0021-972x196/$03.00/0
Journal
of Clinical
Endocrinology
and Metabolism
Copyright
0 1996 by The Endocrine
Society
Vol. 81, No. 6
Printed
an U.S.A.
The Human
Insulin
Analog Insulin
Lispro Improves
Insulin
Binding
on Circulating
Monocytes
of Intensively
Treated Insulin-Dependent
Diabetes
Mellitus
Patients*
PETER M. JEHLE,
ROLF D. FUSSGAENGER,
ULRIKE
KUNZE,
MANFRED
DOLDERER,
WOJCIECH
WARCHOL,
AND
IRMTRAUT
Department
Department
University
KOOP
of Internal Medicine,
University of Ulm (P.M.J., R.D.F., U.K., M.D.), Ulm; and Charitk,
of Internal Medicine,
Humboldt-Universitn,
of Berlin (I.K.), Berlin, Germany; and
of Posen (TV. W.), Posen, Poland
ABSTRACT
The rapidly
absorbed
analog of human
insulin,
insulin
lispro (LPI,
is characterized
by a faster onset of action, a higher peak insulin
level,
and a shorter
duration
of action compared
with regular
insulin
(RI).
The aim of this study was to investigate
whether
intensified
treatment with either LP or RI influences
insulin
receptor
status. Twelve
patients
with insulin-dependent
diabetes
mellitus
(IDDM)
participating in a multicenter
randomized
cross-over
trial were allocated
to
this study.
Four patients
began with LP, whereas
eight patients
started with RI. Each patient
was switched
to the other insulin
after
a S-month
treatment
period.
Competitive
[1251]A-14-insulin
binding
studies
were performed
with isolated
monocytes.
Treatment
with
insulin
lispro increased
the total number
of insulin
binding
sites from
9,400 2 2,200 (RI) to 20,300 ? 3,000 (LPYmonocyte
(P < 0.001). The
insulin
concentration
required
for a 50% competition
of [1251]insulin
binding
(IC,,) decreased
from 0.6 i 0.2 (RI) to 0.1 2 0.03 (LP) nmol&
indicating
significantly
higher affinity
of insulin
binding
sites during
LP treatment
(P < 0.001). In additional
experiments,
the time course
of insulin
binding
was determined
after an oral meal. In LP-treated
IDDM
patients,
the affinity
and capacity
of insulin
binding
showed a
nadir 1 h after insulin
injection
and a regained
binding
affinity
and
capacity
5 h later. These changes observed
after LP treatment
were
comparable
to the effect of endogenous
insulin
secretion
in healthy
control
subjects.
In contrast,
the IDDM
patients
who injected
RI
showed a decreasing
insulin
binding
affinity
and capacity,
most markedly expressed
after 5 h. The corresponding
serum levels of insulin
were inversely
correlated
with the affinity
and capacity
of insulinbinding
sites. Pretreatment
of cultured
human
IM-9 lymphoblasts
with LP or RI yielded no difference
in the down-regulation
of insulin
binding.
In summary,
intensified
conventional
insulin
therapy
with LP
increased
the number
and affinity
of insulin
receptors
on circulating
monocytes
to a level similar
to that observed
in healthy
subjects.
We
conclude
that the improved
insulin
receptor
status observed
during
LP treatment
is caused by its more physiological
pharmacokinetic
profile.
(J Clin Endocrinol
Metab
81: 2319-2327,
1996)
T
HE TREATMENT of diabetes mellitus with intensified
insulin therapy, i.e. administration of intermediate or
long acting (basal) insulin once or twice daily and preprandial injections of fast acting (regular) insulin, can only partially mimic the physiological insulin response. The major
problem of commercially available regular insulin formulations is delayed absorption, resulting in a later onset and a
prolonged duration of action. In combination with basal
insulin, this often contributes to hypoglycemia 3-4 h after a
meal, forcing the patient to eat a snack between major meals.
Delayed absorption is due to the fact that regular human
insulin (RI) contains hexameric insulin complexes. When
injected SC,these insulin hexamers have to dissociate into
dimers and monomers for absorption. Therefore, depending
on the dose, the peak effect of RI occurs from 2-6 h after
injection and lasts aslong as 16 h (1). The new insulin analog,
insulin lispro (LP), has been designed to overcome these
problems. LP differs from regular human insulin with reReceived
May 12, 1995. Revision
received
December
18, 1995. Accepted January
16, 1996.
Address
all correspondence
and requests
for reprints
to: Rolf D.
Fussgaenger,
M.D., University
of Ulm, Department
of Internal Medicine,
Robert-Koch-StraiJe
8, 89081 Ulm, Germany.
* This work was supported
by the Landes-Forschungsschwerpunkt
Baden-Wiirttemberg:
Spltkomplikationen
bei Diabetes
mellitus
and a
grant from the Deutsche
Forschungsgemeinschaft.
spect to an inversion of the amino acids 28 (Pro) and 29 (Lys)
in the B chain of the insulin molecule, resulting in a rapid
dissociation of hexamers directly to monomers and a reduced affinity for reassociation (2, 3). Recent studies have
shown that LP is absorbed more rapidly after SCinjection,
with a faster onset of action, a higher peak insulin level, and
a shorter half-life (4). In vitro studies with LP showed no
differences in receptor binding and [‘4C]glucose uptake compared with human insulin (5-7).
It has been shown that insulin has a direct effect on the
number and affinity of insulin receptors. IYI vitro studies
demonstrated down-regulation of insulin receptors by insulin (8). Insulin resistance in obese patients with noninsulindependent diabetes mellitus (NIDDM) is correlated with
high serum insulin levels and a reduced number of insulin
receptors (9). Thus, insulin resistance occurring during insulin treatment may be partially due to hyperinsulinemiainduced down-regulation of insulin receptors. The more
transient serum insulin concentrations during LP treatment
may, therefore, improve insulin receptor status.
The aim of this randomized 6-month cross-over study was
to investigate whether intensified treatment of IDDM patients with LP or RI, each given for 3 months, influences the
affinity and number of insulin receptors on circulating monocytes.
2319
JEHLE ETAL,.
2320
Subjects
and
Methods
Study design
All patients
participated
in a multicenter,
randomized,
cross-over
study to compare
the effect of a 3-month
treatment
period with LP to a
control period with regular human insulin. Inclusion
criteria were IDDM
according
to WHO criteria, prestudy
insulin therapy
of at least 2 months,
optimum
compliance
with diabetic
diet and insulin
therapy,
and informed consent in accordance
with local regulations.
