52 4.1 Does Metformin Directly or Indirectly Inhibit PTP-IB? ... ........ . . .... ..... 54 4.2 Comparison with a Known Tyrosine Phosphatase Inhibitor Vanadium . ........ . ............ . . ... .. . .. ... .. ...... 55 4.3 Other Future Studies .. .. .. ....... .. .. ... .. ..... ... ... ....... ... .... .. 56 Vll

PAGE 8

5 0 Conclusion ... ...... ........ ... .......... ...... .... ... .................. ...... .. .... 59 References . ....... ............... .. .... ..... ......... .. .... ... ............. .. .... 60 Vlll

PAGE 9

FIGURES Figure 1.3.1 Insulin Receptor Autophosphorylation ...... .......... ... ......... . ..... .. 7 1.4 1 Insulin! Akt Pathway .... .. ... ... .......... ... ........ .. .......... ... ... 10 1.5.1 PTPIB Regulates Insulin Receptor Activity ... . ............ ... ........ 13 2 .4.1 Sample Preparation for Tyrosine Phosphatase Assay . .. ... .... ... ..... 21 3 1 1 Metformin Inhibited Tyrosine Phosphatase Activity in PMCs ........... 33 3 2 1 Metformin Did Not Directly Inhibit PTP-IB In-Vitro .. ..... ... ........... 35 3.3. 1 PMCs Rescued Metformin Inhibition ofPTP-IB .... ........... .. ....... .. 37 3.4 1 HPLC Separation of Regulatory Peptide ....... ......... .... ... ....... ...... 41 3.5 1 Metformin Stimulated PIRK Activity In-Vitro ............. .. ............ .44 3.6.1 PIRK Preferentially Phosphorylated Tyrosine 1150 of the Regulatory Domain ............................ ... ... ... ... .... .... 46 3 .7.1 Metformin Stimulated Tyrosine Kinase Activity inPMCs ... .. ................ . .. ... .......... ............................. 48 3 8 1 Metformin Inhibited Tyrosine Phosphatase Activity Toward the Regulatory Peptide ......... '" .... ................... .. ... ... ... 51 IX

PAGE 10

TABLES Table 2.7 1 A Summary of Regulatory Peptide Nomenclature and Elution Time ... ... 27 x

PAGE 11

1. Introduction 1.1 Diabetes Mellitus Diabetes mellitus is a disease characterized by insufficient synthesis, secretion or response to the metabolic hormone insulin. Diabetes is clinically defined by the American Diabetes Association by fasting glucose levels greater than 126 mg/dL (greater than 20% above normal) (Zimmet et al., 1999). These high glucose levels, or hyperglycemia result in excessive urination and the development of painful and debilitating complications. There are two primary forms of diabetes. Type 1 diabetes, also known as insulin-dependent diabetes mellitus (IDDM) generally strikes in childhood. This juvenile form of diabetes is believed to be the result of autoimmune damage to the insulin-producing cells of the pancreas. Type 1 diabetics, who compose 5-10% of the diabetic population, produce little to no insulin In type 2 diabetes or non insulin-dependent diabetes mellitus (NIDDM) the cells are usually capable of producing some insulin ; however target cells are insulin resistant and insulin production is insufficient to normalize plasma glucose concentrations. Nearly 17 million Americans or 3% of the population, currently suffer from type 2 diabetes. Surprisingly, about 1 / 3 of these cases are undiagnosed (Margolis and Saudek, 2002).

PAGE 12

In the early stages ofNIDDM, many patients produce nonnal to abnonnally high amounts of insulin, but the hormone fails to lower blood glucose levels Thus, NIDDM is associated with insulin resistance. Insulin resistance is defined as a need for an additional 200 units of insulin daily to control hyperglycemia (Anderson, 1994). Although the mechanisms underlying insulin resistance haven't been elucidated, human and animal models indicate that insulin properly binds to its cellular receptor ; however, a defect in the cells' internal signaling results in subnonnal glucose transport (Sredy et aI., 1995). Insulin resistance is characterized by a lack of response to insulin by the body's tissues. It appears to be caused by a defect in the insulin-signaling cascade rather than by a defect in insulin binding to the receptor. In insulin resistance insulin binds its receptor but glucose transporters are not recruited to the plasma membrane (as they are in healthy tissue). The glucose tolerance test is a way of quantifying insulin resistance In this test, patients fast for at least 12 hours and then consume a sugar solution. Blood glucose levels are then measured. Glucose levels remain very high in the bloodstream in diabetics, glucose intolerant or insulin resistant individuals. Current American Diabetes Association guidelines define Type 2 diabetes by glucose levels of greater than 126 mg/dl, whereas insulin resistant individuals have glucose levels of 2

PAGE 13

111-126 mg/dl (Zimmet et aI., 1999). Nonnal individuals have fasting glucose concentrations below 110 mg/dL. 1.2 Metabolic Syndrome X Insulin resistance and hyperglycemia are also symptoms ofthe newly defined metabolic syndrome X (MSX) (Hansen, 1999). MSX is known by many names, including insulin resistance syndrome and the cardiovascular disease risk factor cluster. It is characterized by five major symptoms: obesity, hyperglycemia, dyslipidemia, glucose intolerance, and insulin resistance. An estimated 75 million Americans are believed to suffer from the syndrome; this is greater than one out of every four people (Hansen, 1999)! It is a growing epidemic in the United States and throughout the world. The most evident major symptom of MSX is obesity central or visceral obesity in particular. Central obesity is detennined by a waist to hip ratio. Males with a ratio of greater than 0.90 and females with greater than 0.85 are centrally obese (Peeke and Chrousos, 1995). Central obesity, an excess of peritoneal fat, is a major risk factor for heart disease and diabetes. Obese individuals with this pear-shaped body type are at far greater disease risk than persons with peripheral obesity (Bjorntorp, 2000). It is believed that the nature of central obesity, in this disorder, is due to enlarged adipocytes, high insulin and cortisol levels, and increased 3

PAGE 14

glucocorticoid receptor concentrations in the abdomen. Additionally, decreased leptin sensitivity, often associated with MSX, has been linked to the buildup of adipose tissue (Zimmet et aI., 1999). Leptin is a hormone, released by fat cells, that reduces hunger In MSX, abnormal lipid levels (dyslipidemia) are present in the bloodstream. High triglyceride levels (triglyceridemia) and low high-density lipoprotein (HDL) cholesterol (the "good form") levels are examples of dyslipidemia. Triglyceridemia is defined by fasting triglyceride levels above 200 mg/dl (Zimmet et aI., 1999) and low HDL cholesterol is defined as levels below 35 mg/dI. High total cholesterol and low-density lipoprotein ("bad" form) levels are potent risk factors for heart disease that are often present in the metabolic syndrome; however, they are not required. Relatively little is known about the cause of the dyslipidemia in MSX, but it is believed to be related to altered lipolysis induced by hormonal changes (Bjomtorp and Rosmond, 1999). Hypertension is the least consistently-associated symptom of the syndrome. Blood pressure greater than 140/90 is generally considered sufficient for diagnosis of high blood pressure (Zimmet et al., 1999). Although this symptom of MSX is seen far less often than the five symptoms noted above, hypertension is also linked to cardiovascular disease. 4

