Severe combined immunodeficiencies (SCID)

SCID consists of a group of genetic disorders characterized by a block in T lymphocyte differentiation that is variably associated with abnormal development of other lymphocyte lineages, i.e. B or NK lymphocytes or more rarely of the myeloid lineage [1,2]. At least eight diseases can be distinguished according to phenotype and inheritance pattern (Table 1). The overall frequency is estimated to 1 in 75 000–100 000 births [3,4]. Only those SCIDs with severe T cell depletion will be discussed here, leaving those who usually have normal T cell numbers (e.g. MHC class II and Zap-70 deficiency) for a separate review.

Table 1.

Severe combined immunodeficiencies (SCID)

Disease	Inheritance	Cells affected	Gene
Reticular dysgenesis	AR	Myeloid, T, B, NK	?
Adenosine deaminase deficiency	AR	T, B, NK	ADA
T(–) B(–) SCID	AR	T, B	Rag-1, Rag-2
T(–) B(–) SCID	AR	T, B Increased cell radiosensitivity	?
X-linked SCID	XL	T, NK	γc
T(–) SCID	AR	T, NK	JAK-3
T(–) SCID	AR	T	IL-7Rα
T(–) SCID	AR	T (TCRα,β)	CD45
T(–) SCID	AR	T	?

Clinical presentation

The clinical presentation is fairly uniform and is characterized by early onset of infections, mainly of the respiratory tract and gut. We reviewed the frequency of infections and the diagnosis in 117 patients with SCID who were referred to our centre [3]. Oral candidiasis, persistent diarrhoea with growth impairment and/or interstitial pneumonitis are the most frequent infectious manifestations leading to diagnosis. There are no differences between the various SCIDs, except for an earlier onset of infections in patients with ADA deficiency. The persistence and recurrence of infections in SCID patients rapidly lead to growth impairment and malnutrition. Our findings were similar to an American study of 100 infants [4].

Even common opportunistic organisms such as Pneumocystis carinii and Aspergillus species can cause infections. Intracellular organisms such as listeria and legionella can cause devastating disease, as can viruses, especially those of the herpes group. Infection by Epstein–Barr virus (EBV), although rare in this age group, can lead to uncontrolled B lymphocyte proliferative disorders (BLPD) in B(+) SCID patients, similar to that seen in immunosuppressed transplant recipients. Thirty-one such cases have been reported [5]. Some occurred following therapy of SCID (thymic transplant, fetal liver transplant or T-depleted marrow transplant), which was presumably the source of EBV. BLPD has also developed in untreated SCID patients [6]. Live vaccines can also cause life-threatening infections. We have observed BCG infection in 10/28 vaccinated patients (including two with local infection and eight with disseminated involvement of the liver, spleen and lungs), which was fatal in three cases [3]. Interestingly, three of six patients who received oral attenuated poliovirus had detectable virus in the stools, but none developed poliomyelitis; this was probably either because of slow viral replication in transplant recipients or because of protection by maternal immunoglobulins. It is clear that live vaccine must not be given to children at risk for SCID.

Non-infectious clinical manifestations consist mainly of graft-versus-host disease (GVHD) caused by the patients' inability to reject allogeneic cells. The two possible sources of allogeneic cells are maternal lymphocytes and transfusion.

Maternal T lymphocyte engraftment is frequently detected in SCID patients using molecular tools [7,8]. Circulating maternal T cells are detected in approximately 50% of cases. It is noteworthy that maternal T cells are not found in ADA-deficient patients, possibly because maternal T cells are killed by the raised levels of deoxyadenosine (see below). Maternal T lymphocyte numbers range from 10 to several thousand/μl of blood. They usually have a normal phenotype, with some degree of in vivo activation, as shown by the expression of MHC class II molecules and/or the IL-2 receptor [9]. In some cases, maternal T cells have been reported to be clonal [5,9], a finding suggestive of either transplacental passage of a very small number of T cells or secondary expansion of alloreactive clones in the host.

