Stimulatory Effect of Human Insulin on Erythroid Progenitors
(CFU-E and BFU-E) in Human CD34' Separated Bone Marrow
Cells and the Relationship between Insulin and Erythropoietin
I . Aoki: M . Taniyama; K. Toyama," M . Homori; K. Ishikawa"
"Second Department of Internal Medicine, Kyorin University School of Medicine, Tokyo, Japan; bThirdDepartment
of Internal Medicine, Showa University School of Medicine, Tokyo, Japan; 'First Department of Internal
Medicine, Tokyo Medical College, Tokyo, Japan
Key Words. Insulin Erythropoietin CFU-E BFU-E CD34
Abstract. Erythropoietin is known to be effective
for the treatment of anemia in chronic renal failure,
but the efficacy of erythropoietin for anemia in other
diseases is not so great. Insulin exerts a growth pro-
moting activity in various kinds of cells. In the pre-
sent study, the effects of insulin on erythroid
progenitors (colony forming units-erythroid, CFU-E;
and burst forming units-erythroid, BFU-E) in human
bone marrow were examined at various concentra-
tions of recombinant human erythropoietin (rh-Epo)
to clarify the relationship between erythropoietin
and insulin. Human insulin stimulated the forma-
tion of CFU-E and BFU-E in the presence of three
concentrations (0.25, 5, and 100 Ulml) of rh-Epo.
Stimulatory effects of human insulin on CFU-E and
BFU-E were also observed in the nonphagocytic and
nonadherent bone marrow fraction (NP-NA frac-
tion) and in the NP-NA and T cell-depleted fraction
at each concentration of rh-Epo. Human insulin fur-
ther stimulated the CFU-E and BFU-E growth in
CD34' separated bone marrow cells. These results
indicate that the enhancing effect of human insulin on
erythroid progenitors is not mediated through mono-
cytes and macrophages or T cells, suggesting a direct
action on erythroid progenitors.
Introduction
Erythropoietin has been applied to the
treatment of anemia in chronic renal failure,
Correspondence:Dr. I. Aoki, Second Department
of Internal Medicine, Kyorin University School of
Medicine, 6-20-2, Shinkawa, Mitaka-city, Tokyo 181,
Japan.
Received October 29, 1993; provisionally accepted
November 29, 1993; accepted for publication January
17, 1994. OAlphaMed Press 1066-5099/94/$5.00/0
STEM CELLS 1994;12:329-338
and many effective cases have been reported
[ 1, 21. However, the efficacy of erythropoietin
for anemia in other diseases such as aplastic
anemia [3,4], myelodysplastic syndrome (MDS)
[5, 61, multiple myeloma [7, 81, rheumatoid
arthritis (RA) [9], etc., is not as great as that in
the case of chronic renal failure. Some other
hematopoietic factors besides erythropoietin
are therefore thought t o be needed for the
improvement of anemia in such diseases.
Insulin's capability of sugar transport as a meta-
bolic action is well known. However, insulin is
also known to have a growth promoting activity
in various kinds of cells [lo-151. The action of
insulin on erythroid progenitors in CD34' sep-
arated bone marrow cells and the relationship
between insulin and erythropoietin have been
infrequently documented. In the present inves-
tigation, the effect of insulin on human bone
marrow erythroid progenitors (colony forming
units-erythroid, CFU-E; and burst forming
units-erythroid, BFU-E) in the presence of var-
ious concentrations of recombinant human ery-
thropoietin (rh-Epo) was examined in an attempt
to elucidate the participation of insulin in ery-
thropoiesis. We also investigated the effect of
human insulin on erythroid progenitors in
human CD34' separated bone marrow cells to
clarify the site of action of human insulin.
Materials and Methods
Bone Marrow Cells
Normal bone marrow cells were aspirated
from individuals after obtaining informed
330
consent. Bone marrow mononuclear cells
(BMMC) were isolated by Ficoll-conray den-
sity centrifugation (400 g for 30 rnin), washed
three times and used for cultures. Portions of
the mononuclear cell fraction were treated with
silica in phosphate buffered saline (KAC-2,
Japan Immunoresearch Laboratories, Takasaki,
Japan) at 37°C for 60 rnin for phagocytosis. This
fraction was subjected to Ficoll-conray density
centrifugation (400 g for 30 min), and the cells in
the upper layer containing nonphagocytic cells
were collected. Nonadherent cells were obtained
according to the methods of Messner et al. [ 161
with slight modifications. The nonphagocytic
cell fraction, after washing three times, was incu-
bated in a glass dish for 12 h to remove adherent
cells. Nonphagocytic and nonadherent (NP-NA)
cells were used as the NP-NA fraction. The T
cell-depleted fraction was obtained according
to the procedures of Horwitz and Garrett [ 171
with slight modifications. Sheep red blood cells
(SRBC, I x 109/ml)were treated with 50 U neu-
raminidase at 37°C for 30 min and washed three
times. The NP-NA fraction was incubated with
SRBC, left at room temperature for 10 min, cen-
trifuged at 800 rpm for 10 min, and then left at
room temperature for 60 min for rosette forma-
tion. This fraction was subjected to Ficoll-conray
centrifugation (400 g for 30 rnin), and the cells in
the upper layer containing nonrosette-forming
cells were collected, washed three times, and
then employed as the NP-NA and T cell-depleted
(NT) fraction (NP-NA and NT fraction).
Isolation of CD34+ Cells
Ten ml of Dulbecco's phosphate buffered
saline/Ca*+, Mg2+free (GIBCO, Grand Island,
NY) containing 1 mM ethylenediaminetetraacetic
acid (EDTA)/2Na (D-PBS), were added to tissue
culture flasks containing covalently immobilized
soybean agglutinin (SBA) or anti-CD34 mono-
clonal antibodies (mAb) (AIS MicroCELLector)
[ 18-20], and left at room temperature for 1 h. The
D-PBS was then aspirated, and another 10 ml of
D-PBS were poured into the flasks. The mixture
was shaken well for 30 sec so that it would evenly
cover the surface on which the cells were