PAGE 15

Each of the major symptoms for Metabolic Syndrome X is a risk factor for cardiovascular disease and they are often expressed concurrently. A vast majority of type 2 diabetics have MSX. 1.3 Insulin Receptor Activation The insulin receptor is composed of two extracellular alpha subunits and two transmembrane beta subunits all linked through disulfide bonds (Figure 1.3.1). Each beta subunit possesses a tyrosine kinase domain and a regulatory domain. Each alpha subunit contains a binding site for insulin. When the peptide hormone insulin binds to this membrane receptor it induces a conformation change, and this partially stimulates IRK activity. This weak tyrosine kinase activity is sufficient to allow autophosphorylation of the receptor as the subunits phosphorylate themselves. In autophosphorylation, the kinase domains place phosphate on the regulatory domains. The regulatory domain consists of tyrosine residues 1146, 1150, and 1151. All three tyrosine residues must be phosphorylated to induce full activation of the insulin receptor. Thus, when insulin binds to the receptor it causes a conformational change that allows the Beta subunits to be autophosphorylated by their tyrosine kinase domains. Upon complete phosphorylation of the regulatory domain of the insulin receptor, the receptor becomes a fully active tyrosine kinase. Once fully activated, 5

PAGE 16

the insulin receptor kinase (IRK) can phosphorylate other substrates and induce numerous signaling pathways. 6

PAGE 17

IrWin Rn;p:cr lnsUinlirdlllsite La:tiwWlhtylaine re;iWes 1146, 11S), ad 1151 Figure 1.3.1 Insulin Receptor Autophosphorylation. The insulin receptor activates its tyrosine kinase activity via insulin induced autophosphorylation of the regulatory domain (tyro sines 1146, 1150, 1151). 7

PAGE 18

1.4 Insulin Signaling Upon binding to an insulin receptor, insulin initiates a chain reaction of biochemical events within numerous signaling pathways. Like a domino effect, each reaction causes another to occur. The cascade of events in the Akt pathway leads to the insulin induced uptake of glucose into cells and this lowers blood glucose concentrations. Figure 1.4.1 illustrates the Akt signaling pathway. Autophosphorylated insulin receptor tyrosine kinase will phosphorylate other tyrosine-containing proteins. The primary substrate phosphorylated by the insulin receptor is insulin receptor substrate 1 (IRS-I). Once phosphorylated, IRS-I can serve as a binding site for the p85 subunit of3-phosphoinositide-dependent (PI3) kinase. This binding stimulates the dimerization of this regulatory subunit (P85) with the catalytic p 110 subunit of PI3 kinase. Once p85 binds plIO, the activated PI3 kinase phosphorylates the membrane lipid phosphatidylinositol4,5 bisphosphate (PIP2), producing phosphatidylinositol 3,4,5 trisphosphate (PIP3). PIP3 serves as a binding site for proteins containing a plextrin homology (PH) domain. PIP3 dependent kinase 1 (PDKl) and Akt are both serine/threonine kinases that contain PH domains As its name implies, PDKI is activated when bound to PIP3. Activated PDKI will phosphorylate and activate Akt. Activated Akt causes the movement of Glut 4 from internal stores to the plasma membrane. Acting at the plasma membrane, glut4 allows the movement of glucose 8

PAGE 19

into the cell. Recent evidence has suggested that Akt activation is necessary and sufficient for insulin induced glucose transport to occur (Cohen 1999). 9

PAGE 20

Figure 1.4.1 InsulinlAkt Pathway. Insulin signaling stimulates glucose transport via the Akt pathway. The autophosphorylated insulin receptor phosphorylates IRS-I. IRS-I activates PI3 kinase via binding. PI3 kinase produces PIP3, allowing for PDKI activation. PDKI activates Akt, and Akt stimulates glut 4 translocation by an unknown mechanism 10

PAGE 21

1.5 Protein Tyrosine Phosphatase 18 Protein tyrosine phosphatases (PTPases) are a family of enzymes that regulate the insulin signaling pathway. By removing one or more regulatory phosphates from the insulin receptor and IRS-I, they stop the domino effect within the cell (Figure 1.5.1); that is, PTPases tum off insulin's signaling cascade. Protein tyrosine phosphatase IB (PTP-IB) has been implicated as the primary PTPase that turns off the insulin receptor (Salmeen et aI., 2000, Elchelby et aI., 1999, Cheung et aI., 1999). PTP-I B is known to have at least two regulatory phosphorylation sites. IRK phosphorylates PTP-IB on a regulatory tyrosine, and inactivates phosphatase activity. This will ensure continued IRK activity (Tao et aI., 2001a). In contrast, the cyclic AMP dependent kinase (PKA) phosphorylates PTP-IB on a regulatory serine residue. When phosphorylated on this regulatory serine, phosphatase activity is enhanced (Tao et aI, 2001b). PKA also phosphorylates PTP IB on a threonine residue that appears to show no regulatory activity (Tao et aI., 2001 b). In addition to inhibiting the insulin signal, PTP-IB has also been found to inhibit leptin signaling (Zabolotny et aI., 2002). Leptin is a hormone that induces satiety ("feeling full"). PTP-IB knock-out mice (where leptin signaling would be activated to higher than normal levels) are lean (Zabolotny et aI., 2002), while 1 1

PAGE 22

overproduction of PTP-I B is associated with obesity. 12

PAGE 23

"ON" AUTO PHOSPHORYLATED "OFF" INSULIN INSULIN RECEPTOR RECEPTOR a PTP-IB Figure 1.5.1 PTP-1B Regulates Insulin Receptor Activity. The deactivation of the insulin receptor by a tyrosine phosphatase. PTP-l B removes phosphate from the insulin receptor regulatory domain thereby decreasing insulin receptor kinase activity. 13

PAGE 24

1.6 Metformin Since its April 1995 introduction to the United States, glucophage (metfonnin) has become the most-prescribed diabetes pill in the nation For a drug that has effectively lowered the blood glucose concentrations of European and Canadian diabetics since 1958, metfonnin's recent popularity should not be surprising. Despite the drug's long, proven history, its biochemical mechanisms of action still have not been elucidated. Previous studies suggest that metfonnin doesn't increase insulin levels or insulin binding to the insulin receptor (Prager and Schemthaner, 1983 Vigneri et aI., 1981), therefore it must be stimulating the insulin pathway via a post-binding site modification. Metfonnin increases insulin receptor activity (Stith et aI., 1996 1998). In plasma membrane preparations isolated from whole cells, therapeutic concentrations (10 Ilg/ml) of the drug lead to a 300% increase in tyrosine kinase activity (Stith et aI., 1996). Later, Stith et a1. (1998) reported that metfonnin induces a 25% increase in purified insulin receptor kinase activity with a maximum stimulation at 1.0 Ilg/ml metfonnin. This is a huge discrepancy! Metfonnin's stimulation of tyrosine kinases in the PMC is 12 fold greater than with isolated insulin receptor tyrosine kinase Stith et a1. (1996) suggested that metfonnin must enter the cell and then acts at or near the plasma membrane. This suggestion was based on the finding that 14

PAGE 25

metfonnin acts more rapidly in isolated plasma membrane-cortices (PMCs) than with whole cells. There was a time delay of approximately 90 minutes for maximal metfonnin action in whole cells. This delay may be due to metfonnin entering the cell and building up to effective levels. Since metfonnin can immediately stimulate the insulin receptor in PMCs, metfonnin must enter the cell but act at or near the membrane. It has also been shown that metfonnin increases the phosphotyrosine content of IRK (Stith et aI., 1996) and IRS-l (Stith et aI., 1996, Pryor et aI., 2000) The increase in phosphotyrosine on the insulin receptor suggests that metformin may be regulating the insulin receptor via altered phosphorylation of its regulatory domain. Pryor et aI. (2000) have shown that metformin stimulates the insulin signaling pathway shown in Figure 1.4.1. Metformin stimulates PI3 kinase and Akt activities and increases translocation of the glucose transporter Glut 4 to the plasma membrane in insulin-resistant adipocytes. Insulin was used to induce an insulin-resistant state in the adipocytes. One major question remains about metfonnin's ability to stimulate IRK activity: How does metfonnin stimulate IRK activity? Our theory is that metfonnin must enter the cell then inhibit PTP-I B which results in activation of the insulin receptor 15