The most intriguing observation regarding maternal T cell engraftment in SCID patients is the paucity of clinical manifestations. In the majority of cases, the presence of maternal T cells is entirely asymptomatic, while approximately 30–40% of patients have mild symptoms and signs such as erythema with skin T cell infiltration, eosinophilia, and elevated liver enzymes with periportal T cell infiltration [3]. In recent years there have been no reports of fatal GVHD caused by maternal T cell engraftment in SCID patients. Several explanations have been proposed: they include oligoclonality of maternal T cells with lack of alloreactivity toward the child's antigens, or tolerance of transfused maternal T cells because of associated mild haematopoietic stem cell engraftment (and T cell differentiation). The presence of a few maternal T cells in the periphery should not delay or confuse the diagnosis of SCID. It may be an obstacle to T cell engraftment following T-depleted haplo-identical bone marrow transplantation (BMT), especially if the donor is not the mother and if the patient is not treated with myeloablative and immunosuppressive drugs [10,11]. Following HLA-identical BMT, there is usually a dramatic expansion of donor T cells cytotoxic for maternal cells 10–12 days post-BMT, that results in their rapid elimination [12,13]. This ‘graft versus graft’ reaction can cause transient GVHD symptoms.

In contrast, post-natal inoculation with allogeneic lymphocytes by plasma, erythrocyte, platelet or leusocyte transfusion usually causes a fatal acute GVHD syndrome marked by diffuse necrotizing erythroderma, gut mucosa abrasion and biliary epithelium destruction, sometimes associated with stroma cell lesions in the marrow. This GVHD syndrome can occur within 2–4 weeks and is usually resistant to the most powerful immunosuppressive drugs. In a small number of cases (two out of 11 in our experience) GVHD does not develop, although allogeneic anti-host T cells can be detected and cause resistance to BM engraftment [3].

T(–) B(–) SCID

About 20% of patients with SCID have a phenotype characterized by an absence of mature T and B lymphocytes, while functional NK cells are detectable [3,4]. Usually the thymus is hypoplastic. The condition can be cured by allogeneic bone marrow transplantation. This T(–) B(–) form of SCID has an autosomal recessive inheritance and will be covered in a separate review.

T(–) B(+) SCID

X-linked SCID

X-linked SCID (SCID-X1) accounts for 50–60% of cases of SCID [14]. It is characterized by an absence of mature T and NK lymphocytes, whereas B cells have a normal phenotype and are present in increased numbers. Histologically, the thymus lacks a cortex/medullar differentiation. Lymphoid precursors are scarce, and Hassal's corpuscles are not detectable [15]. Peripheral lymphoid organs are also hypoplastic. These data indicate that there is an early block in the T cell differentiation pathway in this disease. SCID-X1 is curable by allogeneic BMT, indicating that the defect is intrinsic to the lymphoid lineage [14]. Studies of X-chromosome inactivation patterns in obligate carriers have shown a skewed pattern in T and NK cells as well as in B cells, whereas a random pattern was usually detected in the other haematopoietic lineages [16,17]. The SCID-X1 gene product is therefore expressed and involved in the maturation of the T, B, and NK cell lineages. Of note is the observation that the X chromosome inactivation pattern is more skewed in mature than immature B cells [18].

The SCID-X1 locus was mapped to Xq12-13.1 [19]. It was then recognized that the gene encoding the γ-chain of the IL-2 receptor (now renamed γc) was localized to the same region, and mutations of the γc gene were found in SCID-X1 patients [20]. That γc mutations cause SCID-X1 has now been proven in several ways: all patients with SCID-X1 have the γc gene mutation [21,22], in vitro gene transfer of γc into patient's EBV-transformed B cells and marrow cells corrects the high-affinity, IL-2 receptor deficiency and NK cell differentiation block, respectively [23–26]. Canine XL-SCID is also associated with a mutation in the γc gene [27]. Finally, γc(–) mice exhibit a similar, although not entirely identical, phenotype (see below) [28,29].