adsorbed. This maneuver was repeated 4-6 times
for priming. BMMC were suspended in 4 ml of
D-PBS containing 0.5% (w/v) human y-globulin
(Polyglobin-N, Bayer Yakuhin, Ltd., Osaka,
Japan), and the suspension was left to stand at
room temperature for 15 min. The suspended
Stirnulatory Effect of Human Insulin
bone marrow cells were added to the SBA coated
flask, then the flask was gently rolled so that the
cell suspension would spread evenly on the cell
adsorption surface. The flask was left at room
temperature for 1 h, and the flask was then gen-
tly rolled for 10 sec so that the cells which did
not adhere to the surface would float.
Nonadherent cells were collected, and the flask
was rinsed gently two times with 4 ml of D-PBS.
The cells not adhering to the SBA were poured
into the AIS MicroCELLector coated with
anti-CD34 mAb, which was primed using a pro-
cedure as described above. After it was left to
stand for 1 h, the flask was gently rolled for 10
sec to remove nonattached (CD34-) cells, and 4
ml of D-PBS containing 10% fetal bovine serum
(FBS, JR Scientific, Woodland, CA) was then
added. The washing procedure was repeated until
no cells floated, and the flask was loaded with 4
ml of D-PBS containing 10%FBS. The cap was
fastened tightly, and the narrow side of the flask
was tapped firmly 2-3 times to release adherent
(CD34') cells. The suspension of released CD34'
cells was aspirated, and the flask was subse-
quently rinsed with D-PBS containing 10% FBS
several times to recover the additional CD34'
cells. The CD34+ cells were washed two times,
and then adjusted to a final concentration of 3-8
x 103/mlfor culture. The CD34' cell fraction is
enriched approximately 40- to 60-fold in CFU-E
and BFU-E activities as compared with the
mononuclear cell fraction by these procedures.
Colony Assays
CFU-E and BFU-E were measured by the
methylcellulose culture methods according to
the procedures of Iscove et al. [21] with slight
modifications. Bone marrow cells were cultured
in alpha-modification of Eagle's medium
(a-medium, Flow Laboratories, Irvine, Scotland)
supplemented with 30% FBS, 5 pg/ml gentam-
icin, and 0.8% methylcellulose (Dow Chemical,
Midland, MI) stimulated with rh-Epo, (epo-
etin-alpha, Kirin Brewery Pharmaceutical Co.,
Ltd., Tokyo, Japan), under 5% COz in air at 37°C
in a humidified incubator. The specific activity of
the epoetin-alpha is 2 x lo' IU/mg protein. After
7 days of culture, groups of eight or more cells
exhibiting a red or gold color were counted as
colonies derived from CFU-E, and after culture
for 14 days, cell groups consisting of 100 or more
red-colored cells, or three or more subcolonies
with red-colored cells were counted as colonies
Aoki/Tani yama/Toyama/Homori/Ishikawa
derived from BFU-E. CFU-E and BFU-E
colonies were counted in the same plates. An
Epo concentration of 5 U/ml was used for the
standard culture. Culture was also carried out at
Epo concentrations of 100 U/ml (high) and 0.25
U/ml (low).
Insulin
Insulin Human (Humarin-R, Shionogi
Pharmaceutical Co., Ltd., Osaka, Japan) was
the human insulin used, and Lente Insulin
(Lente Iletin, Lilly Pharmaceutical Co., Ltd.,
Kobe, Japan) was the porcine insulin used.
Specific activities of Humarin-R and Lente
Iletin are 28 and 26 U/mg, respectively. Human
insulin level in the FBS used in this study
revealed 0.8 pU/ml (5 x lo-'* M). Each insulin
was dissolved in a-medium and added to each
culture system to achieve final concentrations of
lO-'l to M. Only a-medium was added for
the control culture. Experiments were performed
5-8 times in each of three concentrations of
rh-Epo, and in each of 4 types of separated bone
marrow cells.
Time Dependency of Human Insulin
The time dependency of the addition of
human insulin was investigated. Human insulin
( M) was added to the culture system for
CFU-E assays containing 5 U/ml of rh-Epo,
simultaneously after 24 h, 2 days, 3 days, 4 days
and 5 days in each different plate including the
same numbers of normal BMMC. After 7 days
of culture, each CFU-E colony count was com-
pared with the colony count which was cultured
Table I. Effects of porcine insulin on CFU-E and BFU-E in human bone marrow (rh-Ep: 5 U/ml)
Porcine insulin CFU-E
(M) (2 x 105 BMMC)
(-1 260 f 59
10-1' 315 f 82
10-10 320 f 93
10-9 363 f 103
10-8 391 f 79"
10-7 41 1 f 75b
I 0" 404 f 7@
p < 0.05
bp < 0.01
Significant difference from the (-) value by t-test. rh-Ep: recombinant human erythropoietin; BMMN cells:
BMMC.
331
in addition of only a-medium without exoge-
nous human insulin. In other experiments,
M of human insulin was added to cultures at
the initiation, after 24 h, 4 days, 7 days, 9 days
and 12 days for BFU-E assays. After 14 days,
the BFU-E colony numbers in each plate were
compared with those in the control culture with
the addition of only a-medium without exoge-
nous insulin.
Statistics
The differences between test and control
groups were calculated by using Student's t-test.
Results
Effects of Porcine Insulin on CFU-E and BFU-E
The influence of porcine insulin on the for-
mation of CFU-E and BFU-E in human whole
BMMC was examined in the presence of 5 U/ml
of rh-Epo. The data obtained are shown in Table
I. Porcine insulin increased the CFU-E growth
in a dose-dependent manner, and a peak of
increase was observed at about lo-' M. Porcine
insulin also increased the formation of BFU-E
in a dose-dependent fashion. The increase was
marked at concentrations of M, and peaked
at about M as in the case of CFU-E.
Effects of Human Insulin on CFU-E
The number of CFU-E in the nonseparated
BMMC in normal humans was 148-475 (284 rt
43, mean kSE) in the presence of 5 U/ml of
rh-Epo. Human insulin enhanced the CFU-E
BFU-E
(2 x 105 BMMC)
8 0 f 16
94 f 22
97 f 29
108 f 21"
115f228
134 f 27b
121 f 30"
332 Stimulatory Effect of Human Insulin

