1991 77: 486-492   

 BA Miller, K Foster, JD Robishaw, CF Whitfield, L Bell and JY Cheung proteins in the response of erythroblasts to erythropoietinRole of pertussis toxin-sensitive guanosine triphosphate-binding

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:

reserved.Copyright 2011 by The American Society of Hematology; all rights900, Washington DC 20036.weekly by the American Society of Hematology, 2021 L St, NW, Suite Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published    

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

Role of Pertussis Toxin-Sensitive Guanosine Triphosphate-Binding Proteins in the Response of Erythroblasts to Erythropoietin

By Barbara A. Miller, Karen Foster, Janet D. Robishaw, Carol F. Whitfield, Laurie Bell, and Joseph Y. Cheung

Human progenitor-derived erythroblasts have been recently shown t o respond t o erythropoietin (Epo) with an increase in intracellular free calcium concentration [Ca,]. To explore the role of guanosine triphosphate (GTP)-binding proteins in mediating the rise in [CaJ, single day 10 erythroid burst forming unit (BFU-E)-derived erythroblasts loaded with Fura-2 were pretreated with pertussis toxin (PT), stimulated with Epo, and [Ca,] measured over 18 minutes with fluorescence microscopy coupled to digital video imaging. The [Ca,] increase in day 10 erythroblasts stimulated with Epo was blocked by pretreatment with PT in a dose-dependent man- ner but not by heat-inactivated PT. These observations provided strong evidence that a PT-sensitive GTP-binding protein is involved. To further characterize the GTP-binding protein, day 10 erythroblast membrane preparations were solubilized, electrophoresed, and immunobloted with anti- bodies specific for the known PT-sensitive G-protein sub- units: the three subtypes of G, (1.2, and 3) and Goa. G,,, or G,,

ROWTH FACTORS that influence the proliferation G and differentiation of erythroid cells include erythro- poietin (Epo),',' human granulocyte-macrophage colony- stimulating factor (GM-CSF),3 and human interleukin-3 (IL-3): While the precise stages of erythroid differentiation at which these factors exert their effects is uncertain, Epo has clearly been shown to be required for the proliferation1 differentiation of more mature erythroid progenitors and precursors.',2 The biochemical mechanism through which Epo exerts this effect has been of great interest. Using fluorescence microscopy coupled to digital video imaging, single normal human erythroblasts loaded with the Caz+- sensitive probe Fura-2 have been shown to respond to Epo with an increase in intracellular free calcium concentration ([Ca,]).5,6 This [Ca,] increase is Epo dose-dependent and is related to the stage of erythroid differentiation.6 These observations on single normal cells at defined stages of differentiation are consistent with other reports using

From the Departments of Pediatrics, Medicine and Physiology, The Milton S. Hershey Medical Center, The Pennsylvania State University, Hershey; and The SigF?ed and Janet Weis Center for Research, Geisinger Clinic, Danville, PA.

Submitted March 26, 1990; accepted October 4, 1990. Supported by National Institutes of Health Grant HL40576, and a

grant from the Children's Miracle Network Telethon. B.A.M. is also the recipient of a Junior Faculty Award from the American Cancer Society. J.Y.C. is the recipient of Whitaker Foundation Research grant, American Heart Association (Pennsylvania Affiliate) grant-in-aid, and Juvenile Diabetes Foundation research grant.

Address reprint requests to Barbara A . Miller, MD, Department of Pediatrics, The Milton S. Hershey Medical Center, PO Box 850, Hershey, PA 17033.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 V.S.C. section 1784 solely to indicate this fact.

0 1991 by The American Society of Hematology. 0006-4971 191 17703-001 0$3.00/0

and Gi, were identified but no G, was found. To examine the influence of Epo on adenylate cyclase activity, day 10 erythro- blasts were initially treated with Epo, isolated membrane preparations made, and cyclic adenosine monophosphate (CAMP) production by adenylate cyclase in membrane prepa- rations in the presence of theophylline measured. Epo did not inhibit but significantly stimulated adenylate cyclase activity. However, the mechanism of increase of [Ca,] appears to be independent of adenylate cyclase stimulation because treat- ment of erythroblasts with the cell-permeant dibutryl CAMP failed to increase [CaJ In summary, pertussis toxin blocks the increase in [Ca,] in erythroblasts after Epo stimulation, suggesting that this response is mediated through a pertus- sis toxin-sensitive GTP-binding protein. Candidate PT- sensitive GTP-binding proteins identified on day 10 erythro- blasts were G,, 1.2, or 3, but not Goa. o 1991 by The American Society of Hematology.

murine erythroleukemia (MEL) cells,' Friend virus (FV)- infected cells: or normal human bone marrowg that Epo stimulation causes alterations in cellular Ca'+ homeostasis.

