The impact of an early truncating founder ATM mutation on immunoglobulins, specific antibodies and lymphocyte populations in ataxia-telangiectasia patients and their parents

Abstract

Eleven Norwegian patients (aged 2–33 years, seven males and four females) with Ataxia-telangiectasia (A-T) and their parents were investigated. Five of the patients were homozygous for the same ATM mutation, 3245delATCinsTGAT, a Norwegian founder mutation. They had the lowest IgG2 levels; mean (95% confidence interval) 0·23 (0·05–0·41) g/l versus 0·91 (0·58–1·26) g/l in the other patients (P = 0·002). Among the 11 A-T patients, six had IgG2 deficiency, six had IgA deficiency (three in combination with IgG2 deficiency) and seven had low/undetectable IgE values. All patients had very low levels of antibodies to Streptococcus pneumoniae 0·9 (0·4–1·4) U/ml, while normal levels were found in their parents 11·1 (8·7–13·4) U/ml (P < 0·001). A positive linear relationship between pneumococcal antibodies and IgG2 (r = 0·85, P = 0·001) was found in the patients. Six of 11 had diphtheria antibodies and 7 of 11 tetanus antibodies after childhood vaccinations, while 4 of 7 Hemophilus influenzae type b (Hib) vaccinated patients had protective antibodies. Ten patients had low B cell (CD19+) counts, while six had low T cell (CD3+) counts. Of the T cell subpopulations, 11 had low CD4+ cell counts, six had reduced CD8+ cell counts, and four had an increased portion of double negative (CD3+/CD4-/CD8-) gamma delta T cells. Of the 22 parents (aged 23–64 years) 12 were heterozygous for the ATM founder mutation. Abnormalities in immunoglobulin levels and/or lymphocyte subpopulations were also observed in these carriers, with no correlation to a special ATM genotype.

INTRODUCTION

Ataxia-telangiectasia (A-T) is an autosomal recessive disorder characterized by early onset progressive cerebellar ataxia, oculocutaneous telangiectasias, oculomotor apraxia, dysartria and immunodeficiency. Chromosomal instability and hypersensitivity to ionizing radiation are reflected in cancer susceptibility, especially lymphomas and leukaemia. Malignancies or chronic lung failure with pulmonary infections cause death in early adulthood. The responsible gene, ATM (A-T mutated), maps to chromosome 11q22.23 [1], spans ∼ 150 kb genomic DNA containing 66 exons and was identified in 1995 (MIM ♯208900) [2]. ATM protein kinase is involved in DNA double strand breaks (DSBs) response and repair [3]. Deficiencies in humoral and cellular immunity have previously been reported in A-T [4–6]. The aim of this study was to characterize immunological parameters such as immunoglobulins, specific antibodies and lymphocyte populations in the living Norwegian A-T patients and their parents. Since five patients were homozygous for the Norwegian founder mutation, 3245delATCinsTGAT [7], and 12 parents carried this mutation, we also looked for possible genotype correlations with clinical and immunological features.

MATERIALS AND METHODS

Patients/parents

Eleven patients, seven males and four females (aged 2–33 years), representing all living known Norwegian patients and their parents (aged 23–64 years, mean age 39) were studied. The patients had clinical signs of A-T and increased serum levels of alpha-fetoprotein. For this study, the patients were followed with a clinical examination and blood tests once a year for 2–3 years. Repeated blood samples were also obtained from the parents, and they were interviewed about cancer occurrence and susceptibility to infections.

ATM genotyping

Mutation analyses of DNA prepared from peripheral blood cells were performed using several different techniques like protein truncating test, denaturing gradient gel electrophoresis, heteroduplex analysis, denaturing high performance liquid chromatography followed by sequencing as previously described [7,8]. The ATM mutation results in eight of the patients have been reported [9].

Immunoglobulins and specific antibodies

Quantification of serum immunoglobulins, total IgG, IgA, IgM, IgE and IgG subclasses was performed by nephelometry (Dade Behring, Illinois, US). Lowest detection limits were 0·06 g/l for IgA and 3 kU/l for IgE. IgD was measured by immunodiffusion (Behring) with reference values according to Haraldsson et al.[10].