Exclusion
criteria
were any severe disease (e.g. cancer; adrenal
insufficiency;
known
hemoglobinopathy;
chronic anemias;
or cerebrovascular,
peripheral
vascular, cardiac, renal, or liver disease), known
allergy to insulin, women
who were pregnant
or intended
to become pregnant,
women
who were
lactating,
continuous
SC insulin
infusion
treatment,
total daily dose of
insulin greater than 2.0 U/kg,
body mass index (BMI) greater than 35
kg/m*,
history
of clinically
significant
hypoglycemic
unawareness,
or
clinical signs or symptoms
of drug or alcohol abuse.
The study was approved
by the local ethical committees.
Patients and
control subjects gave informed
consent to participate.
For this study, 12
patients
(mean age, 30 t 8 yr; range, 21-49 yr; male/female,
9/3) with
insulin-dependent
diabetes mellitus
and 12 nondiabetic
subjects (mean
age, 32 t 2.5 yr; range, 24-49 yr; male/female,
6/6) were investigated.
The duration
of diabetes was 11 lr 2.8 yr (range, 2-39 yr). Ten patients
were C peptide
negative,
and 2 male patients,
who had their onset of
disease less than 5 yr previously,
showed a limited
C peptide
response
(0.5 and 1.7 ng/mL,
respectively)
to glucagon
(1 mg).
The insulin
formulations
were supplied
by Lilly Deutschland
(Bad
Homburg,
Germany).
By chance, four patients were randomized
to start
with LP, and eight patients to start with regular
insulin.
LP or RI was
given before each meal containing
more than 20% of the total daily
calories (25 Cal/kg).
Eleven patients used NPH insulin
twice daily to
cover their basal insulin requirements.
One male patient (duration
of
diabetes, 12 yr; C peptide negative)
injected 20 IU NPH insulin once daily
in the evening info The thigh and applied 8-16 IU RI or LP before meals
three times uer dav. Desoite the fact that one iniection
of NPH insulin
is very unusual,
this patient had optimal
diabetes control, with hemoglobin A,, (HbA,,)
values within the normal range. During
both treatment periods, diet and physical activity
were constant. Patients did not
receive any interfering
medications.
All patients were monitored
for insulin
binding
at the end of each
3-month
treatment
period and in some cases additionally
2-4 weeks after
the treatment
cross-over.
Twelve
age- and sex-matched
nondiabetic
subjects served as controls.
The blood samples
were obtained
under
out-patient
conditions
in the morning
during
the fasting state.
In additional
experiments,
we further
investigated
the time course of
insulin binding after an oral meal in IDDM
patients and healthy control
subjects. The six LP-treated
patients, recruited
from the cross-over
study,
were C peptide negative
and had been receiving
LP treatment
for 20 2
4 months (mean age, 34 + 4yr; range, 25-39 yr; male/female,
5/l; HbA,,,
6.7 t 0.3%). The six RI-treated
patients were also C peptide negative
and
had been receiving
RI treatment
for 10 2 3 yr (mean-age,
32 2 3yr; range,
28-32 vr: male/female,
3/3: HbA, -, 7.2 t 0.4%). The effect of endozenous
insulin
secretion after an oral meal was studied
in six normal
weight,
healthy
control subjects (mean age, 30 5 3 yr; range, 25-32 yr; male/
female, 3/3; HbA,,,
5.0 ? 0.3%). Insulin binding
on circulating
monocytes was measured
under fasting conditions
and 1 and 5 h after a
standardized
breakfast
(40 g carbohydrates,
28 g lipids, and 12 g protein). The IDDM
patients injected similar
doses of LP (7 t 3 IU) or RI
(8 -C 2 IU) immediately
before each meal. The doses of NPH injected at
bedtime
(2200 h) on the previous
day were also similar
(LP, 9 + 3 III;
RI, 10 2 2 IU). On the morning
of the study, no NPH insulin was injected.
HbA,,
serum lipids, BMI, and hypoglycemic
episodes
were monitored simultaneously
with the binding
assays. Hypoglycemia
was defined as any episode during which the patient felt (or was told by another
observer)
a sign or symptom
that she/he associated
with hypoglycemia
regardless
of whether
a blood glucose value was obtained
or what the
value of such a reading was. On scheduled
blood glucose tests, regardless of signs or symptoms,
values below 3.5 mmol/L
were defined
as
hypoglycemia.
Insulin requirements
were determined
during the last 24
h before the binding
studies. To determine
free circulating
insulin,
we
used a RIA (CIS, Dreieich,
Germany)
that also recognizes
LP (LP standard, Lilly Deutschland).
I
I
JCE
[125111nsulin
binding
& M . 1996
Vol81*No6
assay
Peripheral
venous
blood (40-50 mL) was collected
in two 25-mL
syringes
each containing
1000 IU sodium
heparin.
Mononuclear
cells
were-separated
by Ficoli-Hypaque
density
gradient
sedimentation
accordine
to the method
of Bovum
(10). The mononuclear
cell laver was
removid
and diluted
in HEI’ES assay buffer (50 mmol/L
HEI’ES, 10
mmol/L
dextrose,
15 mmol/L
sodium
acetate, 5 mmol/L
potassium
chloride,
120 mmol/L
sodium chloride,
1.2 mmol/L
magnesium
sulfate,
10 mmol/L
calcium chloride,
and 0.1% BSA) to a final concentration
of
lo7 mononuclear
cells/ml.
Viability,
as assessed by trypan
blue exclusion, was always
greater
than 95%. To correct
the binding
data per
monocyte,
the percentage
of monocytes
in the final mononuclear
preparation
was determined
by morphological
criteria
in cytocentrifuge
smears stained with May-Griinwald/Giemsa.
In all experiments
the
percentage
of monocytes
ranged from 15-30%, whereas
the variability
of monocyte
number
was small when a given individual
was studied on
separate
*occasions.
To confirm
the accuracy
of the morphological
method.
flow cvtometrv
(EPICS-VCS.
Coulter.
Krefeld,
Germanv)
was
used to’distinguish
monocytes
(CD14positive
cells) from lymphocytes
and granulocytes.
Both methods
yielded
similar results.