Signal transduction mechanisms have been well de- scribed for certain hormonelreceptorleffector pathways, such as adenylate cyclase inhibition and activation."." Receptors for certain hormones activate membrane-bound transducing proteins that bind guanosine triphosphate (GTP) (G-protein)."-" G-proteins can modulate the activ- ity of multiple specific effectors including adenylate cyclase,'' ion channels in the plasma memb~ane, '~ and enzymes that regulate phosphatidyl inositol metaboli~m.".'~ Activation of some effectors results in an increase in intracellular free calci~m.'~. '~ Pertussis toxin (PT) can inactivate the a subunit of certain G-proteins by adenosine diphosphate (ADP)- ribosylation, which decreases or eliminates receptor- mediated responsi~eness.~~ Treatment with PT can block the potentiation of calcium channel currents in response to GTP-$3 in dorsal root ganglian cells'6 and the decrease in phosphatidyl inositol bisphosphate and increase in [Ca,], which follows stimulation of neutrophils with chemotactic peptides." PT-catalysed ADP-ribosylation of G proteins in certain cell types also results in modulation of growth factor-induced cellular responses including change in [Ca,].

To determine whether the [Ca,] increase that follows Epo stimulation is mediated via a G-protein, progenitor- derived erythroblasts at a defined stage of differentiation were pretreated with PT before Epo stimulation. The [Ca,] increase in single erythroblasts that followed Epo stimula- tion was inhibited by preincubation with PT in a dose- dependent manner. This result was not observed after pretreatment with heat-inactivated PT. PT ADP-ribo- sylated membrane proteins of approximately 40 Kd from progenitor-derived erythroblasts. Immunoblotting of eryth- roblast membranes with antibodies that recognize G,, and G, showed the presence of G,- but not Go-proteins in the erythroblast membrane, suggesting that the family of G, proteins is the PT-sensitive substrate.

486 Blood, Vol77, No 3 (February 1). 1991: pp 486-492

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

GTP-BINDING PROTEINS IN EPO SIGNALLING 487

MATERIALS AND METHODS

Preparation of erythroid burst-forming unit (BFU-E)-derived eryth- roidprecursors. Adult blood was obtained in acid citrate dextrose (ACD) according to a protocol approved by The Milton S. Hershey Medical Center Committee on Clinical Investigation. Adult blood BFU-E were partially purified by 2-aminoethylisothio-uronium bromide hydrobromide (AET)-treated sheep red blood cell (RBC) rosetting, adherence to plastic, and panning, as described previ- ously? Partially purified mononuclear cells from adult blood were cultured in 0.9% methylcellulose media with 2 U/mL recombinant Epo (rEpo) (> 10,000 U/mg; Amgen, Thousand Oaks, CA), and rGM-CSF (gift of Dr Steven Clark, Genetics Institute, Cambridge, MA; 25 ng/mL final concentration) at plateau concentratiom6 To study erythroid precursors at specific stages of differentiation, cells from maturing BFU-E-derived colonies were plucked from culture on day 10. Day 10 cells are partially hemoglobinized and respond to Epo with an increase in [Ca,].'

BFU-E- derived cells were removed from culture on day 10 and labeled with antihuman (32 microglobulin.6 Cells were then bound to antimouse Ig-coated glass coverslips by incubating at 37°C for 1 hour. During this 1-hour period at 37"C, cells were exposed to Isove's Modified Dulbecco's Media (IMDM) (control cells) or to 0.05,0.10, 0.25, 1, or 5 pg/mL active or heat-inactivated PT (List Biological Laborato- ries, Campbell, CA). Media was removed and the cells were then incubated in phosphate-buffered saline (PBS) at 37°C for 20 minutes and containing the same concentrations of PT as well as 1 pmol/L Fura-2 acetoxymethyl ester (Molecular Probes, Inc, Eu- gene, OR). Thus, the total PT pretreatment period was 80 minutes. Total time lapse from removal of cells from culture to completion of Fura-2 loading was 3 to 5 hours. Cell viability as judged by txypan blue exclusion was 2 98%. [Ca,] and its changes in response to Epo (2 U/mL), IMDM, or dibutryl cyclic adenosine monophosphate (CAMP) (5 mmol/L; Sigma, St Louis, MO) in Fura-2-loaded cells incubated in PBS with physiologic concentrations of calcium were measured with the digital video imaging system described previ- o~sly. ' .~ Cells were chosen based on visible Fura-2 fluorescence and normal cellular morphology (round cytoplasmic borders and ab- sence of cytoplasmic vacules) by light microscopy. Over 90% of the cells on coverslips met these criteria. The dose of Epo was chosen to elicit maximal [Ca,] response both in terms of amplitude of [Ca,] as well as percent of cells with significant [Ca,] increase over baseline?,' The concentration of dibutryl cAMP used was effective in inhibiting insulin-stimulated glucose transport activity in human adipocytes.'" The gain of the intensified Fairchild charge coupled device video camera was adjusted such that autofluorescence (excited at both 350 and 380 nm) from unloaded cells was not different from background fluorescence, which was subtracted before calculation of [Ca,]. [Ca,] was calculated from Fura-2 fluorescence signals by the in vivo calibration method: assuming a kd of 224 nmoln for Ca" Fura-2 complex.19.zu We have previously shown that using our Fura-2-loading protocol, the fluorescence properties of intracellular Fura-2 and of Fura-2 free acid in solution are similar." In addition, [Ca,] values derived from the Fura-2 free acid calibration cuwe (in vitro calibration) and from in vivo calibration method are indistinguishable.'

