We previously reported the clinical phenotype of two siblings with a novel inherited developmental and immunodeficiency syndrome consisting of severe intrauterine growth retardation and the impaired development of specific lymphoid lineages, including transient CD8 αβ T lymphopenia and a persistent lack of blood NK cells. We describe here the elucidation of a plausible underlying pathogenic mechanism, with a cellular phenotype of impaired survival of both fresh and herpesvirus saimiri-transformed T cells, in the surviving child. Clearly, NK cells could not be studied. However, peripheral blood T lymphocytes displayed excessive apoptosis ex vivo. Moreover, the survival rates of CD4 and CD8 αβ T cell blasts generated in vitro, and herpesvirus saimiri-transformed T cells cultured in vitro, were low, but not nil, following treatment with IL-2 and IL-15. In contrast, Fas-mediated activation-induced cell death was not enhanced, indicating a selective excess of cytokine deprivation-mediated apoptosis. In keeping with the known roles of IL-2 and IL-15 in the development of NK and CD8 T cells in the mouse model, these data suggest that an impaired, but not abolished, survival response to IL-2 and IL-15 accounts for the persistent lack of NK cells and the transient CD8 αβ T lymphopenia documented in vivo. Impaired cytokine-mediated lymphocyte survival is likely to be the pathogenic mechanism underlying this novel form of inherited and selective NK deficiency in humans.
The role of murine NK cells in protective immunity to viruses has been clearly established (1), but that of human NK cells remains unclear, in the absence of a well-defined primary immunodeficiency associated with selective NK deficiency. Four types of SCID, characterized by intrinsically impaired T cell development (2), are associated with a lack of NK cells. Adenosine deaminase deficiency and reticular dysgenesis also result in a lack of B lymphocytes. X-linked and autosomal recessive SCID due to mutations in the common γ-chain and JAK3, respectively (3)–both of which are components of the IL-2, 4, 7, 9, 15, and 21 pathways–are associated with a lack of T and NK lymphocytes but normal B cell development (4). A patient lacking NK cells but with a less severe T cell deficiency has also been described (5). He displayed impaired IL-2R β-chain expression but no genetic lesion was identified. Patients with T and NK cell defects have been reported to suffer from various infections, but the contribution of NK lymphopenia to the development of these infections remains unclear, given the known dominant contribution of T cells, as exemplified by IL-7R-deficient patients (6) who selectively lack T cells and present with a clinical phenotype comparable to that of NK-deficient SCID patients.
Only a couple of cases of apparently selective defects in human NK cell development have been reported (7). The first patient, reported in 1989, was an adolescent girl who initially presented disseminated, life-threatening varicella infection (8). She subsequently developed CMV pneumonitis and cutaneous HSV infections. However, NK cell counts were not determined before viral illnesses and varicella-zoster virus was recently shown to cause a decrease in NK cell numbers in otherwise healthy children (9). The patient remained healthy until the age of 13 and the patient was 17 when reported in 1989. NK deficiency persisted 6 years after the last infection (10), but a more recent follow-up indicated that she subsequently died of myelodysplasia (7), strongly suggesting that she probably suffered from an acquired, as opposed to inherited, hemopoietic deficiency (11). Moreover, this case was sporadic and the parents of the child were not consanguineous. A second patient with NK deficiency and recurrent varicella infections was recently reported; this patient died from one such infection at the age of 2 years, precluding any clinical and immunological follow-up (12). This case was sporadic too, but the parents were consanguineous, suggesting that their child may have had an inherited defect. These two patients (8, 12) may have suffered from an inherited defect of NK cells, suggesting that human NK cells may be involved in antiviral immunity. However, the lack of data for NK cell counts in these two patients before the onset of viral disease, the absence of identified familial cases, and the lack of disease-causing cellular phenotype make it impossible to draw firm conclusions.