Competitive
binding
studies were performed
under equilibrium
conditions at 15 C and pH 7.8 using 10 pmol/L
[‘251]Tyr-A-14-insulin
tracer
(SA, 360 mCi/mg;
labeled
by the chloramine-T
method
and subsequently
purified
by high pressure
liquid chromatography
by Dipl.-Ing.
A. Liebe, Hoechst,
Frankfurt,
Germanv)
in the absence or presence of
unlabeled
insulin over a range of insulin concentrations
from10
pmol/L
to 0.1 pmol/L
(values
determined
in quadruplicate).
After 90 min of
incubation,
replicate
0.2-mL
aliquots
of the incubation
mixture
were
transferred
to microtubes
containing
0.1 mL dibutyl-dinonyl-phthalate
and centrifuged.
The microtubes
were cut in the ester phase to separate
the cell pellet from supernatant.
Bound and free radioactivities
were
counted in a y-counter
(Berthold,
Munich,
Germany).
Under these conditions,
nonspecific
binding,
defined
as the amount
of [‘2511insulin
bound to the cell pellet in the presence of 0.1 pmol/L
unlabeled
insulin
was less than 1% of the total radioactivity.
The binding
specificity
of [ ‘2511insulin
to insulin receptors
on monocytes was further
determined
by the following
methods.
1) By means of
flow cytometry,
we found specific insulin receptors
only on monocytes,
whereas
B and T lymphocytes
lacked significant
binding.
These results
are in agreement
with previous
reports (11). 2) The monoclonal
antibody
MA-20 (12), directed against the o-subunit
of the human insulin receptor
(Amersham-Buchler,
Braunschweig,
Germany),
inhibited
insulin binding up to 85%, whereas
oIR-3 (131, directed
against the o-subunit
of the
human insulin-like
growth
factor I (IGF-I)
receptor
(Oncogene
Science,
Uniondale,
NY), did not affect insulin binding.
For in vitro binding
studies we used IM-9 cells, an established
human
lymphoblastoid
cell l&e (8). Cells were grown in continuous
suspension
culture
in RPM1 1640 medium
(Biochrom,
Berlin, Germanv)
supulemented
with 25 mmol/L
HEPES buffer,
10% FCS (Seromed,
M&&h,
Germany),
200 mmol/L
glutamine,
10,000 E/mL
penicillin
G, and 10
mg/mL
streptomycin.
Medium
was changed
every 3 days.
Presentation
of binding
data
In competition-inhibition
experiments,
the percentage
of total radioactivity
that was specifically
bound was plotted as a function
of the log
total insulin
concentration.
Receptor
affinity
was analyzed
by a selfdeveloped
computer-assisted
curve-fitting
program
based
on the
Cheng-Prusoff
relationship
(14). Further
improvement
of data fitting
was obtained
by adaptation
to binding
of multiple
ligands according
to
McGonigle
et al. (15). A one-site model was used to calculate
the apparent ajfinities
of insulin
receptors
in all experiments,
i.e. the one-site
model (I&,, mol/L):
B/B_ = (B/B, X IC,,)/(IC,,
+ I), where B is bound
hormone,
B, is maximal
bound hormonz,
I is mhibitor
(unlabeled
hormone), and IC,, is the inhibitor
concentration
for half-maximal
tracer
competition.
However,
in most instances,
data for insulin binding
were
fitted with higher
statistical
significance
when a two-site
model was
used to calculate
binding
affinity;
apparent
affinity
(IC,,)
was then
subdivided
into a high affinity
(IC,,,, ) and a low affinity
(IC,,,,)
binding
site, i.e. the two-site
model (ICsa,i vs. ICsO,z, mol/L):
B/B, = [(Bl/B,
X
IC50,1)/IC50,1
+ I)] + I(B2/B,
X IC,,,,/IC,,,,
+ I)]. By plotting
the
bound/free
ratio of [‘251]insulin
(B/F) as a function
of bound hormone
““,
INSULIN
LISPRO IMPROVES
(8) according
to the method
of Scatchard
(16), curvilinear
plots were
obtained.
The total binding
capacity
(R, = total bound)
was derived
from the point where the linear extrapolation
of the curve intersects the
horizontal
axis. The number of insulin-binding
sites was normalized
per
monocyte.
By performing
binding
studies with various
concentrations
of mononuclear
cells from the same donor, we found a linear relationship between
the numbers
of cells and receptors.
To standardize
our
binding
assay, we used a final concentration
of lo7 mononuclear
cells/
mL.
Statistical
analyses
All results are expressed
as the mean t SEM. The Wilcoxon
test was
used when data from all 12 patients were analyzed.
Student’s
t test for
unpaired
data was used for analysis of smaller subgroups.
Correlations
between variables
were assessed using univariate
linear regression
analysis. P < 0.05 was accepted as statistically
significant.
Insulin binding during
(cross-over study)
Results
treatment
with RI or LP
At the end of each 3-month treatment period, all patients
were monitored for [‘251]insulinbinding on isolatedmonocytes.
In individual cases,additional binding studieswere performed
in the first 24 weeks after the cross-overto investigate whether
changesin binding could be observed earlier. The results during eachtreatment period were not different whether binding
data were obtained during the first or the third month, or
whether LP or RI was applied first. Therefore, if not otherwise
indicated, the results are mean values for each patient during
the individual treatment period.
Competition-inhibition curves were analyzed using the computer-assistedcurve-fitting program noted above. IC5,,values
INSULIN
2321
BINDING
decreasedfrom 0.6 k 0.2 (RI) to 0.1 +- 0.03 (LPI nmol/L (P <
0.001).As lessunlabeled insulin was necessaryfor half-maximal
tracer displacement, the affinity of insulin receptors was significantly higher during LP treatment. Figure 1 shows the results in eachindividual patient. The I&, values obtained from
the 12 nondiabetic subjects(females,0.16 k 0.37;males,0.17 k
0.3 nmol/L) did not differ from those in the LP-treated group,
but were significantly lower than those observed during RI
treatment 8 < 0.01).
Due to the limited blood sample volume and the variation in the number of monocytes, we were forced to
perform an assay with only six concentrations of competitive insulin in 50% of the experiments, which were calculated by the one-site model. The other experiments with
more than seven data points additionally allowed the application of the two-site model with satisfying statistical
precision. The IC50,1 representing the high affinity state of
insulin binding increased significantly (P < 0.05) from 55.4
+ 25 (RI) to 1.4 t 1.4 pmol/L (LP). A slightly higher
affinity was observed during LP treatment compared with
F$ 2j
6
y++j
0.01
0.1
1
10
100
0.1
1
10
100
IO
zi
\
b
E
c
0
1
i?