ADP-ribosylation of BFU-E-derived erythroblast membranes with PT. Membranes from BFU-E-derived erythroblasts were ADP- ribosylated using previously described methods"," with the follow- ing modifications. BFU-E-derived colonies were plucked from culture at day 10 until 1 to 5 x 10' erythroid precursors had been harvested and suspended in IMDM at 4°C. All remaining steps

Measurement of [CaJ in early erythroid precursor cells.

until ADP-ribosylation were at 4°C. Cells were then washed with PBS once and spun at 60% for 10 minutes. The pellet was resuspended in 5 mmol/L MgCI,, 1 mmol/L EDTA, 4.4 mmol/L Tris, 5 pg/mL aprotinin, and 5 pg/mL leupeptin at pH 8. The suspension was vortexed vigorously to effect hypotonic lysis and then spun at 650g for 15 minutes to remove nuclei and mitochon- dria. The supernatant was then centrifuged at 41,OOOg for 20 minutes. The membrane pellet containing 2.5 to 10 pg protein was then resuspended in 10 pL of 75 mmol/L Tris-HCI pH 7.5 with 0.12 mg/mL Dnase I. To this was added 4 pL 120 mmol/L thymidine (10 m m o w thymidine final), 1 pL 50 mmoln adenosine triphosphate (ATP) (1 mmol/L final), 1 pL 5 mmol/L GTP (0.1 mmol/L final), 1 pL 50 mmol/L EDTA (1 mmol/L final), 24 pL of PT (25 pg/mL final preincubated in 25 mmol/L dithiothretol at 32°C for 30 minutes) and 10 pL P32 nicotinamide adenine dinucleotide (36 Cimmol, 2 Ci/mL, 11 pmoVL final; New England Nuclear, Boston, MA). This mixture was incubated at 32°C for 45 minutes, then stopped by the addition of 10% cold trichloroacetic acid, washed with diethylether, and resuspended in Laemmli buffer?' Mem- branes were electrophoresed on an 11% acrylamide gel overnight, followed by drying of the gel and autoradiography. In control experiments, either the membrane preparation or PT was not included in the ribosylation mixture.

Membrane pellets were prepared as described above for ADP-ribosylation and resuspended in 20 p.L of 20 mmol/L HEPES, 2 mmol/L MgCI,, and 1 mmol/L EDTA. The pellets were frozen at -70°C until immuno- blot was performed. Seventeen different membrane preparations from progenitor-derived erythroblasts were pooled, electro- phoresed on an 11% acrylamide gel, and blotted as previously described." Blots were probed with antipeptide antibodies specific for G,,,,, (A-56), G,, (A-54), and G, (A-10).= The specificity of these antibodies has been well documented previously?' Antibody binding was detected by incubation of blots with '=I-labeled goat antirabbit IgG. Autoradiographic images of the blots were ob- tained with Kodak XAR-5 film (Kodak, Rochester, NY) after exposure with an intensifying screen.

To determine the influence of Epo on adenylate cyclase activity, day 10 BFU-E-derived erythroblasts were removed from culture and incubated in IMDM at 37°C for 60 minutes. rEpo (4 U/mL) was added where specified during the 60-minute incubation. In some experiments, cells were preincubated with PT (1 p.g/mL) for 60 minutes at 37°C and then incubated with Epo for 20 minutes. Membranes were prepared from cells and adenylate cyclase activity was determined by measuring the formation of cAMP from unlabeled ATP in the presence of theophylline as described by Bonanou-Tzedaki et al.2h Briefly, the reaction was initiated by addition of 50 pL of the membrane fraction to 150 pL of assay mixture followed by incubation at 37°C for 20 minutes. The adenylate cyclase assay mixture contained (final concentration):2 mmolL ATP, 3 mmol/L MgCI,, 10 mmol/L NaCI, 10 mmol/L KC1, and 6 mmol/L theophylline as a phosphodiesterase inhibitor in 50 mmol/L Tris-HCI buffer pH 7.4. The reaction was terminated by heating the tubes at 99°C in a heat block for 5 minutes followed by centrifugation at 3,OOOg for 10 minutes at 4°C. Portions of the supernatant were removed and cAMP measured with cAMP assay kit (Amersham, Arlington Heights, IL). Results are expressed as picomoles per 1 x 106 cells. Protein content was not determined because of insufficient sample size but assuming 200 pg proteinn x lo7 cells, picomole content is similar to that previously reported."

Influence of PT on adenylate cyclase activity was assayed by treatment of intact cells followed by measurement of adenylate cyclase activity in membranes prepared from stimulated cells. This

Identification of PT-sensitive G-proteins with antisera.

Adenylate cyclase activity in progenitor-derived erythroblasts.

