Erythroid cells in immunoregulation: characterization of a novel suppressor factor

Immunology Letters 93 (2004) 171–178

Erythroid cells in immunoregulation: characterization of a novel suppressor factor Galina V. Seledtsova a,∗ , Victor I. Seledtsov a , Denis M. Samarin a , Vladimir V. Senyukov a , Irina P. Ivanova a , Zoja A. Akimenko b , Irina G. Tsyrlova c , Steve S. Wolpe c , Vladimir A. Kozlov a a

Department of Immunohematology, Institute of Clinical Immunology, Russian Academy of Medical Sciences, 14 Yadrintsevskaya street, 630099 Novosibirsk, Russia b SRC VB “Vector”, Koltzovo, Novosibirsk Region, Russia c Pro-Neuron, Inc., Rockville, MD, USA Received 10 July 2003; received in revised form 22 February 2004; accepted 16 March 2004 Available online 20 April 2004

Abstract Nucleated erythroid cells (EC) have been previously reported to possess a potent natural suppressor (NS) activity for B-cell responses. In this study, we demonstrate that murine EC are able to reduce not only lipopolysaccharide (LPS)-driven B-cell proliferation, but also proliferative and cytotoxic T-cell responses generated in a primary allogeneic mixed lymphocyte culture (MLC); and that a soluble low molecular weight factor may be involved in such EC-derived immunoregulation. In addition, the erythroid cell-derived suppressor factor (ESF) was found to be capable of effectively reducing the allergen-driven proliferation of peripheral blood mononuclear cells (PBMC) isolated from allergic patients. From the data presented herein, it appears that ESF is heat-stable (80 ◦ C for 20 min) and has molecular weight (MW) lower or close to 0.5 kDa. ESF activity is resistant to both enzyme (trypsin plus chymotrypsin) proteolysis and action of the enzymes such as lipase and phospholipase C. On the other hand, ESF is effectively inactivated by neuraminidase treatment, suggesting the presence in its structure of sialic residue(s). The neuraminidase-sensitive, ESF-like activity is readily detected in the medium conditioned with normal mouse bone marrow (BM) cells. On fractionation of low MW erythroid products on a reversed-phase C16 column in a linear acetonitrile gradient (5–95%), ESF activity is detected in the first peak alone with the shortest time of its retention by the column. The results suggest that (1) by producing ESF, EC may regulate both B- and T-cell-mediated immune processes and (2) based on its physicochemical and biological characteristics, ESF can be distinguished from each of earlier characterised suppressor mediators of bone marrow origin. © 2004 Elsevier B.V. All rights reserved. Keywords: Erythroid cell; Natural suppressor activity; Erythroid cell-derived suppressor factor

1. Introduction

Abbreviations: Ab, antibody (antibodies); BM, bone marrow; Con, concanavalin; CM, conditioned medium(s); CTL, cytotoxic T-lymphocyte(s); EC, erythroid cell(s); Epo, erythropoietin; ESF, erythroid cell-derived suppressor factor; FCS, fetal calf serum; IMDM, Iscove’s modified Dulbeco’ medium; IL, interleukin; LME, l-leucyl methyl ester; LPS, lipopolysaccharide; MLC, mixed lymphocyte culture; NS, natural suppressor; PBMC, peripheral blood mononuclear cell(s); TGF-␤, transforming growth factor-␤; UV, ultraviolet ∗ Corresponding author. Tel.: +7-3832-28-26-73/49-46-40; fax: +7-3832-22-70-28. E-mail address: [email protected] (G.V. Seledtsova). 0165-2478/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.imlet.2004.03.011

There is compelling evidence suggesting nucleated erythroid cells (EC) to have a significant immunoregulatory potential. The previous studies have documented that EC are able to suppress antibody-mediated responses generated in the body with both thymus-dependent and -independent antigens, and that not only primary, but also secondary immune reactions are susceptible to EC-derived suppression [1–4]. According to the results obtained in our laboratory [5–8], EC release into the medium soluble products capable of immediately suppressing B-cell blastogenesis. One of those products was identified as a transforming growth factor-␤ (TGF-␤) [7]. In this study, we demonstrate that not

172

G.V. Seledtsova et al. / Immunology Letters 93 (2004) 171–178

only B-cell, but also T-cell responses may be controlled by EC and that a novel, neuraminidase-sensitive, low molecular weight (MW) factor may be involved in EC-derived immunoregulation.

