An imbalance of naive and memory/effector subsets and altered expression of CD38 on T lymphocytes in two girls with hyper-IgM syndrome

Abstract

In this report we evaluated CD4⁺ T, CD8⁺ T and natural killer (NK) cell counts, the levels of naive/memory subsets within the CD4⁺ T lymphocyte population, expression of CD38 on T lymphocytes, and CD4⁺ and CD8⁺ T cell cytokine production in two girls with hyper-IgM (HIM) syndrome. Both girls developed recurrent infections early in infancy, presenting a wide spectrum of clinical manifestations, with a strikingly different disease severity between them. CD4⁺ T cell counts were low in both children (patient 1: 214 cells/mm³ and patient 2: 392 cells/mm³), and the CD4/CD8 T cell ratio was 0·4 for patient 1, the patient with the more severe disease, and 1·4 for patient 2. NK cell numbers were low in patient 1 (60 cells/mm³) and borderline (286 cells/mm³) with regard to normal levels in patient 2. An imbalance of naive and memory/effector cell subsets was found in both girls, with the percentage of CD45RA⁺ 27⁺ (naive) CD4⁺ T lymphocytes being 5·8 and 12·4 for patients 1 and 2, respectively. Expression of CD38 on the surface of T lymphocytes was low in patient 1. Detection of intracellular interferon (IFN)-γ and tumour necrosis factor (TNF)-α in CD4⁺ and CD8⁺ T lymphocytes upon PMA-Io stimulus was preserved in both children. In conclusion, we found low numbers of CD4⁺ T lymphocytes and a dramatic redistribution of naive and memory/effector CD4⁺ T lymphocytes in two girls with non-X-linked HIM syndrome. Furthermore, we found low expression of CD38 on T lymphocytes and low numbers of NK cells in the patient with the more severe disease, indicating a possible role for these cells in the pathogenesis of this immunodeficiency.

INTRODUCTION

Hyper-IgM (HIM) syndrome represents a group of distinct entities with similar clinical features, characterized by recurrent infections, low or absent serum IgG, IgA and IgE and normal, or more often elevated, levels of IgM [1]. Approximately 70% of cases are X-linked, and associated with a mutation of the CD40 ligand (CD40L) gene, which encodes a protein expressed on activated T lymphocytes [1–4]. A subset of cases occurs in females displaying a non-X-linked form that arises either sporadically or involves autosomal recessive or dominant transmission [5–7]. Recently, it has been shown that these cases may be due to mutations of either the activation-induced cytidine deaminase (AID) gene or mutations of the CD40 gene [8,9].

Opportunistic infections, such as Cryptosporidium and Pneumocystis carinii, are also seen in affected patients at an abnormally high frequency, indicating that cell-mediated immune functions are compromised, even with normal numbers of CD4⁺ and CD8⁺ T lymphocytes [10,11]. Thus, there is increasing interest in the evaluation of T cell differentiation and function in patients with HIM syndrome. A decreased number of circulating CD4⁺ T cells in a subset of these patients has consistently been reported [12–14], but detailed lymphocyte immunophenotyping of these cells is still lacking.

In this report we evaluated CD4⁺ and CD8⁺ T and natural killer (NK) cell counts, naive/memory CD4⁺ T lymphocyte subset levels, expression of surface activation markers on CD8⁺ T lymphocytes, and CD4⁺ and CD8⁺ T cell cytokine production in two girls with HIM syndrome. These patients had clinical manifestations of different grades, which enabled both a comparison with normal values and with the severity of the disease.

PATIENTS AND METHODS

Patients and healthy volunteers

The Institutional Review Board approved the study and written informed consent was obtained from all participants in accordance with the Brazilian Ministry of Health Guidelines.

Case reports

Patient 1

Clinical manifestations began when she was 3 months old, with adenitis on the right axilla associated with Bacille Calmette-Guérin (BCG) immunization. She was treated with isoniazid for 3 months. At 7 months she was hospitalized with severe varicella. Between the ages of 1 and 5 years she presented with chronic diarrhoea, three episodes of pneumonia, one with pleural effusion, cellulitis in both legs, osteomyelitis in the left tibia and a massive abdominal lymphadenopathy that required exploratory surgery; surgery revealed a non-specific inflammatory reaction. During follow-up she presented with further episodes of otitis and sinusitis. Diagnosis of HIM was made when she was 7 years old; at this time she started on intravenous immunoglobulin (IVIG). At age 9 an inflammatory bowel disease was diagnosed, which improved after specific treatment. In 2000, despite high doses of IVIG (600 mg/kg/every 3 weeks), she was hospitalized with shingles. Laboratory evaluation at the time of diagnosis revealed a plasma IgM level of 432 mg/dl, undetectable levels of IgA and very low IgG levels (164 mg/dl). Post-IVIG treatment, IgM levels ranged from 168 to 1000 mg/dl. Isohaemagglutinins were normal. Specific antibodies (IgG) for vaccine antigens were negative despite regular immunization. At the age of 5 she had a negative reaction to PPD.