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

488

250

200

150

MILLER ET AL

-

-

-

method was chosen because it had been reported that membranes isolated from cells previously treated with Epo exhibited adenylate cyclase activity greater than or equal to that obtained from isolated membrane preparations stimulated with E ~ o . * ~

RESULTS

Eflect of PT on the increase in [CaJ in BFU-E-derived erythroblasts stimulated by Epo. To explore the possibility that a GTP-binding protein may be involved in Epo action, day 10 BFU-E-derived cells preincubated with 0 to 5.0 Fg/mL PT were loaded with Fura-2 and the time course of [Ca,] changes after Epo addition was followed. Table 1 shows the baseline [Ca,], peak [Ca,], and the peak increase in [Ca,] above baseline of cells treated with Epo or IMDM and followed for 18 minutes. Pretreatment with PT inhib- ited the increase in [Ca,] observed with Epo stimulation in a dose-dependent manner. Treatment with IMDM resulted in no significant change in [Ca,].

To demonstrate that the inhibition in [Ca,] increase was not due to nonspecific cytotoxicity of PT, cells were preincu- bated with 5 Fg/mL PT or 5 &mL heat-inactivated PT and subsequently stimulated with Epo. PT was heat-inactivated by incubating at 95°C for 20 minutes?' Figure 1 shows the time course of the change in [Ca,] for BFU-E-derived erythroblasts pretreated with IMDM, PT, or heat-inacti- vated PT and then stimulated with Epo. The increase in [Ca,] observed with Epo treatment followed a similar time course in cells pretreated with media or heat-inactivated PT. The change in [Ca,] in cells pretreated with heat- inactivated PT and stimulated with Epo was significantly different from that observed following pretreatment with PT (P < .Ol) , which was similar to cells not exposed to Epo.

ADP-riboylatwn of erythroblast membrane proteins by PT. To demonstrate the presence of PT-sensitive substrates in human erythroblast membranes, the ability of activated PT to ADP-ribosylate erythroblast membrane preparations was studied. PT consistently ADP-ribosylated membrane proteins of approximately 40 Kd. These results are shown in Fig 2A and were consistently seen in nine different mem- brane preparations. When either PT or membrane prepara- tions were eliminated from the ADP-ribosylation reaction mixture, no bands were observed.

Table 1. Effect of PT on the [Ca.] Response to Epo

Preincubation Stimulated No. of With PT (pg/mL) With Cells Baseline Peak AiCa.) (nmol/L)

0 ~ p o 36 2 0 2 3 2 0 7 2 6 i * 1 8 7 2 6 0 0.05 Epo 7 32 2 6 8 2 k 20+ 51 2 22 0.10 Epo 13 31 2 8 7 6 % 16* 4 5 % 16 0.25 Epo 8 3 7 2 12 7 6 % 2 2 3 9 k 14 1 .o Epo 20 21 2 6 4 2 r 10 21 2 6 5.0 ~ p o 21 2 7 2 4 3 5 2 8 8 2 7 0 IMDM 12 31 k 12 4 5 2 13 1 4 2 7

Day 10 BFU-E-derived erythroblasts preincubated at 37°C for 80 minutes with 0 to 5 pg/mL PT and then stimulated with 2 U/mL Epo. Mean % SEM are shown for 15 experiments.

'Statistically increased above baseline after Epo treatment by Stu- dent's t-test (P < .05).

2 c - 0 m o_ 100

50

I L 1 I I I

1 3 6 9 12 15 18

Minutes

Fig 1. Time course of [Ca.] in Epo-stimulated cells pretreated with or without PT. Day 10 BFU-E4erived cells were pretreated for 80 minutes with IMDM I-.-), PT 5 pg/mL (-04, or heat-inactivated PT 5 Fg/mL ( 4 - 1 and then stimulated with Epo (2 U/mL). [Cas] levels in control cells pretreated with IMDM and stimulated with IMDM (-O-) are also shown. [CaJ was measured at 0, 1, 3, 6, 9, 12, 15, and 18 minutes after Epo stimulation. Mean [CaJ % SE shown. ''Indicates a significant increase above baseline as determined by one-way analy- sis of variance. Data represent mean of eight experiments. Four cells were analyzed for IMDM control; all other data points represent 8 to 11 cells.

Identification of G-proteins in erythroblast membrane with antisera. To confirm that PT-sensitive substrates were G-proteins and to identify which G-proteins were ADP- ribosylated by PT, erythroblast membrane proteins were immunoblotted with antipeptide antibodies specific for the (Y subunits of Gi1,3, Gi2, and Go. The immunoblots in Fig 2 show the presence of a 41-Kd Gi,, and/or Gi, band (Fig 2B) and a 40-Kd Gi, band (Fig 2C). No G, was observed in erythroblast membranes (Fig 2D). Therefore, the sub- strates for ADP-ribosylation by PT appear to be one or more forms of G,.

To determine whether Epo inhibited adenylate cyclase after interaction with a G,-type G-protein in erythroblasts, ade- nylate cyclase activity was measured in membranes pre- pared from day 10 cells treated with Epo for 0 to 60 minutes. Adenylate cyclase activity significantly increased following Epo treatment and the stimulatory effect pla- teaued at 20 minutes after Epo addition (Fig 3). To explore the influence of PT on Epo-stimulated adenylate cyclase

Effect of Epo stimulation on adenylate cyclase activiy.