0.1
0.0'
LP
0.01
FIG. 1. Apparent
affinity
of [12511insulin
binding
on isolated
monocytes from 12 IDDM
patients
that were randomized
for 3 months
of
therapy
using the human
insulin
analog LP or RI, then switched
to
the other insulin.
As less unlabeled
insulin
is necessary
for halfmaximal
tracer displacement
(IC,, value),
the affinity
of insulin
receptors is significantly
higher
during
LP treatment
(P < 0.001).
Total insulin (nmol/ L)
FIG. 2. Competition-inhibition
curves
of three
representative
patients (--,
LP; - - -, RI). Note the marked
shift to the left, as seen in
the first slope of the graphs representing
the increased
high affinity
insulin-binding
state during
LP treatment.
2322
JEHLE
JCE & M
Vol81
ET AL.
l
l
1996
No 6
0.02
A
a,
5
0.01
40,000
e
e
0
r
5
0
;;;
a,
.%
ul
0.05
0.04f3
.-i!
I
F
7
.a
.-c
5
E
c
30,000
20,000
10,000
n
0.15
0
RI
FIG.
4. Significant
monocyte
during
< 0.001).
o,:d
0
0.1
0.2
0
Insulin bound (ng/mL)
3. Scatchard
plots
- - -, RI). Note the higher
concentration
of insulin
FIG.
corresponding
to the data in Fig. 2 (--,
LP;
binding
affinity
(B/F) as well as the increased
bound (B) during
LP treatment.
that in nondiabetic subjects, who showed an IC50,1 of 16.8
+ 10 pmol/L. The changes of the low affinity state (IC50,2)
from 9.3 ? 2.2 (RI) to 5.5 t 2.7 nmol/L (LP) were not
significant (P = 0.18). Again, LP-treated patients showed
values similar to those of nondiabetic subjects (I&,,,, 5.6
+ 1.5 nmol/L).
The competition-inhibition curves of three representative
IDDM patients are shown in Fig. 2. The considerable shift to
the left in the first part of the graphs, most markedly expressed in graphs A and C, indicates the increased high
affinity binding state. The same data, given as Scatchard
plots, demonstrate not only a higher binding affinity (B/F),
but also an increased amount of insulin bound (B) during LP
treatment (Fig. 3). The total number of insulin-binding sites
increased significantly from 9,400 2 2,200to 20,300 + 3,000/
monocyte (P < 0.001). Figure 4 demonstrates the changes in
each individual patient. In nondiabetic subjects,the number
of binding sites averaged 22,500 5 3,50O/monocyte.
Time course of insulin
binding
after an oral
meal
We further investigated the relationship between insulin
binding and endogenous insulin secretion after an oral meal
increase
LP treatment
in
total
shown
LP
insulin
receptor
for each individual
number
patient
per
(P
in normal weight, healthy control subjectsand compared the
effects of LP or RI in IDDM patients. The differences in
insulin-binding capacity and affinity observed between LP
and RI treatments were comparable to those demonstrated in
the cross-over study.
The time course of binding capacity is shown in Fig. 5A.
Under fasting conditions, the numbers of insulin-binding
sites were comparable in nondiabetic controls (27,900 +
2,00O/monocyte) and in LP-treated IDDM patients (24,600
2 5,400), but this value was significantly lower in RItreated IDDM patients (10,200 -C 1,750; P < 0.001). One
hour after breakfast, binding capacity declined significantly in nondiabetic controls (22,400 + 2,300; P < 0.005)
and in LP-treated IDDM patients (14,650 + 2,200; P <
0.005), whereas no decrease was found in RI-treated IDDM
patients (9,500 C 1,000). Five hours after breakfast, the
binding capacity rose to values somewhat higher than
those under fasting conditions in controls (32,600 ? 4,300)
or after the injection of LP (30,300 5 4,000). In contrast,
RI-injected patients showed a nadir of blunted insulinbinding sites after 5 h (5,200 + 1,000; P < 0.05 IIS. fasting
values).
The time course of binding affinity is depicted in Fig. 5B.
Under fasting conditions, the affinity of insulin-binding
sites was not statistically different between nondiabetic
controls (I&,, 0.17 Ifr 0.04 nmol/L) and LP-treated IDDM
patients (IC,,, 0.11 2 0.02 nmol/L). Significantly decreased binding affinity was measured in RI-treated IDDM
patients (IC,,, 0.31 t- 0.03 nmol/L; P < 0.05). One hour
after breakfast, the binding affinity was diminished in
nondiabetic controls (IC,,, 0.4 2 0.09 nmol/L; P < 0.05)
and in LP-treated IDDM patients (IC,,, 0.42 ? 0.06
INSULIN
LISPRO IMPROVES
b
i?
INSULIN
BINDING
0.6
40,000
E
g
30,000
.5
5P
20,000
5. Time course of insulin
binding
capacity
(sites per monocyte;
A), insulin
binding
affinity
(IC,,; B), serum insulin
(C), and serum glucose (D) on circulating mono&es
under fasting conditions
(Oh) and I and 5 h after a standardized
breakfast
measured
in healthy
control
subjects
and IDDM
patients
injecting
similar
doses of LP or RI (8 + 2 IU)
immediately
before a meal without
any
use of NPH insulin.
Each subgroup
was
composed of six subjects of both gender.
2z
o
0.4
0.3
E
0
controls
controls
LP
1
nmol/L;
P < 0.05), whereas RI-treated IDDM patients
showed only a weak decrease (I&,, 0.38 2 0.08 nmol/L).
Five hours after breakfast, there was an increase in binding
affinity in nondiabetic controls (I&,, 0.18 +- 0.07 nmol/L)
and in LP-treated IDDM patients (I&,, 0.22 2 0.06 nmol/
L), regaining values comparable with the fasting condition. In contrast, in RI-treated IDDM patients, the lowest
binding affinity was observed 5 h after breakfast and inP < 0.05 zIs.
sulin injection (I&,, 0.53 2 0.11 nmol/L;
fasting values).
The levels of free circulating insulin are given in Fig. 5C.