2. Materials and methods 2.1. Mice C57BL/6 (B6; H-2b ) (C57BL/6 × CBA)F1 (CBF; H-2b /H-2k ) and (C57BL/6 × DBA)F1 (BDF; H-2b /H-2k ) mice were bred in our own facilities. DBA (H-2d ) mice were obtained from the Institute of pharmacology (Tomsk Scientific Center, Russia). The mice were both males and females and of ages ranging between 3 and 6 months. They received autoclaved food and boiled water. 2.2. Culture conditions Unless otherwise stated, the cells were cultured in RPMI 1640 medium supplemented with 10 mM HEPES, 2 mM l-glutamine, 5 × 10−5 M mercaptoethanol, 30 ␮g/ml gentamycin (all reagents from Sigma, St. Louis, MO), and 5% fetal calf serum (FCS) (ICI, Novosibirsk, Russia), at 37 ◦ C in a moist atmosphere of 5% CO2 /95% air. 2.3. Tumour cell line L1210 lymphoma cell line of DBA (H-2d ) origin has been obtained from Moscow Oncologic Scientific Center of Russian Academy of Medical Sciences and was maintained in RPMI 1640 medium supplemented with 10% FCS, 2 mM l-glutamine, and antibiotics. 2.4. Mouse EC The newborn EC were freshly isolated from the spleens of 1–2 days old mice by the technique earlier described in detail [7,8], including the preparation of single cell suspension, exposure of cells to anti-Thy1.2 monoclonal antibody (Ab) (Cedarlane Laboratory Limited, Ont., Canada) and the following removal of B, T-lymphocytes and non-specifically adherent cells by a procedure of “panning” on anti-Ig Ab-coated plastic dish. As revealed by cytological analysis of the Nocht–Macsimov stained smears, the separated cell population comprised no less than 95% EC mostly with basophilic cytoplasmas. For stimulating erythropoiesis, phenylhydrazine (Sigma), a haemolytic agent, was injected intraperitoneally into mice at a dose of 1.2 mg/ ml per mouse. After 12–20 h, the mice underwent additional injections of phenylhydrazine solution (0.6 mg/0.5 ml per mouse in each injection). The erythropoietic spleens were isolated from the mice at 36 h after the last injection and the single cell suspension was prepared from such spleens exactly as described [9]. For

separation of EC from other cells, the spleen cell suspension was first centrifuged over a Percoll solution (density 1.075 g/ml) at 400 × g for 30 min to deplete the cells with high density and erythrocytes. The cells at interface were then collected, extensively washed and incubated at 107 ml−1 for 30 min in serum-free RPMI 1640 medium with 15 mM l-leucyl methyl ester (LME) (Fluka Chemika, Buchs, Switzerland), a lysosomotropic compound that has been previously established [10,11] to be toxic to large granular lymphocytes and cells of monocyte/macrophage lineage. Following incubation with LME, the cells were extensively washed with a cold RPMI 1640 medium/10% FCS and further cultured at 3 to 5 × 106 /ml with 5 U/ml recombinant human erythropoietin (Epo, Cilag, Switzerland) in the serum-free Iscove’s modified Dulbeco’ medium (IMDM, Sigma) containing 0.5% bovine serum albumin (Sigma), 4 mM l-glutamine, 5 × 10−5 M mercaptoethanol and antibiotics, in a glass flask for 48–72 h. After culture with Epo, the non-adherent cell population comprised no less than 97% EC of various degree of maturity, as determined by cytological analysis of the stained smears and additionally confirmed by immunofluorescence with a monoclonal Ab MAE9 recognising a murine erythroblast antigen (AG–EB) [12]. 2.5. Human EC The execution of the experiments with human fetal cells has been approved by the Scientific Council and Ethics Committee at the Institute of Clinical Immunology of the Russian Academy of Medical Sciences. The liver is well known to be a haemopoietic organ in the fetus. The liver tissue was isolated from a human fetus (gestational age 16–22 weeks) after spontaneous or therapeutic abortion, and then prepared in the form of cell suspension, as described earlier [9]. The cells were cryopreserved in the routine way in RPMI 1640 medium containing 50% FCS and 10% dimethyl sulfoxide (Fluka), and further stored in liquid nitrogen until use. On the day of experiment, the cells were thawed at 37 ◦ C, washed extensively, treated with LME, as described above, and further cultured at 3 to 5×106 /ml with 1–2 U/ml human Epo in IMDM containing 0.5% BSA, 2 mM l-glutamine, 5 × 10−5 M mercaptoethanol and antibiotics, for 2–3 days. As a next step, the viable non-adherent cells were separated from dead cells by centrifugation on a Percoll solution (density 1.070 g/ml) and extensively washed. The cytological analysis, as well as immunofluorescence staining with monoclonal Ab HAE9 recognising human erythroblast antigen [13], indicated that the percent of EC in the prepared cell population was close to 100%. 2.6. Murine bone marrow (BM) cells Normal BM was obtained from the mouse tibias and femurs and prepared in the form of single cell suspension exactly as described [9].