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

GTP-BINDING PROTEINS IN EPO SlGNALLiNG 489

Fig 2. ADP-ribosylation and immunoblots of membrane proteins from day 10 erythroblasts. Pro- teins with or without PT labeling were resolved by sodium dodecyl sulfate-polyacrylamide gel electro- phoresis and transferred from gels to nitrocellulose. The UP-labeled membrane blot was subjected to autoradiography. The unlabeled blots were treated with ’nl-labeled goat antirabbit IgG F(ab’), frag- ments after incubation with primary antiserum at a 1:5W dilution. Numbers to the left (A) mark the location of molecular weight standards. Lane 1, standard lane containing 2.5 pg of a mixture of the 41-Kd u subunits of G,,,. the 40-Kd a subunit of G, and the 39-Kd u subunit of G, from bovine brain. Lane 2 contains 25 pg membrane protein from erythroblasts. (A) 1.4 pg membrane protein ADP- ribosylated by PT; (B) Gh,> peptide antiserum A-56; (C) G, peptide antiserum A-54; (D) G, peptide antiserum A-10. Film was exposed to blots with an intensifying screen at -70°C for 7 days.

activity, day 10 erythroblasts were preincubated with PT or media for 60 minutes and then stimulated with Epo for 20 minutes (Table 2). Stimulation with Epo resulted in a significant increase in CAMP levels (P < .025). Pretreat- ment with PT resulted in a further increase in CAMP levels above that observed with Epo alone but results did not reach statistical significance.

To study the role of CAMP in mediating the [Ca,] increase in Epo-stimulated erythroblasts, day 10 BFU-E- derived erythroblasts were loaded with Fura-2 and then visualized with digital video imaging. [Ca,] measurements were taken at baseline and during 20 minutes of stimulation with the cell-permeant analogue dibutryl CAMP. Results are shown in Table 3. [Ca,] increased significantly above

Fig 3. Measurement of adenylate cyclase activity in BFU-E- derived erythroblasts after Epo stimulation. Day 10 erythroblasts were stimulated for 5 to 80 minutes with Epo (4 UlmL). Membranes were subsequently prepared and adenylate cyclase acthrity assayed in the presence of 6 mmol/L theophylline. Results are expressed as mean picomoles CAMP 2 SEM per 1 x 10‘ cells for eight separate experiments. Indicates a significant increase above unstimulated cells as determined by the paired t-test (P < .02).

control after Epo stimulation (P < .001) but there was no significant change following incubation of cells with dibutryl CAMP.

DISCUSSION

Two general mechanisms of signal transduction for hormones and growth factors are well described: coupling of receptors to effectors through G-protein~’~” or coupling through activation of tyrosine kinase activity intrinsic to the receptor.”*29 G-proteins are GTP-binding heterotrimers that function as intermediaries in transmembrane signaling by linking receptors to a variety of specific effectors includ- ing phosphoinositidase (resulting in inositol lipid break- down),M adenylate cyclase,” and ion channels.’’ Agonist/ receptor interaction for many of these signal transduction mechanisms results in an increase in [Cac].1”’I*16M-32 G -pro- teins are substrates for ADP-ribosylation by PT and/or

Table 2. Influence of PT on Adenylate Cyclase Activity in Epo-Treated Cells

Preincubated With Stimulated With CAMP (pmol)

0.47 & 0.16 IMDM EPO 1.05 f 0.18. PT EPO 1.43 2 0.39* PT iMDM 0.55 f 0.17

Day 10 erythroblasts were pretreated with or without PT (1 pg/mL) at 3PC for 60 minutes and then stimulated with or without Epo (4 U/mL) for 20 minutes. Membranes were prepared and adenylate cyclase activity assayed in the presence of 6 mmolR theophylline. Mean 2 SEM pmol CAMP produced by membranes from 1 x 10’ cells shown for seven experiments.

Differences among means were tested with one-way analysis of variance. A priori comparisons of means of control versus Epo-treated groups were then performed, using F-tests as tests of significance. There were no differences between IMDMAMDM versus PTAMDM groups.

IMDM IMDM

*P < .025 compared with IMDMAMDM.

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

490

Table 3. Effect of Dibutryl cAMP on [CaJ Levels

MILLER ET AL

[Ca,] (nmol/L) No. of Stimulated With Cells Baseline Peak AICa,] (nmol/L)

IMDM 9 2 1 2 4 3 6 ~ 7 15? 4 EPO 13 3 9 ~ 5 203?53* 164252 DiButrylcAMP 10 24 z 6 37 f 5 13 z 3

Day 10 BFU-E-derived erythroblasts stimulated with 2 U/mL Epo, 5 mmol/L dibutryl CAMP, or IMDM for 20 minutes. Mean ? SEM shown for four experiments. Significance of differences among the means of six groups (three controls, three treatments) was determined by one-way analysis of variance. A priori comparisons of means of any n groups (n < 6) were then performed, using F-tests as tests of signfi- cance. There were no statistically significant differences among the means of baseline [Ca,] values.

