1987 69: 1674-1681   

 BA Miller, SP Perrine, G Antognetti, DH Perlmutter, SG Emerson, C Sieff and DV Faller erythroid cellsGamma-interferon alters globin gene expression in neonatal and adult

http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#repub_requestsInformation about reproducing this article in parts or in its entirety may be found online at:

http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#reprintsInformation about ordering reprints may be found online at:

http://bloodjournal.hematologylibrary.org/site/subscriptions/index.xhtmlInformation about subscriptions and ASH membership may be found online at:

Copyright 2011 by The American Society of Hematology; all rights reserved.20036.the American Society of Hematology, 2021 L St, NW, Suite 900, Washington DC Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by    

For personal use only. by guest on July 10, 2011. bloodjournal.hematologylibrary.orgFrom

1674 Blood. Vol 69. No 6 (June). 1987: pp 1674-168 1

‘y-Interferon Alters Globin Gene Expression in Neonataland Adult Erythroid Cells

By Barbara A. Miller, Susan P. Perrine, Giovanna Antognetti, David H. Perlmutter, Stephen G. Emerson,

Cohn Sieff, and Douglas V. Faller

Interferons have the ability to enhance or diminish the

expression of specific genes and have been shown to affect

the proliferation of certain cells. Here. the effect of ‘y-

interferon on fetal hemoglobin synthesis by purified cord

blood. fetal liver. and adult bone marrow erythroid progeni-

tors was studied with a radioligand assay to measure

hemoglobin production by BFU-E-derived erythroblasts.

Coculture with recombinant ‘y-interferon resulted in a

significant and dose-dependent decrease in fetal hemoglo-

bin production by neonatal and adult. but not fetal. BFU-

E-derived erythroblasts. Accumulation of fetal hemoglobin

by cord blood BFU-E-derived erythroblasts decreased up to

38.1 % of control cultures (erythropoietin only). Synthesis

of both G�/ A,,�, globin was decreased, since the � ratio

was unchanged. Picograms fetal hemoglobin per cell was

decreased by fly-interferon addition, but picograms total

hemoglobin was unchanged. demonstrating that a recipro-

cal increase in $-globin production occurred in cultures

T HE DEVELOPMENTAL SWITCH from -y to /3 globin

chain production has been studied extensively as a

model of gene regulation. An understanding of the control-

ling elements of this switch has therapeutic implications,

since continued synthesis of fetal hemoglobin (HbF) into

adult life could ameliorate the clinical manifestations asso-

ciated with hemoglobinopathies, including sickle cell disease

and /3-thalassemia.’FIbF production by progenitor-derived erythroblasts has

therefore been studied to determine factors involved in the

control of ‘y globin and /3 globin gene expression. For

example, the presence of adherent cells may increase the

relative amount of HbF produced by adult progenitors.2

Supernatant from Mo cells, a human lymphoid tumor line,

has been found to enhance hemoglobin A synthesis by cord

From the Division ofPediatric Hematologo’/Oncology. The Mi!-

ton S. Hershey Medical Center. Pennsylvania State University,

Hershey; the Division of Pediatric Hematology/Oncology, Dana-

Farber Cancer Institute, and Children �c Hospital Medical Center;

the Department of Pediatrics. Harvard Medical School. Boston;

and Children’s Hospital. Oakland. CA.

Submitted October 23, /986; accepted January 23. 1987.

Supported by National Institutes of Health (NIH) Grant No.

HL-37l 18.

Dr Miller and Dr Emerson are the recipients of N1H HLBI

Clinical Investigator Awards. Dr Faller is the recipient of an

American (‘ancer Society Research grant and is a John A. Hartford

Fe/low.

Address reprint requests to Barbara A. Miller. MD. Division of

Pediatric Hematology/Oncology. The Milton S. Hershey Medical

Center, P0 Box 850, Hershey. PA I 7033.The publication costs ofthis article were defrayed in part by page

charge payment. This article must therefore be hereby marked

“advertisement” in accordance with 18 U.S.C. §1734 solely to

indicate this fact.

C / 987 by Grune & Stratton, Inc.

0006-4971/87/6906-00021$3.oo/o

treated with ‘y-interferon. No toxic effect of “y-interferon

on colony growth was noted. The addition of ‘y-interferon

to cultures resulted in a decrease in the percentage of HbF

produced by adult BFU-E--derived cells to 45.6% of control.

Fetal hemoglobin production by cord blood. fetal liver. and

adult bone marrow erythroid progenitors, was not signifi-

cantly affected by the addition of recombinant GM-CSF.

recombinant interleukin 1 (IL-i ). recombinant IL-2. or

recombinant a-interferon. Although fetal progenitor cells

appear unable to alter their fetal hemoglobin program in

response to any of the growth factors added here. the

interaction of neonatal and adult erythroid progenitors

with ‘y-interferon results in an altered expression of globin

genes. This supports the concept that developmental gb-

bin gene switching can be regulated by environmental

factors.

S 1987 by Grune & Stratton, Inc.

blood BFU-E-derived erythroblasts.3 In addition, an un-

characterized factor in fetal sheep serum has been shown to

modulate HbF production in cultures of neonatal and adult

BFU-E and of BFU-E from patients with homozygous

/3-thalassemia and hereditary persistence of fetal hemoglobin

(HPFH).45 These observations suggested that soluble factors

elaborated by cells are capable of influencing the type of

globin gene expressed in erythroid progenitors.