γc belongs to the haematopoietic cytokine receptor family, characterized by four conserved cysteines and the repeated WS motif [30]. The γc-chain is constitutively expressed by T cells, B and NK cells, as well as myeloid cells and erythroblasts (reviewed in [31]). γc expression together with the IL-2Rα and β subunits generates the high-affinity receptor for IL-2, and plays a major role in signal transduction through activation of its associated tyrosine kinase JAK-3 [31].

A number of mutations of the γc gene have now been reported in SCID-X1 patients [21,22]. Since the disease is lethal, a 30% rate for new mutations is expected for each generation, accounting for the variety of mutations found. It is remarkable that many single amino acid substitutions in the extracellular domain are sufficient to abrogate T and NK cell differentiation. Some affect conserved cysteines and the WS motif, the structure of which is likely to be required for the overall configuration of the molecule [21]. Others, like an ala→val substitution in position 156, create a molecule that is expressed but fails to bind IL-2 or to transduce signals [32].

γc is a member not only of the IL-2 receptor but also of the IL-4, IL-7, IL-9, and IL-15 receptors [33], augmenting in each case the affinity for the cytokine and participating in signal transduction. The SCID-X1 phenotype appears therefore to be the complex association of defects in these five cytokine/receptor systems. Recent studies in mutant mice generated by homologous recombination have brought significant insight into the role of IL-7 in T cell differentiation. γc(–) mice have a profound immunodeficiency [28,29]. The T cell phenotype of γc mice is virtually identical to the IL-7(–) and IL-7Rα(–) mice [34,35]. These data strongly argue for a major role of IL-7 in inducing survival and proliferation of early T cell progenitors in the thymus [36–38]. This is confirmed by the block in T cell development observed in two patients with IL-7Rα deficiencies [39] (see below). Furthermore, γδ T cells are completely lacking in γc(–) mice.

The NK cell deficiency observed in SCID-X1 is likely to be the main consequence of defective IL-15-induced signalling. Indeed, IL-15 (with SCF) can trigger CD56⁺ NK cell generation from CD34⁺ marrow progenitors [40]. We found that following γc gene transfer into SCID-X1 patients' marrow, functional NK cells (CD56⁺) can differentiate in the presence of SCF and IL-15 [26].

SCID-X1 B cells make IgE in the presence of IL-4 and a CD40-mediated signal [15]. However, SCID-X1 EBV-B cells do not activate JAK-3 and STAT6 in the presence of IL-4 [18]. These results can be explained by the presence of a γc-independent IL-4 receptor able to transduce at least some signals after IL-4 binding. As expected, IL-2 and IL-15 do not induce an immunoglobulin switch in SCID-X1 B cells, in contrast to their effects on control B cells [15]. V(D)J elements of immunoglobulin normally rearrange in SCID-X1 B cells, while most of the J_H are in germ-line configuration, probably reflecting a lack of T cell help [41]. In rare instances, γc gene mutations have been found in patients lacking not only T and NK cells but also B cells. No obvious explanation appears for this ‘atypical’ phenotype [3]. This fact further stresses the lack of demonstrable correlation between genotype and phenotype observed so far. It may very well be that modifier gene(s) could play a role.

JAK-3 deficiency

A non-X-linked form of SCID characterized by a phenotype identical to SCID-X1 has been shown to be the consequence of mutations of the JAK-3 encoding gene [42,43]. JAK-3 is a tyrosine kinase that is bound to the intracellular tail of γc and is activated upon cytokine binding to the multichain receptor. JAK-3 phosphorylates STAT-5 protein. Phosphorylated STAT proteins dimerize and are translocated to the nucleus where they act as transcription inducing factors for several genes involved in progression of cell division [44]. The identical phenotype of γc and JAK-3 deficiencies demonstrate the essential role of JAK-3 in transducing signals triggered by cytokine binding. A number of distinct mutations, most leading to premature stop codons, lead to a similar T(–), NK(–), B(+) phenotype.