We recently reported the first kindred with an inherited selective NK deficiency (13). Four related children from large, inbred Irish kindred were found to have very low counts of NK cells in the blood (below 4% and generally below 100 cells/mm3). One patient presented with EBV-driven lymphoproliferative disorder and two patients with severe pneumonitis of probable viral origin. The fourth patient has remained clinically healthy so far. The NK cell deficiency is a specific and inherited defect in this family, as it is a familial trait that was documented before any clinical infectious disease in at least one of the four patients. Moreover, it segregates as an autosomal recessive trait that is linked to the centromeric region of the chromosome 8 (logarithm of the odds score 4.51). This NK cell defect was therefore registered as a distinct nosological entity in the Online Mendelian Inheritance in Man (OMIM) database (OMIM:609981), unlike the previously reported patients (8, 12). The clinical phenotype of these patients strongly suggests that NK cells in human are also involved in antiviral immunity and also perhaps in antitumoral immunity (13). Although it is clear that the patients from this kindred suffer from a selective deficit of NK cells, including NKT cells, the pathogenic mechanism remains obscure. The disease-causing gene has not been identified and no disease-causing cellular phenotype was identified either.
We previously reported the clinical phenotype of a new developmental and immunological syndrome observed in two sisters, born to nonconsanguineous parents (14). The sisters showed severe prenatal and postnatal growth retardation, with strictly normal psychomotor development. They lacked detectable blood NK cells and displayed transient CD8 αβ T lymphopenia, affecting the CD45RO memory cells in particular. They had neutropenia, but presented no neutropenia-associated infectious diseases. The older sister died of CMV infection at the age of 18 mo (14). The severity of this infection may have resulted from NK cell deficiency, consistent with findings for the mouse model of mouse CMV infection (1, 15, 16, 17, 18, 19). However, NK cell deficiency was diagnosed after the onset of CMV disease; moreover, other phenotypic features, such as low CD8 cell counts, may also have been involved (20, 21). NK cell deficiency was diagnosed in the younger sister in the neonatal period. She is now doing well at 8 years of age and she has not yet been infected by any of the known human-tropic herpesviruses (this report). This familial congenital syndrome thus consisted of a specific NK cell deficiency associated with intrauterine growth retardation revealing the first possible link between NK cell differentiation and in utero development. We describe here a plausible pathogenic mechanism of NK cell deficiency and the probable disease-causing cellular phenotype, characterized by enhanced apoptosis of lymphoid cells due to impaired cell survival in response to IL-2 and IL-15.
Materials and Methods
We previously reported a novel complex syndrome, observed in two French sisters (patients P1 and P2) born to healthy, nonconsanguineous parents (14). This syndrome is characterized by severe pre- and postnatal growth retardation (−3 SDs for height and weight), facial dysmorphism, and immunodeficiency. No retardation of psychomotor development is observed (occipitofrontal circumference at birth: 32 cm for P1 and 33 cm for P2). The immunological phenotype consists of leukopenia with a lack of detectable NK cells (and no detectable NK cytotoxic activity), a transiently small CD8αβ CD45RO T cell fraction and neutropenia (typically 400–700/mm3). The eldest sister (P1) was born in 1993, at 37 wk of gestation. She died due to CMV infection at the age of 18 mo. The second child (P2) was born in 1998, at 38.5 wk of gestation. Bone marrow aspirate, collected 10 mo after birth, was normal, indicating that the leukopenia was a peripheral defect. The small proportion of CD8αβ CD45RO T cells in this patient was found to be a transient deficiency. Indeed, since the age of 3 years, P2 has presented a normal number of CD8αβ CD45RO T cells. T cell proliferation was normal after mitogen (PHA Ag, CD3) and Ag (tetanus toxoid and poliovirus) stimulation, as was the TCR Vβ2 profile. No major abnormality was found in the number of αβ T cells or their activity (14). Since our previous report, P2 has remained healthy and she is now 7 years old and has no prophylactic treatment. However, she remains seronegative for Abs against the herpesviruses commonly associated with illnesses of childhood (HSV, CMV, EBV, varicella-zoster virus). The findings of the most recent immunological investigation conducted for this child, at the age of 7 years, are reported in Results. Growth hormone (GH) treatment was initiated at the age of 4 years, to correct growth retardation. P2 had a height and weight four SDs below the mean at the beginning of treatment. By the age of 6 years, she had attained a height of 108 cm (−2SD) and a weight of 14.8 kg (−2SD). The parents of this patient gave informed consent for all investigations reported here. This work has been approved by appropriate institutional committee (Necker Comités Consultatifs de Protection des Personnes dans la Recherche Biomédicale (CCPPRB)).
