Reduced Expression of the Interferon-Gamma Messenger RNA inIgG2 Deficiency

N. KONDO, R. INOUE, K. KASAHARA, T. FUKAO, H. KANEKO, H. TASHITA & T. TERAMOTO

Department of Pediatrics, Gifu University School of Medicine, Gifu, Japan

(Received 27 June 1996; Accepted in revised form 8 October 1996)

Kondo N, Inoue R, Kasahara K, Fukao T, Kaneko H, Tashita H, Teramoto T. Reduced Expression of theInterferon-Gamma Messenger RNA in IgG2 Deficiency. Scand J Immunol 1997;45:227–230

The specific defect that causes IgG2 deficiency, which is one of the primary immunodeficiencies, is unknown.Recently, it was shown that interferon-g (IFN-g) induces synthesis of human germline Cg2 transcripts. In theauthors’ previous study and the present one, peripheral blood lymphocytes (PBLs) of all five tested patientswith IgG2 deficiency failed to produce enough IFN-g when stimulated with phytohaemagglutinin orconcanavalin A although they produced a sufficient amount of interleukin-2 (IL-2). The low level of IgG2production in pokeweed mitogen-stimulated PBLs of four tested patients was improved by the addition ofrecombinant IFN-g. In this study, the amount of IFN-g messenger RNA showed various degrees of reductionin all five tested patients. Sequence analysis of the IFN-g coding regions and flanking regions revealed neithera point mutation nor a deletion for any of the patients. Thus the results suggest that the reduced expression ofIFN-g messenger RNA may play an important role in the IgG2 deficiency of these patients.

Naomi Kondo MD, Department of Pediatrics, Gifu University School of Medicine, Tsukasa-machi 40, Gifu500, Japan

INTRODUCTION

Differentiation of B cell isotypes (class switching) is a directedprocess that depends on the presence of particular activationsignals such as CD40 ligand and cytokines [1]. It was shown thatlipopolysaccharide (LPS)-stimulated murine B cells switch toIgG2a in the presence of interferon-g (IFN-g) [2]. Recently, itwas shown that IFN-g induces synthesis of human germline Cg2transcripts, which are further increased by Staphylococcusaureus Cowan I (SAC) plus interleukin-2 (IL-2) but not bySAC alone [3].

An IgG subclass deficiency is an antibody deficiencyassociated with the absence or severe reduction in the level ofone or two subclasses but with normal or increased levels ofother subclasses. The specific defect that causes IgG subclassdeficiency is unknown. Deficiency in IgG2 is sometimesassociated with low or undetectable IgG4 and/or IgA levels.Recurrent infections often occur in IgG2-deficient individualsand are manifested by sinopulmonary infections caused by theencapsulated bacteriaPneumococcusand Haemophilus.Wepreviously reported that defects in production of IFN-g play animportant role in IgG2 deficiency [4]. Here we report thatexpression of the IFN-g messenger RNA is reduced in IgG2deficiency.

MATERIALS AND METHODS

Patients. We studied two affected male siblings (Y.S., 3 years old andR.S., 7 years old) whose serum IgG2 levels were extremely low(< 2.7 mg/dl for each), as determined by enzyme-linked immunosorbentassay using monoclonal antibodies against the four human IgG sub-classes (HP6012 for IgG1, HP6014 for IgG2, HP6050 for IgG3 andHP6011 for IgG4, recommended by the World Health Organization), aswell as two affected female siblings (Ya.H., 3 years old and Yi.H., 6years old) and one affected male (S.Se., 2 years old) who showed lowserum IgG2 levels (32.8 mg/dl, 41.2 mg/dl, 38.8 mg/dl, respectively)with very low antibodies to polysaccharide antigens ofPneumococcus,from three unrelated families (Table 1). Middle-ear and upper respira-tory infections have occurred repeatedly in all five of the patients. Theirparents showed normal serum IgG subclass levels and IgA levels. Noneof the patients had significantly reduced percentages of circulatingCD3+, CD4+, CD8+, CD14+, CD19+, or CD20+ cells. Peripheral bloodlymphocytes (PBLs) of all five patients stimulated with phytohaemag-glutinin (PHA) or concanavalin A (ConA) for 24 h failed to produceenough IFN-g, although they produced a sufficient amount of IL-2(Table 1) [4]. Peripheral blood lymphocytes of all five patients stimu-lated with PHA or ConA for 48 h also failed to produce enough IFN-g.Interestingly, low level IgG2 production in pokeweed mitogen (PWM)-stimulated PBLs of four tested patients was improved by the addition ofrecombinant IFN-g [4]. Production of IgG2 from normal controls wasunchanged, or only slightly changed, by the addition of recombinantIFN-g [4].