IL-7Ra deficiency

In two patients with a T(–) B(+) NK(+) SCID phenotype, mutations impairing the expression of the α subunit of the IL-7 receptor have been described [39].

The IL-7 receptor is composed of IL-7Rα and γc. This observation confirms the essential role of IL-7 in the early steps of T cell differentiation, while it can be dispensable, at least in humans, for B cell differentiation.

Some other patients with a same SCID phenotype do not exhibit mutations in the IL-7Rα encoding gene. The mechanism underlying the T cell deficiency remains unknown. A potential IL-7 deficiency should lead to the same phenotype, but this would not be corrected by classical BMT since IL-7 is produced by stromal cells.

Atypical ‘T (–) B (+)’ SCID

A combined X-linked immunodeficiency characterized by progressive loss of T and B cell function leading to death during childhood has been described in several families. In two families, patients' T cells were found to be oligoclonal [45,46]. In one family, the X-chromosome inactivation pattern in obligate carriers together with gene mapping was consistent with a form of X-linked SCID. Analysis of the γc gene in the propositus demonstrated two transcripts, one truncated and one of normal size, which accounted, respectively, for 80% and 20% of total γc mRNA [47]. A single base-pair substitution in the last position of exon 1 was found that probably disturbed splicing of intron 1 (resulting in the abnormal mRNA product), while a less frequent, normal splicing generated the normal sized mRNA encoding a protein with a conservative Asp→Asn substitution in position 39. The γc-chain could be detected in EBV-B cells from the patient. A reduced number of normal, high-affinity IL-2 binding sites was detected. This case suggests that reduced expression of the γc may profoundly disturb T cell differentiation, with only a relatively small number of clones progressing along the T cell differentiation pathway. As mentioned above, a seemingly identical phenotype has been reported with γc mutations that reduced JAK-3 binding and T cell activation [47–49].

In another case presenting with normal numbers of T and B lymphocytes, T cells were shown to proliferate in the presence of mitogens [50]. However, IL-2 binding was reduced. It was postulated that this was the mechanism of the defective antigen-specific T cell responses. More importantly, it was shown that in this boy, a mutation of γc (R222C) caused the reduced binding to IL-2. This case might be the consequence of defective IL-2/IL-2R interaction, while interaction of the γc receptor with other cytokines is preserved. This contrasts with the apparently converse observation of a SCID-X1 phenotype caused by a γc mutation (A156V) affecting IL-7- (and IL-4)-mediated responses, but less so with IL-2- and IL-15-mediated responses [51].

More surprisingly, we described a profound change in SCID-X1 phenotype following an unsuccessful attempt at bone marrow transplantation [52]. Despite a γc gene deletion encompassing most of the intracellular domain of γc, the child developed partially functional T cells that were of host origin. Such T cells were detected over a 8-year period following BMT, though in declining numbers. The mechanism by which, in the absence of possible direct JAK-3 activation, these T cells have differentiated and been partially functional is not understood [52].

Finally, an unusual SCID-X1 patient had 800–2000 μl circulating T cells which were able to respond partially to mitogens and antigens [53]. Although γc expression could not be detected on the patient's B cells, monocytes and granulocytes, while NK cells were not detectable, T cells did express γc. In the B cells, the γc gene was found to be mutated (Cys→Arg substitution at position 115). In T cells however, the mutation could not be found. The mother is a carrier of the mutation. These results could be accounted for by a reverse mutation that occurs in a T lineage-committed cell. This observation provides two interesting pieces of information. Such a rare event restored, at least for a 4-year period, a stable, albeit incomplete, T cell pool [54]. The selective advantage conferred to this cell lineage appears very high, giving support to the feasibility of gene transfer as a treatment for SCID-X1 using currently available vectors. A similar reversion has been recently described in a patient with adenosine deaminase (ADA) deficiency [55].