The in vitro toxin neutralization test on Vero cells in microculture was used for detection of diphtheria antibodies (detection limit: 0·01 IU/ml, protective level ≥0·1 IU/ml, relative protective levels: 0·01 up to 0·1 IU/ml) [11]. Tetanus antitoxin was measured with enzyme linked immunosorbent assay (ELISA) (detection and protective limit: 0·1 IU/ml IgG) [12]. IgG antibodies to Streptococcus pneumoniae were tested against the 23-valent polysaccharide vaccine and measured with ELISA [13], levels given in arbitrary units (U/ml). We compared our results to historical controls of healthy, unvaccinated adults [14], and levels of pneumococcal antibodies below 2·5 U/ml were regarded as nonprotective. Antibodies to the capsular polysaccharide of Hemophilus influenzae type b (Hib) were measured with ELISA using an antigen composed of Hib oligosaccharides conjugated to human serum albumin (HbO-HA) (protective limit: 1·0 µg/ml) [15]. Antibodies to viral antigens were measured using enzyme immunoassay (EIA) for antivaricella-zoster virus (VZV) IgG, antiherpes simplex virus (HSV) IgG, and antimeasles IgG. The microparticle enzyme immunoassay (MEIA) was used for detection of antirubella virus IgG and anticytomegalovirus (CMV) IgG. EIA was employed for detection of the anti-Epstein-Barr virus (EBV) nuclear antigen (EBNA) and anti-EBV virus capsid antigen (VCA) IgG.

Lymphocyte phenotyping and mitogen stimulation

Flowcytometric immunophenotyping of peripheral blood leucocytes was performed using the TruCount technique (Becton Dickinson, San Jose, CA, USA) with lysed heparinized blood. The monoclonal antibodies used were: anti-CD3, anti-CD4, anti-CD8, anti-CD19, anti-HLA-DR, anti-CD16, anti-CD56, anti-CD14, anti-TCR-αβ and anti-TCR-γδ. The normal ranges for the lymphocyte subpopulations were taken from Comans-Bitter et al.[16]. Mitogen stimulation was performed with concanavalin-A (ConA), phytohemagglutinin (PHA) and pokeweed mitogen (PWM). Stimulation index was defined as counts per min (cpm) divided with the spontaneous proliferation.

Data handling and statistics

The Regional and National Committee for Medical Research Ethics and the Norwegian Data Inspectorate approved this study, and signed consent was obtained from each person included. The data were collected in a Microsoft Access 97 database. To meet legal requirement, the patient administrative data were coded when registered. SPSS 10·0 and 11·0 for Windows were used for the statistical analyses. To compare two groups, t-tests were used, while one-way anova models were used to compare more than two groups. Bonferroni corrections were performed to accommodate multiple tests. To study the linear relationship between two continuous variables, Pearson coefficients and linear regression components were computed. Associations between categorical variables were studied with Fisher's exact test.

RESULTS

Genetics

Results of the DNA mutation analyses in the 11 patients are shown in Table 1. Five were homozygous for the founder mutation, 3245delATCinsTGAT, two compound heterozygous for this mutation, one homozygous for another frameshift mutation [9,17–19], and the last three patients were compound heterozygous with other ATM mutations.

Table 1. ATM mutations and infections in patients.

