Peripheral blood haematopoietic progenitor cells in patientswith beta thalassaemia major receiving desferrioxamine ordeferiprone as chelation therapyEfthimia Vlachaki1, Elissavet Ioannidou-Papagiannaki1, Konstantinos Tziomalos2, StylianiHaralambidou-Vranitsa1, Vassilios Perifanis2, Ioannis Klonizakis1, Miranda Athanassiou-Metaxa2

1Second Department of Internal Medicine, Aristotle University of Thessaloniki, Hippokration Hospital, Thessaloniki; 2Thalassaemia Unit,

Hippokration Hospital, Thessaloniki, Greece

Survival expectance of patients with beta thalassaemia

major (TM) has significantly increased during the last

two decades. This can primarily be attributed to regular

blood transfusion and the implementation of chelation

therapy for secondary haemosiderosis. Desferrioxamine

is widely used as a chelation agent but requires subcuta-

neous or i.v. administration for at least 12 h on a daily

basis. Therefore, compliance with chelation therapy is

frequently suboptimal, particularly during adolescence,

resulting in excessive iron accumulation. Although two

orally active chelating agents (deferiprone and deferasi-

rox) are available nowadays in the treatment of haemo-

siderosis, there are still some mechanisms involved,

which are not well understood and studied (1, 2).

For the past few years – in Greece since 2000 – deferi-

prone is increasingly being used as an alternative to des-

ferrioxamine chelating regimen. Deferiprone (1,2-

dimethyl-3-hydroxypyrid-4-one) has the significant

advantages over desferrioxamine that the former is orally

active and chelates iron from the heart more efficiently

than desferrioxamine (2, 3). The main adverse effect of

deferiprone is the development of neutropenia, which

occurs via an unknown mechanism, particularly in non-

splenectomised patients and is reversible upon treatment

discontinuation (2).

Haematopoietic progenitor cells [granulocyte-erythro-

cyte-monocyte-megakaryocyte colony forming units

(CFU-GEMM), granulocyte-macrophage colony forming

Abstract

Objectives: The main adverse effect of deferiprone is the development of neutropenia, which occurs via

an unknown mechanism. We aimed to gain insight into the pathogenesis of deferiprone-induced neutrope-

nia by assessing the peripheral blood haematopoietic progenitor cells. Methods: Sixteen patients with

beta thalassaemia were studied; nine (Group A) were receiving desferrioxamine and seven (Group B) def-

eriprone. Ten healthy individuals comprised the control group (Group C). Results: Granulocyte-erythrocyte-

monocyte-megakaryocyte colony forming units were significantly more in Groups A and B compared with

Group C. Granulocyte-macrophage colony forming units (CFU-GM) were significantly more in Group B

compared with Group C. Macrophage colony forming units were significantly less in Group B compared

with Group C. Granulocyte colony forming units (CFU-G) were significantly more in Group A compared

with Group C. We found a trend in the difference in the number of CFU-G between patients’ groups

(P = 0.123). Adding serum from patients receiving deferiprone to cultures of controls resulted in a matur-

ation arrest of the granulocytic lineage. Conclusion: Our findings point to a maturation arrest at the level

of CFU-GM as a potential mechanism of deferiprone-induced neutropenia.

Key words neutropenia; granulocytic lineage; granulocyte-macrophage colony forming units; deferiprone; desferrioxamine; chelation

therapy; beta thalassaemia

Correspondence Konstantinos Tziomalos, 63 Solonos Street, Thessaloniki 54248, Greece. Tel: 00302310823487; Fax:

00302310992834; e-mail: ktziomalos@yahoo.com

Accepted for publication 20 September 2006 doi:10.1111/j.1600-0609.2006.00773.x

ORIGINAL ARTICLE

European Journal of Haematology ISSN 0902-4441

48ª 2006 The Authors

Journal compilation 78 (48–51) ª 2006 Blackwell Munksgaard

units (CFU-GM), granulocyte colony forming units

(CFU-G), macrophage colony forming units (CFU-M)

and burst-forming units erythroid (BFU-E)] originate

from differentiating stem cells and are committed to dif-

ferentiate to specific mature blood cells; they cannot be

discriminated in peripheral blood smears and they are

identified in short-term cultures by their ability to form

colonies consisting of one, two or even more cell lines

(4).

The aim of the present study was to gain insight into

the pathogenesis of deferiprone-induced neutropenia by

assessing the peripheral blood haematopoietic progenitor

cells of patients with TM receiving different chelators

and age-matched healthy individuals.