These different observations indicate that it is very likely that more ‘atypical’ forms of SCID do exist and are undiagnosed. Therefore, all T cell immunodeficiencies should be tested for the known SCID molecular mechanisms.

Adenosine deaminase deficiency

About 20% of SCIDs are caused by ADA deficiency. ADA is a ubiquitous enzyme that reversibly transforms adenosine to inosine and 2′-deoxyadenosine (dAdo) to 2′-deoxyinosine. It belongs to the salvage pathway of purine metabolism [56]. The mechanism by which ADA deficiency leads to severe T, NK and B lymphocytopenia without seriously affecting other tissues is now understood. The immunodeficiency is the consequence of accumulation of adenosine and dAdo substrates that are indirectly toxic to lymphocytes, especially to immature lymphocytes (thymocytes). dAdo, which can freely diffuse, is phosphorylated into deoxyATP. Immature lymphocytes and, to a lesser extent, mature lymphocytes are poorly able to reversibly degrade dATP into dAdo, in contrast to other cell lineages. dATP blocks cell division by inhibiting ribonucleotide reductase, an enzyme required for generation of the other deoxynucleotides. In their absence, DNA synthesis cannot proceed [57]. DNA synthesis can also be blocked by inactivation of S-adenosyl homocysteine hydrolase, which donates a methyl group to DNA. There is also toxicity for resting cells that may involve induction of chromosomal breaks, due in part to a blockade of endogenous DNA repair processes [58]. Furthermore, dATP indirectly reduces nicotinamide-adenine dinucleotide (NAD) levels.

In most cases (estimated at about 85%), ADA deficiency results in a typical SCID with very low T and B cell counts. Clinical manifestations often occur earlier than in other forms of SCID [3]. In addition to severe infections and failure to thrive, approximately 50% of patients develop skeletal abnormalities with cupping and flaring of the costochondral junction and mild pelvic dysplasia. Some patients have neurological signs, including cortical blindness, and dystonia. It is difficult to exclude a diagnosis of viral encephalitis, but resolution of neurological abnormalities following specific treatment of ADA deficiency suggests a direct metabolic consequence of the enzyme deficiency [59]. Mesangial sclerosis and abnormal renal function have been noted in some patients, as well as cortical adrenal fibrosis [59]. It is not however, proven that the latter lesions are specific to ADA deficiency. The typical, early onset type of ADA deficiency is associated with barely detectable activity in erythrocytes and lymphocytes, erythrocyte dATP levels exceeding 100 nmol/ml packed erythrocytes. ADA gene mutations often affect the active site of the molecule, or there is a deletion within the ADA gene. In some other patients, clinical onset is delayed by several months. T cell lymphocytopenia may not be complete and there is often eosinophilia, with residual ADA activity in lymphocytes. Late-onset ADA deficiency has also been described, with the first clinical manifestations occurring after 2 or 3 years or even later, as described in several patients [60]. In those patients with a milder T cell immunodeficiency, lymphocytopenia develops gradually. Autoimmune manifestations are not uncommon, as in other incomplete T cell immunodeficiencies. For instance, refractory thrombocytopenic purpura at the age of 17 years has been described as the first manifestation of ADA deficiency [61]. In patients with late-onset ADA deficiency, dATP erythrocyte levels are usually less elevated than in the early onset type (< 1000 ng/ml packed erythrocytes). Mutations of the ADA gene may consist of missense mutations or splice site mutations not directly affecting the active site of the enzyme [59,60]. However, heterogeneity is further increased by the fact that most patients are compound heterozygotes.

The ADA gene has been mapped to 20q13-11 and cloned. It consists of 1089 nucleotides divided into 12 exons. Different deletions, missense mutations and splicing mutations of the ADA gene that induce ADA-deficiency SCID have been characterized [59–62]. Further description and study of mutations inducing ADA function will advance the phenotype/genotype correlation.