In healthy controls and LP-treated IDDM patients, serum
insulin was highest after 1 h (P < 0.05 us. 0 h) and had
returned to baseline levels after 5 h (P < 0.05 ZJS.1 h). In
contrast, RI-treated IDDM patients showed 1.5-fold lower
insulin levels after 1 h, whereas after 5 h insulin was still as
high as after 1 h.
In both RI- and LP-injected patients, fasting glucose levels
were elevated (Fig. 5D). This is most likely due to the fact that
the prandial insulin was given 2 h later (0800 h) than usual,
and no NPH insulin was used during the study. Corresponding to the kinetic profiles of the insulin analogs, postprandial
glucose control was faster after injection of LP than of RI (Fig.
5D).
receptors
0.5
10,000
controls
In vitro down-regulation
of insulin
9 lymphoblasts
by RI or LP
i
.
3
E
s
L?
2
FIG.
2323
on human
IM-
To further examine whether the prolonged decrease in
insulin binding observed after the injection of RI might be
LP
In
LP
5h
00hIlhEtI!
controls
LP
I
RI
due to a pharmacodynamic action different from that of LP,
we studied the down-regulation of insulin receptors in vitro.
For this purpose we used IM-9 lymphoblasts that were pretreated for various time periods (1 min to 24 h) with both
analogs in various concentrations (1 pmol/L to 1 pmol/L).
The two analogs yielded comparable dose- and time-dependent down-regulations of specific insulin binding, with a
maximum at 1 pmol/L insulin added for 18 h, whereas IGF-I
was significantly less effective, corresponding to its known
lo- to loo-fold lower binding affinity to insulin receptors
(Fig. 6).
Clinical
data and relationship
with
insulin
binding
Table 1 summarizes the clinical data of this study. LP
treatment resulted in a significant decrease in basal insulin
dosesby 9% (P < 0.05). Prandial insulin, triglycerides, cholesterol, BMI, and hypoglycemic episodes were without significant differences in LP- or RI-treated patients. At the onset
of the study, 5 of 12 patients already had very low HbA,,
levels (6.3 ? 0.4%) and did not further improve with LP. In
contrast, the other 7 patients showed significantly lower
HbA,, values after 3 months of LP treatment (RI, 7.8 2 0.3%;
LP, 6.9 2 0.17%; P < 0.05).
Table 2 shows the linear regression analysis between insulin binding and clinical data (r = coefficient of correlation).
The serum levels of insulin were inversely correlated with
the number of insulin receptors (Fig. 7A: controls: r = -0.56;
P < 0.01; Fig. 7B: LP: r = -0.56; P < 0.01; Fig. 7C: RI: r = -0.5;
P < 0.05). Furthermore, a significant inverse correlation was
2324
JEHLE
lcontrol
=IGF-1
ClRl
EILP
50
25
0
FIG. 6. Down-regulation
of specific insulin
binding
by LP, RI, IGF-I,
or albumin
(control)
in human IM-9 lymphoblasts
(4 x 10” cells). After
incubation
for 18 h with 1 umol/L
of the indicated
peptides.
IM-9 cells
were washed
twice in HEPES
assay buffer
to remove
free insulin.
Binding
of [ 1251]insulin
was subsequently
investigated
as indicated
in
Subjects and Methods.
obtained between HbA,, and receptor number (RI: r = -0.7;
P < 0.005; LP: r = -0.74; P < O.OOS), whereas no correlation
was found with serum glucose. Thus, only long lasting hyperglycemia
may decrease the number of insulin receptors.
Only in the LP-treated
patients was a significant
direct linear
correlation
apparent between the receptor number and the
doses of prandial
insulin injected during 12 h (r = 0.86; P <
0.001) or 24 h (r = 0.77; P < 0.005).
Binding
affinity of insulin was impaired
with increasing
serum insulin levels. This is reflected by the direct correlation
of serum insulin with IC,, values (Fig. 7D: controls: r = 0.5;
P < 0.05; Fig. 7E: LP: r = 0.5; P < 0.05; Fig. 7F: RI: r = 0.66;
P < 0.01). As with insulin, a direct correlation
was observed
between serum glucose and IC,, values. The coefficients of
correlation
were lower for glucose than for insulin (RI: r =
0.4; P < 0.05; LP: r = 0.41; P < 0.05; controls: r = 0.5; P < 0.05).
Hence, at least statistically,
the binding
affinity of insulin
receptors is inversely correlated
with insulin and slightly less
pronounced
with glucose. Furthermore,
there was a direct
correlation
between
triglycerides
and IC,, values in RItreated patients (r = 0.84; P < O.OOl), indicating
lower binding affinity with increasing
serum triglycerides.
Discussion
This study demonstrates
monocytes
change in both
different insulin analogs are
intensified
insulin treatment
that insulin receptors on human
their affinity and number when
used to treat IDDM patients. The
regimen, the “gold standard”
of
ET AL.
JCE C M . 1996
Volt31 . No 6
insulin therapy today, may not completely
prevent an impairment
of insulin binding
when regular human insulin is
used to cover the increse in blood glucose after meal ingestion. In contrast, intensified
treatment
with the novel rapid
acting human insulin analog LP restores the capacity and
affinity of insulin binding
on circulating
monocytes of IDDM
patients to values measured
in healthy nondiabetic
subjects
(9, 11, 17). After only 14 days of LP treatment,
improved
insulin binding
was observed, which remained
stable when
it was restudied
after 3 months (cross-over
study) as well as
after 2 yr (study of postprandial
regulation
of insulin receptors).
Although
the broad advantages of intensified
insulin therapy with a highly significant
reduction
of diabetes-related
complications
have been clearly demonstrated
in several
studies (l&19), a critical issue of daily practice is the demand
that insulin absorption
from SC injection sites should peak as
closely as possible as glucose absorption
from the gut. Selfassociation
to hexamers of commercially
available insulins
delays the absorption
and onset of action, thereby inducing
a prolonged
duration
of action. To prevent postprandial
hyperglycemia,
most patients are recommended
to wait 30-60
min after insulin injection
before eating, depending
on the
preprandial
blood glucose level (20). On the other hand, as
regular
insulin
acts longer than the postprandial
glucose
excursion, a snack is often necessary between major meals to
prevent hypoglycemic
episodes. To overcome these disadvantages, a spectrum of insulin analogs has been developed
over the past years (21).