G.V. Seledtsova et al. / Immunology Letters 93 (2004) 171–178

2.7. Cell-conditioned media and ultrafiltered solutions The washed EC or BM cells at 5 to 10 × 106 /ml were cultured in serum free IMDM supplemented 4 mM l-glutamine, 5 × 10−5 M mercaptoethanol, and antibiotics with for 24 h. The cell-conditioned media (CM) were then separated from cells by centrifugation, exposed to heating (80 ◦ C) for 20 min, and filtered through a 0.22 ␮m membrane. The ultrafiltered solutions were recovered from EC–CM by Amicon ultrafiltration using YM 10 or YM 05 filters at 30–40 psi. The control samples containing the serum-free culture medium alone were treated in the same way. After being filter sterilised, the ultrafiltered solutions were stored at 4 ◦ C until use. Immediately before use in experiments the solutions were heated to 80 ◦ C for 20 min. 2.8. Enzyme treatment The ultrafiltered, test and control samples (500 ␮l) were treated with the below-indicated enzymes according to the manufacture’s instructions (Sigma). After this treatment, the enzymes were inactivated by their exposing to 80 ◦ C for 20 min and then removed from samples by filtration through a Microcon-10 centrifuge unit (Amicon) at 12,000 × g for 30 min. 2.9. Solid phase extraction The test and control solutions were subjected to solid-phase extraction on an 4.6 × 250 mm Diasorb-C16 column (Biochimak, Russia) with particles of 10 ␮m size, using for elution a linear acetonitrile gradient (from 5 to 95%) in 0.1% trifluoroacetic acid. The flow rate was 1 ml/min. The extracts were detected by UV absorption at 280 nm with a spectral detector (Linear, US) and lyophilised immediately after their preparation. 2.10. Proliferation assays Mouse (CBF or BDF) spleen cells intact or depleted of T-cells by treating with Thy1.2 monoclonal antibody/complement (Cedarlane Laboratories Limited, Ont., Canada) [9] were cultured at 3 × 105 well with 12 ␮g/ml lipopolisacharide (LPS) from Escherichia coli (serotype B5:055, Sigma) in a 96-well round-bottom culture plate (BDSL, Ayrshire, England) for 48 h. Normal CBF spleen cells at 2 × 105 per well were cultured with the same number of BDF spleen cells for 48 h. In such culture, there is a strongly pronounced two-directional proliferative response and lymphokine production, preceding the generation of allospecific cytotoxic T lymphocytes [14,15]. The eight patients with allergic diseases (six men, two women; four with bronchial asthma and four with pollinosis) whose peripheral blood mononuclear cells (PBMC) demonstrated a significant proliferative response to the appropri-