Reticular dysgenesis

This is a very rare SCID condition characterized not only by defective lymphoid differentiation but also by a block of myeloid differentiation. Autosomal recessive inheritance is presumed but not strictly proven, owing to the rarity of the syndrome [1]. It is not clear whether some forms of SCID [3] with neutropenia truly represent a different syndrome or whether blocked haemopoiesis could be secondary to persistent viral infection in some cases. Nothing is known about the mechanism of the syndrome but its haematopoietic origin is probable since it is curable by BMT [63]. Of note, SCID has been described in an infant with cyclic haematopoiesis [64] that resolved following BMT.

SCID has been associated with multiple gastrointestinal atresia in two kindreds with absence of T cells and very low B cell counts [65,66]. The inheritance is presumed to be autosomal recessive. The atresias consist of diaphragms. It is not known whether this is a contiguous-gene syndrome or whether a factor is involved in both the development of the ‘hollow stage’ of the gastrointestinal tract and lymphocyte differentiation.

Recently a CD45 membrane-associated tyrosine phosphatase deficiency was reported in a child with a profound T cell lymphopenia, mild NK cell lymphopenia and high B cell counts [67]. This phenotype appears similar to the one of CD45(–/–) mice [68] and confirms the importance of this phosphatase in T lymphocyte development. Interestingly, γδ(+) T cell development was spared.

Treatment of scid

The natural outcome of SCID is poor and most of the patients have died by the age of 1 year. Prophylaxis and treatment of infections with IV Immunoglobulin substitution and cotrimoxazole for P. carinii are required but can at best marginally prolong survival. Other obligatory measures include 25 Gy irradiation of blood products in order to avoid fatal GVHD and avoidance of live vaccines such as bacille Calmette–Guérin (BCG).

The standard treatment is allogeneic BMT as described 32 years ago [69]. The BMT procedure is unusual, in that no myeloablation or immunosuppression is required to achieve engraftment. Such transplantations usually result in a split chimerism, as only T cells (+/− NK cells) are of donor origin. HLA identical BMT is characterized by rapid T cell reconstitution following expansion of the donor memory T lymphocyte pool [11,70], with newly generated T cells being detected 3–4 months after BMT. This time interval appears to be the minimal requirement for efficient development of T cells in host thymus [11,70]. Because of the virtual absence of GVHD, the probability of success is >90% in recent years.

Since the early 1980s, SCID patients lacking an HLA identical donor have been treated with haploidentical stem cell transplantation. GVHD prevention by T cell elimination of the marrow inoculum was the key to success [70–72]. A recent survey showed that, in the absence of GVHD, survival can reach 78% [70]. Myelo-ablation is not a prerequisite, at least for NK(–) SCID conditions [70,73].

Studies on the outcome of marrow transplantation in SCID patients provide some clues to the solution of fundamental problems in the field of haematopoietic cell transplantation, but they also raise questions. For example, it is not known why, in the absence of any conditioning regimen, T lymphocyte engraftment occurs regularly, whereas B cell and myeloid cell engraftment are exceptional. Moreover, it is not known whether pluripotent haematopoietic stem cells from the donor persist and preferentially differentiate into T cells, or whether common lymphoid progenitors [74] in the donor's marrow migrate to the thymus, where they differentiate. If the latter occurs, long-term loss of naive T cells can be anticipated, since common lymphoid progenitors do not have the capacity for self-renewal.