The human insulin analog LP is characterized
by a rapid
dissociation
into monomers
after SC injection, with a greatly
reduced capacity for reassociation,
leading to a more rapid
absorption
(2, 3). The serum concentration
of LP peaks at
more than twice the level and in less than half the time as
regular human insulin
(4). In our study we observed the
highest level of LP 1 h after SC injection,
which was about
1.5-fold higher than that of RI. Whereas after 5 h the serum
level of LP had returned
to baseline, the level of RI was still
2-fold higher.
The purpose of this study was to investigate
whether
the
different
pharmacokinetic
profile of LP may influence
the
binding
characteristics
of insulin receptors. Because the insulin receptors of human liver, adipose tissue, or muscle are
inaccessible
to direct study, we investigated
the insulin receptor of circulating
monocytes
exposed to fluctuations
in
serum insulin. In patients with syndromes
of insulin resistance, such as obesity and acanthosis nigricans, the degree of
impairment
of [iz51]insulin
binding
on circulating
monocytes
correlated
very closely with the severity of insulin resistance
(17, 22). It has been reported
that the levels of circulating
insulin were inversely
correlated
with the number and affinity of insulin receptors. By lowering
basal insulin levels by
hypocaloric
dieting or chronic fasting, the number of insulin
receptors on monocytes
could be restored
toward normal
values (22). In our IDDM subjects we observed no influence
of NPH-insulin
on insulin binding, which might be due to the
relatively
low doses injected SC into the thigh to achieve
maximally
delayed absorption.
On the other hand, the circulating levels of prandial
insulin were inversely correlated
with insulin binding
affinity and capacity. Only for LP was
INSULIN
TABLE
1. Clinical
data
of the
12 IDDM
patients
LISPRO IMPROVES
treated
with
insulin
(III/24
(W/24
(RI
h)
h)
resp.
LP; IU/24
TABLE
insulin
(mmoL’L)
(mmol/L)
2. Linear regression
binding
after an oral
analysis
meal)
between
binding
parameters
IC,,
IC,,
IC,,
IC,,
IC,,
IC,,
IC,,
IC,,
IC,,
IC,,
us.
us.
us.
us.
us.
us.
us.
us.
us.
us.
no.
no.
no.
no.
no.
no.
no.
no.
no.
no.
no.
us.
us.
us.
us.
us.
us.
us.
us.
us.
us.
us.
HbA,,
serum glucose
triglycerides
cholesterol
BMI
serum insulin
prandial
insulin
(III/24
h)
prandial
insulin
(IUI12
h)
basal insulin
(III/24
h)
basal insulin
(III/12
h)
IC,,
HbA,,
serum glucose
triglycerides
cholesterol
BMI
serum insulin
prandial
insulin
(W/24 h)
prandial
insulin
(W/12 h)
basal insulin
(III/24
h)
basal insulin
W/12
h)
a direct relationship apparent between the injected dosesand
the number of insulin receptors, which is most likely explained by the up-regulation of insulin binding sites5 h after
the injection of LP following the nadir at 1 h. Comparable
fluctuations in the affinity and concentration of insulin receptors were also observed for endogenous insulin in healthy
control subjects. In contrast, RI-injected patients showed a
nadir of blunted insulin-binding sites 5 h after injection. As
after injection of S-10 IU RI, insulin receptors were exposed
to insulin levels higher than 50 Q/mL for longer than 5 h,
injecting regular insulin three times daily is likely to promote
insulin receptor down-regulation.
From studies using IM-9 lymphoblasts, a well accepted in
vitro model of the human insulin receptor (81,it is known that
up-regulation of insulin receptors requires twice the time as
receptor down-regulation (8, 23). In our hands, the downregulation induced by LP in vitro was not different from that
causedby RI, according to previous studies with LP showing
no differences in receptor binding compared with regular
human insulin (5). Therefore, the delayed appearance of RI
after SCinjection and its prolonged presence in serum are
considered to decrease the capacity and affinity of insulin
binding.
One could speculate that the gender distribution of the
IDDM patients studied, their wide variation in duration of
diabetes, or residual secretion of C peptide may influence our
data. In a study with 32 IDDM and 26 NIDDM subjects of
both gender, aged 17-70 yr, insulin binding data were not
study
Insulin
% 4.0
k 2.0
2 4.9
k 0.3
24.6
20.4
45.0
6.9
lispro
+
k
2
k
1.9 ir 0.6
1.5 -+ 0.3
5.3 z 0.3
24.3 k 0.8
and clinical
data
(cross-over
study
P value
3.3
2.8
3.5
0.2
5.3 2 0.4
24.3 i- 0.8
Regular
Receptor
Receptor
Receptor
Receptor
Receptor
Receptor
Receptor
Receptor
Receptor
Receptor
Receptor
2325
cross-over
insulin
25.8
22.5
48.3
7.2
h)
HBA,, (%I
Triglycerides
Cholesterol
BMI (kg/m’)
BINDING
LP or RI in the randomized
Regular
Preprandial
NPH insulin
Total insulin
INSULIN
and
0.25
0.04
0.04
0.07
0.75
0.75
0.94
study
on the time
insulin
Insulin
course
lispro
r
r
r
r
r
r
r
r
r
r
r
=
=
=
=
=
=
=
=
=
=
=
-0.7,
-0.34,
-0.05,
-0.07,
0.09,
-0.5,
0.01,
0.32,
~0.34,
-0.18,
0.22,
P < 0.005
NS
NS
NS
NS
P < 0.05
NS
NS
NS
NS
NS
r
r
r
r
r
r
r
r
r
r
r
=
=
=
=
=
=
=
=
=
=
-0.74,
P < 0.005
-0.2
0.04, NS
0.04, NS
0.2, NS
-0.56, P < 0.01
0.77, P < 0.005
0.86, P < 0.001
0.17, NS
0.3, NS
0.25, NS
r
r
r
r
r
r
r
r
r
r
=
=
=
=
=
=
=
=
~0.05, NS
0.4, P < 0.05
0.84, P < 0.001
0.16, NS
-0.1, NS
0.66, P < 0.01
-0.25,
NS
-0.35,
NS
r
r
r
r
r
r
r
r
r
r
=
=
=
=
=
=
=
=
=
=
-0.05,
NS
0.41, P < 0.05
-0.08,
NS
-0.39,
NS
-0.08,
NS
0.5, P < 0.05
0.15, NS
0.32, NS
-0.19,
NS
0.1, NS
= -0.28,NS
= -0.23,
NS
=
of
correlated with the duration of diabetes, age, gender, menstrual cycle, or residual secretion of C peptide; however, an
inverse correlation was obtained between insulin receptor
number and HbA,, (data not shown).