PAGE 26

1.7 The Xenopus Cell Model Many factors make the Xenopus laevis oocytes an excellent cell model for our studies. First, insulin signaling pathways have been well-characterized and well studied in Xenopus. Second oocytes are easily maintained and attainable year-round. Additionally, oocytes are very large cells (approximately 1.2 mm in diameter) that can be manipulated very easily. For example we can manually isolate the plasma membrane-cortex (PMC). The ability to isolate a membrane is very unusual but the Xenopus oocyte has an unusual plasma membrane that is highly reinforced. 16

PAGE 27

2. Materials and Methods 2.1 Xenopus Maintenance Wild type Xenopus laevis, the African clawed frog, were obtained from Xenopus Express (Ft. Lauderdale Florida). Each female frog was fed 2 grams of ground beef heart three every other day. To prevent infections food was withheld for two days after arrival at the University of Colorado at Denver. Frogs were kept in 30 gallon plastic tanks. The water was replaced with 15 gallons of fresh water about 4 hours after feeding 10 ml of NovA qua (Novatek, Haywood, CA) was added to the tank after refilling with fresh cold tapwater. NovAqua removes CO2 and prevents infection. 2.2 Xenopus Oocytes Xenopus oocytes were acquired from frogs primed 3-10 days before experiments with 50 IU Pregnant mare serum gonadotropin (PMSG) (Sigma, St. Louis, MO). Growth of the ovarian follicles and the corpus luteum was stimulated by priming with PMSG. This decreased the time required for oocytes to respond to insulin (Stith et al., 1985). The animals were placed in ice water for 45 minutes and then sacrificed with a rat guillotine. Ovaries were removed, blotted to remove excess blood and placed in room temperature OR2 solution (83 mM NaCl, 0.5 mM CaC12, 1 mM MgC12, 10 mM HEPES). The ovaries were then sectioned into smaller segments 17

PAGE 28

and placed in fresh OR2 solution. Healthy, stage VI oocytes (1.1 nun or greater) were manually isolated from follicles and maintained in OR2. 2.3 Plasma Membrane-Cortices Plasma membrane-cortices (PMCs) are manually isolated from Xenopus oocytes as previously described by Sadler and Maller (1981) Each PMC is about 1050 run thick and contains approximately 1 J..lg of total protein. PMCs contain the plasma membrane, integral membrane proteins, peripheral membrane proteins, few pigment granules, and the cortex (the outer 5 microns or more of the cytoplasm). All proteins present in the PMC may retain normal function as insulin can stimulate the insulin receptor and enzymes down-stream of the insulin receptor (Stith et aI., 1996). Before preparing PMCs, petri dishes (5 cm) were filled with Sadler's isolation buffer (SIB) (10 mM NaCI, 10 mM HEPES, pH 7.2) and placed on a bed of crushed iced for 30 minutes. Groups of 15 defolliculated oocytes were placed in the ice-cold SIB. Oocytes were opened by piercing the cells with closed forceps and allowing the forceps to open. The tom cells were then rotated, plasma membrane side up and flattened against the bottom of the petri dish by gently depressing with forceps. The PMC preparations were then stored on ice for 45 minutes to allow the yolk and intracellular material to dissociate from the plasma membrane. PMCs were then teased away from the cytosol and transferred to a 1.7 ml microfuge tube using a wide-18

PAGE 29

mouthed transfer pipette. The preparation can be washed more thoroughly. The more extensively washed PMCs are clearer as more pigment granules and cytoplasm are removed. 2.4 Analysis of Endogenous Tyrosine Phosphatase Activity To examine the effect of metformin on phosphatase activity a phosphatase assay in PMCs kit (kit 2 Upstate biotechnology Lake Placid NY) was used. These experiments were conducted by a graduate student (Thomas Morrison) with my assistance. The kit measures the release of phosphate from a tyrosine phosphopeptide (T-S-T-E-P-Q-pY-Q-P-G-E-N-L). The phosphate release is corrected to determine the moles of phosphate released. To measure phosphate the assay utilizes the fact that the free phosphate binds to molecules of malachite green When bound to phosphate, malachite exhibits increased absorbance at 650 run Absorbance was measured with a Milton Roy Spectronic 1001 spectrophotometer set at 650 run. After isolating PMCs in groups of 15, the microfuge tubes were "flash spun" by holding the "spin button" on the Beckman microfuge for about 1 second to pellet the PMCs. With one treatment group, the supernatant was removed and replaced with 100ui SIB with 150 mM NaCI for 5 minutes. This is the salt-washed group. The PMCs were then flash-spun again. The supernatant (containing peripheral membrane proteins removed from the PM C) was removed from all groups and PMCs were 19

PAGE 30

resuspended in 80 ul SIB. Phospho peptide (20 ul of 1 mM) was added, samples were gently vortexed and incubated at 15 0 C for 15 minutes. 30 seconds before the time point, samples were flash-spun and 20 ilL of supernatant was removed and placed in a separate 0.5 ml microfuge tube The 20 ilL of supernatant was analyzed for free phosphate and the remaining sample was treated with either 2 ilL of deionized water (dH20) or 2 ilL of 40 Ilg/ml metforrnin (l0 Ilg/ml). The samples were vortexed to resuspend the PMCs and incubated for another 15 minutes at 150 C Then, the samples were spun again and 20 ilL of supernatant was removed for phosphate analysis. Figure 2.4.1 summarizes the sample preparation 20

PAGE 31

15 min I + p No Drug 15 min 1 Figure 2.4.1 Sample Preparation for Tyrosine Phosphatase Assay. 4 samples were prepared. 1) samples of 15 untreated PMCs were incubated 15 minutes with 200 IlM phosphopeptide. 2) samples of 15 untreated PMCs were incubated an additional 15 minutes (30 minutes total) with PMCs. 3) samples of 15 salt washed PMCs were treated with 10 Ilg/ml metformin after 15 minutes of phosphate release and incubated for an additiona115 minutes (30 minutes total). 4) samples of 15 PMCs (not salt washed) were treated with 10 Ilg/ml metformin after 15 minutes of phosphate release and incubated for an additional 15 minutes (30 minutes total). 21

PAGE 32

2.5 Analysis ofPTP-lB Activity In Vitro Purified human recombinant PTP-I B was purchased from Upstate Biotechnology (Lake Placid, NY). The effect of met form in on PTP-IB activity was analyzed using a modification of the procedure suggested by the manufacturer of the recombinant PTP-l B. This measurement does not require the difficult assay of phosphate but records the amount of product (para-nitrophenol; pNP) by spectrophotometric means. PTP-I B removes the phosphate from para-nitrophenol phosphate (pNPP) (Upstate Biotechnology 20-106) in this non-specific phosphatase assay, and this produces pNP. pNP, the product of the dephosphorylated pNPP, absorbs strongly at 405 nm after phosphate is cleaved. PTP-IB activity was assayed in a mixture ofPTP-IB Assay Buffer (25 mM HEPES, 50 mM NaC!, 5 mM dithiothreitol, 2.5 mM ethylene-diamine-tetraacetic acid (EDTA), pH 7.2) and bovine serum albumin (BSA) solution (100 f.1g1ml in assay buffer). Before preparing samples, 2 f.1L of PTP-l B stock was diluted with 198 f.1L of BSA solution. To prepare each sample, 90 f.1L BSA solution, 60 f.1L Assay buffer, 10 f.1L dilute enzyme, 20f.1L 100 f.1g1ml metformin or deionized water, and 20 f.1L 50 mM pNPP in PTP-IB assay buffer were combined Samples were allowed to incubate with metformin for 20 minutes prior to pNPP addition. This incubation was performed to allow maximal interaction of met form in and PTP-IB. Each sample contained 0.075 f.1g PTP-IB and a final concentration of 5 mM pNPP. After pNPP 22