Identity	ATM mutations* Paternal/Maternal	ATM†	Sex	Age (years)	Skin	Respiratory	Others
NOAT1	3245delATCinsTGAT(fs)/3245delATCinsTGAT(fs)	R/R	M	20·5	Periungual warts, molluscum contagiosum, paronychia/onychomycosis, skin infections/impetigo, balanitis, cheilitis	Pneumonia, otitis media, sinusitis, tonsillitis	Aphtous lesions/mouth ulcers, gastroenteritis, diarrhoea
NOAT10	3245delATCinsTGAT(fs)/3245delATCinsTGAT(fs)	R/R	M	9·5	Periungual warts, molluscum contagiosum, paronychia/onychomycosis, cheilitis	Otitis media	Gastroenteritis, diarrhoea
NOAT11	3245delATCinsTGAT(fs)/3245delATCinsTGAT(fs)	R/R	F	7	Cheilitis, seborrhoic eczema	Pneumonia, interstitial lung disease
NOAT14	3245delATCinsTGAT(fs)/3245delATCinsTGAT(fs)	R/R	M	7	Cheilitis
NOAT16	3245delATCinsTGAT(fs)/3245delATCinsTGAT(fs)	R/R	M	4	Cheilitis, balanitis	Pneumonia	Urinary tract infections
NOAT17	6890 A > C(ms)/3245delATCinsTGAT(fs)	R/a	M	12	Molluscum contagiosum, cheilitis		Aphtous lesions/mouth ulcers
NOAT18	3245delATCinsTGAT(fs)/4110delG(ss)	R/a	M	7	Periungual warts, molluscum contagiosum, seborrhoic eczema, cheilitis		Diarrhoea
NOAT4	8978delGAAAinsAT(fs)/7875TG > GC(ds)	a/a	F	32	Herpes zoster, cheilitis		Diarrhoea, conjunctivitis/chalazion
NOAT13	8432delA(fs)/8432delA(fs)	a/a	M	6	Onychomycosis, cheilitis
NOAT15	4110delG(ss)/Not defined	a/a	F	15	Cheilitis	Chronic secretory otitis	Urinary tract infections
NOAT20	5932G > T(ns)/2880delC(fs)	a/a	F	3	Cheilitis	Otitis media	Diarrhoea

^*Mutations in ATM gene are designed according to recommended nomenclature [17–19]. Abbreviations: Ins, insertion; del, deletion; fs, frameshift; ds, double substitution; ms, missense; ns, nonsense and ss, splice site mutation;

^†R/R, homozygous for the Norwegian founder mutation; R/a, compound heterozygous for the founder mutation; a/a, other mutations in ATM, founder mutation excluded

Twelve parents were carriers of the Norwegian founder mutation (Table 2). Ten parents were carriers of seven other known ATM mutations and one yet undefined mutation.

Table 2. Parents' ATM mutations and results of immunological investigations.

Identity	Gender	ATM mutation*	Exon	ATM†	Age (years)	Immunoglobulins	Lymphocyte populations × 109/l	Mitogens SI-PHA/ConA/PWM‡
NOAT1	M	3245delATCinsTGAT (fs)	24	R	47		CD3+ 0·6 L, CD8+ 0·1 L
NOAT1	F	3245delATCinsTGAT (fs)	24	R	46		N	ND
NOAT10	M	3245delATCinsTGAT (fs)	24	R	43	IgD 278 IU/ml H	CD4-/CD8-/CD3+ 0·3 H	N
NOAT10	F	3245delATCinsTGAT (fs)	24	R	34	N	CD16+/CD56+/CD3–0·07 L	SI-PWM 3·3 L
NOAT11	M	3245delATCinsTGAT (fs)	24	R	38	IgM 0·34 g/dl L, IgA 0·66 g/dl L, IgE < 3 kU/l L	N	N
NOAT11	F	3245delATCinsTGAT (fs)	24	R	37	IgG3 0·14 g/dl L
NOAT14	M	3245delATCinsTGAT (fs)	24	R	35	IgG3 0·09 g/dl L	N	ND
NOAT14	F	3245delATCinsTGAT (fs)	24	R	41	N	N	ND
NOAT16	M	3245delATCinsTGAT (fs)	24	R	30	N	N	N
NOAT16	F	3245delATCinsTGAT (fs)	24	R	29	N	N	SI-ConA 2·3 L, SI-PWM 2·1 L
NOAT17	M	6890 A > C (ms)	49	a	50	N	N	ND
NOAT17	F	3245delATCinsTGAT (fs)	24	R	34	IgA < 0·06 g/dl L, IgE < 3 kU/l L	N	ND
NOAT18	M	3245delATCinsTGAT (fs)	24	R	38	N	N	ND
NOAT18	F	4110delG (ss)	30	a	39	IgD 210 U/ml H	N	ND
NOAT4	M	8978delGAAAinsAT (fs)	64	a	64	IgE < 3 kU/l L	CD3+ 0·67 L, CD19+ 0·08 L	ND
NOAT4	F	7875TG > GC (ds)	55	a	59	IgG3 0·15 g/dl L	N	ND
NOAT13	M	8432delA (fs)	60	a	36	N	N	ND
NOAT13	F	8432delA (fs)	60	a	33	N	N	ND
NOAT15	M	4110delG (ss)	30	a	41	N	N	ND
NOAT15	F	Not defined		a	38	IgG3 0·17 g/dl L	N	N
NOAT20	M	5932G > T(ns)	42	a	27	IgG3 0·13 g/dl L	N	ND
NOAT20	F	2880delC(fs)	21	a	23	N	N	ND