Scand. J. Immunol.45, 227–230, 1997

227q 1997 Blackwell Science Ltd

Cell preparation.Peripheral blood lymphocytes were isolated fromheparinized blood by gradient centrifugation in Ficoll-Paque (Pharma-cia, Uppsala, Sweden). A solution of RPMI-1640 buffered with 10 mM ofN-2-hydroxyethyl-piperazine-N0-2-ethanesulphonic acid (Hepes) andsupplemented with 2 mM of L-glutamine, 100 units/ml of penicillin,100 mg/ml of streptomycin, and 10% heat-inactivated fetal calf serumwas used as a culture medium. The PBLs were suspended to give adensity of 1.5× 106/ml in the culture medium.

Complementary DNA (cDNA) and RNA polymerase chain reaction(RT-PCR).The PBLs were cultured with PHA for 6 h in a humidifiedatmosphere of 5% CO2 in air. Total cellular RNA was isolated fromcultured cells using a guanidine isothiocyanate method with caesiumchloride modification, and was quantified by absorbance at 260 nm. Forpreparation of IFN-g cDNA, total RNA (5mg) was reverse-transcribedusing 20 pmol of primer F8 (Fig. 1) and Moloney murine leukaemiavirus (M-MLV) reverse transcriptase (400 U, Gibco BRL, Life Tech-nologies, Gaithersburg, MD, USA) in 20ml of reaction mixture. Aliquotsof 5 ml of reaction mixture were subjected to cycling conditions of 1 mindenaturation at 948C, 2 min annealing at 548C, and 2 min polymerizationat 728C. The forty cycles were repeated using primers F7 and F8 (inFig. 1) using a DNA thermal cycler (Perkin Elmer, Norwalk, CT, USA).The PCR products were size-fractionated by 1.2% agarose gel andstained by ethidium bromide. The UV photograph was taken. Thepredicted size of the PCR product of the cDNA is 501 bp (Fig. 1).

Dot blot analysis.Dot blot analysis was carried out on 5mg, 0.5mg and0.05mg RNA samples [5]. After hybridization, the filter was washed twiceby boiling in 15 mM NaCl/1.5 mM sodium citrate/0.1% SDS, and it wasconfirmed that previous hybridized probes were completely washed out.

Southern blot analysis.High molecular weight DNA was extractedfrom peripheral neutrophils, then digested with PvuII, EcoRI, HindIII, orMspI. Digested DNA was size-fractionated by 1% agarose gel electro-phoresis, transferred to a nylon membrane and hybridized with IFN-g

probe.Probes.The IFN-g probes were prepared as follows. Primers were

chemically synthesized as shown in Fig. 1b. PCR using DNA from

healthy neutrophils and primers F1 and F2, or using healthy cDNA andprimers F7 and F8, was carried out with Taq DNA polymerase. Theb-actin probe (770 bp) was purchased from Oncor (Gaithersburg, MD,USA). The probes were labelled by a random primer using dig-system.

228 N. Kondo et al.

q 1997 Blackwell Science Ltd,Scandinavian Journal of Immunology, 45, 227–230

Table 1. IFN-g production in patients with IgG2 deficiency

Y.S. R.S. Ya.H. Yi.H. S.Se.Subjects (3 year old male) (7 year old male) (3 year old female) (6 year old female) (2 year old male) Controlsd

Serum concentrations, mg/dlIgG1 678 995 446 439 402 315–657IgG2 < 2.7 < 2.7 32.8 41.2 38.8 81–228

IL-2 production, U/mla

— 2.6 2.4 2.3 2.2 2.5 1.9–3.0ConA 16.6 20.9 20.5 16.0 25.6 15.2–31.6

INF-g production, U/mla

— 0.1 1.3 0.1 0.2 0.2 0.2–1.3ConA 1.1 2.6 1.2 1.7 2.1 13.7–29.5PHA 1.0 12.8 3.8 5.8 12.0 27.5–126.0

RPMI-1640 buffered with 10 mM of N-2-hydroxyethyl-piperazine-N0-2-ethanesulphonic acid (Hepes) and supplemented with 2 mM of L-glutamine,100 units/ml of penicillin, 100mg/ml of streptomycin, and 10% heat-inactivated fetal calf serum was used as a culture medium. The PBLs weresuspended to give a density of 1 or 1:5 × 106/ml in the culture medium.

aPBLs (1:5 × 106/ml) were cultured in culture-test tubes with or without ConA (25mg/ml) or PHA (10mg/ml) for 24 h. IL-2 with IFN-gconcentrations in supernatants were measured by radioimmunoassay (Medgenix, Fleurus, Belgium and Centocor, Malvern, PA, USA, respectively).