Methods

Sixteen patients with TM were included in our study.

Ten patients were male and six were female, with a med-

ian age of 26.0 yr (range: 20–30 yr). All patients were

being regularly transfused in order to keep the pretrans-

fusion haemoglobin level at approximately 9.5 g/dL.

Nine patients (Group A) were receiving chelation ther-

apy with desferrioxamine 45 mg/kg/d s.c. for 5 d/wk and

seven patients (Group B) with deferiprone 75 mg/kg/d in

three divided doses. In Group A, three patients had

undergone splenectomy and one was HCV-RNA posit-

ive, whereas in Group B four patients had undergone

splenectomy and none was HCV-RNA positive. Median

white blood cell count was 12 300/lL (range: 5000–

19 500/lL) and median ferritin values were 2274 lg/L(range: 618–4000 lg/L). Ten healthy individuals com-

prised the control group (Group C).

All subjects gave written informed consent. Venous

blood (10 mL) was drawn from all subjects and collected

into sterile tubes containing unfractionated conventional

heparin as anticoagulant. Mononuclear cells (2 · 105/

mL) were isolated from blood samples by centrifugation

with Ficoll-Hypaque (Sigma, St Luis, MO, USA) and

were plated on culture dishes coated with methylcellulose

(Methocult HC 4435; Stem Cell Technologies, Van-

couver, Canada). We also studied the effects of adding

100 lL of serum from patients from Groups A and B to

cultures of controls. Finally, we tested for formation of

erythroid colonies in the absence of erythropoietin

(Methocult HC 4533; Stem Cell Technologies). The cul-

tures were incubated at 37�C for 14 d in 5% CO2. Two

independent investigators evaluated the number of hae-

matopoietic progenitor cells per well; samples were

viewed with an inverted fluorescent microscope (Axiovert

25; Zeiss AG, Gottingen, Germany).

All data were analysed by using the statistical software

package spss (version 10.0; SPSS Inc., Chicago, IL,

USA). The Mann–Whitney test was used for compari-

sons between groups. A two-tailed P-value <0.05 was

considered statistically significant.

Results

Results are shown in Table 1. BFU-E were significantly

more in Groups A and B compared with Group C

(P = 0.001 and P = 0.015, respectively). CFU-GEMM

were significantly more in Groups A and B compared

with Group C (P < 0.001 and P = 0.007, respectively).

In contrast, CFU-GM were significantly more in Group

B compared with Group C (P = 0.05), but did not differ

significantly between Groups A and C. CFU-M were sig-

nificantly less in Group B compared with Group C

(P < 0.05), but did not differ significantly between

Groups A and C. CFU-G were significantly more in

Group A compared with Group C (P < 0.05), but did

not differ significantly between Group B and C.

Patients from Groups A and B did not differ signifi-

cantly in the number of any kind of colonies; neverthe-

less, we found a trend in the difference in the number of

CFU-G between patients’ groups (P = 0.123) that did

not reach statistical significance, possibly because of the

small number of patients studied (Fig. 1). We also

observed that adding serum from patients receiving def-

eriprone to cultures of controls resulted in a maturation

arrest of the progenitor cells of the granulocytic lineage,

whereas adding serum from patients receiving desferriox-

amine did not have such an effect. More specifically, in

the first case, we observed an increase in the number of

Table 1 Number of colony forming units

(CFU) of haematopoietic progenitor cells

in patients with thalassaemia on desferrioxam-

ine (Group A) or deferiprone (Group B)

treatment and in controls (Group C)

ColoniesGroup A(n = 9)

Group B(n = 7)

Group C(n = 10)

Burst-forming units erythroid 124 (62–205)1 129 (32–195)2 66.5 (55–80)

Granulocyte-erythrocyte-monocyte-megakaryocyte

colony forming units

3 (2–6)1 2 (1–7)2 1 (0–2)

Granulocyte-macrophage colony forming units 10 (5–35) 15 (6–17)2 10 (8–15)

Macrophage colony forming units 2 (0–4) 1 (1–2)2 2.5 (0–7)

Granulocyte colony forming units 10 (3–49)1 4 (2–13) 4.5 (2–7)

Number of colonies is expressed as median (range).1 Significant difference in the comparison between Groups A and C.2 Significant difference in the comparison between Groups B and C.