173

ate allergen (dust mite or birch pollen) were selected for study. PBMC were isolated from heparinize venous blood by Ficoll–Verographin (density 1.078 g/ml) sedimentation in the standard way, extensively washed, and further cultured at 3 × 105 per well with the appropriate allergen (100 PNU/ml, Allergen, Stavropol, Russia) for 48 h. The test both cell and soluble additions were introduced into the cultures at the time of plating. The controls included the cells responding to a proliferative stimulus in the complete culture medium alone or in the presence of control cell or soluble additions. DNA synthesis was measured by pulsing individual wells with 0.75 ␮Ci of [3 H]thymidine during last 5 h of culture. The percentage suppression of cell proliferation was calculated as follows:
 cpm test 1− 100 cpm control 2.11. Cytotoxic T lymphocyte (CTL) activity assay Normal CBF spleen cells (3 × 106 per well) were stimulated with irradiated (20 Gy) DBA spleen cells (6 × 106 ) in a 24-well plate (Linbro), for 120 h. The EC cells (3 × 106 per well) from newborn CBF mice were introduced in the cultures at the time of plating. Instead of EC, irradiated B6 spleen cells were added into the control cultures. After being cultured the viable effector cells were mixed with [3 H]thymidine-labelled, L1210 cell targets (2×104 per well) at the below-indicated ratios in a 96-well round-bottom culture plate and CTL activity was further was determined by a 6 h, [3 H]thymidine residue assay exactly as previously described in detail [7,9,15]. The percentage of cytolysis was calculated by the formula analogous to that indicated above for calculating percentage suppression. 2.12. Statistics Experiments were performed at least three times and representative data (mean of triplicate cultures ±S.D.) are shown. Student’s t-test was used to determine the significance of the differences between experimental values. The differences pointed in the following text were statistically significant (P < 0.05). 3. Results 3.1. EC and ESF in suppressing lymphoblastogenesis It has been previously reported that EC freshly isolated from the spleens of phenylhydrazine-treated mice were able to markedly suppress LPS-driven B-blastogenesis, while not exerting a significant effect on T-cell proliferative response, as well as the generation of allospecific cytotoxic T lymphocytes in a murine allogeneic MLC [7]. The data presented in Fig. 1, however, indicate that murine newborn EC, but also Epo-expanded “phenylhydrazine” EC, were able

174

G.V. Seledtsova et al. / Immunology Letters 93 (2004) 171–178

According to the previously reported data [5–7], soluble molecules with MW lower or close to 10 kDa may be involved in EC-derived immunosuppression. Actually, as shown in the Fig. 3, the pattern of EC-derived suppressor activity was readily reproduced by EC–CM, as well as ultrafiltered solution containing erythroid products with MW lower or close to 0.5 kDa. In addition, the data presented in Fig. 4 indicate that a human low MW ESF was able to markedly reduce the allergen-driven proliferation of PBMC isolated from the allergic patients. Similar results were obtained in the experiments with a murine ESF (data not shown). Thus, the results suggest that, through releasing ESF, EC may regulate mitogen/antigen-induced immune processes which are based on both B-cell and T-cell proliferative responses. 3.2. Physicochemical characteristics of ESF

Fig. 1. Suppression by EC of lymphoblastogenes. The EC of B6 mouse, freshly isolated from newborn spleens (A) or expanded in vitro with Epo from adult erythropoietic spleens (B), were cultured at 1.5 × 105 per well with either T-cell-depleted spleen cells (3 × 105 per well) from CBF1 mice upon LPS (12 ␮g/ml) stimulation or MLC consisting of normal CBF and BDF spleen cells (each 2 × 105 per well) for 48 h. The control proliferative LPS and MLC response was close to the value between 38,000 and 43,000 cpm, respectively.

to considerably reduce not only LPS-driven B-cell blastogenesis, but also proliferative MLC response, although the latter in a somewhat lesser degree. In addition to their anti-proliferative activity, the murine newborn EC were found to be active in suppressing CTL generation occurring in a primary allogeneic MLC (see Fig. 2). CONTROL MLC MLC WITH EC

70 CYTOLYSIS (%)

60 50 40 30 20 10 0 -10

10

30

90

EFFECTOR-TO-TARGET CELL RATIO

Fig. 2. Suppression by EC of CTL generation. Normal CBF responder spleen cells (3 × 106 per well) were cultured for 120 h with irradiated (20 Gy) DBA spleen cells (6 × 106 per well) in the absence (control) or the presence newborn EC (3 × 106 per well) from CBF mice. The CTL activity was determined by a 6 h [3 H]thymidine residue assay using L1210 cells as targets. The control [3 H]thymidine incorporation in the labelled tumour targets incubated without cell effectors was 7200 cpm.