N, Normal; ND, Not done; Abnormal values are marked H for high and L for low according to reference values for adults (Supplement).

^*Mutations in ATM gene are designed according to recommended nomenclature; Ins, insertion; del, deletion; fs, frameshift; ds, double substitution; ms, missense; ns, nonsense and ss, splice site mutation;

^†R/R, homozygous for the Norwegian founder mutation; R/a, compound heterozygous for the founder mutation; a/a, other mutations in ATM, founder mutation excluded.

^‡SI, Stimulation index

Clinical history

Of the 11 A-T patients, 10 had minor to moderately increased susceptibility to infections, usually mucocutaneous and respiratory tract infections (Table 1). Six families reported that their child with A-T had more infections than classmates or siblings at similar age. Three patients had been hospitalized because of infection. Episodes with fever of unknown cause were reported in five patients. One patient (NOAT11) received prophylactic antibiotics, while none were substituted with immunoglobulin during the study period. Six patients had had chickenpox infection, three becoming severely ill, including one with protracted herpes zoster. Three of five patients homozygous for the founder mutation had recurrent pneumonia, while none of the others had this problem. Two patients (NOAT1, NOAT11) had acute lymphatic leukaemia (γδ T cell subtype) at the age of three and seven years, respectively. One of them (NOAT1) had a CNS relapse with another type of T cell expansion. Both patients' malignancies were successfully treated without irradiation and with reduced use of radiomimetic drugs. NOAT11 developed interstitial lung disease one year after chemotherapy, and was still on systemic steroids during our investigation. None of the others were treated with immunosuppressive drugs while included in this study. One patient (NOAT4) had diabetes mellitus, anaemia, bleeding tendency (epistaxis, nail haemorrhages) and lower limb pitting oedema. Another patient (NOAT20) had transient erytroblastopenia at the age of two years. Just after the end of this study the two oldest patients died, one (NOAT1) of respiratory failure complicated with pneumonia. The other (NOAT4) died of widespread thrombotic microangiopathy in the brain, and liver cirrhosis was also found on autopsy.

As part of the national vaccination programme, all patients had received diphtheria and tetanus vaccinations and seven the Hib vaccination. Seven had received the measles-mumps-rubella (MMR) vaccine at 15 months of age, and two patients had received measles vaccine (the only available before 1983), without adverse reactions. None had received pneumococcal vaccines.

Three parents carrying the founder mutation reported an increased susceptibility to infections, mainly of the respiratory tract such as sinusitis. All parents had followed the national vaccination program in childhood. None of the 22 parents had been treated for cancer.

Immunoglobulins

The patients homozygous for the founder mutation had significantly lower amount of IgG2 than the other patients; mean (95% confidence interval) 0·23 (0·05–0·41) g/l versus 0·91 (0·58–1·26) g/l (P = 0·002, Independent samples t-test, 2-tailed). No other immunoglobulin class was significantly related to this ATM gene variant (Table 3). Six patients had low IgG2 levels, and six had IgA deficiency. The patients with IgG2 deficiency had more respiratory infections including pneumonias, and most of those with IgA deficiency had recurrent diarrhoea. Variations over time in the immunoglobulin concentrations were found in each patient, particularly for IgD (See ranges Table 3). There were significant (P < 0·01, anova, Multiple comparisons, post hoc test, Bonferroni adjusted), not age or sex dependent, differences between the patients in the levels of IgG1, IgG2, IgG3, IgG4 and IgD.

