Expression of Pax5 gene in human haematopoietic cells and tissues: comparison with immunodeficient donors

Abstract

In mice, Pax5 gene is indispensable for B cell development. Pax5-deficient mice fail to produce mature B cells owing to complete arrest of B cell development at a precursor stage. However, the lineage and stage of human Pax5 gene expression have remained elusive. In this investigation expression of the human Pax5 gene was studied. Pax5 gene expression was detected in B cell lines but not in myeloma cell lines. CD19 expression was correlated with Pax5 gene expression. Adult spleen and bone marrow and fetal spleen and liver showed strong Pax5 gene expression, as did the corresponding mouse tissues, as reported previously. In common variable immunodeficiency (CVID) peripheral blood lymphocytes (PBL) with a decreased number of B cells, no Pax5 gene expression was detected. Some CVID PBL stimulated with IL-2, IL-10 and anti-CD40 monoclonal antibody, expressed the Pax5 gene. Defect of Pax5 gene expression in CVID may be caused by regulatory T cell disorder.

INTRODUCTION

The Pax5 gene encodes a B cell-specific activator protein (BSAP), which has been identified as a transcriptional factor that is expressed at early, but not late, stages of B cell differentiation [1]. Various binding sites for Pax5 in the promotors of B cell-specific genes have been identified, including sites in the promotors of the genes encoding CD19 [2], VpreB [3] and Blk [4], as well as multiple sites within the IgH locus, including a region upstream of Sγ2a and Sε [5]. Furthermore, Pax5 was identical to Sα-bp, which binds to two sites upstream of Cα, as well as to Sμ binding proteins [6–8]. Over-expression of Pax5 in splenic B cells stimulates proliferation of these B cells [9]. Experiments performed using mice that were Pax5-deficient due to targeted gene disruption revealed that Pax5-deficient mice fail to produce small preB, B and plasma cells, owing to a complete arrest of B cell development at an early precursor stage [10]. The expression and function of the Pax5 gene in mouse cells and tissues have been extensively investigated [11–13]. It is assumed that the Pax5 gene is specifically expressed in mouse B cell lineage haematopoietic cells. However, the lineage specificity of Pax5 gene expression in human cells remains obscure.

Common variable immunodeficiency (CVID) is characterized by recurrent infection and decreased serum immunoglobulin levels [14]. Some cases of CVID are associated with a complete absence of surface IgM⁺ B cells and immunoglobulins of all three major isotypes, as in Pax5-deficient mice [10]. In this study we analyse the expression of the human Pax5 gene in various haematopoietic cell lines and tissues and in CVID peripheral blood lymphocytes (PBL). In human cell lines the Pax5 gene was specifically expressed in B lineage cells, which expressed CD19. Myeloma cell lines did not express the Pax5 gene. In CVID with a decreased number of mature B cells among PBL, Pax5 gene expression was not detected. Stimulation with anti-CD40 MoAb and cytokines induced Pax5 expression in some CVID PBL. We discuss the role of Pax5 in human B cell differentiation and proliferation.

MATERIALS AND METHODS

Cell lines and human tissues

Human cell lines were cultured in RPMI medium with 10% fetal calf serum (FCS). The cell lines used in this study are listed in Table 1. PBL were prepared from heparinized blood using Ficoll–Hypaque. Human adult tissues were obtained at autopsy and fetal tissues at abortion before 24 weeks of pregnancy. Informed consent was obtained before the tissues were collected.

Table 1.

Expression of Pax5 gene and CD marker in various haematopoietic cell lines [16–28]

Immunophenotypic analysis

Cells were stained with a panel of diagnostic reagents in suspension using the direct and indirect immunofluorescence technique [15]. Briefly, cells were first incubated with a specific MoAb, in excess, for 30 min. In cases in which cytoplasmic immunoglobulin was stained, the cells were preincubated with ice-cold methanol. After being washed, cells were stained with FITC-conjugated goat anti-mouse F(ab′)₂ isotype-specific antisera (Cappel, CA). Testing for terminal deoxy-transferase (TdT) was done on methanol-fixed mononuclear cell smears with rabbit anti-calf thymus TdT antiserum followed by staining with FITC-conjugated goat anti-rabbit IgG (Cappel).