Cell purification and culture
PHA-activated T cell blasts.
PBMCs were activated by incubation with PHA (1/700; BD Biosciences) for 3 days and then reisolated on a Ficoll density gradient and resuspended in Panserin 401 (Biotech), 5% SAB, 2 mM l
Herpesvirus saimiri-infected T cell lines.
Transformed T cell lines were generated as previously described (22) and cultured in Panserin 401/RPMI 1640 (v/v), 2 mM l-glutamine, 10% SAB, and 40 IU/ml IL-2.
PBMCs were plated in 96-well plates at a density of 1066 cells/ml and activated by incubation with IL-2 or IL-15. Herpesvirus saimiri-infected T cell lines were plated at a density of 5 × 105/ml and activated by incubation with IL-2.
Whole blood and PBMCs were stained with Annexin VFITC, 7-aminoactinomycin D (7AAD)3 (559925) and Abs against human CD4-PE (555347), CD8-PE (555367), CD14-PE (555398), CD19-PE (555413), or CD15-PE (555402). All these Abs were supplied by BD Biosciences/BD Pharmingen. From whole blood, RBC were lysed with BD PharM Lyse (555899) (BD Biosciences/BD Pharmingen). DNA fragmentation was assessed in PBMCs, PHA-activated T cell blasts, and herpesvirus saimiri-infected T cells, by washing cells in 0.9% NaCl and incubating the cells in a hypotonic solution containing propidium iodide (50 μg/ml), 0.1% sodium citrate, and 1/100 Triton X-100. We also conducted annexin V and propidium iodide staining (Annexin VFITC Detection kit II (BD Biosciences/BD Pharmingen)) on PHA-activated T cell blasts, according to the manufacturer’s instructions. The staining patterns obtained were analyzed by flow cytometry (FACScan: FACS; BD Biosciences). For Fas-induced apoptosis, PHA-activated T cell blasts were stimulated on day 4 with a human anti-Fas mAb (APO-1-3; Coger). For this stimulation, 2 × 105/ml cells in a 96-well plate were activated by incubation for 1 h in medium supplemented with 20 IU/ml IL-2 and the anti-Fas Ab (at a concentration of 100 or 250 ng/ml). We then added 10 μg/ml rabbit anti-mouse IgG Fcγ (Jackson ImmunoResearch Laboratories). Cells were incubated for 24 h and stained.
In vitro proliferation assays
Cell proliferation was assayed by adding 1.0 μCi [3H]thymidine to the medium for the last 18 h of culture. Radioactivity was counted with a Skatron apparatus (OSI). Cell proliferation was also measured by CFSE labeling: 3 × 105
473-Akt (9271; Cell Signaling), Akt1 (sc-5298; Santa Cruz Biotechnology), phosphorylated-Tyr705-Stat3 (sc-8059; Santa Cruz Biotechnology), Stat3 (sc-7179; Santa Cruz Biotechnology), phosphorylated-Tyr694-Stat5 (9351; Cell Signaling), and Stat5b (sc-1656; Santa Cruz Biotechnology). Blots were incubated with the secondary Ab (rabbit anti-mouse IgG Fcγ and mouse anti-rabbit IgG Fcγ (Jackson ImmunoResearch Laboratories)) for 1 h at room temperature. Immune complexes were detected by ECL, using an ECL kit (Amersham International) according to the manufacturer’s instructions.
Bcl-2 and Bcl-xL induction
PHA-activated T cell blasts were left unstimulated or were stimulated on day 6 with 10 or 100 IU/ml IL-2 or with 1 or 10 ng/ml IL-15. On day 9, cells were stained with FITC-conjugated anti-human Bcl-2 oncoprotein Ab (clone 124; DakoCytomation), anti-human Bcl-xL Ab (clone BXL03; Chemicon International), and PE-goat anti-mouse (Caltag Laboratories) or with FITC-mouse IgG1 (555748; BD Biosciences/BD Pharmingen), purified IgG1 (BD Pharmingen) as isotype controls, using the IntraStain kit (K2311; DakoCytomation), according to the manufacturer’s instructions.