Table 3. Mean values of Immunoglobulins in A-T patients.

Identity	ATM*	Number of tests	IgM g/l	IgD IU/ml	IgG3 g/l	IgG1 g/l	IgA† g/l	IgG2 g/l	IgG4 g/l	IgE‡ kU/l
NOAT1	R/R	2	1·72	695 H	0·19	4·7	<0·06 L	0·28 L	0·02 L	<3 L
NOAT10	R/R	2	1·41	81	0·35	9·0	<0·06 L	0·29 L	0·02 L	<3 L
NOAT11	R/R	3	1·22	275	0·16	8·2	0·78	0·07 L	0·12	3
NOAT14	R/R	2	0·72	166	1·16	8·9	2·15	0·41 L	0·19	3
NOAT16	R/R	2	0·83	60	0·12 L	7·2	0·63	0·10 L	0·02 L	<3 L
NOAT17	R/a	3	1·49	300 H	0·42	7·6	1·40	0·89 L	0·07	<3 L
NOAT18	R/a	3	1·09	370 H	0·29	7·4	<0·06 L	0·55 L	0·08	<3 L
NOAT4	a/a	2	7·99 H	370 H	0·12 L	28·0 H	<0·06 L	1·03 L	0·34	<3 L
NOAT13	a/a	4	1·53	58	0·70	12·4 H	<0·06 L	1·23	0·05	<3 L
NOAT15	a/a	4	1·88	79	0·46	8·6	2·39	1·27	0·01 L	5
NOAT20	a/a	3	1·61 H	89	0.11 L	7·1	0·07 L	0·53	0·09	5
Total mean		2·7	1·95	210	0·36	9·95	0·76	0·61 L	0·08	<3 L
Median individual range		3	0·40	55	0·04	1·65	0·04	0·17	0·02	0
Lowest-highest range		2–4	0·10–1·04	7–530	0·01–·31	0·20–4·50	0·00–·70	0·00–·75	0·00–·13	0–6

Immunoglobulin(Ig) subclasses are listed in table columns according to the Ig heavy chains constant region's genetic reading frame. Abnormal values are in bold, and marked H for high and L for low according to reference values for age (Supplement). Borderline values are given in Italics

^*See Table 1.

^†Lowest detection limit for IgA: 0·06 g/L.

^‡Lowest detection limit for IgE: 3 kU/L.

Ten of 22 parents had immunoglobulin levels outside the 90–95% central ranges, five of them were carriers of the founder mutation (Table 2). Five parents had low IgG3 (mean values 0·09–0·17 g/l), and three undetectable IgE (<3 kU/ml). One mother had IgA deficiency (<0·06 g/l) with anti-IgA antibodies. One father had low IgA (0·66 g/l) and low IgM (0·34 g/l). Two parents had increased IgD levels (mean value 278 and 210 IU/ml). These abnormalities were found in repeated tests. IgG2 levels were normal in all parents, 2·99 (2·35–3·62) g/l; mean (95% confidence interval).

Specific antibodies

The A-T patients had antibody levels to Streptococcus pneumoniae 0·9 (0·4–1·4) U/ml; mean (95% confidence interval) which were lower than their parents (P < 0·001, see data below), and also lower than expected for children and unvaccinated adults [14,20]. Patients homozygous or compound heterozygous for the founder mutation, had the lowest levels of pneumococcal antibodies, NOAT20 excluded (Table 4). A significant positive linear relationship existed between pneumococcal antibodies and IgG2 levels (r = 0·85, P = 0·001). Diphtheria antitoxin showed protective levels in six, relative protective levels in three and no protection in two patients (Table 4). Tetanus antitoxin showed protective levels in seven patients, while two had levels at the detection limit and in two no antitoxin was found. A significant positive linear relationship between diphtheria and tetanus antitoxin levels was found (r = 0·9, P < 0·0001). Antibodies to CMV, EBV (EBNA + VCA), HSV, VZV, measles and rubella were found in 6, 6, 1, 6, 6 and 7 patients, respectively (Table 4). No association to genotype was found. The patients who had experienced chickenpox (NOAT1, 4, 14, 15, 17 and 18), had specific antibodies. Of the eight who had received the MMR vaccine, antibodies to measles and rubella were found in five and six, respectively. Of the seven patients who had received Hib vaccines, four had protective levels of Hib antibodies (Table 4).