Patient 2

This patient presented with her first bacterial infection at age 2. During the following 2 years she presented with six other episodes of pneumonia, recurrent suppurative otitis and sinusitis. At age 3 she presented with a non-specific inflammatory epitrochlear lymph node enlargement, at which time she was referred to our division, and a diagnosis of HIM was established. In 1999, she presented with a skin lesion on the right arm. Biopsy showed chronic granulomatous inflammation, with caseous necrosis, suggestive of tuberculosis. Apart from the skin lesion, she has been in good health ever since. Laboratory evaluation at the time of diagnosis revealed plasma IgM of 834 mg/dl and undetectable IgA and IgG levels (IgA < 6 mg/dl and IgG < 33 mg/dl). Isohaemagglutinins were normal. She started on IVIG 400 mg/kg every 4 weeks, which resulted in plasma IgM levels ranging from 111 to 356 mg/dl.

Both patients were immunized according to Brazilian guidelines, which included the following vaccines: BCG, measles, oral polio vaccine (OPV) and diptheria–pertussis–tetanus (DPT). Except for the above-described satellite lymph node reaction after the BCG (patient 1), no other adverse effects were observed.

Neither patient had suggestive family history for hereditary immunodeficiency nor consanguinity parents. The first patient's four sisters and two brothers, as well as the second patient's brother, are all healthy.

Two age-matched girls were included as representative of normal values for lymphocyte immunophenotyping.

Blood sampling

Five millilitres of blood were collected from patients and healthy volunteers in heparin- and EDTA-treated vacuum tubes (Becton Dickinson, Plymouth, UK) and transferred immediately to the immunology laboratory. Patients were sampled again after 2 years of follow-up.

CD4⁺, CD8⁺ T cell and natural killer (NK) cell counts

For blood CD4⁺ and CD8⁺ T cell and NK cell counts, a TriTest and TrueCount reagent kit [Becton Dickinson Immunocytometry Systems (BDIS), San Jose, CA, USA] were used according to the manufacturer's instructions. Briefly, TrueCount tubes received appropriate amounts of TriTest reagents, containing CD3/CD4/CD45, CD3/CD8/CD45 or CD3/CD16⁺ 56/CD45 and 50 µl of EDTA-treated blood, and were incubated for 15 min at room temperature. Five hundred millilitres of lysing solution were added and incubated for 15 min before sample acquisition. Samples were analysed using Multiset software (BDIS).

Immunophenotyping of T lymphocytes

CD45RA fluorescein (FITC, clone L48), CD27 phycoerythrin (PE, clone L128), CD8 peridin chlorophyll protein (PerCP, clone SK1), CD4 PerCP (clone SK3) and CD3 allophycocyanin (APC, clone UCHT1) monoclonal antibodies were obtained from (BDIS), and CD38 fluorescein isothiocyanate (FITC, clone HIT2) was obtained from PharMingen (San Diego, CA, USA). One hundred microlitres of EDTA-treated blood were incubated at room temperature with a combination of monoclonal antibodies for 15 min in the dark, and then treated with haemolysis buffer for a further 10 min. Cells were washed and resuspended in phosphate saline buffer supplemented with 1% sodium azide for cytometric analysis.

Detection of intracellular tumour necrosis factor (TNF)-α and interferon (IFN)-γ in T lymphocytes