Reverse transcriptase-polymerase chain reaction for detecting mRNA expression

RNA was extracted from cell and tissues using Isogen (Nippon Gene, Toyama, Japan). cDNA was synthesized with MMTV reverse transcriptase using 1 μg or 100 ng or 10 ng of RNA and oligo dT-primer.

The polymerase chain reaction (PCR) primers were as follows: Pax5 sense 5′- AATGACACCGTGCCTAGCGT-3′, Pax5 antisense 5′-GGTGGTGAAGATGTCTGAGT-3′; CD19 sense 5′-TAAGTCATTGCTGAGCCTAGA-3′, CD19 antisense 5′-TCGCTGCTCGGGTTTCCATAA-3′; β-actin sense 5′-TGACGGGGTCACCCACACTGTGCCCATCTA-3′, β-actin antisense 5′-CTAGAAGCATTTGCGGTGGACGATGGAGGG-3′.

PCR was performed for 15–35 cycles (Fig. 1) in a PCR thermal cycler (Takara, Ootsu, Japan) at a denaturation temperature of 94°C for 1 min and extension at 72°C for 1 min; the annealing temperature was varied accordingly to the melting temperature for each specific pair of primers. The PCR products (10 μl of a total of 50 μl) were electrophoresed in a 8% acrylamide gel for Pax5 gene or 2% agarose gel for the other genes, stained with ethidium bromide and visualized using UV light.

Fig. 1.

Condition of reverse transcriptase-polymerase chain reaction (RT-PCR) for detection of Pax5 gene. cDNA was synthesized using: A, 10 ng mRNA prepared from Epstein–Barr virus (EBV)-transformed cell line; B, 100 ng mRNA; C, 1 μg mRNA. Lane 1, 15 cycles; lane 2, 20 cycles; lane 3, 25 cycles; lane 4, 30 cycles; lane 5, 35 cycles.

Stimulation with anti-CD40 MoAb, IL-2 and IL-10

PBL from case 3 or 5 were cultured in RPMI medium containing 10% FCS, 1% anti-CD40 MoAb (B-B20; Serotec, Oxford, UK), 50 U/ml of human IL-2 (Shionogi, Osaka, Japan) and 100 ng/ml human IL-10 (Genzyme, Cambridge, MA). After 3 days of culture, cells were washed with PBS three times and collected by centrifugation. Cells were then resuspended in the medium containing anti-CD40 MoAb, IL-2 and IL-10, or IL-2 and IL-10 and incubated for 4 days. After the cells were washed, RNA was extracted from them.

RESULTS

Condition of reverse transcriptase-PCR for detection of Pax5 gene

Semiquantification of Pax5 gene expression was performed by amplification of 10 ng, 100 ng and 1 μg mRNA prepared with 15–35 cycles from Epstein–Barr virus (EBV)-transformed cell line (Fig. 1). The faint band was visible when 100 ng mRNA were amplified with 25 cycles. In this study positive expression of Pax5 gene with the PCR conditions of 1 μg mRNA and 30 cycles was defined as showing the band intensity stronger than the PCR conditions with 100 ng mRNA and 30 cycles or with 1 μg mRNA and 25 cycles. Reverse transcriptase (RT)-PCR was then done using 1 μg mRNA and 30 cycles.