As indicated above, ESF is heat-stabile and has MW lower or close to 0.5 kDa. The data presented in Fig. 5A show that a human ESF activity inhibitory for B-cell blastogenesis was not diminished upon action of the enzymes such as trypsin plus chymotrypsin, lipase and phospholipase C. At once this activity was found to be totally inactivated by a neuraminidase chromatographically purified from Clostridium perfringens (Sigma product number 3001, activity of 2–5 units per mg protein), suggesting the presence in ESF of functionally significant a sialic residue(s). Similar results were obtained in the experiments with a recombinant neuraminidase (Sigma product number 7771) (data not shown). BM is well known to be an essential place of erythropoiesis in adult mammalians. Therefore, first of all, ESF-mediated regulation must cover BM cells. Consistent with this proposition, we found that <0.5 kDa ultrafiltered solution, recovered from murine BM cell-CM, contained the heat-stable, neuraminidase-sensitive factor suppressive for B-cell blastogenesis (see Fig. 5B). The picture of fractionating human, low MW, erythroid products on a C16 column in acetonitrile (5–95%) gradient is presented in Fig. 6. The ESF activity was stably associated with one UV-absorbing peak alone having the shortest time retention by the column. When added at 30% of its concentration occurring in the culture supernatant, ESF recovered in this fraction was able to provide not less than 25% suppression of B-cell blastogenesis. The remaining other fractions exhibited no ESF activity.

4. Discussion In line with the previous data [1–7], the results reported herein suggest EC to possess a natural suppressor (NS) activity that is defined as the ability of unprimed cells to exert non-specific suppression of various immune-mediated responses. Both B- and T-cell proliferative reactions may be susceptible to EC-derived suppression. These results

G.V. Seledtsova et al. / Immunology Letters 93 (2004) 171–178

< 0,5 kDa

80

80 70

70

60

60

SUPPRESSION (%)

SUPPRESSION (%)

EC-CM

MLC

LPS

50 40 30 20

50 40 30 20

10

10

0

0 7,5

15

30

7,5

SOLUBLE ADDITION CONCENTRATION (%)

(A)

175

15

30

SOLUBLE ADDITION CONCENTRATION (%)

LPS

MLC

80 80 70

70 60 SUPPRESSION (%)

SUPPRESSION (%)

60 50 40 30 20

50 40 30 20 10

10

0

0 7,5

15

30

7,5

15

30

-10

(B)

SOLUBLE ADDITION CONCENTRATION (%)

SOLUBLE ADDITION CONCENTRATION (%)

Fig. 3. Suppression of lymphoblastogenesis by ESF. LPS response of CBF1 spleen cells and MLC response of CBF and BDF spleen cells were generated in the presence of the indicated concentrations of mouse (A) or human (B) EC–CM or ultrafiltered (<0.5 kDa), ESF-containing solutions. The control proliferative LPS and MLC response was close to the value between 47,000 and 43,000 cpm, respectively.

Fig. 4. Suppression of allergen-driven, lymphocyte proliferation by ESF. PBMC isolated from the patients with allergic disease were stimulated with the appropriate allergen in the presence of 30% human ESF-containing, ultrafiltered (<0.5 kDa) solution. The control cell proliferations were in a range of 8000–14,000 cpm.

seem to be somewhat discrepant to our own previous data indicating that EC freshly isolated from the spleens of phenylhydrazine-treated mice were able to markedly reduce LPS-driven B-cell blastogenesis, while not significantly affecting T-cell proliferation and CTL generation in allogeneic primary MLC [7]. In our view, such inconsistency in EC action may be explained by difference in levels of functional activity of EC populations which were used in experiments. Like proliferating lymphocyte, which are much more functionally active than resting ones, Epo-stimulated EC might more broadly and significantly affect immunogenesis, compared with EC not receiving a growth stimulus. Indeed, in our experiments, after being stimulated with Epo, EC population greatly increased its own proliferative activity, as determined by a [3 H]thymidine uptake assay. The unmanipulated newborn EC found to be active in reducing not only B, but also T-cell responses also incorporated

176

G.V. Seledtsova et al. / Immunology Letters 93 (2004) 171–178

Fig. 5. The human ESF activity (A) and the murine BM cell-derived, ESF-like activity (B) after its enzyme treatment. The samples were treated with the indicated enzymes as described under Section 2. CBF spleen cells were stimulated with LPS in the presence of 30% ultrafiltered control or test samples. The proliferative responses into the control cultures (without ESF) were close to 50,000 cpm.