Determination of cytokines by ELISA
At the age of 7 years, P2 had a lymphocyte count of 2200/mm3 (data not shown). There were 10% CD19+ cells and B cells had a normal phenotype, with CD27-switched B cells present in normal numbers. The Ab response to recall protein and polysaccharide Ags was normal. We found that 85% of the cells were CD3+, 52% CD4+, and 15% CD8+ (44% of which were memory CD45RO CD8+ T cells). Proportions of αβ and γδ T cells were not strictly normal, with γδ T cells somewhat overrepresented (20%). T cell proliferation in response to recall Ags (tetanus toxoid and poliovirus) was normal. Moreover, regulatory T cells, identified based on intracellular Foxp3 expression (23), were detected in normal numbers. Normal CD56 expression was observed on a subset of T cells (2% CD3+CD56+) but only 0.2% of the cells were NK cells (CD3−CD56+). The expression of other NK markers (CD94, NKG2A, NKRP1, CD62L, and KIR2DL1/2DS1, KIR2DL2/2DL3/2DS2, KIR3DL1/3DS1, KIR2DS4, KIR3DL2) on CD3+ cells was normal (Table I⇓). P2 almost completely lacked NKT cells (0.01% for normal values between 0.05 and 0.15%), as defined by flow cytometry (CD3+, TCRVα24+/Vβ11+, or CD1d-tetramer+). Finally, CD14+ and CD16+ monocytes were present in normal numbers, as were myeloid dendritic cells and plasmacytoid dendritic cells. Polymorphonuclear neutrophil (PMN) levels were consistently low, between 400 and 700/mm3. We detected no autoantibodies against PMNs. Platelet and erythrocyte counts were also normal. Thus, overall, P2 lacked detectable NK cells, had very low counts of NK-T cells, and had low numbers of PMNs at 7 years of age.
Proliferation of fresh T cells
The persistent lack of peripheral NK, NKT cells, and the transient lack of CD8 memory T cells suggested that our patient might present impaired lymphoid stimulation by IL-15. Indeed, mice lacking IL-15 or the α- or β-chains of its receptor present very low levels of NK cells, NKT cells, and memory CD8 αβ T cells (24, 25, 26, 27). As IL-15 and the closely related cytokine IL-2 induce the proliferation of T and NK cells in vitro (28, 29, 30, 31, 32, 33), we assessed the cellular responses of P2 to these two cytokines. T cell proliferation was normal following the stimulation of PBMC (71% lymphocytes and 11% monocytes) by incubation for 3 days with PHA (data not shown) or anti-CD3 (Fig. 1⇓a), as demonstrated by assessing [3H]thymidine incorporation. T cell proliferation was almost normal following stimulation with a CD3-specific Ab plus rIL-2 (Fig. 1⇓a). However, PBMC from our patient stimulated with IL-2 alone showed lower levels of proliferation than did PBMC from six controls, including one age-matched healthy child (Fig. 1⇓b). The dose-dependent proliferation of control blood cells induced by IL-2 in this assay is primarily due to expansion of the NK cell and CD8 T cell populations (data not shown). No IL-2-driven proliferation of CD8 T cells was detected by measuring CFSE incorporation in the patient (data not shown), even though the patient had normal numbers of these cells at the time of the experiment, suggesting that the patient’s CD8 T cells did not respond well to the stimulation of PBMC by IL-2. Following CD3 plus IL-15 stimulation, the patient’s PBMC displayed levels of proliferation similar to those observed for the six controls (Fig. 1⇓c). The patient’s PBMC also responded poorly to stimulation with IL-15 alone (Fig. 1⇓d). Impaired proliferation was also observed if PBMC from a healthy control were depleted of NK cells and stimulated with IL-15, despite the proliferation of CD4 and CD8 T cells (data not shown). The defect in lymphocyte proliferation upon IL-15 stimulation observed in P2 therefore results largely from the absence of NK cells. Overall, no apparent defect in the proliferation of the patient’s T cells was observed if PBMC were stimulated with PHA and CD3, alone or with IL-2 or IL-15. In contrast, proliferation in response to IL-2 or IL-15 was found to be impaired, due to the absence of NK cells (IL-2 and IL-15) or to a poor response of CD8 cells (IL-2).