Table 4. Specific antibodies in A-T patients.

Identity	ATM*	Anti-Hib (µg/ml)	Hib + vaccine	Anti-diphtheria toxin IU/ml	Anti-tetanus toxin IU/ml	Anti-pneumococcal polysaccaride U/ml	Anti- CMV	Anti-† VZV†	Anti- HSV	Anti-EBV EBNA	Anti-EBV VCA	Anti- measles	Anti- rubella	MMR/ Me‡
NOAT1	R/R	0·73 L	N	0·01 L	<0·1 L	0·1 L	–	+	+	+	NT	–	–	Me
NOAT10	R/R	1·83	Y	0·64	1·0	1·2 L	–	–	–	–	+	+	+	MMR
NOAT11	R/R	0·30 L	Y	0·02 L	0·1 L	0·6 L	–	–	–	–	–	–	–	MMR
NOAT14	R/R	1·11	Y	0·04 L	0·2	0·5 L	+	+	–	+	NT	+/–	+	MMR
NOAT16	R/R	0·76 L	Y	0·16	0·7	0·2 L	+	–	–	–	+	–	–
NOAT17	R/a	0·54 L	N	0·01 L	<0·1 L	0·4 L	+	+	–	+	NT	+	+	MMR
NOAT18	R/a	2·83	Y	0·64	0·5	0·3 L	–	+	–	–	–	–	+/–	MMR
NOAT4	a/a	1·8	N	1·2	1·2	1·5 L	–	+	–	–	NT	+	+	Me
NOAT13	a/a	1·03	Y	0·16	0·2	2·1 L	+	–	–	–	–	+	+	MMR
NOAT15	a/a	0·67 L	N	2·56	8·9	2·1 L	+	+	–	–	NT	+	+	MMR
NOAT20	a/a	0·47 L	Y	0·04 L	0·1 L	0·5 L	+	–	–	–	+	+	+	MMR
Total mean		1·27		1·2	1·2	0·9 L

NT, Not tested; N/Y, No/Yes. Low values are in bold, and marked L for low. Borderline values, corresponding to relative protective levels, are given in Italics.

^*See Table 1.

^†Those patients who had had chickenpox were positive.

^‡MMR, Combination vaccine against measles/mumps/rubella; Me, Single vaccine against measles.

The parents had normal levels of pneumococcal antibodies, 11·1 (8·7–13·4) U/ml; mean (95% confidence interval) and protective levels of diphtheria and tetanus antitoxin antibodies showing a significant positive linear relationship (r = 0·7, P = 0·001). All had antibodies to VZV and measles (data not shown). Twelve, 20, 17 and 20 parents had antibodies to CMV, EBV (EBNA or VCA), HSV and rubella, respectively.

Lymphocytes

B cell (CD19+) counts were severely reduced in 10 patients and often T cell (CD3+) numbers were decreased, mainly CD3+/CD4+/CD8- but also CD3+/CD8+/CD4- cells (Table 5). Four patients had increased relative counts of double negative (CD3+/CD4-/CD8-) T cells, with a ratio of γδ positive to αβ positive cells exceeding 6 : 1. Also, seven patients had a high relative count of HLA DR + T-cells. Five of 10 patients showed reduced proliferative response to one or several mitogens (Table 5). Four of these were homozygous for the founder mutation. No correlation to genotype was found for the other lymphocyte parameters. The number of NK (CD16+/CD56+/CD3-) cells was increased in two patients, and NK cell counts in all patients were positively correlated (r = 0·8, p = 0·005) to their total lymphocyte counts, 1·50 (1·06–1·93) × 10⁹/l [mean, 95% confidence interval].

Table 5. Mean values of lymphocyte subsets in A-T patients.