*P i .001.

cholera toxin, which modulate their function. Multiple forms of PT- or cholera toxin-sensitive G-proteins have been For some receptor/G-protein/effector complexes, bacterial toxins inhibit the ability of agonists to stimulate a change in [Ca,]. For example, preincubation of mast cells and neutrophils with PT blocked the increase in [Ca,] usually observed following treatment with compound 48/80 or chemotactic peptide^.^^.^^ Cholera toxin is capable of inhibiting the T-cell antigen receptor-mediated increases in inositol bisphosphate and cytoplasmic free calcium.34

The signal transduction pathway through which Epo exerts its effects was explored here. Treatment of single BFU-E-derived and normal human bone marrow cells9 with Epo results in an increase in intracellular calcium. This increase is Epo dose dependenP9 and is specific for the stage of differentiati~n.~.” In addition, using optical sectioning microscopy and three-dimensional image reconstruction, we have demonstrated a disproportionate increase in nuclear Ca2+ when compared with cytosolic Ca2+ in Epo-stimulated human erythroid progenitor^.'^ We pos- tulated this large increase in nuclear Ca2+ activates Caz+- dependent endonucleases and may commit the cell to terminal differentiation.

To determine whether the [Ca,] increase that follows Epo stimulation is mediated through a GTP-binding pro- tein, progenitor-derived erythroblasts were pretreated with PT before Epo stimulation. Pretreatment with PT inhibited the increase in [Ca,] observed following Epo stimulation. This inhibition was PT dose dependent and specific, be- cause pretreatment with heat-inactivated PT did not inhibit the increase in [Ca,]. This observation provides strong evidence that the [Ca,] increase in erythroblasts in response to Epo is mediated through a PT-sensitive G-protein. In a preliminary report, Hilton et al” have shown that GTP-yS inhibited binding of [I”’]-Epo to its receptor on FV- infected spleen erythroblasts. In addition, PT inhibited the 2-day Epo-dependent proliferation of an Epo-responsive cell line 32D C1 23.” These results using different ap- proaches also support the concept that G-proteins are involved in Epo signal transduction.

To further identify PT-sensitive substrates in day 10 erythroblasts, membrane preparations were ADP-ribosy- lated by PT. At least four different polypeptides have been

reported to be ADP-ribosylated by PT. By denaturing gel electrophoresis, these polypeptides have apparent molecu- lar masses of 39 to 41 Kd.’03’’ PT-sensitive substrates of approximately 40 Kd were reproducibly observed, consis- tent with the molecular weight of previously described G, (Y

subunits. Resolution of the ADP-ribosylation band was in- sufficient to distinguish the precise number of bands present. If sufficiently low amounts of membrane were loaded it may have been possible to resolve 41-Kd and 40-Kd bands on gels by silver stain, but the resolution would be partially lost due to 32P-~catter. Therefore, immunoblots were performed with antibodies specific for the different PT substrates2’ to define more precisely the PT-sensitive a protein subunits present. Immunoblot analysis showed that the substrates for ADP-ribosylation by PT present on erythroblast mem- branes are G,,, G,,, or G,3. The G, subunit, which has a molecular mass of 39 Kd, was not detected.

The nature of the effector system to which the Epo receptor is coupled has not been explored here. With re- spect to Epo-induced [Ca,] increase, phosphoino~i t idase~~~~’~~ or direct G-protein gating of calcium channel^"^'^ are two likely possibilities because both are associated with G-pro- teins and with change in [Ca,]. Although G, is classically associated with direct G-protein gating of ion channel^,'"^' recent evidence shows that the family of G, proteins can also modulate Ca2+ currents. For example, in the adreno- cortical cell line Y1, angiotensin I1 stimulated voltage- dependent Ca2+ currents that were blocked by PT.” Y1 cells also contain G,-type G-proteins but no Go, suggesting that G,-type G-proteins can mediate stimulation of calcium currents.” The possibility was considered that Epo inter- action with a G,-type G-protein might inhibit adenylate cyclase, resulting in a decrease in CAMP levels. This decrease in turn might relieve tonic inhibition (by CAMP) of phosphoinositide hydrolysis or affect activity of ion chan- nels, resulting in a secondary increase in [Ca,]. This hypothesis appears unlikely because this study and others have demonstrated the ability of Epo to stimulate adenylate cyclase and cause an increase in cAMP levels.26 The influence of Epo on CAMP levels is controversial because other investigators have not demonstrated an increase in cAMP levels in response to EPO.~’.~’ This controversy may partially be due to the use of different cell lines or erythroleukemia cells. However, no studies to our knowl- edge have demonstrated a decrease in CAMP levels in response to Epo stimulation. In addition, our current results show that direct stimulation with dibutryl cAMP resulted in no appreciable increase in [Ca,] in day 10 erythroblasts. Our data suggest that the PT-sensitive and Epo-stimulated increase in [Ca,] is not mediated via suppres- sion of adenylate cyclase activity.41 In this light, it is interesting to note that other investigators have shown that dibutyl CAMP and other CAMP-enhancing agents were not effective in inducing cell growth in Epo-dependent cell lines,39 MEL cells,@ or immature rabbit erythroblast^.^^ These investigators concluded that CAMP does not play a direct role as the second messenger in Epo signal transduc- tion. Indeed, Tsuda et al found that Epo-stimulated growth was inhibited by CAMP enhancing agent^.'^