Apoptosis of fresh T cells
IL-2 and IL-15 have also been reported to induce the survival of T cells and NK cells in vitro and in vivo in the mouse model and in vitro in humans (34, 35, 36). We assessed apoptosis in total PBMC by flow cytometry analysis of propidium iodide incorporation. The patient’s PBMC presented a much higher rate of apoptosis than control cells when cultured in serum-supplemented medium devoid of cytokines (Fig. 2⇓a). This experiment was repeated five times on cells from P2 and we also tested six controls included one age-matched control. Mean apoptosis rates in control PBMC were close to 12% (7.4–17.1%) on day 2 and reached 22% (17–28.1%) by day 3. For P2, >30% (28–32%) of PBMC were apoptotic on day 2 and >38% (34.7–47.4%) were apoptotic on day 3, these proportions being much higher than the upper limit of the normal range. PBMC were also stained with an Ab against a lineage-specific surface marker and annexin V/7AAD, to enable us to determine which populations were apoptotic. More cells from CD8 and CD4 T cells subsets were apoptotic in the patient than control lymphocytes (Fig. 2⇓b), despite the normal CD4 T cell counts of P2. The patient’s B cells also presented a higher rate of apoptosis than control B cells at days 2 and 3 (data not shown). This suggested that the apoptotic phenotype documented on T cells was broader, possibly involving IL-2 and/or IL-15 (37, 38). PBMC apoptosis was also assessed by flow cytometry analysis of propidium iodide incorporation upon stimulation. The stimulation of PBMC with IL-2, with or without anti-CD3 Ab, induced a dose-dependent decrease in the percentage of apoptotic control cells (Fig. 2⇓, c and d). A dose-dependent decrease in apoptosis was also observed for the patient’s PBMC, but the decrease (in percentage) was lower than that for control cells (Table II⇓). The highest concentrations of IL-2 were almost associated with a normalization of the phenotype, suggesting that the patient’s cells may respond better to high doses of IL-2. Similar results were obtained with IL-15, with a less profound phenotype for the patient’s cells (Fig. 2⇓, e and f, and Table II⇓). We did not succeed in generating NK cells in vitro by stimulating PBMC with high doses of IL-2, possibly reflecting the poor response of the patient’s NK cells to this cytokine. In conclusion, we observed excess spontaneous and IL-2/IL-15-antagonized apoptosis of the patient’s T cells in assays conducted with fresh PBMC.
Proliferation and apoptosis of PHA-activated T cell blasts
We generated PHA-activated T cell blasts and assessed their proliferation following stimulation with IL-2 and IL-15. PBMC were stimulated with PHA, isolated by Ficoll density centrifugation on day 3, and cultured in medium supplemented with IL-2 (40 IU/ml). PHA-activated T cell blasts were composed of a mixture of CD4 and CD8 T cells (data not shown). On day 6, cells were washed and IL-2 (10, 20, 40, or 100 IU/ml) or IL-15 (0.5, 1, 5, or 10 ng/ml) was added. T cell proliferation was assessed on day 9 by evaluating [3H]thymidine incorporation. PHA-activated T cell blasts from the patient proliferated normally in response to IL-2 or IL-15 stimulation (Fig. 3⇓, a and b), confirming the results obtained with fresh PBMC stimulated with anti-CD3 Ab and a cytokine. The percentage of apoptotic T cell blasts was also determined in these experiments. On day 9, in the absence of cytokine stimulation, the number of apoptotic T cells was normal for the patient (∼70%). Following the addition of low concentrations of IL-2 (10 IU/ml) on day 6, control T cells displayed a marked decrease in apoptosis on day 9, whereas the patient’s T cells did not. Addition of 10 IU/ml IL-2 resulted in a lower decrease of apoptosis in the patient’s than in the control’s cells (Fig. 3⇓c, Table III⇓). The percentage of apoptotic cells decreased to levels similar to those for control PHA-activated T cell blasts following stimulation with high concentrations of IL-2 (100 IU/ml) (Fig. 3⇓c). Control cells responded much better than the patient’s cells to low concentrations of IL-15 (0.5 and 1 ng/ml), with a marked decrease in apoptosis (Fig. 3⇓d, Table III⇓). At high concentrations of IL-15 (5 and 10 ng/ml), the patient’s T cell blasts responded almost normally. We compared the apoptosis levels of CD4 and CD8 T cell blasts. These two subpopulations presented similar defects in their responses to IL-2 and IL-15 (data not shown). In conclusion, the patient’s T cell blasts required concentrations of IL-2 or IL-15 at least 10 times higher than normal to prevent cytokine-starvation apoptosis and to ensure survival. In contrast, the level of apoptosis was not greater in the patient’s cells than in control cells following Fas stimulation, and may even have been lower (Fig. 3⇓e). Thus, the pathways of starvation-mediated apoptosis and cytokine-mediated cell survival are affected in the patient, whereas death-domain-dependent-Fas-mediated apoptosis is not affected (39, 40, 41).