growth with an increasing dose of human insulin in a manner similar to that of 5 U/ml or 100
(Fig. 1A). The CFU-E number was 170-580 U/ml of rh-Epo (Fig. 2C). A plateau of enhance-
(353 f 63) in the presence of a high concentra- ment was observed at M. N o BFU-E
tion (100 U/ml) of rh-Epo. Human insulin also colonies were detected without rh-Epo at any
increased the formation of CFU-E at 100 U/ml concentration of human insulin.
of rh-Epo (Fig. 1B). At a low concentration of
rh-Epo (0.25 U/ml), the CFU-E number was Site of Action on CFU-E and BFU-E by
68-210 (152 f 22). and human insulin stimu- Human Insulin
lated the formation of CFU-E at this low con- The influences of human insulin on CFU-E
centration of rh-Epo (Fig. 1C). It was therefore growth in the presence of three concentrations of
inferred that the increase of CFU-E growth rh-Epo in the NP-NA fraction are demonstrated
induced by human insulin was independent of in Table IIA, in order to investigate the site of
the rh-Epo concentration. No CFU-E colonies action. CFU-E formation in the NP-NA frac-
were detected without rh-Epo at any concen- tion was stimulated by the addition of human
tration of human insulin. insulin as in the case of whole BMMC inde-
pendently of rh-Epo concentrations. BFU-E for-
Effects of Human Insulin on BFU-E mation of the NP-NA fraction was also
The effect of human insulin on the BFU-E enhanced by human insulin in each of three con-
formation of human bone marrow whole centrations of rh-Epo as in the case of CFU-E
mononuclear cells in the presence of 5 U/ml of formation (Table IIB). These data suggest that
rh-Epo is illustrated in Figure 2A. Human human insulin acts on CFU-E and BFU-E by a
insulin increased the number of BFU-E with an mechanism which does not involve monocytes
increasing dose of human insulin, and a peak or macrophages included in the NP-NA frac-
increase was observed at or lo-' M. At a tion. To elucidate the site of action of human
high concentration of rh-Epo (100 U/ml), human insulin on CFU-E and BFU-E more clearly, a
insulin also increased the growth of BFU-E in a similar study was performed with the NP-NA
dose-dependent fashion (Fig. 2B). At a low con- and NT fraction of bone marrow. Results are
centration (0.25 U/ml), the BFU-E formation shown in Figures 3A and B. The CFU-E growth
was enhanced by the addition of human insulin was also enhanced by human insulin in all three