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

GTP-BINDING PROTEINS IN EPO SIGNALLING 49 1

The evidence presented here suggests that in day 10 human BFU-E-derived erythroblasts, signal transduction of Epo occurs through a PT-sensitive GI-type GTP-binding protein. Recently, the effects of two other hematopoietic growth factors, IL-3 and GM-CSF, on myeloid cells, have also been shown to be mediated through G-proteins.2’~~~

Determination of the precise identity of the G-protein(s) and effector system(s) involved in the response of different hematopoietic lineages at distinct stages of differentiation to hematopoietic growth factors will be required for a fuller biochemical understanding of the variety of proliferative and differentiation pathways observed.

REFERENCES

1. Jacobs K, Shoemaker C, Rudersdorf R, Neil1 SD, Kaufman RJ, Mufson M, Sukra J, Jones SS, Hewick R, Fritsch EF, Kawakita M, Shimizu T, Miyake T Isolation and characterization of genomic and cDNA clones of human erythropoietin. Nature 313:806,1985

2. Spivak J: The mechanism of action of erythropoietin. Int J Cell Cloning 4:139, 1986

3. Sieff CA, Emerson SG, Donahue RE, Nathan DG, Wang EA, Wong GG, Clark S C Human recombinant granulocyte-macro- phage colony-stimulating factor: A multi-lineage hematopoietin. Science 2301171,1985

4. Yang Y-C, Ciaraletta AB, Temple PA, Chung MP, Kovack S, Witek-Giannotti JS, Leary AC, Kriz R, Donahue RE, Wong GG, Clark S C Human IL-3 (multi-CSF): Identification by expression cloning of a novel hematopoietic growth factor related to murine IL-3. Cell 47:3, 1986

5. Miller BA, Scaduto RC, Tillotson DL, Botti JJ, Cheung JY: Erythropoietin stimulates a rise in intracellular free calcium concentration in single early human erythroid precursors. J Clin Invest 82:309,1988

6. Miller BA, Cheung JY, Tillotson DL, Hope SM, Scaduto RC Jr: Erythropoietin stimulates a rise in intracellular free calcium concentration in single BFU-E derived erythroblasts at specific stages of differentiation. Blood 73:1188,1989

7. Levenson R, Houseman D, Cantley L: Amiloride inhibits murine erythro-leukemia cell differentiation: Evidence for a cal- cium requirement for commitment. Proc Natl Acad Sci USA 77:5948, 1980

8. Sawyer ST, Krantz SB: Erythropoietin stimulates 45Ca2+ uptake in Friend virus-infected erythroid cells. J Biol Chem 259:2769,1984

9. Mladenovic J, Kay NE: Erythropoietin induces rapid in- creases in intracellular free calcium in human bone marrow cells. J Lab Clin Med 11223,1988

IO. Spiegel AM: Signal transduction by guanine nucleotide binding proteins. Mol Cell Endocrinol49:1,1987

11. Gilman AG: G proteins: Transducers of receptor-generated signals. Ann Rev Biochem 56:615, 1987

12. Lochrie MA, Simon MI: G protein multiplicity in eukaryotic signal transduction systems. Biochemistry 27:4957, 1988

13. Brown AM, Birnbaumer L: Direct G protein gating of ion channels. Am J Physiol23:H401, 1988

14. Meyer T, Holowka D, Stryer L Highly cooperative opening of calcium channels by inositol 1,4,5-trisphosphate. Science 240 653,1988

15. Milligan G: Techniques used in the identification and analysis of function of pertussis toxin-sensitive guanine nucleotide binding proteins. Biochem J 255:1,1988

16. Scott RH, Dolphin AL: Activation of a G protein promotes agonist responses to calcium channel ligands. Nature 330:760,1987

17. Okajima F, Ui M: ADP-ribosylation of specific membrane proteins by islet-activating protein, pertussis toxin, associated with inhibition of a chemotactic peptide-induced arachidonate release in neutrophils. J Biol Chem 259:13863,1984

18. Kashiwagi A, Huecksteadt TP, Foley JE: The regulation of glucose transport by CAMP stimulators via three different mecha- nisms in rat and human adipocytes. J Biol Chem 258:13685,1983

19. Tsien RY, Rink RJ, Poenie M: Measurement of cytosolic free Ca2+ in individual small cells using fluorescence microscopy with dual excitation wavelengths. Cell Calcium 6:145, 1985

20. Grynkiewicz G, Poenie M, Tsien RY: A new generation of Caz+ indicators with greatly improved fluorescence properties. J Biol Chem 260:3440,1985