PAGE 33

addition, samples were then incubated for 10 minutes at 370 before stopping with 100 ilL of 2M K2C03 The final volume was raised to 1000 ilL with dH20 and absorbance was measured at 405 nm. PTP-1 B activity was calculated using the equation: PTP-1B activity= [O.OOIL cm)]/ T/ In this equation, 0.001 L is the final volume (1 mL), is the absorbance of PTP-IB at 405 nm (absorbance is 0.015 for 0 05 Ilg ofPTP-IB), c is the extinction coefficient for pNP (1.78 X 101\4 1M cm), cm is the pathlength oflight (1.0 cm), "T" is the time for this assay (10 minutes), and "Ilg" PTP-1B per sample (0.75 2.6 Analysis ofPTP-lB Activity with PMCs PTP-IB activity was also assayed in the presence ofPMCs. PTP-1B activity measurement and PMC isolation were procedures described in sections 2.3 and 2.4 Calculations were performed to demonstrate that the phosphatase activity from PMCs was negligible in comparison to the exogenous recombinant PTP-I B activity. PTP-IB activity was assayed in a mixture ofPTP-IB Assay Buffer (25 mM HEPES, 50 mM NaCl, 5 mM dithiothreitol, 2.5 mM EDTA, pH 7.2) and bovine serum albumin (BSA) solution (100 in assay buffer). Samples were prepared by aliquotting from prepared enzyme cocktails. This was done to help reduce the 23

PAGE 34

between samples error due to pipetting. PTP-IB (4 ilL) enzyme was diluted into 1596 ilL ofBSA solution and 960 ilL ofPTP-IB assay buffer was added to the diluted enzyme. To prepare final enzyme cocktails, 640 ilL of this dilute enzyme mixture was then aliquoted into 2 microfuge tubes and treated with 80 ilL of 100 Ilg/ml metformin or 80 ilL of dH20 for control samples. Metformin was allowed to incubate for 20 minutes with the enzyme solution (SoC). PMCs were flash-spun and 170llL of the enzyme cocktail was removed. Samples were then incubated for 20 minutes on ice and warmed to 37 0 for 5 minutes. Reactions were started by the addition of 20 ilL 50 mM pNPP. Each sample contained 0.1725 Ilg PTP-l B, 20 PMCs and a final concentration of 5 mM pNPP. Samples were incubated for 15 minutes at 370 before stopping with 100 ilL of 2M K2 C03. The final volume was raised to 1000 ilL with dH20 and absorbance was measured at 405 nm. Enzyme activity was calculated as before. 2.7 Development of a New HPLC Peptide Analysis for Insulin Receptor Autophosphorylation To examine the phosphorylation sites on the insulin receptor, many researchers use radioactive 32p to label various phosphotyrosines on the intact insulin receptor. Then the intact insulin receptor is cut up into peptide fragments, and the 24

PAGE 35

peptide fragments (some of which contain the regulatory phosphorylated amino acids) are separated by high-performance liquid chromatography (HPLC). While we have not yet used an intact insulin receptor for our studies, we have set up a "peptide" separation method through modification of an HPLC method by Morris White (1990). In other experiments, a peptide representing the insulin receptor regulatory domain was added to PMCs. The use of this peptide would enable in-vitro analysis of the phosphorylation content of the insulin receptor regulatory domain. Analysis of insulin receptor phosphopeptides required that we set up a high performance liquid chromatography (HPLC) procedure that would be able to separate the peptides High-pressure liquid chromatography (HPLC) separates molecules (in this case phosphopeptides) according to subtle differences in their solubility characteristics All HPLC analysis was performed using a RaininIV arian HPXL model HPLC system consisting of two isocratic pumps, a high pressure mixer, UV detector and Macintosh control interface. The UV detector was set to 270 nm (the optimum wavelength for the detection of tyrosine residues). The HPLC method uses a flow rate of 1 mllminute. A Supe1co LC-318 column provided efficient separation of all phosphopeptides of interest. A solvent gradient consisting of water and acetonitrile both acidified with 0.05% TF A, is ramped from 5% acetonitrile to 19.1 % acetonitrile over 40 minutes. The gradient 25

PAGE 36

then ramps to 25% acetonitrile in 5 minutes then returns to 5% acetonitrile by the end of the 50-minute run. The HPLC method was characterized using insulin receptor phosphopeptides (lRO, IR5, IR9 IR 10, IR 5 ,9 10) from Biomol (Plymouth Meeting Pennsylvania). Table 2 .7.1 indicates the phosphorylation state of each peptide. The elution time for each phosphorylated peptide was determined by running the standards individually and in combination 26

PAGE 37

Table 2.7.1 A Summary of Regulatory Peptide Nomenclature and Elution Time. The Biomol regulatory peptides consist of the insulin receptor residues 1142-1153. This sequence of amino acids is known to regulate the insulin receptor based on its phosphorylation state. The elution order for the 3 bisphosphorylated peaks has not yet been determined, but they elute around 21-25 minutes. Phosphorylated Tyrosine Biomol Name HPLC Elution (amino acid number in the Time (min) human insulin receptor) None IRO 36 Monophosphory lated: 1146 IR5 32.7 1150 IR9 34.6 1151 IRlO 33.7 Bisphosphorylated: 21-25 1146 1150 IR 5 9 1146,1151 IR 5, 10 1150,1151 IR 9,10 Trisphosphorylated: 1146 1150 ,1151 IR5,9,1O 19 27

PAGE 38

I would like to describe the HPLC methodology in more general terms For example, we note that the insulin receptor peptide with all three phosphates elutes at 19 min on our HPLC gradient. However, when the peptide is incubated with tyrosine phosphatases (either from the Xenopus PMC or human PTP-l B itself), dephosphorylation takes place (e.g., phosphate on sites 1146, or 1150 or 1151 is removed from the peptide). Thus, we see new peaks appearing on the HPLC recorder (occurring at 33-36 minutes); these new peaks include the peptide with only one phosphate present. By comparing the size of peak areas versus time, we can determine which phosphate is removed first, and also follow the rate at which phosphate is removed from the specific regulatory sites on the insulin receptor peptide. Perhaps metformin will inhibit the removal of phosphates located on a certain tyrosine from the insulin receptor peptide. We can also examine the action of met form in upon human PTP-IB. Our model predicts that metformin, acting through an intermediate present in our plasma membrane preparation, will inhibit PTP-IB. While the PMC contains endogenous tyrosine phosphatases, the endogenous levels of tyrosine phosphatase activity are relatively low, and we can add human PTP-l B to flood out the endogenous tyrosine phosphatase activity. Thus, we can record the activity of a human tyrosine kinase (PTP-IB) and measure the effects of met form in on both PTP-IB and the level of phosphof) T lation of the human insulin receptor regulatory peptide. 28

PAGE 39

Conversely, we can add the insulin receptor peptide without any phosphates on it to PMCs, and record phosphate addition to the peptide. Perhaps metformin will stimulate the addition of the phosphates to the insulin receptor peptide. With the insulin receptor regulatory domains and PMCs, we can quantify phosphatase and kinase activities simultaneously (as we only record the final sum total of both kinase and phosphatase action). 2.8 Analysis of the Phosphorylation State of Insulin Receptor Regulatory Domain Peptides with The effect of metformin on in-vitro [3IRK (the commercially-available purified beta subunit ofthe insulin receptor) activity was assayed using the new HPLC methodology. Samples were prepared by diluting 1 f.1g of [3IRK into 68 f.1L of [3IRK assay buffer with or without 10 f.1g/ml metformin. The reaction was started by the addition of 32 f.1g IRO and carried out for 20 minutes at 370 Reactions were stopped via the addition of 100 f.1L 10% trichloro-acetic acid (TCA). 150 ilL (75%) of each sample was injected onto the HPLC for analysis of kinase activity. 29