Vlachaki et al. Haematopoietic cells under deferiprone

ª 2006 The Authors

Journal compilation 78 (48–51) ª 2006 Blackwell Munksgaard 49

CFU-GEMM, whereas all other colonies of progenitor

cells of the granulocytic lineage showed a remarkable

decrease; in striking contrast, adding serum from patients

receiving desferrioxamine to cultures of controls did not

affect the number of any colony of progenitor cells of

the granulocytic lineage. This was uniformly observed

when we repeated this experiment in three separate con-

trols and the results are shown in detail in Table 2.

Development of BFU-E was not affected in either case.

Finally, we examined cultures from three patients from

Group A, four patients from Group B, and three con-

trols for endogenous erythroid colonies (EEC) formation.

EEC formation was apparent in cultures from all studied

patients from Group A (development of 45, 30 and 41

EEC, respectively), as well as in cultures from all studied

patients from Group B (development of 30, 15, 6 and 38

EEC, respectively), but in none of the studied controls.

When all 16 patients with TM were analysed together,

we found significantly less BFU-E in splenectomised

patients [median number, 94 (range: 32–129)] compared

with non-splenectomised patients [median number, 155

(range: 115–205)] (P = 0.002). In a separate analysis by

patient group, BFU-E were significantly less in splenec-

tomised patients compared with non-splenectomised

patients in both Groups A and B (P = 0.02 and

P = 0.03, respectively). We did not find any other differ-

ences in CFU between splenectomised and non-splenec-

tomised patients.

Discussion

Progenitor cells of the erythroid lineage are increased in

patients with TM, possibly because of the presence of

haemolysis and expanded erythropoiesis (5–7). In our

study as well, BFU-E were significantly more in patients

with TM. In addition, EEC formation was also apparent

in cultures from patients with TM regardless of the

applied chelation treatment; according to our knowledge,

this phenomenon has not been reported this far. It is well

known that EEC formation occurs in sickle cell disease

and sickle cell and beta thalassaemia compound hetero-

zygotes and is attributed to the presence of expanded

erythropoiesis (6, 7).

In patients with TM, the number of progenitor cells of

the erythroid lineage significantly correlates with the

number of progenitor cells of the granulocytic lineage

(5). In our study as well, CFU-GEMM, the more imma-

ture progenitor cells of the granulocytic lineage, were sig-

nificantly more in patients with TM, regardless of the

applied chelation treatment. Nevertheless, patients trea-

ted with deferiprone had significantly less CFU-M than

controls and did not show an increase in the number of

7 9 n = Group B Group A

CFU

-G (

mea

n ±

SD

)

40

30

20

10

0

–10

Figure 1 Mean number (±SD) of granulocyte colony forming units

(CFU-G) according to patient group. There is a trend in the difference

in the number of CFU-G between patients’ groups (P = 0.123).

Table 2 Number of colony forming units (CFU) of haematopoietic progenitor cells after adding serum from patients receiving desferrioxamine or

deferiprone to cultures of controls

Granulocyte-erythrocyte-monocyte-megakaryocytecolony forming units

Granulocyte-macrophagecolony formingunits

Granulocytecolonyforming units

Macrophagecolony formingunits

Burst-formingunitserythroid

1st in vitro test Control 1 0 9 2 5 80

+ desferrioxamine1 0 8 2 4 73

+ deferiprone2 1 1 0 1 76

2nd in vitro test Control 2 0 10 4 3 60

+ desferrioxamine1 0 8 4 3 65

+ deferiprone2 1 1 0 0 69

3rd in vitro test Control 3 1 15 6 1 55

+ desferrioxamine1 1 13 7 2 50

+ deferiprone2 2 2 0 0 61

1 After adding serum from a patient receiving desferrioxamine.2 After adding serum from a patient receiving deferiprone.

Haematopoietic cells under deferiprone Vlachaki et al.

50ª 2006 The Authors

Journal compilation 78 (48–51) ª 2006 Blackwell Munksgaard

CFU-G, in striking antithesis with patients receiving des-

ferrioxamine. Furthermore, our finding that patients

receiving deferiprone did exhibit more CFU-GEMM and

CFU-GM compared with controls, points to a matur-

ation arrest at the level of CFU-GM; we can only

hypothesise that this might represent a deferiprone-

induced effect. The effects of adding serum from patients

receiving deferiprone to cultures of controls further sup-

port this hypothesis, as we observed a remarkable

decrease in all colonies of progenitor cells of the granulo-

cytic lineage except CFU-GEMM; in contrast, adding

serum from patients receiving desferrioxamine did not

have such an effect. Neutropenia is one of the most

important side effects of deferiprone, occurring in 2.1%

of the patients, whereas agranulocytosis is encountered

in 0.4% of the patients; prompt recovery of white blood

cell number occurs once deferiprone is withdrawn (2).