21. Gilmore W, Weiner LP: The effects of pertussis toxin and cholera toxin on mitogen-induced interleukin-2 production: Evi- dence for G-protein involvement in signal transduction. Cell Immunol113:235,1988

22. Ribeiro-Net0 FAP, Mattera R, Hildebrandt JO, Codina J, Field JB, Bimbaumer L, Sekura RD: ADP-ribosylation of mem- brane components by pertussis and cholera toxin. Methods Enzy- mol 109:566,1985

23. Laemmli UK: Cleavage of structural proteins during the assembly of the head of the bacteriophage T4. Nature 277:680,1970

24. Robishaw JD, Smigel MD, Gilman AG: Molecular basis for two forms of the G protein that stimulates adenylate cyclase. J Biol Chem 261:9587,1986

25. Foster KA, McDermott PJ, Robishaw JD: Expression of G Proteins in rat cardiac myocytes: Effect of KCl depolarization. Am J Physiol259:H432,1990

26. Bonanou-Tzedaki SA, Setchenska MS, Arnstein HRV: Stim- ulation of the adenylate cyclase activity of rabbit bone marrow immature erythroblasts by erythropoietin and haemin. Eur J Biochem 155:363,1986

27. He Y, Hewlett E, Temeles D, Quesenbeny P: Inhibition of interleukin 3 and colony-stimulating factor I-stimulated marrow cell proliferation by pertussis toxin. Blood 71:1187, 1988

28. Williams LT: Signal transduction by the platelet-derived growth factor receptor. Science 243:1564,1989

29. Yarden Y, Ulrich A Molecular analysis of signal transduc- tion by growth factors. Biochemistry 27:3113, 1988

30. Berridge MJ: Inositol trisphosphate and diacyglycerol: Two interacting second messengers. Ann Rev Biochem 56:159, 1987

31. Mattera R, Graziano MP, Yatani A, Zhou 2, Graf R, Codina J, Birnbaumer L, Gilman AG, Brown AM: Splice variants of the (Y subunit of the G protein G, activate both adenylyl cyclase and calcium channels. Science 243304, 1989

32. Nakamura T, Ui M: Simultaneous inhibitions of inositol phospholipid breakdown, arachidonic acid release, and histamine secretion in mast cells by islet-activating protein, pertussis toxin. J Biol Chem 260:3584,1985

33. Ohta A, Okajima F, Ui M: Inhibition by islet-activating protein of a chemotactic peptide-induced early breakdown of inositol phospholipids and Cat+ mobilization in Guinea pig neutro- phils. J Biol Chem 260:15771, 1985

34. Imboden JB, Shoback DM, Pattison G, Stobo JD: Cholera toxin inhibits the T-cell antigen receptor-mediated increases in inositol trisphosphate and cytoplasmic free calcium. Proc Natl Acad Sci USA 83:5673,1986

35. Imagawa S, Smith BR, Palmer-Crocher R, Bunn HF: The effect of recombinant erythropoietin on intracellular free calcium in erythropoietin-responsive cells. Blood 73:1452,1989

36. Yelamarty RV, Miller BA, Scaduto RC, Yu FTS, Tillotson DL, Cheung JY: Three-dimensional intracellular calcium gradi-

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

492 MILLER ET AL

ents in single human BFU-E derived erythroblasts induced by erythropoietin. J Clin Invest 851799,1990

37. Hilton CY, Tan AS, Berridge Mv: Evidence for direct involvement of G-proteins in erythropoietin signal transduction. Blood 74:193a, 1989 (abstr, suppl 1)

38. Hescheler J, Rosenthal W, Hinsch K-D, Weilfern M, Traut- wein W, Schultz G: Angiotensin 11-induced stimulation of voltage- dependent Ca++ currents in an adrenal cortical cell line. EMBO J 7:619,1988

39. Tsuda H, Sawada T, Sakaguchi M, Kawakita M, Takatsuki K Mode of action of erythropoietin (Epo) in an Epo-dependent murine cell line. I. Involvement of adenosine 3', 5'-cyclic monophos- phate not as a second messenger but as a regulator of cell growth. Exp Hematol17:211,1989

40. Fukumoto H, Matsui Y, Obinata M: Mechanism of erythro- poietin action on the erythroid progenitor cells induced from murine erythroleukemia cells (TS A8). Development 105:109,1989

41. Katada T, Oinuma M, Kusakabe K, Ui M: A new GTP- binding protein in brain tissues serving as the specific substrate islet-activating protein, pertussis toxin. FEBS Lett 213:353, 1987

42. Bonanou-Tzedaki SA, Sohi MK, Arnstein HR: The role of CAMP and calcium in the stimulation of proliferation of immature erythroblasts by erythropoietin. Exp Cell Res 170276,1987

43. McCall SE, Kreis C, DiPersio JF, Borgeat P, Naccache P Involvement of guanine nucleotide binding proteins in neutrophil activation and priming by GM-CSF. Blood 73:588, 1989

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