A B
rh Ep 5Ulml low

I
rh Ep 0 25Ulml

low

0 ' " " " ' 0

0
(.I 10.'' 10.10 10.' 104 10.' 104 (4 lo-" lO.'O 10.' 104 10.' 10' I4 10"' d o 10.0 104 10.' 104
HUMAN INSULIN (M) HUMAN INSULIN (M) HUMAN INSULIN (M)

Fig. 1. Effect of human insulin on CFU-E in whole BMMC at three concentrations of recombinant human ery-
thropoietin (rh-Epo: A) 5 U/ml; B) 100 U/ml; C) 0.25 U/ml). Human insulin enhanced the CFU-E growth in a dose-
dependent fashion, at all three concentrations of rh-Epo. Significant difference from (-) value by t-test A) M
and lod M, p < 0.05; B) 10%M and M, p < 0.05; C ) lo-' M and M, p < 0.05.
Aoki/Taniyama/Toy ama/HomoriDshikawa 333

h Ep 5Ulml
B m-Ep 1WUlml
C rbEp 025Ulml
2w 2w

1.1 10"' 10.'' 10'' 104 10' 10d

HUMAN INSULIN (M)

Fig. 2. Effect of human insulin on BFU-E in whole BMMC at three concentrations of recombinant human erythro-
poietin (rh-Epo: A) 5 U/ml; B) 100 U/ml; C) 0.25 U/ml). Human insulin enhanced the BFU-E growth in a dose-
dependent manner, at all three concentrations of rh-Epo. Data are expressed as mean fSE. Significant difference from
(-) value by f-test (*p < 0.05, **p < 0.01). BMMN cells: BMMC.

concentrations of rh-Epo even in the NP-NA stimulated the BFU-E formation in the NP-NA
and NT fraction (Fig. 3A). Human insulin also and NT fraction in each of three concentrations

Table 11. Effects of human insulin on CFU-E (A) and BFU-E (B) in human nonphagocytic and nonadherent bone
marrow at three doses of recombinant human erythropoietin
A. CFU-E in NP-NA fraction (5 x lo4cells)
Human insulin Recombinant human erythropoietin (U/ml)
(M) 0.25 5 100
(-1
10-10
46 5 * 134 f 36 173 f 18
61f9 156 f 35 221 f 29
10-9 6 7 f 10 180 f 56 253 f 36
10-8 73 f 9" 186 f 52" 263 f 31'
10-7 70 f 8" 205 f 63a 270 f 24b
10-6 70 f 7b 200 f 4Sb 253 f 25"

B. BFU-E in NP-NA and NA fraction (5 x lo4cells)

Human insulin Recombinant human erythropoietin (U/ml)
(M) 0.25 5 100
(-1 llf2 33 f 3 45 f 5
10-10 12f 1 34 f 3 56f6
10-9 15 f 2 41f3 62 f 5"
10-8 15 f 2 46 f 66 f 7'
10-7 17 f 2a 52 f 4b 70 f 6b
10" 18 f 2a 49 f 5" 69 f 5b
' p < 0.05
b p < 0.01

Significant difference from the (-) value by t-test. NP-NA fraction: nonphagocytic and nonadherent bone marrow
334 Stimulatory Effect of Human Insulin

A rh.Ep B
120 rh-Ep
550

+ 5Ulml
5W + 5Ulml
t O25UIml
1W --t 025ulml
450

--* 400

3 350
f
300
I
-:250
aw
20 150

1W
M
50 &-+-+--

0 0
(.I 10.10 10.0 104 10-1 104

HUMAN INSULIN (M)

Fig. 3. Effect of human insulin on CFU-E and BFU-E in the nonphagocytic, nonadherent and T cell-depleted (NP-
NA, NT) bone marrow cells in three concentrations (0.25, 5 and 100 U/ml) of recombinant human erythropoietin
(rh-Epo). Human insulin stimulated the CFU-E (A) and BFU-E (B) growth at each of the three concentrations of
rh-Epo. Data are shown as mean fSE. Significant difference from (-) value by t-test (*p < 0.05, **p < 0.01).