Fig. 6. ESF activity after solid phase extraction. The fractionation of low MW erythroid products was introduced on an C16 column as described under Section 2. The elution time was 30 min. The peak with ESF is hatched.

significantly more [3 H]thymidine, than did unstimulated EC prepared from the anaemic adult mice (data not shown). The cells with NS activity are believed to play an important role in the regulatory mechanism providing normal formation of the immune system. It has been previously suggested that the cells of lymphoid [16,17], macrophage [18,19] and uncertain [20] differentiation lineage may be responsible for NS activity occurring in the fetus and the newborn body. However, throughout early ontogenesis erythropoesis passes ahead of the development of all other haemopoietic lineages. According to the available data [21,22], mature erythroid progenitors, by themselves, can secret Epo and thereby stimulate their own growth and differentiation in an autocrine manner. The majority of the nucleated cells in the haemopoietic fetal liver and about 50% cells in the newborn spleen pertain to an erythroid lineage [19]. These data together with the findings directly indicating NS activity of fetal and newborn EC suggest that EC, more than other cells with NS activity, might play a leading role in controlling the formation of the immune system.

G.V. Seledtsova et al. / Immunology Letters 93 (2004) 171–178

The erythroid cell-derived suppression is not associated with impaired viability of antigen/mitogen lymphocytes ([4], but also our own unpublished data). Like other NS cells, EC mediate their immunoregulatory activity through releasing into the medium the soluble molecules capable of reducing lymphocyte proliferation. As previously reported [7], TGF-␤ may mediate in part NS activity of EC. The results have suggested involvement of more than one mediator in EC-derived immunosuppression. In agreement with this suggestion, we described herein ESF, a mediator, other than TGF-␤, which is capable of reducing both Band T-cell responses, including those that undergo allergic diseases. ESF was found to have MW lower or close to 0.5 kDa and to preserve its suppressive activity after heating at 80 ◦ C for 20 min or boiling for 5 min. This factor is resistant to the action of the enzymes, such as lipase and phospholipase C, but also to proteolysis by trypsin plus chymotrypsin. Moreover, in our own experiments not presented herein, the enzymes such as papain, tripain, protease from Streptomyces griseus and protease from Streptomyces pitosus, each alone or in complex, also failed to inactivate ESF. On the other hand, ESF was found to be highly sensitive to neuraminidase treatment, thereby suggesting the presence in ESF structure of a functionally significant sialic residue(s). BM is known to be a main place of erythropoiesis in adult mammalians. In our own experiments, a neuraminidase– sensitive, ESF-like activity was readily recovered in the medium conditioned with unmanipulated murine BM cells. These data support the suggestion that EC resident in normal BM may be important elements in the immunoregulatory mechanisms, not excluding, however, the possibility that BM cells, other than EC, may elaborate a ESF-related factor(s). ESF is characterised by a short time of its retention by the a reversed-phase column. This may suggest significant hydrophilic properties of this mediator. Sialic residue-containing gangliosides (sialoglycolipids) are well known to be possess a considerable suppressor activity that partly accounts for immunodepression in tumour-bearers [23,24]. ESF can be distinguished from gangliosides, based on its physicochemical characteristics including its MW. However, it is not excluded that ESF might be a sialic residue-containing product of metabolic cleavage of a ganglioside. In its physicochemical and functional characteristics ESF can be distinguishable from each of earlier identified immunosuppressor mediators of BM origin, including, a bone marrow-derived suppressor factor (BDSF), that has been reported to be a glycolipid with MW 1–1.5 kDa [25,26]. The primary structure of ESF is now under identification in our laboratory. Not only soluble mediators, but also a contact cell–cell interactions might be involved in EC-mediated immunohaemoregulation. The published data [27] evidence, that with differentiation process, EC become able to express