An important shortcoming of BMT in SCID is the frequent deficiency of B cells. Most patients in whom donor B cells engraft acquire a repertoire of functional B cells [75]. By contrast, the majority of patients with host B cells after transplantation are unable to produce normal amounts of immunoglobulins and have to be treated long-term with immunoglobulins. Perhaps, the genetic defect of the host B cells (especially with regard to a deficiency of γc chain or JAK-3) explains this phenomenon. However, in a number of cases host B cells, including those without γc-chain expression or JAK-3, can make IgM and IgG antibodies after marrow transplantation. It may be that in SCID patients who have persistent hypogammaglobulinaemia after transplantation, the lymphoid-organ matrix, long devoid of lymphocytes, is unable to reconstitute the architecture required for normal immune B cell function. If this is true, transplantation early in life should increase the likelihood of normally functioning host B cells. The SCID condition can also affect outcome as, in patients with T(–) B(–) SCID, the rates of engraftment and survival after transplantation of HLA-haplo-incompatible marrow are low [76]. It is believed that host NK cells account for the high rate of graft failure in the absence of myelo-ablative conditioning. It has also been postulated that defective DNA repair in a subgroup of patients with deficiencies in T and B cells, but normal NK cells, accounts for severe complications caused by GVHD or infection. Perhaps this subgroup of patients should be treated with a protocol that includes the inhibition of host NK cells.

Is there room for further progress in the treatment of SCID? More rapid appearance of T cells and improved B cell function are two important goals that could limit early mortality from viral infections (due to T cell deficiency) and later morbidity (mostly due to B cell deficiency). The use of donor T cell infusions (provided that the risk of GVHD can be minimized) and the use of cytokines involved in progenitor cell proliferation (e.g. IL-7) are possibilities worth considering.

ADA deficiency can also be treated by enzymatic substitution with ADA coupled to polyethylene glycol (PEG–ADA). Weekly intramuscular administration usually results in normalization of dATP levels in blood cells. The T cell rise occurs within weeks with evidence of specific T cell immune responses. PEG–ADA treatment is effective in most patients (approx. 90% of cases [77,78]. The failures appear often to correspond to the most severe phenotype associated with disruption of the enzyme's active site. The dose of PEG–ADA has to be increased in some patients, particularly in those who develop antibodies to ADA, while quality of T cell immunity may decline with time. It is presently difficult to delineate precisely indications of PEG–ADA treatment versus haploidentical stem cell transplantation.

Gene therapy

Lethality of SCIDs, insufficient success of BMT, and the discovery of SCID genes led to an interest in the potential application of gene therapy in this field. The selective advantage expected to be conferred to transduced cells, because SCID gene products provide survival, growth or differentiation signals to lymphocyte progenitors, were grounds for optimism. The occurrence of reverse mutations in γc and ADA deficiency [53–55], and demonstration of a selective advantage conferred to transduced progenitor T cells in JAK-3(–) mice [79], provide further encouragement. Repeated injections of transduced peripheral T cells from ADA-deficient patients after ex vivo infection with a retroviral vector containing the ADA gene led to persistent detection of functional transduced T cells over an 8-year period [80]. However, PEG–ADA treatment was not discontinued. ADA gene transfer into CD34 selected marrow or cord cells resulted at best in the detection of low numbers of transduced cells [81–84]. Inadequate gene transfer methodology and concomitant PEG–ADA treatment, which limits the potential growth advantage of transduced cells, are possible explanations for these results.

More recently, we achieved correction of the T and NK cell immunodeficiency in three patients with γc deficiency following γc gene transfer into CD34⁺ cells using a MFG vector [85]. These results demonstrate a selective advantage can be conferred to transduced cells leading to clinical benefit. However, long-term assessment is required to evaluate fully the potential of this therapy for this and other SCID conditions.

Prenatal diagnosis

Prenatal diagnosis and in utero therapy

Gene identifications for most of the SCID conditions enable prenatal diagnosis by molecular methods in cases at risk by chorionic villous biopsy at week 8–10 of gestation. In the other cases, immunological methods can be used to assess fetal lymphocyte populations in the cord blood taken under echography at week 20–22 of gestation. Recently, two fetuses with XL SCID were successfully treated by in utero injection of haploidentical CD34⁺ cells [86,87]. Although elegant, this approach merits caution, as there is about the same risk of abortion to treating the infant at birth [68], and it may not work in NK(+) SCID.