of rh-Epo, as in the case of CFU-E (Fig. 3B).
These findings suggest that monocytes or
macrophages and T cells are not involved in the
mechanism of the stimulatory effects of human
insulin on CFU-E and BFU-E.
Effect of Human Insulin on C F U - E and BFU-E
in CD34' Separated Bone Marrow Cells
Table I11 shows the effects of human
insulin on CFU-E and BFU-E in CD34' sepa-
rated bone marrow cells. Results of five differ-
ent experiments are presented as actual numbers

Table 111. CFU-E and BFU-E colony numbers in CD34' separated bone marrow cells in the presence of human
insulin (rh-Epo: 5 U/ml)
Human - 10-10 I 0-9 10-8 I 0-7 10" CD34' cells
insulin (M) cultured
Exp I CFU-E 12f2 13f3 15f2 18f4 17f2 20f 1 (3 x 103)
BFU-E 9fl 10f2 10f3 12f2 13f2 13 f 1 s
ExpII CFU-E 88f6 96f8 109f5 128 f 9 167 f 12 143 f 88 (8 x 10')
BFU-E 35f3 34f4 4053 46 f 5 51 f 3 a 53 -I 48
ExpIII CFU-E 71f4 85f4 98f7 107 f 5" 108 f 7n 115 f 68 (5 x lo3)
BFU-E 30f4 36f6 43f2 46 f 4 48 f 5 43 f 2
ExpIV CFU-E 46f5 51f4 62f3 62 f 2 66 f 4 64f3 ( 6 x lo3)
BFU-E 1652 1853 19f4 22 f 3 25 f 2" 24 f 4
EXPV CFU-E 58f3 62f4 73f7 88 f 6 92 f 5" 95 f 3" (4 x 103)
BFU-E 18f2 21f3 24f5 25 f 3 30 f 2p 27 f 3
&p< 0.05