Pax5 gene expression in B cell lines and non-B cell lines

Most B cell lines we examined showed Pax5 gene expression (Fig. 2). In two myeloma cell lines, RPMI8226 and U-266, Pax5 gene expression was not detected (Table 1). In non-B cell lines, including myelomonocytic and megakaryocytic cell lines such as HL60 and MOLM-1, Pax5 gene expression was detected. In connection with CD19 surface expression, all cell lines showing CD19 surface expression we examined showed Pax5 gene expression. Expression of Tdt, immunoglobulin, CD13 and CD34 did not correlate with Pax5 gene expression.

Fig. 2.

Pax5 and CD19 gene expression in various human cell lines. Lane 1, LAZ221; lane 2, NALM-1; lane 3, NALM-6; lane 4, KOPN-8; lane 5, DND-41; lane 6, NALM-19; lane 7, NALM-20; lane 8, KG-1; lane 9, MOLM-1; lane 10, HL-60.

Pax5 gene expression in adult and fetal cells and tissues

In adult PBL, bone marrow and spleen Pax5 gene expression was detected (Fig. 3). In some cord blood samples, Pax5 gene expression was scarcely detected, although cord blood samples showed variation in Pax5 PCR products (data not shown). In fetal spleen and liver, Pax5 gene expression was detected. In fetal and adult brain Pax5 gene was scarcely detected.

Fig. 3.

Pax5 gene expression in human tissues. Lane 1, adult PBL; lane 2, cord blood; lane 3, adult bone marrow; lane 4, adult liver; lane 5, adult kidney; lane 6, adult spleen; lane 7, adult brain; lane 8, fetal liver; lane 9, fetal kidney; lane 10, fetal spleen; lane 11, fetal brain.

Pax5 gene expression in CVID PBL

In all CVID cases except case 7, decreased serum IgM, IgG and IgA levels were detected (Table 2). Some patients showed immunological phenotype-like X-linked aggamaglobulinaemia such as male and absence of surface immunoglobulin-positive cells, but all patients had no family history for immunodeficiency. Btk protein was detected in cases 1 and 2 (data not shown) [29]. Therefore we tentatively classified the patients as CVID. In case 7 the serum IgM level was almost normal but the serum IgG and IgA levels were decreased. No Pax5 gene expression was detected in PBL from any CVID cases except case 7 (Fig. 4a). Deduced from the comparison of the number of control surface IgM⁺ and CD19⁺ cells with CVID, CVID B cell number of PBL was not less than 10-fold dilution of control B cell number. Pax5 gene mRNA was detected in a sample of 10-fold diluted control PBL mRNA (Fig. 4b). This suggests that the decreased Pax5 gene expression in CVID PBL was due not only to the decreased number of B cells among the CVID PBL, but also to decreased Pax5 gene expression in a single CVID B cell compared with that in a single control B cell.

Table 2.

Immunological feature in immunodeficient patients

Fig. 4.

(a) Pax5 gene expression in common variable immunodeficiency (CVID) PBL. Lanes 1 and 2, control PBL; lanes 3–9, CVID cases 1–7. (b) cDNA was synthesized using: lane 1, 1000 ng mRNA prepared from control PBL; lane 2, 100 ng mRNA; lane 3, 10 n g mRNA; lane 4, 1 ng mRNA; lanes 5–9, CVID cases 1–5 PBL 1000 ng mRNA.

Induction of Pax5 gene expression in some CVID PBL stimulated with anti-CD40 MoAb, IL-2 and IL-10

CVID PBL were stimulated with anti-CD40 MoAb, IL-2 and IL-10. Arpin et al. reported that such stimulation induced generation of memory B cells and plasma cells in vitro [30]. In case 5, the stimulation induced Pax5 gene expression in CVID PBL (Fig. 5). Pax5 gene expression was stronger in CVID PBL incubated with anti-CD40 MoAb, IL-2 and IL-10 for 7 days than in CVID PBL incubated with anti-CD40 MoAb, IL-2 and IL-10 for 3 days and then IL-2 and IL-10 for 4 days. In case 3, Pax5 gene expression was slightly induced when incubated with anti-CD40 MoAb, IL-2 and IL-10 for 7 days.