The generative mechanism of this adverse effect remains

largely unresolved, but our results allow us to postulate

that deferiprone might induce a maturation arrest of the

progenitor cells of the granulocytic lineage. Whether it

also induces an increase in apoptosis remains to be

explored. Recent studies have provided evidence that cer-

tain novel chelating agents could evoke cancer cell apop-

tosis. Furthermore, chelators have also been shown to

bind trace elements that often play pivotal roles in the

function of enzymes, transcriptional factors and genes

(8). In an elegant series of experiments, Cunningham

et al. (9) studied the effects of deferiprone on bone mar-

row myeloid progenitors using the CFU-GM system and

showed that the toxicity of deferiprone to CFU-GM was

abrogated by addition of sufficient iron to saturate the

chelator, suggesting that, in vitro, free deferiprone is

toxic, at least in part, by depriving the cultured cells of

iron. They also proposed that, alternatively, iron might

prevent the chelator entering the cell by complexing it

(9). We did not assess the effects of iron in our study

and we cannot therefore provide further insight in this

putative pathogenetic mechanism. However, Cunning-

ham showed that deferiprone is approximately 16 times

less toxic than desferrioxamine to normal bone marrow

CFU-GM (9), in contrast to our findings. This apparent

discrepancy could be attributed to the fact that, while

Cunningham added deferiprone to bone marrow (9), we

have used serum from patients receiving deferiprone,

which also contains deferiprone metabolites that might

play a causative role in the induction of the maturation

arrest of the progenitor cells. Finally, Cunningham also

studied a patient with reversible deferiprone-induced

agranulocytosis and did not find evidence for increased

in vitro sensitivity of his progenitors to deferiprone (9).

Finally, we must emphasise that deferiprone did not

seem to affect the more immature forms of the granulo-

cytic lineage, in spite of the decrease in the number of

CFU-G and CFU-M. Furthermore, it is well known that

granulocytopenia rapidly resolves once treatment is

stopped. Therefore, we can conclude that deferiprone is

relatively safe and reduces complications in patients with

thalassaemia by improving compliance and cardiac per-

formance.

References

1. Maggio A, D’Amico G, Morabito A, et al. Deferiprone

versus deferoxamine in patients with thalassemia major: a

randomized clinical trial. Blood Cells Mol Dis 2002;28:196–

208.

2. Ceci A, Baiardi P, Felisi M, et al. The safety and effective-

ness of deferiprone in a large-scale, 3-year study in Italian

patients. Br J Haematol 2002;118:330–6.

3. Pennell DJ, Berdoukas V, Karagiorga M, et al. Random-

ized controlled trial of deferiprone or deferoxamine in beta-

thalassemia major patients with asymptomatic myocardial

siderosis. Blood 2006;107:3738–44.

4. Kralovics R, Prchal JT. Haematopoietic progenitors and

signal transduction in polycythaemia vera and primary

thrombocythaemia. Baillieres Clin Haematol 1998;11:803–

18.

5. Chen JS, Lin KH, Tsao CJ. Peripheral blood hematopoiet-

ic progenitor cells in beta-thalassemia major. Int J Cell Clo-

ning 1992;10:338–43.

6. Perlingeiro RC, Costa FF, Saad ST, Arruda VR, Queiroz

ML. Spontaneous erythroid colony formation in Brazilian

patients with sickle cell disease. Am J Hematol 1999;61:40–

5.

7. Bincoletto C, Perlingeiro RC, Saad ST, Costa FF, Queiroz

ML. Hydroxyurea promotes the reduction of spontaneous

BFU-e to normal levels in SS and S/beta thalassemic

patients. Hemoglobin 2001;25:1–7.

8. Le NTV, Richardson DR. Potent Iron Chelators Affect

Multiple Genes Involved in the p53 Tumor Suppressor Path-

way and Cell Cycle Control: A Gene Array Study. 13th

International Conference on Oral Chelation in the Treatment

of Thalassaemia and Other Diseases, July 12–15. Prague:

SOSTE Notiziario: 2003:33.

9. Cunningham JM, al-Refaie FN, Hunter AE, Sheppard LN,

Hoffbrand AV. Differential toxicity of alpha-keto hydroxy-

pyridine iron chelators and desferrioxamine to human hae-

mopoietic precursors in vitro. Eur J Haematol 1994;52:176–

9.

Vlachaki et al. Haematopoietic cells under deferiprone

ª 2006 The Authors

Journal compilation 78 (48–51) ª 2006 Blackwell Munksgaard 51