Fig. 1. The human IFN-g gene region containing three introns. (a) The50- and 30-untranslated regions of the IFN-g gene are shown as cross-hatched boxes and coding regions of exons (475–588, 1828–1896,1992–2174, 4600–4734 for each) are represented by solid boxes.Three introns (vertical arrows) are represented by open boxes. Thesense (→) and antisense (←) primers (F1–F8) used for PCR in thisstudy are showed by horizontal arrows. The open triangle shows PvuIIsite. The numbers of nucleotide sequences were as referred to in Ref.6. (b) The nucleotide sequences of sense and antisense primers (F1,201–221; F2, 639–659; F3, 1659–1680; F4, 2263–2287; F5, 444–4464; F6, 4912–4935; F7, 475–498; F8, 4711–4734) are represented.

Nucleotide sequencing.As shown in Fig. 1, the human IFN-g genecontains three introns and four exons [6]. The PCR-amplified DNAproducts of the IFN-g coding regions and flanking regions using primers(F1–F6) as shown in Fig. 1 were cloned into suitable vectors and fullysequenced in each case. Nucleotide sequences of amplified DNA weredetermined using DNA sequencer (Applied Biosystems, ABI, FosterCity, CA, USA).

RESULTS

To identify the basis for the failure of production of IFN-g in thepatients, we analysed IFN-g messenger RNA in PBLs of bothpatients and controls stimulated with PHA for 6 h. The RNAfrom all five patients was reverse-transcribed into cDNA andanalyzed by PCR. Use of primers designed to examine the entirecoding region of IFN-g (Fig. 1) gave PCR products of normallength for all five patients (Fig. 2, A-I). However, in dot blotanalysis, the amount of IFN-g messenger RNA showed variousdegrees of reduction among the patients compared with that of

healthy controls (Fig. 2, A-II). Subsequently, the PCR-amplifiedcDNA of the IFN-g coding region was fully sequenced in eachcase. Sequence analysis of the cDNA revealed neither a pointmutation nor a deletion in 12 or more clones from each patient.

Genomic DNA extracted from neutrophils was subjected toSouthern blot analysis using an IFN-g probe. The resultsobtained using several restriction enzymes revealed a uniformpattern of restriction fragments (Fig. 2B, and data not shown),indicating that no large deletion was present in the IFN-g generegion. Moreover, sequence analysis of the PCR-amplified DNAproducts of the IFN-g coding regions and flanking regionsrevealed neither a point mutation nor a deletion in 10 or moreclones from each product for each patient.

Thus our results indicate that the defective production ofIFN-g is due to reduced expression of IFN-g messenger RNA,although no mutation or deletion was found in the IFN-g codingregions and the flanking regions.

DISCUSSION

Mosmann & Coffman [7] reported that mouse helper T cellclones fall into two main groups (Th1 and Th2), definedprimarily by differences in the pattern of lymphokines synthe-sized. The Th1 clones synthesize IL-2, IFN-g and lymphotoxin,whereas expression of these lymphokines is not detectable in Th2clones. Conversely, only Th2 clones synthesize detectableamounts of IL-4 and IL-5. Wierengaet al.[8] presented evidenceof the presence of distinct functional subsets of CD4+ T cells inhumans on the basis of the production of IL-4 and IFN-g. In ourprevious study and the present one, PBLs of all five testedpatients with IgG2 deficiency stimulated with PHA or ConAfailed to produce enough IFN-g, although they produced asufficient amount of IL-2. These findings suggest that defectiveIFN-g production in IgG2 deficiency may be caused by a specificabnormality in the process of IFN-g production rather than areduction of a specific clone. Recently, Daltonet al.reported thatresting splenic natural killer cell activity is reduced in IFN-g-deficient mice [9]. Although none of the five patients hadsignificantly reduced percentages of circulating CD16+,CD56+, or CD57+ cells, natural killer cell activity was reducedin one of the five patients and was at the low end of the range ofnormal values in two (data not shown).