Here, the effects of recombinant growth factors, particu-

larly those elaborated by lymphoid cells like Mo, on HbF

production by purified adult, cord blood, and fetal progeni-

tors were studied to determine whether any of these factors

are capable of modulating hemoglobin production. Partially

purified cord blood progenitor cells were cultured withrecombinant interleukin I (IL-I), IL-2, ‘y-interferon, or

GM-CSF, and HbF production by BFU-E--derived erythro-

blasts was assessed by a radioligand assay. The addition of

‘y-interferon resulted in a significant decrease in HbF pro-

duction by cord blood BFU-E-derived erythroblasts to

38.1% ± 2.3% of control, but total hemoglobin production

was unaffected. Interferon also decreased the amount of

HbF produced by adult progenitor-derived erythroblasts but

not those derived from fetal liver. This pattern of modulation

of globin expression is similar to that described in fetal sheep

sera.4 Recombinant ‘y-interferon specifically i nteracts with

neonatal and adult progenitor cells by altering globin gene

expression from predominantly ‘y to /3 production.

MATERIALS AND METHODS

Cord blood, bone marrow, andfeta! liver samples. Cord bloodsamples were obtained from normal full-term newborn deliveries,and fetal liver samples were obtained at 18 to 21 weeks’ gestationunder a protocol approved by The Brigham and Women’s Hospital

Human Investigation Committee and The Milton S. Hershey Mcdi-cal Center Committee on Clinical Investigation. Bone marrowsamples were procured from normal volunteer donors after informed

consent obtained within institutional guidelines.

For personal use only. by guest on July 10, 2011. bloodjournal.hematologylibrary.orgFrom

T ri�111

9.1

:,

U-

�0�0u20(0

0��o�.

.0IC)->

0 �I-C,

00

0

P

100F

�Hi�

-y-INTERFERON ALTERS GLOBIN GENE EXPRESSION 1675

Purification oferyihroid progenitor cells. Cord blood and bonemarrow mononuclear cells were separated on Ficoll-Paque (Pharma-

cia Fine Chemicals, Piscataway, Ni), centrifuged, and washed as

previously described.6 Fetal liver was treated with 50 mg/dL colla-

genase, type IV (Sigma Chemical Co. St Louis), and centrifuged

over Ficoll-Paque twice as previously described.7 Adherent cellsfrom all samples were depleted by overnight adherence to I 00 x15-mm plastic culture dishes in Iscove’s modified Dulbecco’s

medium (IMDM) with 20% fetal calfserum at 37#{176}C.Nonadherent

cells were then labeled for one hour at 4#{176}Cwith optimal concentra-

tions of antibodies that recognized precursors and mature myeloidand lymphoid cells to select for antigen-negative stem cells. Antibod-

ies used to label cord blood cells included anti-Leu-I, -Sb, -9, -10,

-12, -Ml, -M2, and -M3 (Becton Dickinson, Oxnard, CA); TG-l

(gift of Peter Beverley); and 10 F7 (gift of Dr William Bigbee).Antibodies to label adult bone marrow cells included anti-Leu-l,

anti-Leu-5b, Campath I, TG-l, anti-Leu-Ml, MO! (Ortho Phar-maceutical Corp. Raritan NJ), My-8 (gift of Dr J. Griffin), YTH89. 18 (gift of Dr Herman Waldman), and 10 F76; antibodies to labelfetal liver cells included anti-Leu-l, anti-Leu-5b, anti-.Leu-I0,

anti-Leu-Ml, anti-BA-I (Hybritech, San Diego) anti-Calla (Bec-

ton Dickinson), TG- I , and 10 F7.’ Mononuclear cells were washed to

remove excess antibody and then incubated for one hour at 4#{176}Con

Petri dishes coated with affinity-purified rabbit antimouse IgG, IgA,

and 1gM. Nonadherent antibody-negative cells were removed by

swirling the plates gently. The frequency of cord blood BFU-E

present in nonadherent cells from this purification step was 1% to

4%, that of bone marrow BFU-E was 1%, and that of fetal liverBFU-E was 1% after adherence to plastic alone and up to 30% after

panning. The final fraction cultured contained fewer than 1%lymphocytes and 0% to 5% monocytes.

Analysis of hemoglobin production by radioligand assay. Par-tially purified mononuclear cells from cord blood, bone marrow, or

fetal liver were cultured in 0.9% methylcellulose media containing

30% fetal calf serum, 9.0 mg/mL deionized bovine serum albumin(fraction V, Sigma), 1.4 x l0� mol/L f3-mercaptoethanol, and 2

�z/mL erythropoietin (Step I, Terry Fox Laboratories, Vancouver,

Canada, a source without detectable burst-promoting activity).

Recombinant generated human p1 7 IL-I (gift ofCistron Technolo-

gy, Pinebrook, Ni), recombinant IL-2 (gift of Biogen, Cambridge,

MA), recombinant ‘y-interferon (gift of Biogen), recombinant a-

interferon (gift of Schering Corp. Bloomfield, Ni), recombinant

GM-CSF (Genetics Institute, Cambridge, MA), or Mo cell-

conditioned medium (Mo-CM, provided by Dr David Golde) was

also added where indicated. In some experiments, ‘y-interferon wasadded to culture dishes at day 0, 4, 6, 8, 10, or 12 of the culture

period. Bone marrow and cord blood progenitors were plated at 5 x

l0� cells/mL and fetal liver at 1.5 x l0� cells/mL (adherent celldepleted) or 5 x l0� cells/mL (panned). Cultures were performed inhumidified 4% CO2 at 37#{176}Cfor 14 days, and erythroid colonies were

then counted.Approximately 50 to 100 BFU-E--derived colonies were plucked

per sample and the average number of erythroid cells per colony

determined. When samples were plucked in duplicate, the number ofcells per colony differed from the average count by ±5% to 20%. Aradioligand assay of total hemoglobin or HbF in picograms per

BFU-E-derived cell was performed as previously described.8’#{176}Cal-culations of hemoglobin content per BFU-E-derived cell were basedon the assumption that all such cells are potential F cells. Thepercent HbF was calculated from picograms of HbF and the totalhemoglobin cell. Reproducibility of this assay has been discussed

previously,’0 but in general, duplicate dishes or multiple assaysperformed on samples from one individual on different days yielded

results within 10% of the mean.