Apoptosis and proliferation of herpesvirus saimiri-transformed T cells
A herpesvirus saimiri-transformed T cell line was generated as previously described (22). Herpesvirus saimiri-transformed T cells were cultured in medium supplemented with IL-2 (40 IU/ml) every 3 days. To measure proliferation and apoptosis, the cells were washed and incubated with (20 or 200 IU/ml) or without IL-2. Proliferation was assessed on day 4 poststimulation by IL-2 by evaluating [3H]thymidine incorporation. Proliferation of the patient’s saimiri-transformed T cells was normal, like that of the patient’s PBMC (Fig. 4⇓a). Apoptosis was measured every day until day 6 poststimulation with IL-2. At day 6, nearly 40% of the patient’s saimiri-transformed T cells were apoptotic, compared with <20% of control saimiri-transformed T cells. Like the patient’s fresh T cells, the patient’s saimiri-transformed T cells thus presented with an excess of apoptosis. A partial complementation of the phenotype was observed with the addition of 20 IU/ml IL-2 and a complete normalization was observed with 200 IU/ml IL-2 (Fig. 3⇑b). Like with P2 PBMC, herpesvirus saimiri-transformed T cells from P2 are therefore excessively apoptotic in the absence of the survival factor IL-2 and respond only to supplementation with high doses of rIL-2.
Normal IL-2 and IL-15 signaling pathways
We investigated the molecular basis of impaired responses to IL-2 and IL-15 in our patient by assessing levels of these cytokines and their signaling pathways. The production of IL-2 and IL-15 and the expression of their receptors (CD132, CD122, CD25, IL-15Rα) were found to be normal on T cell blasts (CD132, CD122, CD25) or monocytes (IL-15Rα) by flow cytometry (data not shown). The primary structure of the β-chain of these receptors (CD122) was also normal, as demonstrated by sequencing of the coding region of the cDNA (data not shown). We then investigated the IL-2- and IL-15-signaling pathways downstream from the receptors. We stimulated PHA-activated T cell blasts from a healthy control and from the patient with 10–1000 IU/ml IL-2. Western blots showed that Stat5, Stat3 (involved in T cell proliferation (42)) and Akt-1 (involved in T cell survival (43, 44)) were normally phosphorylated in the patient, even after stimulation with a low dose of IL-2 (10 IU/ml), which is known to decrease survival rates (Fig. 5⇓a). Ly294002, an inhibitor of Akt-1 phosphorylation, served as a control. A similar result was obtained following stimulation with 1–10 ng/ml IL-15 (data not shown). We detected no defects in signaling between the IL-2 and IL-15 receptors and the Stat and Akt-1 proteins. We then studied the induction by IL-2 and IL-15 of Bcl-2 and Bcl-xL, crucial antiapoptotic factors in T lymphocytes (45, 46, 47). We used flow cytometry to determine intracellular Bcl-2 and Bcl-xL levels in PHA-activated T cell blasts following stimulation with IL-2 (10 or 100 IU/ml) or IL-15 (1 or 10 ng/ml) or in the absence of cytokine. In the absence of cytokine stimulation, both Bcl-2 and Bcl-xL levels were similar in the patient’s cells and in control cells. Dose-dependent Bcl-2 induction in response to IL-2 and IL-15 was observed with PHA-activated T cell blasts from both the patient and controls, even if concentrations of IL-2 known to impair survival were used (10 IU/ml) (Fig. 5⇓b). The primary structure of Bcl-2 was also normal, as shown by sequencing of the coding region of the cDNA. The level of Bcl-xL is poorly induced in response to cytokine stimulations but comparable levels were observed in control and patient PHA-activated T cell blasts (data not shown). The mammalian target of rapamycin (mTOR), a central regulator of cell growth and proliferation, is activated by growth factors such as IL-2, via Akt. Rapamycin is an immunosuppressant that induces cell cycle arrest in G1 phase (48). Therefore, we treated PHA-activated T cell blasts for 72 h with rapamycin. Similar decreases in proliferation were observed for the patient’s and control cells (data not shown), with no additional apoptosis induced by rapamycin in either (data not shown). Thus, neither the induction of Bcl-2 nor that of mTOR seems to be responsible for the low level of lymphocyte survival in our patient.