PAGE 40

2.9 Analysis of the Phosphorylation State of Insulin Receptor Regulatory Domain Peptides with PMCs The effect of metfonnin on kinase and phosphatase activity contained in the PMCs was assayed using the new HPLC method. Samples were prepared by adding 84 ilL of I3IRK assay buffer to 15 PMCs. Reactions were started by the addition of 16 ilL (16 Ilg) IR9 phosphopeptide. The reaction proceeded for 40 minutes at room temperature before stopping it with 100 ilL 10% TCA. 75% (1501lL) of each sample was injected onto the HPLC for analysis 30

PAGE 41

3. Results 3.1 Metformin Inhibited Tyrosine Phosphatase Activity in PMCs With the malachite green assay for tyrosine phosphatase and our PMC preparations, metformin (10 Jl.glml) significantly decreased phosphatase activity (Figure 3.1.1). Metformin was added after a 15 minute control period. As opposed to a control rate of removal of phosphate from tyrosine located on a phosphopeptide (see line labeled "C"), metformin strongly inhibited the rate of removal of phosphate. The use ofPMCs enables us to reduce the amount of background noise (error caused by interference from cytoplasmic kinases and phosphatases) by emphasizing dephosphorylation taking place at the membrane. This is where the insulin receptor is located (as opposed to dephosphorylation events taking place deeper in the cytoplasm). For these tyrosine phosphatase assays, all groups of 15 PMCs were incubated with a phosphopeptide substrate. After 15 minutes, a sample was taken from each sample and the amount of free phosphate was measured. The experimental samples were allowed to incubate another 15 minutes with or without metformin addition. In Figure 3.1.1, phosphate release (a measure of tyrosine phosphatase activity) is strongly inhibited by metfonnin: compare control release ("C") with that in the 31

PAGE 42

presence of metfonnin ("M"). These data lend support to our idea that metfonnin acts by inhibiting a tyrosine phosphatase. Salt washing of the PMCs prior to treatment with metfonnin ("M+W") nullified the inhibiting effects of metfonnin. The tyrosine phosphatase activity in washed samples did not significantly differ from unwashed controls, indicating that tyrosine phosphatases were not removed by the salt wash. The results of the salt wash indicate that an intennediate, necessary for the effects of metfonnin was removed by the salt wash. So metfonnin may be inhibiting PTP-IB indirectly. 32

PAGE 43

Figure 3.1.1 Metformin Inhibited Tyrosine Phosphatase Activity in PMCs. Metformin (M) inhibited phosphatase activity in plasma membrane-cortex preparations. Metformin did not inhibit PTPase activity ifthe PMC was first washed with 150 mM NaCI first (W+M). Compare the release of phosphate from a tyrosine phosphatase peptide substrate in the control group "C" with a 10 Ilg/ml metformin group .... -> t;c 400 w_ wCl)U 300 .... wo:: o::w ::J:WQ. Q. .... W 200 CI)..J ::J:O Q.CI)Q. 100 WO-Z::J: Cl)Q. 0 0 0:: > 0 5 10 15 20 25 30 35 .... TIME (MIN) 33

PAGE 44

3.2 Metformin Did Not Inhibit PTP-IB Activity In Vitro We next wanted to see ifmetformin would directly inhibit human PTP-IB (a commercially-available tyrosine phosphatase known to inhibit the insulin receptor). Our prior work (Figure 3.1.1) suggested that metformin acts indirectly. We combined various concentrations of met form in with purified PTP-IB. We found that metformin did not inhibit the in vitro PTP-IB activity (Fig. 3.2.1) These data suggest that metformin does not act upon PTP-l B directly and this supports the idea that metformin acts through an intermediate to inhibit PTP-IB. The intermediate would only be present in our plasma membrane-cortex preparation (unwashed PMCs) or the intact cell. 34

PAGE 45

l""'4 2.0 -,--------------------, I E-4 e: 1.5 o .0 1.0 ... ..... < Q) 0.5 > ... ..... = 0.0 0.1 1.0 10 100 Metformin Concentration (ug/ml) Figure 3.2.1 Metformin Did Not Directly Inhibit PTP-IB In Vitro. Metformin was unable to inhibit the activity of human recombinant PTP-IB in an in-vitro assay. Phosphate removal from the enzyme substrate pNPP was measured. Human PTP-IB activity was measured with pNPP (a phosphatase substrate). The number of determinations for each sample is 12. Groups were not significantly different. 35

PAGE 46

3.3 PMCs Allowed Metformin Inhibition of PTP-IB We then determined whether metformin would inhibit human PTP-IB when plasma membrane-cortex was present. The PMC may provide the unknown intermediate that would enable metformin to inhibit PTP-I B. Whereas metformin did not directly inhibit PTP-IB (Figure 3.2.1), after PMCs were added (Figure 3 3.1) metformin inhibited PTP-IB by 50% (P less than 0.01) Metformin was effective at concentrations ranging from 1 to 100 Ilg/ml (10 Ilg/ml is the therapeutic concentration) The conclusion of these two experiments is that metformin acts through an intermediate in order to inhibit PTP-l B. 36

PAGE 47

30r----------------__________________ 25 .0 20 .... -.:: = <.5 15 = E I = Q.. E Eo-C 10 Q..-0+---"'-' Denotes p < 0.005 Metformin Concentration (uglml) Figure 3.3.1 PMCs Rescued Metformin Inhibition of PTP-IB. The effect of metformin on Human PTPI-B was assayed in the presence of PMCs Fifteen PMCs were added to human recombinant PTP-IB, and the phosphatase activity was measured using pNPP. Compare to the lack of effect of metformin on PTP-I B shown i n 3.2.1. For each sample N=12 An asterisk denotes significance P less than 0.005. 37

PAGE 48

3.4 Use of a Peptide Containing the Human Insulin Receptor Regulatory Domain One additional line of evidence that would support our theory that metfonnin acts through inhibition of a tyrosine phosphatase would be to describe any metfonnininduced changes in individual phosphorylation sites on the regulatory domain of the insulin receptor. Our model predicts that, in the presence of metfonnin, certain regulatory phosphorylation sites on the insulin receptor would show higher levels of phosphorylation (and this would result in the measured higher insulin receptor activity). Thus, we wanted to switch from artificial substrates, such as pNPP which was used in Figure 3.1.1, to a peptide that contains the insulin receptor regulatory domain. The insulin receptor is regulated through phosphorylation of tyro sines located in a domain of the receptor from about amino acid 1142 to amino acid 1153. In the experiments described below, we use a peptide derived from the insulin receptor that contains these residues (amino acid sequence: 1142-1153: TRDIYETDYYRK). Under conditions where tyrosines in the insulin receptor regulatory domain are phosphorylated, the insulin receptor would be stimulated. We have purchased this peptide with and without various phosphates already attached to tyrosines. One peptide lacks phosphate on any tyrosine on the peptide, and we purchased other peptides with one phosphate on the peptide, and another with three phosphates on the 38