Analysis of hemoglobin accumulation, Gy/.4y ratio. and globinsynthetic ratios in BFU-E-derived erythroblasts by urea Triton

X-lOO gel. Hemolysates from BFU-E-derived erythroblasts wereprepared. The ratio of y to fi chain accumulation in these erythro-

blasts and the � ratio with and without the addition of

-y-interferon to cultures was determined by urea-Triton X- I 00 gel

electrophoresis as previously described.” Assessment of hemoglobinsynthesis was performed by incubating plucked I 3-day-old BFU-

E-derived cells with 50 �zCi 3H-leucine-free minimum essentialmedium with 10% dialyzed fetal calf serum overnight, electrophore-sis on Triton-urea gels, and autoradiography. Globin synthetic ratioswere quantitated by densitometry.

RESULTS

Effect of recombinant growth factors GM-CSF. IL-I,

IL-2, ‘y-interferon, and a-interferon on HbF synthesis by

cord blood BFU-E-derived erythroblasts. Cord blood

mononuclear cells were cultured in methylcellulose with and

without Mo-CM to determine whether the addition of this

source of burst-promoting activity had an effect on the

quantity of HbF produced per BFU-E--derived erythroblast.

In three out of four cord blood samples, the percentage of

HbF measured per BFU-E derived erythroblast decreased in

the presence of Mo-CM (Fig I).

Further experiments were performed to determine which

factor in Mo-CM was responsible for decreased HbF synthe-

sis. Cord blood BFU-E were enriched in these experiments so

that factors added would interact directly with progenitors

rather than with contaminating T cells or macrophages.

Either 5% Mo-CM, 5% recombinant GM-CSF from COS

cell supernatant (100 zg/mL, 1/100 dilution), recombinant

IL-I (700 ng/mL final concentration), recombinant IL-2

(10 U/mL final concentration), recombinant y-interferon

(200 U/mL final concentration) or a-interferon (2,000 U/mL final concentration) was added to methylcellulose cul-

tures on day 0. The concentration of GM-CSF was maximal

-‘! UnL�LL � -Control MO- GM- ILl 1L2 �IFN odFN

CM CSF

Recombinant Growth Factor

Fig 1 . Addition of 5% Mo-CM. 5% recombinant GM-CSF (100�g/mL. 1 /100 dilution). recombinant IL-i (700 ng/mL). recombi-nant 11-2 (10 U/mL), recombinant ‘y-interferon (200 U/mI) ora-interferon (2000 U/mI) to methylcellulose cultures of purified

cord blood BFU-E plated at 6 x i0� cells/mi. HbF content of1 4-day-old BFU-E-derived erythroblaets was measured by aradioligand assay. Results are expressed as a percentage ofcontrol HbF (cultured with erythropoietin only). and the mean ±1SD of four experiments is shown.

For personal use only. by guest on July 10, 2011. bloodjournal.hematologylibrary.orgFrom

50 100 150 200

Recombinant �IFN (u/mb)

1676 MILLER ET AL

9.1LI-u)

.0

0-c

tO>.

0#{149}000)�LI�.00)1c3

Fig 2. Addition of 0 to 200 U/mi recombinant ‘y-interferon

added to cultures of purified cord blood BFU-E. The percentage ofHbF/BFU-E-derived erythroblast was calculated from a radio-ligand assay for fetal and total hemoglobin.

for stimulation oferythroid colony growth. IL-l was used at

two different concentrations, 700 ng/mL and 2800 ng/mL,

both at plateau concentrations for the stimulation of IL-

I-dependent cells. Results were identical, and data are

shown for the addition of 700 ng/mL. The concentration of

IL-2 was on the plateau for support of IL-2-dependent cells,

and -y-interferon (at 200 U/mL) was added at five times the

amount required for maximal induction of class I major

histocompatibility complex proteins on K562 cells. Results of

experiments on four cord blood samples are summarized in

Fig I . To allow for comparison of four separate experiments,

the percentage of HbF measured in all experimental groups

was expressed as a percentage of the control (erythropoietin

only). Mo-CM in the culture medium decreased HbF pro-

duction to 75% of control, but the standard deviation was

large. The addition ofGM-CSF, IL-I, IL-2, or a-interferon

to cultures resulted in the production of HbF that was not

significantly different from the control. However, the per-

centage of HbF measured in BFU-E-derived erythroblasts

incubated with ‘y-interferon was reduced to 38.1% ± 2.3% of

control, a significant reduction.

To determine whether this effect was dose related, recom-

binant -y-interferon at 0 to 200 U/mL was added to methyl-

cellulose cultures of partially purified progenitors. Results of

two experiments are shown in Fig 2. The dose response to

‘y-interferon was greatest in the range of 0 to 20 U/mL, and

maximal effect was nearly achieved at 100 U/mL.

The number of BFU-E-derived colonies and the number

of cells/BFU-E--derived colony were determined in experi-

ments to which recombinant growth factors were added.