rh-Epo: recombinant human erythropoietin. Significant difference from (-) value by t-test.
Aoki/Taniyama/Toyama/Homori/Ishikawa
in duplicate cultures (mean f S E ) . Human
insulin enhanced the CFU-E growth in CD34'
separated bone marrow cells as in the case of
the NP-NA, or NP-NA and NT fraction. Peaks
were observed at about M of human insulin.
Human insulin also stimulated BFU-E forma-
tion in CD34' separated bone marrow cells, as
in the case of CFU-E. The enhancement showed
peaks at about M. Based on these data,
including the results for the NP-NA, or NP-NA
and NT fractions, the effects of human insulin
on erythroid progenitors were thought to rep-
resent direct actions on erythroid progenitors.
Time Dependency of Human Insulin
Results of the time dependency in addition
of human insulin on the CFU-E and BFU-E for-
mation are represented in Figure 4. Data are
shown by the percent increment against the con-
trol value. The enhancing effects on CFU-E
declined with the delay of addition of human
insulin, and the increase rate reached 11% at
day 5 of addition. In BFU-E, the stimulatory
rates also decreased according to the delay of
the addition of human insulin, as in the case of
CFU-E. The increase rate decreased to only 8%
of the control at day 12 of addition.
Discussion
A good example of growth stimulation in
vivo by insulin is seen in the newborn babies
of diabetic mothers. Their organs become over-
grown because of hyperinsulinemia, and poly-
cythemia is also often observed in such babies
[22]. As an explanation of this, Perrine et al.
[23] reported that insulin stimulates cord blood
erythroid progenitors. It has also been reported
that insulin stimulates erythroid progenitors for
mouse fetal liver [24], for both human periph-
eral blood [ 2 5 , 261 and human bone marrow
[27, 281. As reported previously, it is well
known that serum Epo levels differ in various
types of hematological diseases [29-331. Few
studies on the relationship between insulin and
erythropoietin have been reported. We have
therefore performed the culture studies on
insulin stimulated with various concentrations of
rh-Epo. As a result, human insulin demonstrated
a stimulatory effect on erythroid progenitors,
with high concentrations of rh-Epo and also
even with low concentrations of rh-Epo. This
335
I T
Fig. 4. Time dependency of human insulin on CFU-E
and BFU-E. The stimulatory effect of human insulin
(lo-' M) on CFU-E and BFU-E was diminished with
the delay of addition of human insulin. Data are
shown by the percent increment f S E against the con-
trol value. Significant difference from control by
r-test (*p < 0.05, **p < 0.01, ***p < 0.001).
finding suggests that human insulin acts on ery-
throid progenitors independently of the ery-
thropoietin level. Kurtz et al. [24] reported that
insulin enhanced CFU-E proliferation in mouse
fetal liver cells in the absence of Epo, but we
could not find any CFU-E or BFU-E colonies
even with high concentrations of human insulin
without the presence of Epo.
A peak of increase by insulin was noted at
about M at any concentration of rh-Epo in
the present study. Since the physiological con-
centration of insulin is between 10 and 50 pU/ml
(6 x 10-I' to 3 x 1O-Io M), the maximum
response of porcine and human insulin occurred
at considerably higher concentrations than the
physiological concentrations. The action of
insulin as a growth factor has been reported to
occur at concentrations much higher than the
physiological concentrations in several reports
[lo-121. However, insulin is also known to act
not only at superphysiological concentrations
but also at concentrations near the physiologi-
cal concentrations in cell proliferations [ 13-151.
Insulin may thus act over relatively higher and
wider ranges of concentrations as a growth factor.
To determine whether the action of human
insulin is manifested through direct actions on
CFU-E and BFU-E or via some action on other
cells, cultures were performed after removing
phagocytic cells and adherent cells or T cells.
Since human insulin stimulated CFU-E and
BFU-E growth in the NP-NA fraction at all
three concentrations of rh-Epo, the action of
human insulin is considered not to be mediated
336
by monocytes or macrophages included in the
NP-NA fraction. The presence of T cells exerted
no influence on the stimulatory effects of human
insulin at each of the three concentrations of
rh-Epo, so that the mechanism of the stimulatory
effects of human insulin is thought to lack any
involvement by monocytes, macrophages or T
cells. Moreover, human insulin also enhanced
the CFU-E and BFU-E growth even in CD34'
separated bone marrow cells. In Sawadds report
[28], it was demonstrated that human CFU-E
require a direct interaction with insulin-like
growth factor I (IGF-I) and/or insulin for ery-
throid development. Similar to their investiga-
tion, t h e d a t a o b t a i n e d h e r e i m p l y that t h e
stimulatory effects of insulin on erythroid prog-
enitors may be through direct actions on CFU-E
and BFU-E.
In the study of time dependency, no rapid