Identity	ATM*	Number of tests	CD3+ (× 109/1)	CD4+ (× 109/1)	CD8+ (× 109/1)	CD4-/CD8- /CD3+ (% of CD3+)	CD4-/CD8- /TCR γδ + (% of CD3+)	CD4-/CD8- /TCR αβ + (% of CD3+)	HLA-DR + (% of CD3+)	CD19+ × 109/1	CD16+/ CD56+/CD3 × 109/1	SI-PHA†	SI-ConA†	SI-PWM†
NOAT1	R/R	2	0·24 L	0·17 L	0·06 L	2·9	4·8	0·8	24·0 H	0·04 L	0·37	4·3 L	2·8 L	2·7 L
NOAT10	R/R	2	0·3 L	0·21 L	0·08 L	6·0	1·5	3·2	25·2	0·12 L	0·42	2·4 L	6·8	2·9 L
NOAT11	R/R	2	0·2 L	0·16 L	0·03 L	5·9	12·5 H	5·0 H	46·0 H	0·71	0·27	3 L	3 L	3 L
NOAT14	R/R	2	0·97	0·39 L	0·46	14·6 H	13·1 H	0·8	28·0	0·09 L	0·52	25	26	11
NOAT16	R/R	2	1·31	0·30 L	0·83	13·7 H	11·6 H	1·0	23·3 H	0·18 L	0·72	19 L	3·7 L	1·5 L
NOAT17	R/a	2	0·46 L	0·27 L	0·15 L	13·3 H	12·8 H	0·2	32·4 H	0·05 L	0·35	ND	ND	ND
NOAT18	R/a	2	0·18 L	0·12 L	0·4	8·0	10·0	0·6	17·1	0·07 L	0·12	35	38	20
NOAT4	a/a	2	1·44	1·26	0·12 L	0·6	0·8	1·1	13·4	0·05 L	0·19	71	3·7 L	6·5
NOAT13	a/a	3	0·81	0·30 L	0·39	16·7 H	21·4 H	3·3	41·4 H	0·13 L	1·52 H	44	17	8
NOAT15	a/a	2	0·82	0·44 L	0·30	5·0	4·5	1·3	62·6 H	0·09 L	0·23	39	12	5·3
NOAT20	a/a	2	0·8 L	0·25 L	0·37 L	9·0	9·3	1·8	27·1 H	0·35 L	1·55 H	36	6·4	7
Total mean		2	0·68 L	0·35 L	0·29 L	8·1	9·3	1·7	31·0	0·17 L	0·57	27·9	11·9	6·8

ND, Not done. Abnormal values are in bold and marked H for high and L for low according to reference values for age (Supplement). Borderline values are given in Italics.

^*See Table 1.

^†SI, Stimulation index

Five out of 22 parents showed results outside the 90% central ranges, four of them were carriers of the founder mutation. Two fathers had low levels of CD3+ cells (mean counts 0·59 and 0·67 × 10⁹/l). One of them had low levels of CD8+ cells (0·1 × 10⁹/l), and the other low number of CD19+ cells (0·08 × 10⁹/l). Another father had a high level of double negative (CD3+/CD4-/CD8-) alpha beta positive cells (0·3 × 10⁹/l) in two of three samples, collected more than six months apart. Lymphocyte proliferation to mitogens was tested in seven parents. Two mothers, both heterozygous of the founder mutation, had moderately reduced responses. One of them also had low number of NK cells (0·07 × 10⁹/l). No significant differences were found between carriers of the founder mutation and those with other ATM mutations (Table 2). In the parents there were no clinical correlates to the various immunological findings, and the results in each parent differed from their child's immunological data.

DISCUSSION

The patients homozygous for the founder mutation had significantly lower amount of IgG2 than the other patients. They often suffered from recurrent pneumonias, and as the rest of the patients, they had low levels of pneumococcal antibodies. However, in general there was a discrepancy between the sparse clinical history of infections and the abundance of immunological findings in our A-T patients. Hence, their immune response seemed to be more quantitative than qualitative defective, as has been discussed by others [21,22]. Another characteristic finding was the great variations of the immunological results between the patients, even after multiple testing, as has been seen in siblings with A-T [23].