Results of three experiments are shown on Table 1 . None of

these growth factors affected progenitor proliferation at the

concentration at which they were used here.

The picograms of HbF and total hemoglobin measured by

radioligand assay of lysates from BFU-E-derived erythro-

blasts cultured with and without -y-interferon are shown for

experiments on three cord blood samples in Table 2.

Although the quantity of HbF in picograms per cell

decreased with the addition of ‘y-interferon, no change in

total hemoglobin content per cell was noted, which indicated

a reciprocal increase occurred in the amount of /3 globin

produced as the amount of ‘y globin decreased.

Effect of �-interferon on globin synthesis by cord blood

BFU-E-derived erythroblasts measured by 3H-leucine

incorporation. Additional experiments were performed to

determine whether the effect of ‘y-interferon on the synthesis

of HbF by BFU-E-derived erythroblasts occurred through-

out colony maturation and hemoglobin synthesis. Partially

purified cord blood BFU-E were cultured for 14 days with

and without the addition of 200 U/mL ‘y-interferon added at

day 0. The percentage of HbF produced over the entire

culture period was measured by radioligand assay and by

globin chain electrophoresis, and the percentage of ‘y chains

synthesized on days 13 and 14 was measured by ‘H-leucine

incorporation.

Results of three experiments are shown in Table 3. A

decrease in ‘y globin accumulation for the 14-day culture

period was noted after y-interferon addition by measurement

with both urea-Triton X-lOO gel electrophoresis and radio-

ligand assay. The mean percentage of HbF after -y-interferon

addition on day 0, measured by radioligand assay, was

60.6% ± 2.9% of control cultures, less of an effect than noted

in initial experiments. These differences may result from

differences in individual cord blood samples or different

‘y-interferon lots. A marked decrease in ‘y globin synthesis

was also noted by measurement of 3H-leucine incorporation

by BFU-E erythroblasts on days I 3 and I 4 of culture. The

TabI e 1 . Effect of Re combin ant Growth Fa ctors on Cord Blood B FU-E-Derived Colony Growth

Expi Exp2 Exp3

Cells! Cells! Cells/Factor BFU-E BFU-Et BFU-E BFU-Et BFU-E BFU-Et

IMDM Control 232 ± 3 8.7 47 ± 4 2.5 104 ± 4 7.2

Mo-CM 290 ± 9 5.3 50 ± 5 6.8 120 ± 12 4.3

GM-CSF 250 ± 9 8.9 50 ± 3 3.5 121 ± 9 6.6

IL-i, 700ng/mL 274 ± 25 9.3 44 ± 10 3.7 136 ± 8 6.6

IL-2, 20 U/mL 272 ± 30 9.6 48 ± 7 2.6 83 ± 3 8.5

yIFN, 200 U/mL 262 ± 12 6.8 63 ± 14 2.0 1 10 ± 16 8.0

Abbreviations: Exp. experiment; yIFN, y-interferon.

Mean number ± i SD of BFU-E-derived colonies/5 x iO’ enriched cord blood mononuclear cells.

tNumber of cells x 103/BFU-E-.derived colony.

For personal use only. by guest on July 10, 2011. bloodjournal.hematologylibrary.orgFrom

1 23

A’y-

a-

______Day of Addifion of �fFN

Fig 4. Assessment of the percentage of HbF per cord bloodBFU-E-derived cell after the addition of 200 U/mi ‘y-interferon tocultures on day 0. 4. 6. 8. 1 0. or 1 2. BFU-E-derived colonies wereharvested on day 14 and the percentage of HbF measured byradioligand assay.

y-INTERFERON ALTERS GLOBIN GENE EXPRESSION 1677

Table 2. Effect of ‘ylFN on Fetal and Total Hemoglobin

Accumulation in Cord Blood BFU-E-Derived Erythroblasts

HbF!CeII)pg)

Total Hb!Cell(pg)

Percentage ofHbF/CeII

Percentage ofControl

Exp 1

Control 10.5 31.1 33.8

yIFN 4.57 34.4 13.3 39.3

Exp 2

Control 29.5 43.3 68.2

-yIFN 18.9 69.9 26.9 39.4

Exp 3

Control 16.69 32.2 51.8

-ylFN 7.98 39.0 20.5 39.6

Results from three different cord blood samples are shown. Control

cultures contained erythropoietin only. A quantity of 200 U/mL ‘yIFN was

added to the experimental group. Picograms fetal and total hemoglobin

were measured by radioligand assay of lysates from 14-day-old BFU-

E -derived erythroblasts.

results of one experiment measuring 3H-leucine incorpora-

tion by BFU-E-derived erythroblasts with and without incu-

bation with -y-interferon is shown in Fig 3.

The Gy/A.y ratio was measured in two ofthese experiments

by globin chain electrophoresis. The G.y/A.y was 69.5% (cord

blood sample I) and 61.0% (cord blood sample 2) in control

studies and 67.7% (sample I) and 65.6% (sample 2) with

coculture of BFU-E with -y-interferon.