Peripheral blood mononuclear cells (PBMC) were obtained by Fycoll-Paque density gradient (Pharmacia), and after washing with NaCl 0·9% they were resuspended in RPMI medium, supplemented with 10 IU/ml penicillin G, 10 µg/ml streptomycin and 200 mm l-glutamine and 10% fetal calf serum (FCS), at a concentration of to 1 × 10⁶ cell/ml. One millilitre of cell suspension was incubated in the presence of phorbol myristate acetate and ionomycin (PMA-Io, 20 ng/ml and 1 mm, respectively) or with medium alone (control) for 1 h at 37°C at 5% CO₂. Monensin was added (2 µm) and the samples were further incubated for 5 h. Cells were surface stained with CD8-PerCP and CD3-APC, and incubated for 15 min at 4°C. After washing with 2 ml of staining buffer [phosphate buffered saline (PBS) with 1% FCS (Gibco, Gaithersburg, MD, USA), 0·1% sodium azide (Sigma), pH = 7·4–7·6], cells were resuspended in 1 ml of fixation buffer [PBS and 4% paraformaldehyde (Polyscience, Warrington, PA, USA), pH = 7·4–7·6], and left for 20 min at 4°C in the dark. The tubes were centrifuged, the supernatants discarded, and the cells resuspended in 1·5 ml staining buffer [PBS with 1% FCS (Gibco, Gaisthersburg, MD, USA), 0·1% sodium azide (Sigma), pH = 7·4–7·6] and stored overnight at 4°C in the dark. The cells were then centrifuged again and the supernatant was discarded. The samples were stained with 50 µl of diluted monoclonal anti-TNF-α-FITC or anti-IFN-γ-FITC (1 : 250) in permeabilization buffer for 30 min at 4°C in the dark, washed in 2 ml of permeabilization buffer, resuspended in 0·3 ml staining buffer, and analysed using a FACSCalibur flow cytometer (BDIS).

Flow cytometry analysis

Cell samples were processed in a FACSCalibur flow cytometer (BDIS) equipped with an argon and a diode laser for four-colour detection. Data acquisition and analyses were performed using CellQuest software (BDIS). Fluorescence voltages and compensation values were determined by using singly fluorochrome stained cells from a healthy volunteer. For lymphocyte acquisition, a forward and side scatter plot was used to establish a gate over the low side scatter and low to moderate forward scatter. The subpopulations of naive, effector and memory cells were analysed by the absolute number and percentage of cells expressing different combinations of CD45RA and CD27 on the CD4⁺ and CD8⁺ T lymphocytes. The state of activation of CD4⁺ and CD8⁺ T lymphocytes was analysed by measuring the surface expression of CD38. The percentage of CD4⁺ and CD8⁺ T lymphocytes producing either IFN-γ or TNF-α was determined by using dot-plots. In all cases quadrants were established in samples with control isotype antibodies.

RESULTS

T cell subset analysis revealed a pronounced decrease the in absolute numbers of CD4⁺ T, and a slight decrease of CD8⁺ T lymphocytes, compared to the estimated values for this age [15,16]. The percentage of CD4⁺ T lymphocytes was below normal levels in both patients, while the percentage of CD8⁺ T cells was high in patient 1 and low in patient 2. The CD4/CD8 T cell ratio was 0·4 for patient 1 and 1·4 for patient 2. NK cell numbers were extremely low in patient 1 (60 cells/mm³, 4%), while the absolute number was within the normal range (286 cells/mm³) and the percentage was high (20%) even in patient 2 [15,16] (Table 1).

Table 1.

T lymphocytes [cells/mm³ (%^*)] subsets and NK cells counts in two girls with HIM

	Patient 1		Patient 2
T lymphocyte	2000	2002	2000	2002	Reference value for age ([16])
CD3	1059 (92)	957 (94)	819 (60)	950 (63)	1800 (70%)†
CD4	214 (18)	324 (27)	392 (30)	379 (25)	800 (37%)
CD8	553 (49)	512 (54)	267 (20)	324 (21)	800 (30%)
CD4/CD8 ratio	0·39	0·63	1·47	1·17	1·00
NK	60 (4)	33 (3)	286 (20)	451 (29)	300 (12%)

^*Percentage of lymphocytes;

^†median values.

Naive and memory/effector cell subset levels were greatly altered in both girls compared to the described values for this age [15,16]. The percentage of CD45RA⁺ CD27⁺ cells (naive) in the CD4⁺ T lymphocyte population was 5·8 for patient 1 and 12·4 for patient 2 (Table 2). Effector/memory (CD45RA^-CD27⁺ and CD45RA^-CD27^-) cell levels were, in contrast, extremely high (76·2%/17·2% and 57·4%/24·9%), respectively, for patients 1 and 2.

Table 2.