Fig. 5.

Pax5 gene expression in common variable immunodeficiency (CVID) PBL stimulated with CD40, IL-2 and IL-10. Lane 1, CVID case3 PBL, no stimulation; lane 2, case 3 PBL, stimulation with CD40, IL-2 and IL-10; lane 3, case 3 PBL, stimulation with IL-2 and IL-10; lane 4, CVID case 5 PBL, no stimulation; lane 5, case 5 PBL, stimulation with CD40, IL-2 and IL-10; lane 6, case 5 PBL, stimulation with IL-2 and IL-10.

DISCUSSION

In this study it was revealed that the human Pax5 gene is expressed in B cell lines and some non-B cell lines. Moreover, CD19 expression is correlated with Pax5 gene expression; that is, all cell lines expressing CD19 we examined expressed the Pax5 gene. Myelomonocytic or megakaryocytic cell lines, such as HL60 and MOLM-1, showed Pax5 gene but not CD19 expression. Expression of Tdt, immunoglobulin, CD13 and CD34 did not correlate with Pax5 gene expression. These results support the hypothesis that Pax5 gene expression is coupled with CD19 expression [2]. In human adult or fetal tissues, where B lymphocytes were given a differentiation and proliferation including fetal liver, Pax5 gene was expressed. This observation is consistent with data obtained in experiments on mice which indicated that the Pax5 gene is expressed in two waves during mouse embryogenesis, with the first wave of expression resulting from transcription of the Pax5 gene in the developing central nervous system (CNS) and the second wave resulting from transcription of the Pax5 gene in fetal liver with B lymphopoiesis [1]. In our samples of fetal and adult brain, Pax5 gene expression was scarcely detected. This may suggest that in the human CNS, as in the mouse CNS, the Pax5 gene shows a spatial and temporal expression pattern in the developing CNS, and the fetal and adult brain tissues examined in this study were prepared when Pax5 gene expression was decreased, although further investigation to test this hypothesis is required.

In CVID PBL, Pax5 gene expression was not detected. This seems to reflect the decreased number of B cells among the PBL. It is of interest that cord blood from some normal infants scarcely expressed Pax5 gene. CVID patients often have immature B lymphocytes, similar to newborn infants. They have in common a markedly reduced expression of CD40 ligand and often low production of interferon-gamma (IFN-γ) and IL-4 [31–33]. In both instances, circulating B cells seem to be naive rather than defective. In some CVID PBL stimulated with anti-CD40 MoAb, IL-2 and IL-10, Pax5 gene expression was detected. Although the number of B cells among the CVID PBL was very small, these B cells had the ability to respond to anti-CD40 MoAb and cytokine stimulation. Consistent with our data, Eisenstein et al. reported that anti-CD40 MoAb and IL-10 induced differentiation of CVID B cells [34]. Thus, it seems that the possible explanation for lack of Pax5 gene expression in PBL of CVID patients and the induction of Pax5 gene expression by activated B cells of CVID patients reflects the immature status of B cells of these patients, probably due to a T cell defect that induces circulating B cells to activation or maturation [35].

Vorechovsky et al. analysed the Pax5 gene of mutation in the affected individuals from each multiplex family, 10 patients with sporadic CVID and 15 patients with sporadic IgA deficiency using PCR-single-strand conformation polymorphism (SSCP) and Southern blotting analyses, and detected no mutations [36]. Meffre et al. reported a human non X-linked agammaglobulinaemia immunodeficiency disease characterized by blockage of B cell development at early proB cell stage, which showed that Pax5 cDNA was normal [37]. It is known that lesions in human Pax3 and Pax6 genes cause Waardenburg's syndrome and aniridia [38, 39]. This evidence indicates that the Pax proteins are important regulators in early development. In humans, Pax5 deficiency might be involved in the development and implicate human neoplasia of B cell [40, 41] or neurological disorder other than immunodeficiency.