177

on their surfaces Fas ligand that can induce apoptosis of Fas-sensitive BM cells. The regulatory role of EC in haemopoiesis does not appear to be limited by immunosuppression. In fact, EC has been previously reported to be capable of elaborating a range of immunohaemoregulatory cytokines, such as granulocyte-macrophage colony-stimulating factor, interleukin 1 (IL-1), IL-4, IL-6, and interferon-␥ [28–30]. Consistent with these data, we have earlier established that EC-derived IFN-␥ may stimulate producing nitric oxide in macrophages [8], and that EC–CM is able to provide the growth of macrophage, granulocyte–macrophage and erythroid colonies generated in semisolid metylcellulose culture in the presence of Epo (data not shown) Thus, on the one hand, EC, through both production of a suppressor factor(s) and cell-to-cell contact, might prevent superfluous lymphocyte growth in the haemopoietic tissue, on the other hand, EC might create a balanced mediator microenvironment providing a multiple signalling required for normal development of different haemopoietic lineages, including lymphoid one. Erythropoiesis is well known to be an important element in the mechanisms providing adaptation of the body to changing conditions of environment. By enhancing their proliferative and functional activity, EC quickly react to various haemopoiesis-perturbing impacts, including antigen-mediated. Because approaches to erythropoiesis modulation are well founded, it is reasonable to anticipate that further investigations of EC-mediated immunohaemoregulation will significantly enrich an arsenal of clinically relevant procedures aimed at preventing the development of immunopathologic states.

Acknowledgements This study has been supported by the Regional Public Foundation of Russian Medicine Support.

References [1] DeMacario CE, Macario AJL. A new kind of immunosuppression associated with erythropoiesis. Immunol Lett 1979;1:23–6. [2] Macario AJ, DeMacario C, Dugan CB. Erythroblasts can generate immunosuppression in vivo. Medicina (B Aires) 1981;41(Suppl):83– 90. [3] Tsyrlova IG, Kashlakova NV, Kozlov VA. Influence of erythropoietic spleen-derived cells on proliferation of T and B lymphocytes. Immunologiya 1986;4:27–9 (in Russian). [4] Cheglyakova VV, Tsyrlova IG, Kozlov VA. Erythroid nature of bone marrow-derived natural suppressor cells. Immunologiya 1989;3:52–5 (in Russian). [5] Mitasov AV, Tsyrlova IG, Kiselev SV, Kozlov VA. Species-nonspecific action of soluble immunosuppressive factor produced by murine immature erythroid cells. Ann NY Acad Sci 1991;628:399– 409. [6] Samarin DM, Seledtsova GV, Seledtsov VI, Taraban VYa, Kozlov VA. Suppressor effect of immature erythroid cells on B-cell proliferation. Biull Eksp Biol Med 1997;123:66–70 (in Russian).