PAGE 49

peptide. The phosphorylation can occur on tyrosine 1146, or tyrosine 1150 or on tyrosine 1151 (or on all three) of the peptide. Experiments with these peptides a new high perfonnance liquid chromatography (HPLC) procedure that separates the different phosphopeptides. To examine the phosphorylation sites on the insulin receptor, many researchers use radioactive 32p to label various tyrosines on certain amino acids of the insulin receptor Then the insulin receptor is cut up into peptide fragments, and the peptide fragments (some of which contain the regulatory amino acids) are separated by HPLC. While we have not yet used an intact insulin receptor for our studies, we have set up just such a peptide" separation method through modification of an HPLC method by Morris White (1990) Figure 3.4 1 illustrates a typical UV trace from our insulin receptor peptide separation method With our HPLC method, we can separate the peptides that have no phosphate from those that have one or more phosphates ; we can even separate the peptides based on which tyrosine has the phosphate. Thus we can examine whether metfonnin increases or decreases the level of the phosphate on the insulin receptor peptide. Our model would predict that metfonnin would increase the level of phosphorylation on certain regulatory sites known to stimulate the insulin receptor. 39

PAGE 50

Thus, we have developed a new method that should be of great interest to diabetes researchers. The advantages of this system are that we can examine: the human insulin receptor kinase activity the phosphorylation of specific regulatory sites on a human insulin receptor peptide or on the proteolyzed insulin receptor the action of metformin on receptor tyrosine kinase and tyrosine phosphatase activity metformin and insulin action at the plasma membrane (through use of the plasma membrane-cortex or "PMC" preparation) combined action of both tyrosine kinase and tyrosine phosphatase regulation on the insulin receptor peptide (through the total phosphorylation changes of receptor peptide) human tyrosine phosphatase PTP-IB, human insulin receptor (the intracellular portion of the human insulin receptor or "BIRK"), and metformin action. 40

PAGE 51

Figure 3.4.1 HPLC Separation of Regulatory Peptide. Unphosphorylated regulatory domain peptide (32 was incubated with PIRK for 40 minutes. The reaction was stopped with trichloro-acetic acid and the resulting phosphopeptides were separated via HPLC using our new method Four groups of regulatory domains (circled below) were separated by HPLC and quantified by UV detection of tyrosine (See also Table 2.7.1) 1. The trisphosphorylated regulatory domain peptide 2. The bisphosphorylated regulatory domain peptides 3. The monophosphorylated regulatory domain peptides 4 The nonphosphorylated regulatory domain peptide 3 4 ELUTION TIME 41

PAGE 52

3.5 Metformin Stimulated Activity In Vitro We have noted that metformin directly stimulates in-vitro activity (Stith et aI., 1998). In these earlier metformin experiments, radioactive A TP was added to purified and radioactivity on a peptide substrate was recorded. We wanted to confirm these results with the insulin receptor regulatory domain peptide and our new HPLC method. We have now confirmed that metformin does have a small (11 %) stimulatory effect on the insulin receptor (Figure 3.5.1). We combined: the intracellular portion of the human insulin receptor (BIRK) metformin IRO insulin receptor peptide without phosphate We added all constituents under optimal conditions (it took us weeks to determine these conditions; we altered enzyme amounts, substrate, time course, etc.). We then determined that BIRK was able to phosphorylate the insulin receptor peptide without phosphate (refer to section 3.6). That is, catalyzes IRO (unphosphorylated peptide) phosphorylation to IR 9 (one phosphate at the 1150 position). Metformin was able to stimulate the in vitro phosphorylation ofIRO (Figure 3.5.1) however, like the prior in vitro studies of Stith et al. (1998), this stimulation was not large (only 11 %). With our PMCs or whole cells, we achieved a much larger 42

PAGE 53

stimulation (300%) of the tyrosine kinase activity (Stith et aI., 1996). As is discussed elsewhere, we believe that this larger stimulation may be through indirect inhibition of PTP-I B, and the PMC supplies an inhibitory intennediate between metfonnin and the tyrosine phosphatase. This difference in stimulation may be due to the use of different substrates that would reflect different tyrosine kinases being measured. 43

PAGE 54

8 0 QI 7.5 "0 .... ... 7.0 0 -= 6.5 -= 6.0 5 == ) 5.0 cS ... QI Figure 3.5.1 Metformin Stimulated JHRK Activity In Vitro. 0 5 J,lg PIRK was incubated with 200 J.lM IRO for 20 minutes at 370 w ith or without 10 J,lglml metfonnin The asterisk denotes P less than 0 .05. N = 6 44

PAGE 55

3.6 The Human Insulin Receptor Kinase Preferentially Added Phosphate to the 1150 Tyrosine We next wanted to examine the purified beta subunit of the human insulin receptor in our new assay. As noted in Figure 3.6.1, we determined that prefers to add phosphate to tyrosine 1150. That is, the purified beta subunit ofthe insulin receptor preferred to add phosphate to the 1150 (not the 1146 or 1151) position. Next preferred to add phosphate to 1151 then to 1146. In these experiments the regulatory peptide without any phosphates on it (lRO) was added to the human insulin receptor (BIRK) and this produced a new peak on our HPLC gradient corresponding to IR9 (the peptide with tyrosine 1150 phosphorylated) Relative to phosphorylation of tyrosine 1150, the insulin receptor adds less phosphate to tyrosine 1151 and then less yet to tyrosine 1146. It is well known (White 1990) that the human insulin receptor autophosphorylates tyrosine 1150 (and other sites) and that this stimulates the receptor kinase activity. Does metformin change the ability of the human insulin receptor to phosphorylate tyrosine 1150? In other words, does metformin change the preference of the human insulin receptor to phosphorylate this 1150 to 1146? Our data (Figure 3.6.1) demonstrate that metformin does not change the preference of the insulin receptor. Thus we do not suggest that metformin alters the substrate specificity for the human insulin receptor. 45

PAGE 56

Figure 3.6.1 PIRK Preferentially Phosphorylated Tyrosine 1150 of the Regulatory Domain. Metformin had no significant effect on the site specificity of For each sample, N=4. 6.4 Q) 6.2 '0 .Q. 8. 6.0 o .t: C. tn 0.2 Q. C) :l 0.1 u .... t:'1 Control Metformin 46

PAGE 57

3.7 Metformin Stimulated Tyrosine Kinase Activity in PMCs We also examined whether metforrnin could stimulate the phosphorylation of the insulin receptor peptide in a system containing both tyrosine phosphatase and tyrosine kinase activities (the PMC preparation). Thus, we switched from PIRK to PMCs for these experiments. As previously noted, increased levels of phosphorylation of the regulatory peptide would be associated with increased insulin receptor activity in an intact cell. To be able to simultaneously record both tyrosine kinase and phosphatase activities, we used the peptide with one phosphate at the 1150 position (lR9). After incubation with 15 PMCs, we examined the appearance of peaks representing peptides with two phosphates (kinase action ; this section) or without phosphate (phosphatase action; next section). In the absence of metforrnin, we noted that phosphate was added to the IR9 peptide at a rate of 0.18 J.lg/40 minll5 PMCs (see "control" bar in Figure 3 7.1). Metformin significantly increased the rate of a second phosphorylation of this insulin receptor peptide by 31 %. So, metforrnin stimulates tyrosine kinase activity that phosphorylates the first tyrosine (1150) by 11% (Figure 3.5.1) and metforrnin stimulates the addition of a second phosphate to the regulatory peptide by 31 % (Figure 3.7.1) 47

PAGE 58

QI 0.30 ..... QI 025 QI ..... 0 20 o .c Q. o -a 0 .15 .,Q Oll = -o -.. = o u O. 10 Figure 3.7.1 Metformin Stimulated Tyrosine Kinase Activity in PMCs. Fifteen PMCs were incubated 40 minutes with 200 IR9 phosphopeptide. Peak areas of all three bisphosphorylated peptides were combined. N=6. An asterisk indicates P less than 0.05. 48