Immunophenotyping of, and intracellular cytokine detection within, T lymphocytes from two children with HIM

	Patient 1		Patient 2
T lymphocyte	2000	2002	2000	2002	Reference values
CD4+ T lymphocyte (%)
CD45RA+ CD27+	5·8	9·9	12·4	10·0	55–67a
CD45RA−	93·4	89·1	82·3	79·0	4–55·3b1
CD38+	51·7	42·7	79·9	74·1	74·1–100b
IFN-γ	34·8		43·2		18·7 ± 5·6 (7·8–34·5)c
TNF-α	27·4		38·4		30·4 ± 5·5 (17·1–59·2)c
CD8+ T lymphocyte (%)
CD38+	64·9	44·0	88·7	80·3	64·7–100b
IFN-γ	66·5		68·2		36·9 ± 6(11·6–59·5)c
TNF-α	31·7		60·4		29·9 ± 6·3(10·4–59·8)c

^aErkeller-Yuksel et al. [16]: 25th to 75th percentile range of CD45RA⁺ cells.

^bMcCloskey et al. [15]: lower and upper limits to normal

^b1refers to CD45RO⁺.

^cMean ± s.d. (range) from 10 adult healthy volunteers obtained in our laboratory (unpublished data).

The expression of CD38 on CD4⁺ T cells was low in patient 1 (51·7%) and within the normal range in patient 2 (79·9%). The same trend for CD38 expression was seen in CD8⁺ T cells, with levels of 64·9% in patient 1 and 88·7% in patient 2 (Table 2). Studies performed after 2 years of follow-up showed similar results (Table 2).

Naive and memory/effector cell level within the CD4⁺ T lymphocyte population and expression of CD38 on CD4⁺ and CD8⁺ T lymphocytes were within the normal range for this age in two age-matched control girls. A representative result is depicted in Fig. 1.

Fig. 1.

CD4⁺ T lymphocyte differentiation and activation in two girls with HIM and an age-matched control. The subpopulations of naive (as defined by CD45RA⁺/CD27⁺) and effector/memory cells (CD45RA^–) were determined as the percentage of cells expressing these surface markers within the CD4⁺ T lymphocyte population (a). The activation of CD4⁺ and CD8⁺ T lymphocytes was determined by the expression of CD38 (b,c) within these two cell populations. Events were acquired by using a gate for lymphocytes based on forward versus side scatter parameters. T lymphocytes were characterized as either CD3⁺ CD4⁺ or CD3⁺ CD8⁺ cells. Expression of CD38 is represented by histograms and naive/effector and memory cells by dot-plots.

The percentage of IFN-γ producing lymphocytes in response to PMA-Io was 34·8% in patient 1 and 43·2% in patient 2 in CD4⁺+ T lymphocytes, and 66·5% in patient 1 and 68·2% in patient 2 in CD8⁺ T lymphocytes. The detection of TNF-α in CD4⁺ T cells was 27·4% in patient 1 and 38·4% in patient 2, and in CD8⁺ T cells detection was 31·7% in patient 1 and 60·4% in patient 2 (Table 2).

DISCUSSION

Hyper-IgM syndrome is less frequent in girls, and the immune dysfunction underlying the disease is not as well characterized as in the X-linked form seen in male counterparts. Cellular immune dysfunction, recurrent bacterial infections and other opportunistic diseases have been reported [6, 12, 13, 14,17, 18]. Both cases reported here developed recurrent infections in early infancy, presenting a wide spectrum of clinical manifestations, with a strikingly different severity of disease between both. Both presented upper and lower respiratory tract infections, which have been described previously as the most common clinical manifestations of the disease in a multi-centre study [19]. A high frequency of viral infections, including mild to severe herpes zoster infection, has been reported [14, 17, 20]. Patient 1 presented with severe varicella when she was 7 months old and a subsequent herpes-zoster at age 8, despite IVIG therapy.

Immunization with live organisms is not recommended in primary immunodeficiency (PID) patients due to the risk of disease progression. However, BCG immunization is recommended for all Brazilian children in the first month of life. As a consequence of such immunization, patient 1 developed axillary adenitis. A purified protein derivative (PPD) skin test performed some years after immunization was non-reactive in both patients, possibly reflecting the HIM-related T cell dysfunction. Such clinical complications following BCG immunization have been rarely reported [6, 19, 20].