Addition ofy-interferon to cord blood BFU-E cultures at

various times during incubation. ‘y-Interferon at 200 U/

mL was added to cultures of cord blood BFU-E on day 0, 4,

6, 8, 10, or I 2. BFU-E-derived erythroblasts were harvested

on day 14, and a radioligand assay was performed to

determine the quantity of l-lbF produced. Results are shown

in Fig 4. A marked decrease in HbF accumulation was noted

Table 3. Comparison of HbF Synthesis/Cord Blood

BFU-E-Derived Erythroblasts by Different Methods

Percentage of HbF .y Globin,’BFU- E-Derived Cells

Cord Blood Cord Blood Cord Blood

Method 1 2 3

Radioligand Assay

Control 72.0 47.0 79.0

+yIFN 44.0 30.0 44.9

Accumulation by urea-Triton gelt

Control 57.0 40.0 -

+‘yIFN 36.0 30.0 -

3H-leucine synthesisl

Control 53.3 42.5 39

+ylFN 38.6 29.6 0

The percentage of HbF is calculated from the measurement of fetal

and total hemoglobin present in BFU-E-derived erythroblasts by radioli-

gand assay.

tThe percentage of y globin is calculated from the measurement on aurea-Triton X- 100 gel of ‘y and /3 globin chains synthesized during a

1 4-day culture of BFU-E -derived erythroblasts.

*The percentage of y globin is calculated from the measurement of‘H-leucine incorporation by y and /3 globin chains synthesized during an

overnight incubation of day-13 BFU-E-derived erythroblasts, a time

during colony maturation when /3 globin synthesis is normally predomi-

nant.

Fig 3. Autoradiograph of gbobin produced by cord blood BFU-E. Lanes 1 and 3 represent BFU-E cultured without ‘y-interferon;lane 2 shows BFU-E from the same cord blood sample culturedwith ‘y-interteron.

when -y-interferon was added on day 0, 4, or 6. This effect on

I-IbF accumulation decreased dramatically by day I 0, proba-

bly because the majority of HbF production had already

occurred in progenitor-derived cells at that time.’2 After day

12, effects of addition of ‘y-interferon on HbF production

could no longer be detected by measurement of accumula-

tion.

Effect of ‘y-interferon on HbF production by individual

cord blood BFU-E-derived colonies. It is possible that

‘y-interferon inhibits the growth of a hypothetical “fetal”

BFU-E population. To test this possibility, partially purified

cord blood BFU-E were cultured with and without ‘y-

interferon. On day 14 of culture, individual BFU-E from

both groups were plucked, and the hemoglobin content was

assessed by radioligand assay. Results are shown in Fig 5.

When -y-interferon was added to the cultures, HbF produc-

tion was reduced from a mean ± I SD of 68.5% ± 25.5% to

35.8% ± 21 .2% (P < .0001 ).Few ‘y-interferon-treated BFU-

E-derived colonies produced >50% ‘y-globin, and many

produced <25% ‘y globin. No such BFU-E populations

For personal use only. by guest on July 10, 2011. bloodjournal.hematologylibrary.orgFrom

I.

SS

S

S

-a--

S

S

S

S

I

SS

S

S

S

iS

S

S

S

S

S

S

100

90

>‘ 80C0

� :

a� 50

V00

� 40V

00

� 30

.0I�t 20

10

control #{247}YIFN

Fig 5. The percentage of HbF is measured in individual cord

blood BFU-E-derived colonies cultured with erythropoietin only(control) or with 200 U/mi ‘y-interteron. BFU-E-derived colonieswere harvested on day 14 and the percentage of HbF measured byradioligand assay.

w� � 1

� 120-m

00- 00

�u,1OO-.� (0

�80�

�I�i 60

‘3.� 40-

!:� 2O�

a� Confrof �IFN

Recombinant Growth Factor

Fig 6. Addition of 5% Mo-CM. 5% recombinant GM-CSF (100

pg/mi. 1 /100 dilution). recombinant li-i (700 ng/mL). recombi-nant li-2 (10 U/mi). and recombinant ‘y-interferon (200 U/mi or400 U/mi) to cultures of adherence-depleted (�. two experi-ments) or -purified (0, one experiment) fetal liver BFU-E. The HbF

content of BFU-E-derived erythroblasts was measured by radio-ligand assay. Results are expressed as a percentage of the control(erythropoietin only) HbF. An average of results for two experi-ments with adherence-depleted progenitors is shown.

MO- GM-CM CSF

ILl !L2

1678 MILLER ET AL

producing <23% -y globin were seen in the absence of

‘y-interferon. These results suggest a decrease in HbF pro-

duction by all BFU-E-derived colonies.

Effect of recombinant growth factors GM-CSF, IL-I,

IL-2, and ‘y-interferon on hemoglobin synthesis by fetal liver

BFU-E-derived erythroblasts. Fetal liver mononuclear

cells prepared by adherence depletion of macrophages (two

samples) or by partial purification (panning, one sample)

were cultured with and without the addition of 5% Mo-CM,

5% recombinant GM-CSF (100 �g/mL, 1/100 dilution),

recombinant IL-I (700 ng/mL), recombinant IL-2 (10 U/

mL final concentration), or recombinant ‘y-interferon (200

or 400 U/mL). BFU-E--derived erythroblasts were har-

vested on day 14 and the hemoglobin content determined by

radioligand assay. Results are shown in Fig 6. None of these

factors significantly affected l-IbF production from fetal liver

progenitors. When ‘y-interferon was added in a range of 0.2

to 400 U/mL, no effect was seen (data not shown). Doses of

200 U/mL or 400 U/mL ‘y-interferon, which produced a

significant effect on HbF accumulaion in cord blood BFU-

E-derived cells, had no effect on HbF production by fetal

liver.

Effect of recombinant growth factors GM-CSF, IL-i,

IL-2, and ‘y-interferon on hemoglobin synthesis by adult

BFU-E-derived erythroblasts. Partially purified BFU-E

from normal adult bone marrow were prepared as described

earlier and were cultured with 5% Mo-CM, 5% recombinant

GM-CSF (100 �g/mL, 1/100 dilution), recombinant IL-I

(700 ng/mL, final concentration), recombinant IL-2 (10

U/mL, final concentration), or “y-interferon (200 U/mL).