PAGE 59

3.8 Metformin Inhibited Tyrosine Phosphatase Activity that Removes Phosphate from the Insulin Regulatory Peptide Our sample analysis of the last set of experiments (Figure 3 7 .1) also recorded the amount of IR9 was dephosphorylated to IRO. PMCs and insulin receptor phosphopeptide (IR9) were added with or without 10 /-lg/ml metformin. With time, the insulin receptor peptide without any phosphate (IRO) increased in amount (Figure 3 8 1 ) Metformin inhibited total tyrosine phosphatase activity in PMCs by 100% in our pre v ious experiment (Figure 3.1.1). Total tyrosine phosphatase activity would include those phosphatases that act upon the insulin receptor and others presumably not related to insulin signaling As noted in Figure 3.3.1 metformin in the presence ofPMCs inhibited human PTPIB (a known regulator of the insulin receptor) by approximately 50%. In this new experiment, metformin inhibited the appearance of the dephosphorylated form of the peptide that would be associated with an inactive insulin receptor (see bar on the right side of the Figure 3.8.1). In Figure 3.8.1 (next page) metformin inhibited (by 50%) a tyrosine phosphatase located in the PMC, that specifically removes phosphate from the human insulin receptor regulatory domain. This inhibition of the removal of phosphate from the receptor regulatory domain would stimulate the insulin receptor kinase activity (and thus, fight diabetes). 49

PAGE 60

These last results are strong support for our model of action for metfonnin since it utilizes the human insulin receptor peptide 50

PAGE 61

2.0 t .... c. 1.5 C. t .... -1.0 0 J: C. '11 0 J: C. 0.5 = := := 0.0 Figure 3.8.1 Metformin Inhibited Tyrosine Phosphatase Activity Toward the Regulatory Peptide. In these experiments, the monophosphorylated IR 9 was added to 15 PMCs (containing endogenous tyrosine kinases and tyrosine phosphatases). After 40 minutes the peptide (lRO) in the supernatant was measured N=6, asterisk denotes P less than 0 .05. 51

PAGE 62

4. Discussion The work of Stith et aI. (1996) reported that metformin may fight diabetes via its ability to stimulate insulin receptor tyrosine kinase activity. It was also suggested that metformin stimulates receptor tyrosine kinase activity in response to an increase in phosphorylation on the regulatory domain of the receptor (Stith et aI., 1996) In this present work we have shown that metformin inhibits a tyrosine phosphatase (Figure 3.8 .1) that removes phosphates from a peptide corresponding to the regulatory domain of the human insulin receptor. Thus, we have evidence that metformin fights diabetes (and stimulates insulin receptor tyrosine kinase) through the inhibition of tyrosine phosphatase activity. Although very different methods were used, the in-vitro stimulation of by metforrnin (Figure 3.5.1) (15%) was similar to that previously published (25%) (Stith et aI., 1998). However, the increase in tyrosine kinase activity (31 %) in PMCs (Figure 3.7.1) was vastly lower than the 300% increase in PMC tyrosine kinase activity reported by Stith et aI. (1996). This difference is not yet explained; however, the fact that different substrates (regulatory peptide IRO used here: RR-Src used in Stith et al., 1996) were used may account for the discrepancy. The insulin receptor peptide should be largely phosphorylated by the insulin receptor, however many tyrosine kinases will phosphorylate by RR-Src. Since 52

PAGE 63

metfonnin stimulates other tyrosine kinases besides the insulin receptor (Stith et aI., 1998) this may explain why the use of RR -Src recorded a larger metformin stimulation. That is the insulin receptor may have a relatively higher affinity for the regulatory peptide: thus use of the regulatory peptide may quantify insulin receptor kinase activity as opposed to other tyrosine kinases. Thus Metformin may stimulate the insulin receptor by about 30%, but the drug may stimulate another tyrosine kinase by about 250%. In addition RR-Src should have many more phosphorylation sites as compared with regulatory peptides. Previous work by Chavanieu et al (1992) indicated that the regulatory peptide IRO can act as a competitive inhibitor of IRK and PTP-1 B. High concentrations (2 mM; compared to 200 IlM in our assays) of the nonphosphorylated peptide inhibited insulin receptor autophosphorylation some 70%. The trisphosphorylated peptide inhibited phosphatase activity toward the receptor. Although the monophosphorylated form IR9 has not previously been indicated to have effects as a competitive inhibitor it is possible (if not likely) that the phosphopeptide is effecting rates of auto phosphorylation of the IRK. Thus, the substrate we used may inhibit the tyrosine kinase activity and negate some of met form in s effects. We will search for this possible inhibition in future experiments. Assays of tyrosine phosphatase activity in PMCs revealed a strong inhibition of phosphatase activity (100%) by metformin (Figure 3 1.1); however this result may 53

PAGE 64

be slightly misleading because there may be an A TP regeneration system present in the PMCs that would remove phosphate. Thus, free phosphate (measured in this assay) that is produced by the action of a phosphatase may then be recycled by enzymes present in the PMC and artificially lower the free phosphatase measured. Thus, inhibition of tyrosine phosphatase by metformin is probably less than 100%. Our other phosphatase assays did not measure free phosphate, thus free phosphate removal would not artificially alter the results of the other assays. Two more direct assays of phosphatase activity using HPLCIUV and pNPP both indicated that metformin inhibited PTP-IB phosphatase activity by about 50% (Figure 3.3.1). 4.1 Does Metformin Act Directly or Indirectly on PTP-IB? We have two lines of evidence indicating that metformin inhibits tyrosine phosphatase activity through an intermediate. First, salt washed PMCs (where the putative intermediate, but not tyrosine phosphatase, would be removed) are unresponsive to metformin (Figure 3.1.1). Second, there is a lack of a direct action by metformin on PTP-IB (Figure 3.2.1), but the addition ofPMCs (not salt washed, thus containing the supposed intermediate) rescues this inhibition (Figure 3.3.1). Additional, future studies will attempt to identify the intermediate. We will determine whether PTP-l B is inhibited via dephosphorylation and if metformin can stimulate this dephosphorylation. We have not yet examined whether metformin can 54

PAGE 65

stimulate a phosphatase that removes phosphate from PTP-IB (this would stimulate PTP-IB activity) Another set of experiments that we propose is to see if metformin stimulates a protein that then binds to PTP-IB and this binding would inhibit the tyrosine phosphatase. A third possibility would be if the intermediate within the PMCs is a kinase. That is, metformin would stimulate a protein kinase that would then phosphorylate and inhibit PTP l B. However when we reproduced the conditions shown in Figure 3.3 1 in the presence of radioactive 32p ATP we did not detect any metformininduced increase in radioactive 32p on PTP-l B (results not shown). That is autoradiographs of polyacrylamide gel-purified PTP-l B indicate that PTP-l B is not phosphorylated in the presence of metformin. However PTP-l B was poorly separated from the bovine serum albumin in the assay buffer (both have molecular weights of 65 kDa) and these experiments should be repeated without BSA to decrease the background radiation at the 65 kDa bands. 4.2 Comparison with a Known Tyrosine Phosphatase Inhibitor Vanadium We also briefly examined whether metformin inhibits tyrosine phosphatases in a manner similar to vanadium. Our results (not shown) indicate that Xenopus 55