Patients with non-X-linked HIM syndrome present a range of immune dysfunctions, which seem to be disassociated with abnormalities in CD40 or CD40 ligand expression [12]. However, mutations of the CD40 gene have been described recently in connection with this disease [9]. Both children in this report had decreased CD4⁺ T lymphocyte counts, in absolute and percentage numbers, confirming previous results [5,21]. Absolute CD8⁺ T cell numbers were also low in both patients yet, in contrast to CD4⁺ lymphocytes, numbers were higher in patient 1 than in patient 2 [16]. Moreover, the CD4⁺/CD8⁺ ratio was below 1 in patient 1 and greater than 1 in the child with the less severe disease, i.e. patient 2. Apart from the low cell numbers, we observed a striking imbalance in cell differentiation within the CD4⁺ T lymphocyte subset. Only a minor proportion of these cells were naive in both girls, a finding that is in striking contrast with both the normal values previously reported (Table 2) [15,16] and that were also found in the two age-matched control children. The expression of CD95 was higher in patient 1 than in both patient 2 and a healthy volunteer (data not shown). It is difficult, however, to associate the decreased number of CD4⁺ T lymphocytes with CD95 expression, as both patients presented low numbers of these cells and both had a similar imbalance of naïve and memory/effector cells. The low ratio of naive cells may result either from a low thymus output of naive CD4⁺ T lymphocytes or from a constant antigen priming of these cells as a result of the recurrent infections. Although the underlying mechanism is not understood yet, this finding coupled with the decreased CD4⁺ T lymphocyte counts may underscore the cellular immune deficiency seen in these patients.

NK cells play a role in the defence against varicella-zoster virus [22,23] and mycobacterial [24] infections. Patient 2, with the milder disease, presented NK cell absolute counts within the normal range, while patient 1 had a severe depletion of NK cells in the peripheral blood.

The state of cell activation was evaluated by the expression of surface activation markers on CD4⁺ and CD8⁺ T lymphocytes. Expression of CD38 was very low on CD4⁺ and borderline on CD8⁺ T lymphocytes in patient 1. Expression of CD38 on haematopoietic cells has been described as discontinuous: medullar thymocytes are CD38⁺, while peripheral blood T lymphocytes are mostly CD38^-, and activated T cells are strongly positive for CD38 [25]. CD38 is thought to be an important lymphocyte immunoregulatory molecule with adhesion and, importantly, ectoenzyme properties capable of catabolizing nicotinamide adenine dinucleotide (NAD⁺) to cyclic ADP-ribose (cADPR), and then hydrolizing cADPR to adenosine diphosphoribose (ADPR) [26–28]. Expression of CD38 on CD8⁺ T lymphocytes is a useful tool for evaluating the state of cellular activation. Expression has been shown to be up-regulated in HIV-infected patients as compared to negative controls and to be a predictor of disease progression [29]. There is a paucity of data in regard to age-related expression of CD38 on T lymphocytes. According to McCloskey et al. [15] the expected values for children aged 24–130 months are 90% (range 74·1–100%) and 86% (range 64·7–100%) for CD4⁺ and CD8⁺ T lymphocytes, respectively. We found values within these ranges in two age-matched control children; however, CD38 expression was down-regulated in patient 1, the patient with the more severe disease. Although the role of CD38 expression in the setting of HIM is not clear, the suppression of such a multi-functional glycoprotein suggests that it may be implicated in the described immune dysfunction.

It is well known that cellular functions are regulated by cytokines, and that these mediators play a major role in linking innate and specific immune responses. In the cytokine network, TNF-α and IFN-γ are known to be pivotal in the cellular immune response. No changes were noted in the ability of either CD4⁺ or CD8⁺ T lymphocytes to produce either TNF-α or IFN-γ in response to PMA-Io, in contrast to the evident cellular immunity impairment. There are few reports on cytokine production in HIM syndrome. A previous case report [21] demonstrated low production of IL-2 and IFN-γ. However, in a more appropriate approach, the preserved production of cytokines was shown in patients with X-linked HIM syndrome [30]. As in the presently reported cases, a potent stimulus (PMA-Io) was used and intracellular cytokine was detected at the single cell level by flow cytometry. Although these results suggest a preserved T cell cytokine response they need to be interpreted cautiously. Indeed, despite the conserved production of IFN-γ and TNF-α in response to mitogens, the decreased number of circulating T lymphocytes may lead to insufficient cytokine production. Further, it may not extend to the preserved response to recall and alloantigens; impairment in these responses has been shown to precede the derangement of mitogen response in HIV patients [31].

In conclusion, we found low numbers of CD4⁺ T lymphocytes and a dramatic redistribution in the subsets of naive and memory/effector T lymphocytes in two girls with non-X-linked HIM syndrome. Furthermore, we found low expression of CD38 on T lymphocytes and low numbers of NK cells in the patient with the more severe disease, indicating a possible role for these cells in the pathogenesis of this immunodeficiency.