178

G.V. Seledtsova et al. / Immunology Letters 93 (2004) 171–178

[7] Seledtsov VI, Seledtsova GV, Samarin DM, Taraban VYa, Sennikov SV, Kozlov VA. Characterization of erythroid cell-derived natural suppressor activity. Immunobiology 1998;198:361–74. [8] Seledtsova GV, Seledtsov VI, Taraban VYa, Samarin DM, Kozlov VA. A role for interferon-␥ and transforming growth factor-␤ in erythroid cell-mediated regulation of nitric oxide production in macrophages. Immunology 1997;91:109–13. [9] Seledtsov VI, Avdeev IV, Morenkov AV, Seledtsova GV, Kozlov VA. Anti-proliferative effect of bone marrow cells on leukemic cells. Immunobiology 1995;192:205–17. [10] Thiele DL, Lipsky PE. Regulation of cellular function by products of lysosomal enzyme activity: elimination of human natural killer cells by a dipeptide methyl ester generated from l-leucyl methyl ester by monocytes or polymorphonuclear leukocytes. Proc Natl Acad Sci USA 1985;82:2468–72. [11] Thiele DL, Lipsky PE. The immunosuppressive activity of l-leucyl-leucine methyl ester: selective ablation of cytotoxic lymphocytes and monocytes. J Immunol 1986;136:1038–44. [12] Shipova LY, Baranov VN, Mechetner EB, Rosinova EN. Expression of erythroblast antigen (AG–EB) in haematopoietic and non-haematopoietic mouse cells during pre- and postnatal ontogenesis. Immunocytochemical study using monoclonal antibody. Exp Toxicol Pathol 1994;46:409–19. [13] Mechetner EB, Tonevitsky AG, Ievleva ES, Rozinova EN, Popova ON. Identification of a human erythroid cell surface antigen by monoclonal antibody HAE9. Exp Hematol 1987;15:355–9. [14] Ramseiier H. Early recognition of transplantation antigens by T lymphocytes. I. Test system and cellular parameters. Exp Cell Biol 1985;53:138–46. [15] Seledtsov VI, Avdeev IV, Seledtsova GV, Samarin DM, Prokopenko IV, Kozlov VA. Bone marrow cells as cytostatic effectors responsible for suppressing leukemia growth in vitro. Biomed Pharmacother 1995;49:293–9. [16] Mosier DE, Mathieson BJ, Campbell PS. Ly phenotype and mechanism of action of mouse neonatal suppressor T cells. J Exp Med 1977;146:59–73. [17] Argyris BF, Waltenbaugh C. Role of I–J in neonatal suppression. Cell Immunol 1983;80:267–78. [18] Snyder DS, Lu ChY, Unanue ER. Control of macrophage Ia expression in neonatal mice—role of a splenic suppressor cell. J Immunol 1982;128:1458–62.

[19] Piquet P-F, Irle C, Vassdalli P. Immunosuppressor cells from newborn mouse spleen are macrophages differentiating in vitro from monoblastic precursors. Eur J Immunol 1981;11:56–61. [20] Jadus MR, Parkman R. The selective growth of murine newbornderived suppressor cells and their probable mode of action. J Immunol 1986;136:783–92. [21] Stopka T, Zivny JH, Stopkova P, Prchal JF, Prchal JT. Human haematopoietic progenitors express erythropoietin. Blood 1998;91:3766– 72. [22] Sato T, Maekawa T, Watanabe S, Tsuji K, Nakahata T. Erythroid progenitors differentiate and mature in response to endogenous erythropoietin. J Clin Invest 2000;106:263–70. [23] Ravindranath MH, Irie RF. Gangliosides as antigens of human melanoma. Cancer Treat Res 1988;43:17–43. [24] Ravindranath MH, Morton DL. Role of gangliosides in active immunotherapy with melanoma vaccine. Int Rev Immunol 1991;7:303– 29. [25] McGarry RC, Anderson R, Singhal SK. Regulation of humoral immune responses by a bone marrow-derived glycolipid-like molecule. Cell Immunol 1982;71:293–300. [26] Saffran DC, Singhal SK. Suppression of mixed lymphocyte reactivity by murine bone marrow-derived suppressor factor inhibition of proliferation due to a deficit in IL-2 production. Transplantation 1991;52:685–90. [27] DeMaria R, Testa U, Luchetti L, Zeuner A, Stassi G, Pelosi E, Ricconi R, Felli N, Samoggia P, Peschie C. Apoptotic role of Fas/Fas ligand system in the regulation of erythropoiesis. Blood 1999;93:796– 803. [28] Sennikov SV, Eremina LV, Samarin DM, Avdeev IV, Kozlov VA. Cytokine gene expression in erythroid cells. Eur Cytokine Netw 1996;7:771–4. [29] Sennikov SV, Krysov SV, Injelevskaya TV, Silkov AN, Kozlov VA. Production of cytokines by immature erythroid cells derived from human embryonic liver. Eur Cytokine Netw 2001;12:274–9. [30] Sennikov SV, Krysov SV, Silkov AN, Injelevskaya TV, Kozlov VA. Production of IL-10, TNF-␣, IFN-␥, TGF-␤1 by different populations of erythroid cells derived from human embryonal liver. Cytokine 2002;17:221–5.