decrease of stimulatory effects on CFU-E and
BFU-E was observed when human insulin was
added at early stages of cultures. The stimulatory
effects on CFU-E and BFU-E were decreased
according to the delay of the addition of human
insulin, and they almost vanished when insulin
was added at the termination of culture. These
findings indicate that human insulin acts on the
various stages of erythroid progenitors widely.
The mechanism of the stimulatory effects
of human insulin on human bone marrow CFU-E
and BFU-E is unclear. One concept is that the
action of human insulin is mediated through the
IGF-I receptor. The action of IGF-I is known to
be initiated by insulin [12, 341. However, DNA
synthesis is due to interaction of insulin or IGF-I
with t h e insulin r e c e p t o r [35], a n d insulin
increases the entry of [3H]thymidine into DNA in
human skin fibroblasts even after pretreatment of
IGF-I receptors with monoclonal antibodies
a I R - 3 [36]. The insulin receptor itself is thus
considered to be involved in cell proliferation.
The most widely accepted theory at present is
that activation of tyrosine kinase in the insulin
receptor and/or IGF-I receptor could be essential
for the growth promoting action of insulin as a
molecular mechanism [37-401.
I n t h e present s t u d y , w e w e r e a b l e t o
demonstrate in vitro that insulin stimulates the
proliferation of erythroid progenitors almost
independently of t h e rh-Epo d o s e , a n d t h e
enhancing effect of human insulin is observed
even in CD34' separated bone marrow cells.
Human insulin may thus play a stimulatory role
Stimulatory Effect of Human Insulin
in human erythropoiesis, and further studies
may b e of worth in the treatment of various
types of anemia.
Acknowledgments
W e are grateful to Kirin Brewery Pharma-
ceutical Co., Ltd., for providing recombinant
human erythropoietin.
References
1 Winearls CG, Oliver DO, Pippard MJ, Reid C,
Downing MR, Cotes PM. Effect of human ery-
thropoietin derived from recombinant DNA on
the anaemia of patients maintained by chronic
haemodialysis. Lancet 1986;ii:1175-1177.
2 Eschbach JW, Egrie JC, Downing MR, Browne
JK, Adamson JW. Correction of the anemia of
end-stage renal disease with recombinant human
erythropoietin. Results of a combined phase I
and I1 clinical trial. N Eng J Med 1987;316:73-78.
3 Bessho M, Jinnai I, Matsuda A, Saito M,
Hirashima K. Improvement of anemia by recom-
binant erythropoietin in patients with myelodys-
plastic syndromes and aplastic anemia. Int J Cell
Cloning 1990;8:445-448.
4 Yoshida Y, Anzai N, Kawabata H, Kohsaka Y,
Okuma M. Serial changes in endogenous ery-
thropoietin levels in patients with myelodysplas-
tic syndromes and aplastic anemia undergoing
erythropoietin treatment. Ann Hematol 1993;66:
175-180.
5 Cazzola M, Ponchio L, Beguin Y, Rosti V ,
Bergamaschi G , Liberato NL, Fregoni V, Nalli
G , Barosi G , Ascari E. Subcutaneous erythro-
poietin for treatment of refractory anemia in
hematologic disorders. Results of a phase 1/11
clinical trial. Blood 1992;79:27-37.
6 Zeigler ZR, Jones D, Rosenfeld CS, Shadduck
RK. Recombinant human erythropoietin
(rHuEPO) for treatment of myelodysplastic syn-
drome. Stem Cells 1993;11:49-55.
7 Wolfgang 0, Herrmann F, Gamm H, Zeile G ,
Lindernann A, Muller G , Brune T, Kraemer H-P,
Mertelsmann R. Erythropoietin for the treatment
of anemia of malignancy associated with neo-
plastic bone marrow infiltration. J Clin Oncol
1990;8:956-962.
8 Ludwig H, Fritz E, Kotzmann H, Hocker P,
Gisslinger H, Barnas U. Erythropoietin treatment
Aokipaniy ama/Toy ama/Homori/Ishi kawa
of anemia associated with multiple myeloma. N
Engl J Med 1990;322:1693-1699.
9 Tauchi T, Ohyashiki JH, Fujieda H, Lin KY,
Ohyashiki K, Toyama K. Correction of anemia
in rheumatoid arthritis by recombinant human
erythropoietin. Brit J Rheum 1990;29:235-236.
10 Temin HM. Studies on carcinogenesis by avian
sarcoma viruses. VI. Differential multiplication of
uninfected and of converted cells in response to
insulin. J Cell Physiol 1967;69:377-384.
11 De Asua LJ, Surian ES, Flawia MM, Torres HN.
Effect of insulin on the growth pattern and adeny-
late cyclase activity of BHK fibroblasts. Proc
Natl Acad Sci USA 1973;70:1388-1392.
12 Morel1 B, Froesch ER. Fibroblasts as an experi-
mental tool in metabolic and hormone studies.
11. Effects of insulin and nonsuppressible
insulin-like activity (NSILA-S) on fibroblasts in
culture. Eur J Clin Invest 1973;3:119-123.
13 Lawrence JC, Larner J Jr, Kahn CR, Roth J.
Autoantibodies to the insulin receptor activate
glycogen synthase in rat adipocytes. Mol Cell
Biochem 1978;22:153-157.
14 Koontz JW, Iwahashi M. Insulin as a potent, spe-
cific growth factor in a rat hepatoma cell line.
Science 198 1;211:947-949.
15 Massague J, Blinderman LA, Czech MP. The
high affinity insulin receptor mediates growth
stimulation in rat hepatoma cells. J Biol Chem
1982;257: 13958- 13963.
16 Messner HA, Till JE, McCulloch EA. Interacting
cell populations affecting granulopoietic colony
formation by normal and leukemic human mar-