In NIDDM patients, we demonstrated that insulin resistance was correlated with reduced binding affinity of insulin
receptors (24). With regard to the effect of higher insulin
concentrations, insulin resistancehas been reported not only
in NIDDM, but also in IDDM patients (25-27). Normal binding affinity and capacity of insulin receptors, attributed to
physiological pulsatile kinetics of insulin, are a prerequisite
to maintain insulin sensitivity. A recent study reported that
insulin oscillations promote a more efficient glucose utilization than constant insulin delivery (28). In the present study
the lower insulin requirement during LP treatment may reflect a higher sensitivity to insulin. Moreover, in the LPtreated IDDM patients, slightly lower levels of triglycerides
were measured. In the RI-treated patients, an inverse correlation was found between insulin binding affinity and triglyceride levels.
A further important determinant of insulin sensitivity is
glycemia itself. Direct proof of the ability of hyperglycemia
per seto induce insulin resistancehas been obtained in studies
in patients with IDDM (29,30), referred to as glucose toxicity
(31). As LP decreases glucose excursion from meals more
rapidly, subsiding glucose toxicity may additionally contribute to the improved insulin receptor binding observed during LP treatment. This is further underlined by the inverse
2326
JEHLE
JCE
ET AL.
OP....
Ol
0
Endogenous
20
30
Endogenous
@U/mL)
40
. . . . m...50
60
insulin
@U/ml)
30
40
70
40,000 -
Correlation of insulin binding
capacity (A-C) and affinity (D-F) on circulating monocytes with serum insulin
levels. Experiments
were performed
under fasting conditions (0 h) and 1 and
5 h after a standardized breakfast in
healthy control subjects (control), LPinjected IDDM patients (LPI, and RIinjected IDDM patients (RI). In all subgroups, the serum levels of insulin were
inversely correlated with the number of
insulin receptors (A-C) and with the insulin binding affinity (D-F). As higher
IC,, values indicate a lower binding affinity, a direct correlation was described.
FIG.
Insulin
10
& M l 1996
Vol81*No6
7.
0
10
20
Insulin
E
::
s
E
30
lispro
40
50
60
0
10
20
Insulin
&U/ml)
lispro
50
60
(N/ml)
C
15,000-
10
Regular
correlation
between serum glucose and binding
affinity, on
the one hand, and HbA,, and receptor numbers, on the other
hand. Therefore, elevated glucose levels for a short time may
rapidly decrease the affinity of insulin receptors, whereas a
longer time period of hyperglycemia
will be necessary to
change receptor number.
A large multicenter
trial concluded
that LP injected immediately
before the meal improved
postprandial
glucose
control as well as overall glycemic control compared
to RI
given at least 30 min before the meal (32). The purpose of the
cross-over study was to test the safety of LP in IDDM patients
and not to improve glycemic control in patients showing
an
unsatisfactory
diabetes control with RI. The frequency
of
hypoglycemic
episodes was not significantly
changed by LP.
Treatment
with LP did not improve
HbA,, levels in the
patients who already had very low values (6.3 t 0.4%; n =
20
insulin
30
f&l/mL)
40
0
10
Regular
20
30
Insulin
40
(,uU/mL)
5; normal, <6.1%). However,
the other patients showed significantly decreased HbA,, levels after 3 months of LP treatment. At the end of the cross-over study, 10 of the 12 patients
decided to stay on LP treatment.
In summary,
intensified
conventional
insulin
therapy
with LP increased
the number
and affinity of insulin
receptors on circulating
monocytes
to levels similar to those
observed in healthy subjects. Our experiments
support the
concept of an inverse relationship
between
long lasting
serum insulin
levels, on the one hand, and the lowered
number
and affinity of insulin-binding
sites, on the other
hand. We conclude
that the improved
insulin
receptor
status during
LP treatment
may be caused by the quick
pharmacokinetic
profile
of the analog.
The more rapid
absorption
of LP leads to a higher
peak insulin
level, a
faster onset of action, and a shorter duration
of action. This
INSULIN
LISPRO IMPROVES
resembles first phase insulin secretion in nondiabetic control subjects, preserving the high binding affinity and capacity of insulin receptors. Whether the improved insulin
receptor status correlates with long term benefit in terms
of complications in insulin-dependent diabetes mellitus
must be examined in further studies.
14.
15.
16.
Acknowledgments
We thank Mrs. Esther Ribber for her excellent technical
assistance in
performing
the binding
studies, Mrs. Rosa Herzog
for performing
the
RIAs, and Mrs. Kathy Ann Koskowicz
for aid in the preparation
of this
manuscript.
17.
18.
19.
References
1. Bhaskar
R, Chou MCY, Field JB. 1980 Time-action
characteristics
of regular
and NPH insulin
in insulin-treated
diabetics.
J Clin Endocrinol
Metab. 50:475479.
2. DiMarchi
RD, Mayer
JP, Fan L et al. 1992. Synthesis
of a fast-acting
insulin
based on structural
homology
with insulin-like
growth factor 1. In: Smith JA,
Rivier
JE, eds. Peptides:
chemistry
and biology,
Proceedings
of the Twelfth
American
Peptide
Symposium.
Leiden:
ESCOM;
26-28.
3. Ciszak
E, Beals JM, Frank BH, Baker JC, Carter ND, Smith GD. 1995 Role of
C-terminal
B-chain
residues
in insulin
assembly:
the structure
of hexameric
insulin.
Structure.
2615-622.
LYS B’XProB29-human
4. Howey
DC, Bowsher
RR, Brunelle
RL, Woodworth
JR. 1994 ILys(B28),
Pro
(B29)1-human
insulin.
A rapidly
absorbed
analogue
of human
insulin.
Diabetes. 43:396-402.
5. Slieker
LJ, Sundel
K. 1991 Modification
in the 28-29 position
of the insulin
B-chain
alters binding
to the IGF-1 receptor
with minimal
effect on insulin
receptor
binding
[Abstract].
Diabetes.
4O(Suppl
l):A168.