We report here the probable disease-causing cellular phenotype of a patient from the first reported kindred with a well-documented inborn and familial syndrome consisting of intrauterine growth retardation and a lack of NK cells without T lymphopenia. The NK deficiency is not strictly isolated among lymphoid lineages, as our patient also lacked the related NKT cells. Blood γδ T cells and the known subsets of conventional αβ T cells–CD4 Th cells, CD8 CTLs, and Foxp3-expressing regulatory T cells–were present, but CD8 cell maturation was delayed, in particular that of CD45RO memory CD8 T cells. Obviously, we could not investigate the missing NK cell subsets directly. However, by studying conventional αβ T cells, we demonstrated excessive apoptosis and impaired cell survival in response to IL-2 and IL-15. This defective response to IL-2 probably accounted for our unsuccessful attempts to generate NK cells from PBMC in vitro. These data provide a probable molecular mechanism underlying the lack of NK and NKT cells, consistent with the known role of these cytokines in the development of these subsets in the mouse (30). The transient lack of CD8 αβ T cells, notably of memory CD45RO CD8 T cells, is also consistent with previous studies and indicates that these cells progressively compensated for poor stimulation by IL-2 and IL-15. This is also consistent with the near-normal complementation documented in vitro with high doses of the cytokines. Our data finally demonstrate that the cellular defect is broad and affects lymphocytes present in normal numbers in the blood, such as αβ T cells, and even B cells, some of which also express the receptors for IL-2 and IL-15 (37, 38). There is thus both global and partial impairment of lymphoid lineage survival, most pronounced in NK and NKT cells, suggesting that these cells are more dependent on IL-2 and IL-15 for survival than conventional αβ T cells and B cells in humans. Nevertheless, until a disease-causing genotype is identified in this kindred, the cellular phenotype herein reported cannot be strictly proven to be disease causing. Although unlikely, it may be a mere consequence, rather than the cause, of the primary defect.
The impaired survival of lymphocytes probably accounts for the key immunological feature of our patients, the lack of NK cells. The small number of PMNs may reflect an impaired response to survival signals, with an enhanced apoptosis, or it may be a consequence of the lymphoid anomalies. The severe pre- and postnatal growth retardation clearly did not result from lymphoid or myeloid cytopenia. Instead, it may reflect a global impairment of cell survival. GH levels and responses to this hormone in vivo were found to be normal in our patient, and GH is not a major factor affecting development in utero. In contrast, insulin is known to be a major growth factor in utero. Its signaling pathway involves Akt and target genes also involved in the IL-2 signaling pathway (44, 49). Several mutant mice were shown to display in utero growth retardation, including IGF-1-, IRS-1-, AKT-1-deficient mice (50, 51), whose deleted genes are involved in the insulin-signaling pathway. Thus, impaired cell survival with increased apoptosis may account for the developmental and lymphoid phenotypes of our two patients. Identification of the disease-causing gene may shed new light on the survival pathways involving both cytokines and developmental growth factors. In any event, the cellular survival defect leading to excessive apoptosis probably accounts at least for NK and NKT cell deficiency. This in turn is the most threatening clinical feature associated with this syndrome, as P1 died of CMV infection. The survivor (P2) has not yet been infected by herpesviruses.