The ability of B cells to express immunoglobulins with identical antigen specificity, but different effector functions results from the cells' capacity to undergo class switch recombination (CSR). The constant (C) heavy chain genes are located in their reading frame, 5′ to 3′, as follows: Cµ, Cδ, Cγ3, Cγ1, Cα1, Cγ2, Cγ4, Cɛ, Cα2. In our study the A-T patients' levels of immunoglobulin isotypes situated at the beginning of the constant region's reading frame seemed to be increased and then successively reduced downstream the genetic reading frame (Table 3). This is compatible with the role of ATM protein in the end joining repair machinery in CSR. Recently, isotype switch recombination junctions in A-T patients were found to be aberrant [24,25]. The presence of hypergammaglobulinemia and oligoclonal-/monoclonal gammopathy in A-T is frequent as in our patient NOAT4, especially increase in IgM or both IgM and IgG [26].

Antibody response to Hib vaccine as well as to diphtheria and tetanus toxoid vaccines may give indirect information about the T cell function. We found a sufficient antibody response to diphtheria and tetanus vaccines and a partly successful response to Hib conjugate vaccine. This is promising for similarly constructed vaccines in A-T patients, for instance the new pneumococcal conjugate vaccine. All our patients had low antibody levels to S. pneumoniae. Others have also found a low amount of pneumococcal antibodies in A-T patients before and even after pneumococcal polysaccharide vaccine was administered [27]. Pneumococcal antibodies are normally oligoclonal and of IgG2 and IgA isotypes [28], and among our patients those with low IgG2 levels had the lowest levels of pneumococcal antibodies. This may explain the increased susceptibility to respiratory infections.

Both CD8+ and CD4+ T lymphocytes were reduced in number as in [4], not only the CD4+ cells as in [29,30]. The patients had an increased percentage of HLA DR+ T cells, while the total counts of these cells were normal to low (Table 5) [16]. The impaired proliferative response to mitogens that was found in half of our patients, has been seen by others [5], as well as the increased portion of double negative (CD3+/CD4–/CD8–) γδ T cells [4, 29, 31].

The abnormalities in the parents' lymphocyte population pattern and/or immunoglobulin classes seem too frequent to represent simple biological variation, even though low IgA or IgG3 is not rare in nonsymptomatic adults [32]. Low IgE [33] and selective IgA deficiency [34] have been found in A-T relatives as has decreased IgG4 and IgG3 [23], while others have not found antibody abnormalities in A-T parents [35]. Gatti and coworkers found a low mean T cell number in their A-T parents [36]. There was no occurrence of cancers in our small group of A-T parents. This could be due to their young age, or that half (15/22) of them had truncating mutations since ‘missense’ mutations are reported to be more related to malignancies in A-T carriers [37].

We show that the Norwegian founder mutation correlated with a specific immunological phenotype in our small cohort of patients. Others have also reported low IgG2 levels in their A-T patients [36,38]. It has been suggested that patients homozygous for a single mutation, especially near the N-terminal of the gene, before exon 25, have a severe clinical A-T phenotype [39]. Hence, an alternative explanation is that our finding reflects a phenotype of an early exon mutation rather than a specific founder mutation phenotype. Another of our patients (NOAT13) with a less severe immunodeficiency, was double homozygous for a different truncating mutation located closer to the C-terminal. An aberrant function of a truncated ATM protein or the existences of modifying genes coinherited with the founder mutation are other possible explanations. Genotype-phenotype relationships of neurological symptoms, radiosensitivity and life span have been shown by others in milder A-T variants and compound heterozygous with missense mutations involved [40–42]. However, there are reports of different missense mutations causing classical A-T [43,44] which are similar in onset, radiosensitivity and cancer incidence [45].

In conclusion, among the 11 A-T patients, the five patients homozygous for the founder mutation had the lowest IgG2 levels. Other genotype correlations were not observed. The immunological findings in the A-T patients may be valuable in planning the care, including vaccinations, of these patients and their families.