The number of BFU-E-derived colonies and the number of

cells per colony were reduced in adult samples without the

addition of burst-promoting activity (present in Mo-CM and

GM-CSF). Therefore, both GM-CSF and -y-interferon were

added to some of the cultures where indicated. BFU-E-

derived erythroblasts were plucked on day 14 and the

hemoglobin content assessed by radioligand assay. Results

are shown in Fig 7. The addition of-y-interferon significantly

decreased the quantity of HbF produced in cultures of adult

BFU-E-derived erythroblasts, as did the addition of both

-y-interferon and GM-CSF (P < .005 compared with GM-

CSF alone, calculated by one-way analysis of the variance).

The addition of lL-2 resulted in a slight but not significantdecrease in I-lbF synthesis. This may have resulted from

-y-interferon production by residual T cells stimulated by

IL-2. The group to which IL-I was added was not signifi-

cantly different (P > .05) in HbF production from those to

which Mo-CM or GM-CSF was added.

DISCUSSION

The mechanisms controlling the fetal globin switch and ‘y

and fi globin gene regulation remain poorly understood.

Analysis of DNA sequences in the ‘y region of patients with

HPFH provides strong evidence that promoter regions exist

in the first 200 base pairs, 5’ to the ‘y globin gene,”’6 and

experiments with human hybrid globin genes in mouse

erythroleukemia cells (MEL cells) suggest that DNA

sequences regulating �3 globin gene expression are located 5’

For personal use only. by guest on July 10, 2011. bloodjournal.hematologylibrary.orgFrom

�y-lNTERFERON ALTERS GLOBIN GENE EXPRESSION 1679

�12O� -CO100.- I

QI- T

�-.� Control MO- GM- ILl 1L2 SIFN GM-CSF

CM CSF +SIFN

Recombinant Growth Factor

Fig 7. Addition of 5% Mo-CM. 5% recombinant GM-CSF (100

ag/mi. 1 /100 dilution). recombinant li-i (700 ng/mi). recombi-nant li-2 (10 U/mi). recombinant ‘y-interferon (200 U/mi). or

GM-CSF plus y-interferon to cultures of purified adult bonemarrow BFU-E. The HbF content was measured by radioligandassay. Results are expressed as a percentage of the control(erythropoietin only) HbF. The mean ± 1 SD of five experiments isshown.

and 3’ of the transcription initiation site.’7”8 The evidence for

trans- and cis-acting factors that interact with these pro-moter regions has recently been reviewed,’9’20 but none of

these factors has yet been identified.

Several in vivo transplantation experiments suggest that

hemoglobin switching is largely controlled by mechanisms

intrinsic to the stem cell.2123 Other in vivo studies demon-

strate environmental influences do play a role. For example,

transplantation of fetal cells into adult sheep results in

synthesis of adult hemoglobin.24 The fly-to-fl globin switch is

delayed in 80% of infants of diabetic mothers.25 Several

factors, including erythroid potentiating activity3 and factors

produced by adherent cells,2 can affect HbF synthesis in

culture. A factor present in fetal sheep sera was found to

inhibit HbF synthesis in cultures of BFU-E from f3�-

thalassemia patients,4 HPFH patients,5 normal adults, and

neonatal blood4 but had no effect on fetal liver progenitors.4

The �y-to-f3 globin gene switch in BFU-E-derived erythro-

blasts from infants of diabetic mothers and premature

infants is inhibited by the addition of sodium butyrate to

cultures.26 These observations show that interactions

between progenitors and the inductive environment can

affect HbF synthesis.Here, we confirm earlier reports that a factor in Mo-CM

has the ability to reduce HbF synthesis in culture of BFU-

E-derived erythroblasts. We tested the recombinant growth

factors GM-CSF, IL-I, IL-2, -y-interferon, and a-interferon

to determine whether one of these cytokines was responsible.

Recombinant -y-interferon significantly decreased HbF pro-

duction in cord blood BFU-E-derived erythroblasts in a

dose-dependent manner. This decrease resulted from an

absolute decrease in HbF production and a reciprocal

increase in adult hemoglobin production. HbF production

was most dramatically affected when -y-interferon was pres-

ent early in the culture period (Fig 4), as would be expected

since HbF is synthesized first in culture. However, even at I 3

days of culture, HbF synthesis was reduced, as shown by

‘H-leucine incorporation studies (Table 3). Assays of mdi-

vidual BFU-E-derived colonies after rny-interferon additionsuggest that ‘y-interferon decreases the ability of all progeni-

tor-derived colonies to produce HbF (Fig 5). The possibilitythat a “fetal” population is suppressed by -y-interferon and

an “adult” population recruited seems unlikely because after

‘y-interferon addition the number of BFU-E was unchanged

and all BFU-E-derived colonies still produced greater than

10% HbF, which is above the normal adult range (<10%

HbF).

These data suggest that -y-interferon may play a role in the

fetal globin switch. -y-Interferon production has been shown

to be decreased in neonatal cells relative to adult cells due to

both a decreased intrinsic capacity for -y-interferon produc-

tion and to differences in regulatory mechanisms.27’28

Whether the increased ability to produce ‘y-interferon duringmaturation of the fetus correlates with the decrease in HbF

synthesis is unknown. “y-Interferon had no effect on HbF

production by BFU-E-derived erythroblasts from fetal liver.