row cells. Blood 1973;42:701-710.
17 Horwitz DA, Garrett MA. Distinctive functional
properties of human blood lymphocytes: a com-
parison with T lymphocytes, B lymphocytes, and
monocytes. J Immunol 1977;118:1712-1721.
18 Hatzfeld J, Li M-L, Brown EL, Sookdeo H,
Levesque L-P, O’Toole T, Gurney C, Clark SC,
Hatzfeld A. Release of early human hematopoi-
etic progenitors from quiescence by antisense
transforming growth factor p l or R b oligonu-
cleotides. J Exp Med 1991;174:925-929.
19 Lebkowski JS, Schain LR, Okrongly D, Levinsky
R, Harvey MJ, Okarma TB. Rapid isolation of
human CD34 hematopoietic stem cells-purg-
ing of human tumor cells. Transplantation
1992;53:1011-1019.
20 Migliaccio G, Migliaccio A R , Druzin ML,
Giardina PJV, Zsebo KM, Adamson JW.
Long-term generation of colony-forming cells in
337
liquid culture of CD34+ cord blood cells in the
presence of recombinant human stem cell factor.
Blood 1992;79:2620-2627.
21 Iscove NN, Siber F, Winterhalter KH. Erythroid
colony formation in cultures of mouse and human
bone marrow: analysis of the requirement for
erythropoietin by gel filtration and affinity chro-
matography on agarose-concanavalin A. J Cell
Physiol 1974;83:309-320.
22 Oski FA, Naiman JL. Polycythemia and hyper-
viscosity in the neonatal period. In: Hematologic
Problems in the Newborn, 3rd Ed. Philadelphia:
W. B. Saunders, 1982:87-96.
23 Perrine SP, Greene MF, Lee PDK, Cohen RA,
Faller DV. Insulin stimulates cord blood ery-
throid progenitor growth: evidence for an aetio-
logical role in neonatal polycythemia. Brit J
Haematol 1986;64:503-511.
24 Kurtz A, Jerkmann W, Bauer C. Insulin stimu-
lates erythroid colony formation independently of
erythropoietin. Brit J Haematol 1983;53:311-316.
25 Bersch N, Groopman JE, Golde W. Natural and
biosynthetic insulin stimulates the growth of
human erythroid progenitors in vitro. J Clin
Endocrinol Metab 1982;55:1209-1211.
26 Geffner ME, Kaplan SA, Bersch N, Lippe BM,
Smith WG, Nagel RA, Santulli TV Jr, Li CH,
Golde DW. Leprechaunism: in vitro insulin
action despite genetic insulin resistance. Ped Res
1987;22:286-291.
27 Dainiak N, Kreczko S . Interactions of insulin,
insulin like growth factor 11, and platelet-derived
growth factor in erythropoietic culture. J Clin
Invest 1985;76: 1237-1242.
28 Sawada K, Krantz SB, Dessypris EN, Koury ST,
Sawyer ST. Human colony-forming units-ery-
throid d o not require accessory cells, but do
require interaction with insulin-like growth fac-
tor I and/or insulin for erythroid development. J
Clin Invest 1989;83:1701-1705.
29 Urabe A, Saito T, Fukamachi H, Kubota M,
Takaku F. Serum erythropoietin titers in the ane-
mia of chronic renal failure and other hemato-
logic states. Int J Cell Cloning 1987;5:202-208.
30 Aoki I, Higashi K, Homori M, Chikazawa H,
Ishikawa K. Responsiveness of bone marrow ery-
thropoietic stem cells (CFU-E and BFU-E) to
recombinant human erythropoietin (rh-Ep) in
vitro in aplastic anemia and myelodysplastic syn-
drome. Am J Hematol 1990;35:6-12.
31 Vadhan-Raj S, Hittelman W, Lepe-Zuniga JL,
Gutterman JU, Broxmeyer H. Regulation of
endogenous erythropoietin levels in anemia
338
associated with myelodysplastic syndromes.
Blood 1990;75:1749-1750.
32 Nielsen OJ,Brandt M, Drivsholm A. The secretory
erythropoietin response in patients with multiple
myeloma and Waldenstrom’s macroglobulinemia.
Scand J Clin Invest 1990;50:697-703.
33 Hochberg MC, Arnold CM, Hogans BB, Spivak
JL. Serum immunoreactive erythropoietin in
rheumatoid arthritis: impaired response to ane-
mia. Arth Rheum 1988;31: 1318-132 1.
34 Clemmons DR. Multiple hormones stimulate the
production of somatomedin by cultured human
fibroblasts. J Clin Endocrinol 1984;58:850-856.
35 Mottola C, Czech MP. The type I1 insulin-like
growth factor receptor does not mediate increased
DNA synthesis in H-35 hepatoma cells. J Biol
Chem 1984;259: 12705-12713.
36 Flier JS, Usher P, Moses AC. Monoclonal anti-
body to the type I insulin-like growth factor (IGF-I)
receptor blocks IGF-I receptor-mediated DNA syn-
thesis: clarification of the mitogenic mechanisms of
IGF-I and insulin in human skin fibroblasts. Proc
Natl Acad Sci USA 1986;83:664-668.
Stimulatory Effect of Human Insulin
37 Grigorescu F, White MF, Karn CR. Insulin bind-
ing and insulin-dependent phosphorylation of the
insulin receptor solubilized from human ery-
throcytes. J Biol Chem 1983;258:13708-13716.
38 Ebina Y, Ellis L, Jarnagin K, Edery M, Graf L,
Clauser E, Ou JH, Masiarz F, Kan YW, Goldfine
ID, Roth RA, Rutter WJ. The human insulin
receptor cDNA: the structural basis for hor-
mone-activated transmembrane signalling. Cell
1985;40:747-758.
39 Sasaki N, Rees-Jones RW, Zick Y, Nissley SP,
Rechler MM. Characterization of insulin-like
growth factor I-stimulated tyrosine kinase activ-
ity associated with the P-subunit of type I
insulin-like growth factor receptors of rat liver
cells. J Biol Chem 1985;260:9793-9804.
40 Yu KT, Peters MA, Czech MP. Similar control
mechanisms regulate the insulin and type I
insulin-like growth factor kinases. J Biol Chem
1986;261: 1 1341-1 1349.