What genotype corresponds to this cellular phenotype of impaired cell survival? We provide compelling evidence that our patients’ T cells present a defect in survival in response to IL-2 and IL-15. Our results indicate that the genetic defect is downstream from, or parallel to, Stat-3, Stat-5, and Akt-1, and downstream from, or parallel to, mTOR and Bcl-2. Many genes have been reported to be induced by IL-2 and IL-15 stimulation, or deprivation, and may be involved in cell survival. Interestingly, it was recently reported that the deletion of the T-box transcription factor Eomesodermin (Eomes) in mice is responsible of a developmental defect resulting in early embryonic lethality (52). Mice that lack the T-bet transcription factor, heterozygotes for a loss-of-function mutation in Eomes present an immunological phenotype that most strikingly resembles that of our patients. In this specific genetic context, the deletion of one Eomes allele causes a deficiency of IL-15-dependent lymphoid lineages, due to an impaired regulation of CD122. The mice present with NK cell deficiency and an impaired development of cytotoxic memory CD8+ T cells (53). The human EOMES gene and the fine regulation of CD122 should therefore be examined in detail in our patient.
We also recently used pangenomic microarrays to assess the expression pattern of IL-2 and IL-15 target genes following the stimulation of PHA-activated T cell blasts from the patient and a control. The microarray approach may provide a selection of candidate genes. An alternative approach is to recruit other kindreds with a similar phenotype. This should pave the way for the identification of the causal gene by means of a genome-wide scan. In conclusion, we report a plausible pathogenic mechanism underlying the first description of a familial form of inherited selective NK cell deficiency in humans, differing from the previously reported two suggestive descriptions (8, 12) by its familial occurrence and its inborn nature. Our identification of a cellular phenotype, which is probably disease causing, lends further credence to the description of this kindred as a distinct medical entity and paves the way for the identification of the disease-causing gene. The identification of the disease-causing gene will increase our understanding of cell survival and apoptosis in lymphoid and nonlymphoid tissues.
We thank all family members for their kind willingness to participate in this study. We thank Laurent Abel for helpful discussions and critical reading of the manuscript and all members of the Laboratory of Human Genetics of Infectious Diseases for helpful discussions. We thank M. C. Fondanèche, F. Selz, C. Hivroz, S. Kayal, C. Soudais, N. Cerf-Bensussan, S. Latour, S. Guillaume, F. Geissmann, G. Elain, B. Senechal, S. Weller, J. C. Weill, and S. Cholet-Martin for helpful discussions and experimental input.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 C.E. was supported by La Fondation pour la Recherche Médicale, France, European Union (EU). A.A. was supported by L’Assistance-Publique-Hopitaux de Paris, France, EU. The Laboratory of Human Genetics of Infectious Diseases was supported by the Schlumberger Foundation, the Banque Nationale de Paris-Paribas Foundation, the Institut Universitaire de France, and EU Grant QLK2-CT-2002-00846. J.-L.C. is an International Scholar of the Howard Hughes Medical Institute.
↵2 Address correspondence and reprint requests to Dr. Jean-Laurent Casanova, Laboratoire de Génétique Humaine des Maladies Infectieuses, Université de Paris René Descartes-Institut National de la Santé et de la Recherche Médicale Unité 550, Faculté de Médecine Necker, 156 rue de Vaugirard, 75015 Paris, France, European Union. E-mail address:
↵3 Abbreviations used in this paper: 7AAD, 7-aminoactinomycin D; GH, growth hormone; SAB, AB serum; PMN, polymorphonuclear neutrophil; mTOR, mammalian target of rapamycin.
- Received February 15, 2006.
- Accepted September 19, 2006.
- Copyright © 2006 by The American Association of Immunologists