This suggests that a property intrinsic to the fetal liver

BFU-E or cells derived from it determines the early HbF

program. However, in the neonate and adult, progenitor-

derived cells are now susceptible to environmental interac-

tions such as those with -y-interferon, thereby leading to a

decrease in I-lbF production. Because purity of the progeni-

tor cells is not absolute, the possibility that y-interferon acts

indirectly through nonprogenitor cells, which differ in the

fetus compared with the neonate and which are responsible

for differences observed, could not be ruled out. The effects

of-y-interferon on HbF production are remarkably similar to

those of fetal sheep sera previously reported by Papayanno-

poulou et al.4’5 Both decrease HbF production by neonatal

and adult but not fetal BFU-E--derived cells, and both have

an effect on G.y and A� production.5 Whether -y-interferon is

partly or wholly responsible for the switching activity in fetal

sheep sera is not clear but seems unlikely because interferons

have been shown not to work across species barriers.29

The mechanism(s) by which -y-interferon decreases ‘�‘

globin production is not yet clear. Interferons exert effects on

cells by binding to specific high-affinity receptors located on

the cell surface, and this interaction is followed by activation

ofcertain genes and decreased expression ofothers.3#{176}’32 It has

recently been shown that endogenous interferon produced

during the differentiation of mouse erythroleukemia cells is

in part responsible for decreased c-myc RNA synthesis and

the arrest of cell proliferation observed during differentia-

tion.32 Production of and response to endogenous interferon

takes place during the colony stimulating factor (CSF)-

1-induced differentiation of normal bone marrow precursors

to macrophages.’2 A role of endogenous interferon in ery-

throid differentiation may also exist. Recombinant -y-inter-

feron has been reported to inhibit the proliferation of

hematopoietic progenitors in a dose-dependent manner in

short-term and long-term cultures, although the plateau was

reached at much higher concentrations than those used

here.33 Recombinant fly-interferon has also been reported to

For personal use only. by guest on July 10, 2011. bloodjournal.hematologylibrary.orgFrom

REFERENCES

1680 MILLER ET AL

synergistically inhibit CFU-GM in combination with lym-

photoxin.’4 The role of -y-interferon as a natural regulator of

hematopoietic differentiation and HbF production in vivo

requires further investigation, particularly into whether

changes in chromatin structure are involved or required for

this “environmental” effect.’5

In summary, coculture with recombinant “y-interferon

results in decreased HbF synthesis by neonatal and adult but

not by fetal BFU-E-derived erythroblasts, and this decrease

is dose dependent. Synthesis of G.y A..� and /3 globin isaffected throughout the period of hemoglobin production.

During late fetal development, erythroid progenitors appear

I. Weatherall Di, Clegg JB, Wood WG: A model for thepersistence or reactivation of fetal hemoglobin production. Lancet

2:660, 1976

2. Javid J, Pettis PK: Fetal hemoglobin accumulation in vitroeffect of adherent mononuclear cells. J Clin Invest 7 1 : I 356, 1983

3. Testa U, Vainchenker W, Guerrasio A, Beuzard Y, Brenton-

Gorius J, Rosa J, Lusis AJ, Golde D: Hb switching in neonatalcultures. Increase of HbA synthesis in presence of an erythroid

potentiating activity (EPA). J Cell Physiol I 10:196, 1982

4. Papayannopoulou T, Kurachi 5, Nakamoto B, Zanjani ED,Stamatoyannopoulos G: Hemoglobin switching in culture: Evidencefor a humoral factor that induces switching in adult and neonatal but

not fetal erythroid cells. Proc NatI Acad Sci USA 79:6579, 19825. Papayannopoulou T, Tatsis B, Kurachi 5, Nakamoto B, Sta-

matoyannopoulos G: A hemoglobin switching activity modulates

hereditary persistence of fetal hemoglobin. Nature 309:7 1 , 19846. SieffCA, Emerson SG, Donahue RE, Nathan DG, Wang EA,

Wong GG, Clark SC: Human recombinant granulocyte-macro-

phage colony-stimulating factor: A multilineage hematopoietin.

Science 230:1 171, 1985

7. Emerson SO, Sieff CA, Wang EA, Wong GG, Clark SC,Nathan DG: Purification of fetal hematopoietic progenitors anddemonstration of recombinant multipotential colony-stimulating

activity. J Clin Invest 76:1286, 19858. Javid J, Pettis PK, Miller JE: Radio-ligand immunoassay for

human hemoglobin variants. J Immunol Methods 41:277, 19819. Miller BA, Salameh M, Ahmed M, Wainscoat J, Antognetti

G, Orkin 5, Weatherall D, Nathan DG: High fetal hemoglobinproduction in sickle cell anemia in the Eastern Providence of Saudi

Arabia is genetically determined. BlOOd 67:1404, 1986

10. Friedman AD, Linch DC, Miller B, Lipton JM, Javid J,Nathan DG: Determination of the hemoglobin F program in human

progenitor derived erythroid cells. J Clin Invest 75: 1 359, 1985

I I . Rovera G, Magarian C, Bown TW: Resolution of hemoglobinsubunits by electrophoresis in acid urea polyacrylamide gels contain-

ing Triton X-lOO. Anal Biochem 85:506, 1978

12. Friedman A, Linch D, Miller B, Lipton J, Javid J, NathanDG: Determination of the hemoglobin F program in progenitor

derived erythroid cells. J Clin Invest 75: 1 359, 1985

13. Giglioni B, Casini G, Mantovani R, Merli 5, Comi P.

Ottdenghi 5, Saglio G, Camaschella C, Mazza U: A molecular studyof a family with Greek hereditary persistence of fetal hemoglobinand fl-thalassemia. EMBO J 3:2641, 1984

14. Gelinas R, Endlich B, Pfeiffer C, Yagi M, Stamatoyannopou-

los G: G to A substitution in the distal CCAAT box of the A -y-globin

gene in Greek hereditary persistence of fetal hemoglobin. Nature313:323, 1985

15. Collins FS, Metherall JE, Yamakawa M, Pan J, Weissman

to acquire the capacity to respond to factors in their environ-

ment by altering their pattern of globin gene expression.