In our previous study, the low level of IgG2 production inPWM-stimulated PBLs of four tested patients was improved bythe addition of recombinant IFN-g [4]. Wilson et al. [10]reported that IFN-g messenger RNA levels in human neonatalblood mononuclear cells were markedly lower than those inhuman adult cells after incubation with ConA and PMA, com-pared with IL-2 messenger RNA levels. From these findings,they suggested that reduced IFN-g messenger RNA levels inhuman neonatal T cells were due to an intrinsic defect at thepretranslational level and indicated that the magnitude of IL-2and IFN-g gene expression could be independently modulatedpretranslationally [11]. In this study, the amount of IFN-g

messenger RNA showed various degrees of reduction in all

Reduced IFN-g RNA in IgG2 Deficiency 229

q 1997 Blackwell Science Ltd,Scandinavian Journal of Immunology, 45, 227–330

Fig. 2. (A-I) UV photograph of RT-RNA PCR products of IFN-g

transcript amplified from cDNA prepared from PHA-stimulated PBLsof patients and healthy controls. Lanes 1–5, patients (Y.S., R.S.,Ya.H., Yi.H. and S.Se., respectively); C, healthy controls. Molecularmass markers (M) are shown on both sides. The size marker (bp) isshown on the right. (A-II) Expression of IFN-g gene of PHA-stimulated PBLs on patients and healthy controls by dot blot analysis.Samples of 5mg, 0.5mg and 0.05mg of RNA were used. The PCRproduct (501 bp) from primers F7 and F8 was used as IFN-g cDNAprobe. 1–5, patients (Y.S., R.S., Ya.H., Yi.H. and S.Se., respectively);C, healthy controls. Expression ofb-actin was investigated as controls.(B) The IFN-g hybridization patterns of the PvuII-digested DNA oftwo patients and the family members. Lane 1, grandfather; lane 2,grandmother; lane 3, mother; lanes 4 and 5, patients (Y.S. and R.S.,respectively); lane 6, healthy control. The PCR product (459 bp) usingprimers F1 and F2 was used as the IFN-g probe. The position of theDNA size is given in Kb on the right.

five patients with IgG2 deficiency compared with that of healthycontrols, although the PBLs produced a sufficient amount of IL-2. Thus our results suggest that the reduced expression of IFN-g

messenger RNA may play an important role in the IgG2deficiency of these patients, although a change in IFN-g mes-senger RNA stability might also explain our results.

REFERENCES

1 Armitage RJ, Macduff BM, Spriggs MK, Fanslow WC. Human cellproliferation and Ig secretion induced by recombinant CD40 ligandare modulated by soluble cytokines. J Immunol 1993;150:3671–80.

2 Snapper CM, Peschel C, Paul WE. IFN-g stimulates IgG2a secretionby murine B cells stimulated with bacterial lipopolysaccharide. JImmunol 1988;140:2121–7.

3 Kitani A, Strober W. Regulation of Cg subclass germ-line transcriptsin human peripheral blood B cells. J Immunol 1993;151:3478–88.

4 Inoue R, Kondo N, Kobayashi Y, Fukutomi O, Orii T. IgG2deficiency associated with defects in production of interferon-gamma; comparison with common variable immunodeficiency.Scand J Immunol 1995;41:130–4.

5 Kondo N, Motoyoshi F, Kasahara K, Inoue Y, Mori S, Orii T. Failureof c-myc gene expression in B cells of some patients with commonvariable immunodeficiencies. Exp Clin Immunogenet 1992;9:109–16.

6 Gray PW, Goeddel DV. Structure of the human immune interferongene. Nature 1982;298:859–63.

7 Mosmann TR, Coffman RI. Th1 and Th2 cells: different patterns oflymphokine secretion lead to different functional properties. AnnuRev Immunol 1989;7:145–73.

8 Wierenga EA, Snoek M, de Groot Cet al. Evidence for compart-mentalization of functional subsets of CD4+ T lymphocytes in atopicpatients. J Immunol 1990;144:4651–6.

9 Dalton DK, Pitts-Meek S, Keshav S, Figari IS, Bradley A, StewartTA. Multiple defects of immune cell function in mice with disruptedinterferon-g genes. Science 1993;259:1739–42.

10 Wilson CB, Westall J, Johnston L, Lewis DB, Dower SK, Alpert AR.Decreased production of interferon-gamma by human neonatal cells,intrinsic and regulatory deficiencies. J Clin Invest 1986;77:860–7.

11 Lewis DB, Larsen A, Wilson CB. Reduced interferon-gammamRNA levels in human neonates, evidence for an intrinsic T celldeficiency independent of other genes involved in T cell activation. JExp Med 1986;163:1018–23.

230 N. Kondo et al.

q 1997 Blackwell Science Ltd,Scandinavian Journal of Immunology, 45, 227–230