6. Shaw WN, Su KSE. 1991 Biological
aspects of a new human
insulin
analog:
[Lys(B28),Pro(B29)1-human
insulin
[Abstract].
Diabetes.
4O(Suppl
l):A464.
7. Radziuk
J, Davies
J, Pye R, DiMarchi
R, Shields
J, Chance
R. 1992 Subcutaneous absorption
of fast-acting
insulin
analogs:
kinetics
and bioeffectiveness
[Abstract].
Diabetologia.
35(Suppl
l):A3.
8. Gavin
III JR, Roth J, Nevil e
Jr DM, DeMeyts
P, Buell DN. 1974 Insulindependent
regulation
of insulin
receptor
concentrations.
A direct demonstration in cell culture.
Proc Nat1 Acad Sci USA. 71:84-88.
9. Olefsky
JM. 1976 The insulin
receptor.
Its role in insulin
resistance
of obesity
and diabetes.
Diabetes.
25:1154-1157.
10. BSyum A. 1968 Separation
of leukocytes
from blood and bone marrow.
Stand
J Clin Invest. 21:7?-89.
11. Schwartz
RH, Bianco
AR, Handwerger
BS, Kahn RC. 1975 Demonstration
that monocytes
rather than lymphocytes
are the insulin
binding
cells in preparations
of human peripheral
blood mononuclear
leukocytes:
implications
for
studies
of insulin
resistant
states in man. Proc Nat1 Acad Sci USA. 72:474-478.
12. Forsayeth
JR, Montemurro
A, Maddox
BA, DePirro
R, Goldfine
ID. 1987
Effect of monoclonal
antibodies
on human
insulin
receptor
autophosphorylation,
negative
cooperativity,
and down-regulation.
J Biol Chem. 262:41344140.
13. Jacobs S, Cook S, Svoboda
ME, Van Wyk JJ. 1986 Interaction
of the mono-
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
INSULIN
BINDING
2327
clonal antibodies
alphaIRand alphaIRwith insulin
and somatomedin-C
receptors.
Endocrinology.
118:223-226.
Cheng YC, Prusoff
WH. 1979 Relationship
between
the inhibition
constant
(Ki)
and the concentration
of inhibitor
which
causes 50 per cent inhibition
(IC50)
of an enzymatic
reaction.
Biochem
Pharmacol.
22~3099-3108.
McGonigle
P, Neve KA, Molinoff
PB. 1986 A quantitative
method
of analyzing
the interaction
of slightly
selective
radioligands
with multiple
receptor
subtypes.
Mol Pharmacol.
30:329-337.
Scatchard
G. 1949 The attractions
of proteins
for small molecules
and ions. Ann
NY Acad Sci. 51:660-672.
Bar RS, Gordon
P, Roth J, Kahn CR, DeMeyts
P. 1976 Fluctuations
in the
affinity
and concentration
of insulin
receptors
on circulating
monocytes
of
obese patients.
J Clin Invest. 58:1123-1135.
Reichard
P, Nilsson
BY, Rosenqvist
U. 1993 The effect of long-term
intensified
insulin
treatment
on the development
of microvascular
complications
of diabetes mellitus.
N Engl J Med. 329:304-309.
Diabetes
Control
and Complication
Trial Research
Group.
1993 The effect of
intensive
treatment
of diabetes
on the development
and progression
of longterm complications
in insulin-dependent
diabetes
mellitus.
N Engl J Med.
329:977-986.
Lean MJE, Ng LL, Tennison
BR. 1985 Interval
between
insulin
injection
and
eating in relation
to blood
glucose
control
in adult diabetics.
Br Med J. 290:
105-108.
Heinemann
L, Starke
AAR,
Heding
L, Jensen
1, Berger
M. 1990 Action
profiles
of fast onset insulin
analogues.
Diabetologia.
33:384-386.
Kahn CR, Flier JS, Bar RS, et al. 1976 The syndromes
of insulin
resistance
and
acanthosis
nigricans.
Insulin-receptor
disorders
in man. N Engl J Med. 294:
739-745.
Kosmakos
FC, Roth J. 1980 Insulin-induced
loss of insulin
receptor
in IM-9
lymphocytes.
J Biol Chem. 255:9860-9869.
Jehle PM, Ho11 RW, FuPglnger
RD, et al. 1995 Correlation
of insulin
resistance
in type-11 diabetes
with reduced
binding
affinity
of insulin
receptors
[Abstract].
Diabetologia.
38(Suppl
l):A79.
DeFronzo
RA, Hendler
R, Simonson
D. 1982 Insulin
resistance
is a prominent
feature
of insulin-dependent
diabetes.
Diabetes.
31:795-801.
DeFronzo
RA, Simonson
D, Ferannini
E. 1982 Hepatic
and peripheral
insulin
resistance:
a common
feature
of type 2 (non-insulin-dependent)
and type 1
(insulin-dependent)
diabetes
mellitus.
Diabetologia.
23:313-319.
Del Prato S, Nosadini
R, Tiengo
A, et al. 1983 Insulin-mediated
glucose
disposal
in type I diabetes:
evidence
for insulin
resistance.
J Clin Endocrinol
Metab.
57904.910.
Sturis J, Scheen AJ, Leproult
R, Polonsky
KS, Van Cauter
E. 1995 24-hour
glucose
profiles
during
continuous
or oscillatory
insulin
infusion.
Demonstration
of the functional
significance
of ultradian
insulin
oscillations.
J Clin
Invest. 95:1464-71.
Yki-JBrvinen
H, Helve E, Koivisto
VA. 1987 Hyperglycemia
decreases
glucose
uptake
in type 1 diabetes.
Diabetes.
36:892-896.
Vuorinen-Markkola
H, Koivisto
VA, Yki-JHrvinen
H. 1992 Mechanisms
of
hyperglycemia-induced
insulin
resistance
in whole body and skeletal
muscle
of type 1 diabetic
patients.
Diabetes.
41:571-580.
Rossetti
L, Giaccari
A, DeFronzo
RA. 1990 Glucose
toxicity.
Diabetes
Care.
13:610-630.
Vignati
L, Anderson
JH, Brunelle
RL, Jefferson
FL, Richardson
M. 1994
Improvement
of glycemic
control
with the rapidly
absorbed
lispro
insulin
analog in type 1 diabetes
[Abstract].
Diabetologia.
37(Suppl
l):A79.