ACKNOWLEDGMENT

The authors wish to thank Dr David G. Nathan for his excellent

advice and support. We also wish to thank Dr Steve Clark of theGenetics Institute, Cambridge, MA, for the gift of recombinantGM-CSF, the physicians and nurses of the Department of OB-GYNat the Brigham’s and Women’s Hospital, Boston, for their assistancein obtaining cord blood and fetal liver samples, and the adult bonemarrow donors who made the complete study possible.

SM, Forget BG: A point mutation in the A-y-globin gene promoter in

Greek hereditary persistence of fetal hemoglobin. Nature 313:325,I985

16. Miller BA, Oliveri N, Salameh M, Ahmed M, Wainscoat J,

Antognetti G, Nathan D, Orkin 5: Molecular analysis of the high Fphenotype in Saudi Arabian sickle cell anemia patients. Blood 66:62,

1985 (suppl I) (abstr)

17. Wright 5, Rosenthal A, Flavell R, Grosveld F: DNAsequences required for regulated expression of /3-globin genes inmurine erythroleukemia cells. Cell 38:265, 1984

I 8. Charnay P, Treisman R, Mellon P. Chao M, Axel R, Mania-tas T: Differences in human ‘y and -y globin gene expression in mouse

erythroleukemia cells: The role of intragenic sequences. Cell 38:25 1,I 984

19. Kollias G, Wrighton N, Hurst J, Grosveld F: Regulated

expression of human A’y and fi- and hybrid -y/3-globin genes intransgenic mice: Manipulation of the developmental expression

patterns. Cell 48:89, 1986

20. Baron MH, ManiatisT: Rapid reprogrammingofglobin gene

expression in transient heterokaryons. Cell 46:591, 198621. Zanjani E, Lim G, McGlave PB, Clapp JR. Mann LI,

Norwood TH, Stamatoyannopoulos G: Adult hematopoietic cells

transplanted to sheep fetuses continue to produce adult globins.Nature 295:244, 1982

22. Papayannopoulou T, Nakamoto T, Agonstinelli F, Manna M,Lucarelli G, Stamatoyannopoulos G: Fetal to adult hemopoietic celltransplantation in humans: Insights into hemoglobin switching.Blood 67:99, 1986

23. Bard H, Prosmanne J: Postnatal fetal and adult hemoglobinsynthesis in preterm infants whose birth weight was less than 1,000

grams. J Clin Invest 70:50, 198224. Zanjani ED, McGlave PB, Bhaktavathsalan A, Stamatoyan-

nopoulos G: Sheep fetal haematopoietic cells produce adult haemo-

globin when transplanted into the adult animal. Nature 280:495,I 986

25. Perrine SP, Greene MF, Faller DV: Delay in the fetal globinswitch in infants ofdiabetic mothers. N EngI J Med 312:334, 1985

26. Perrine SP, Miller BA, Greene MF, Cohen RA, Faller DV:

Investigation of developmental globin gene switching in infants ofdiabetic mothers. Proceedings ofthe UCLA Symposia on Molecularand Cellular Approaches to Developmental Biology. (in press)

27. Wilson CB, Westall J, Johnston L, Lewis DB, Dower 5K,Alpert A: Decreased production of interferon-gamma by neonatal

cells. J Clin Invest 77:860, 198628. Lewis DB, Larsen A, Wilson CB: Reduced interferon-gamma

MRNA levels in human neonates. J Exp Med 163:1018, 198629. Stuart WE II: The Interferon System. New York, Springer,

1979, p 135

For personal use only. by guest on July 10, 2011. bloodjournal.hematologylibrary.orgFrom

‘y-INTERFERON ALTERS GLOBIN GENE EXPRESSION 1681

30. Friedman RL, Manly SP, McMahon M, Kerr IM, Stark GR: effect of recombinant interferon ‘y in short term and long-term

Transcriptional and post-transcriptional regulation of interferon- marrow culture. Br J Haematol 63:517, 1986

induced gene expression in human cells. Cell 38:745, 1984 34. Murphy M, Loudon R, Kobayastin M, Trinchieri G: y

3 1 . Lamer AG, Jonak G, Cheng Y-SE, Korant B, Knight E Jr, Interferon and lymphotoxin released by activated T cells synergizeDarnell JE: Transcriptional induction of two genes in human cells by to inhibit granulocyte/monocyte colony formation. J Exp Med

164:263, 1986interferon-�3. Proc NatI Acad Sci USA 81:6733, 1984

35. Groudine M, Peretz M, Nakamoto B, Papayannopoulou TH,32. Resnitzky D, Yarden A, Zipori D, Kimchi A: Autocrine Stamatoyannopoulos G: The modulation of HbF synthesis in adult

/J-related interferon controls c-myc suppression and growth arrest erythroid progenitor (burst-forming unit) cultures reflects changes

during hematopoietic cell differentiation. Cell 46:3 1 , I 986 in -y-globin gene transcription and chromatin structure. Proc NatI

33. Coutinho LH, Testa NG, Dexter TM: The myelosuppressive Acad Sci USA 83:6687, 1986

For personal use only. by guest on July 10, 2011. bloodjournal.hematologylibrary.orgFrom