PAGE 66

phosphatases were also inhibited by vanadium compounds like human phosphatases. However, since our evidence suggests that metformin acts through a mechanism dependent upon an unknown intermediate found in the PMCs, we believe that the mechanism used by metformin must be different from that used by vanadium. That is, as opposed to metformin vanadium acts directly upon the phosphatase, can inhibit in-vitro PTP-IB activity and does not require an intermediate for this inhibition 4.3 Other Future Studies There are several studies that need to be performed. We would like to analyze the effect of metformin on PIRK activity with the monophosphorylated (IR9) peptide, and on PMC tyrosine kinase activity using IRO. These experiments would serve to compliment the PMC experiments with IR9 (Figure 3.8.1) and the PIRK experiments with IRO (Figure 3.7.1). Addition ofPIRK and IR9 would examine the rate of second site phosphorylation as a function of metformin The future experiment with PMCs and IRO may lend insight. As previously mentioned, the incubation ofPMCs with IR9 resulted in low level metformin stimulation of tyrosine kinase activity as compared with prior PMC experiments (using RR-Src as the substrate) (31 % versus 300%) (Stith et aI., 1996). Depending upon whether IRO acts as an inhibitor or a better (more phosphorylation possible with 56

PAGE 67

IRO than with IR9) substrate, assaying the tyrosine kinase activity in the PMC with IRO might result in a lesser or greater stimulation by metformin. In addition, to determine possible inhibitory effect ofthe receptor peptides on PMC kinase activity, we would also like to assay insulin stimulated PMCs in the presence of IR9 or IRO. In these experiments, RR-Src would be used as a substrate for tyrosine kinase activity, or we might measure insulin receptor phosphorylation (via Western blot) while IRO and IR9 would serve as potential inhibitors. We will also micro inject IRO or IR9 into whole oocytes to see if they can inhibit insulin induced meiosis. Additionally, we would like to determine the effect of metformin on PTP-l B activity using the trisphosphorylated form of the receptor peptide. This experiment would allow us to determine ifmetformin affects PTP-IB specificity (which sites are preferred?) as well as effects on overall activity. We still need to show that metformin increases the phosphorylation state of the insulin receptor regulatory domain in an intact insulin receptor in a whole cell. This can be done via two techniques. The first step is to allow autophosphorylation to occur in the presence of32p ATP, thus allowing the regulatory domain to be phosphorylated with radiolabeled phosphate. We would then purify the insulin receptor via wheat germ affmity, immunoprecipitation or similar techniques. Then, the receptor must be cleaved with trypsin to form the same regulatory peptides used 57

PAGE 68

in our HPLC analyses. The final step is to separate by HPLC and quantify (via liquid scintillation counting) the amount of phosphate on the insulin receptor regulatory peptides. Instead of peptide separation by HPLC, we may use two-dimensional thin layer electrophoresis. Unfortunately both methods have failed (so far) to yield good results because of poor separation. 58

PAGE 69

5. Conclusion In summary, we have evidence that metformin fights diabetes by acting through an intermediate to inhibit PTP-IB. Inhibition ofPTP-lB would block phosphate removal from the insulin receptor and stimulate receptor tyrosine kinase activity. 59

PAGE 70

REFERENCES Anderson, Kenneth N. Mosby's Medical, Nursing, and Allied Health Dictionary. St. Louis: Mosby-Year Book, 1994. Bevan, P., (2001). Insulin signaling. Journal of Cell Science 114(8): 1429-30. Bjorntorp P, (2000). Metabolic implications of body fat distribution. Diabetes Care 14: 1l32-1143. Bjorntorp P, Rosmond R, (1999) Hypothalamic origin of the metabolic syndrome X. Ann NY Acad Sci 892: 297-307. Chavanieu, A." Calas B., Vaglio, P., Grigorescu, F., (1992). Kinase inhibition by a phosphorylated peptide corresponding to the major insulin receptor autophosphorylation domain. Eur J Biochem 208(2): 367-373. 60

PAGE 71

Cheung, A., Kusari, J., Jansen, D., Bandyopadhyay, D., Kusari, A., Bryer-Ash, M., (1999). Marked impairment of protein tyrosine phosphatase 1 B activity in adipose tissue of obese subjects with and without type 2 diabetes mellitus. J Lab Clin Med 134(2): 115-23. Cohen, P., (1999). The croonian lecture 1998. Identification of a protein kinase cascade of major importance in insulin signal transduction. Philos Trans R Soc Lond B BioI Sci 354(1382): 485-95. Elchebly, M, et aI., (1999): Increased insulin sensitivity and obesity resistance in mice lacking the protein tyrosine phosphatase-l B gene. Science 283(5407): 1544-8. Hansen PC, (1999). The metabolic syndrome X. Ann NY Acad Sci 892:1-24. Margolis, Simeon, and Christopher Saudek. The Johns Hopkins White Papers: Diabetes Mellitus. Baltimore: The Johns Hopkins Medical Institutions, 2002. Meir, R., Hemmings, R. A., (1999). Regulation of protein kinase B. Receptor Signal Transduction Research 19( 1-4): 121-8. 61

PAGE 72

Peeke PM, Chrousos GP, (1995) Hypercortisolism and obesity. Ann NY Acad Sci 774: 665-76. Pellegrini, M. C., Liang H., et aI., (1998). Mapping the subsite preferences of protein tyrosine phosphatase PTP-IB using combinatorial chemistry approaches. Biochemistry 37 (45): 15598-606 Prager, R., and Schemthaner J., (1983). Insulin receptor binding to monocytes, insulin secretion, and glucose tolerance following metformin treatment. Results of a double-blind cross-over study in type II diabetics. Diabetes 32( 12): 1 083-6. Sadler, S.E., and Maller, J.L., (1981). Progesterone inhibits adenylate cyclase in Xenopus oocytes: action on the guanine nucleotide regulatory protein Journal of Biological Chemistry 256: 6368-6373. Salmeen, A., Andersen, J., Myers, M.P., Tonks, N.K., Barford, D., (2000). Molecular basis for the dephosphorylation ofthe activation segment of the insulin receptor by protein tyrosine phosphatase 1 B. Molecular Cell 6: 1401-1412. 62

PAGE 73

Sredy, 1., Sawicki D.R., Flam, B R., Sullivan, D., (1995) Insulin resistance is associated with abnonnal dephosphorylation of a synthetic phosphopeptide corresponding to the major autophosphorylation sites of the insulin receptor Metabolism 44(8): 1074-81. Stith, B. J Woronoff K. and Wiernsperger, N., (1998). Stimulation of the intracellular portion of the human insulin receptor by the antidiabetic drug metfonnin. Biochemical Pharmacology 55 : 533-536. Stith B. J., Goalstone M. L., Espinoza R. Mossel, C. Roberts D. and Wiemsperger N., (1996) The antidiabetic drug metfonnin elevates receptor tyrosine kinase activity and inosital 1,4,5-trisphosphate mass in Xenopus oocytes. Endocrinology l37: 2990-2999. Tao, J., Malbon, C.C., Hsien-yu W (200la). Insulin stimulates tyrosine phosphorylation and inactivation of protein-tyrosine phosphatase IB in Vivo The Journal of Biological Chemistry 276(31): 29520-25 63

PAGE 74

Tao, J., Malbon, e.e., Wang, H., (2001b). Gai2 enhances insulin signaling via suppression of protein-tyrosine phosphatase 1 B. The Journal of Biological Chemistry 276(43): 39705-12. Vigneri, R., Gullo, D., Pezzino, V., (1981). Metformin and insulin receptors. Diabetes Care 7 Suppl 1: 113-7. White M.F., (1990). The insulin receptor tyrosine kinase. Peptide Hormone Action 223-250. Zabolotny JM, Bence-Hanulec KK, Stricker-Krongrad A, Haj F, Wang Y, Minokoshi Y, Kim YB, Elmquist JK, Tartaglia LA, Kahn BB, Neel BG., (2002). PTP1B regulates leptin signal transduction in vivo. Developmental Cell 2(4): 489-95. Zimmet P, et aI., (1999). Etiology of the metabolic syndrome X: potential role of insulin resistance, leptin resistance and other players. Ann NY Acad Sci 892: 2544. 64