Human congenital T-cell receptor disorders

Mature T lymphocytes detect antigens with a variable surface TCR heterodimer, which is either αβ or γδ (Figure 1).

During early T-cell development, the invariant pre-Tα chain is used instead of TCRα. In humans, the variable TCR heterodimers form a complex with 2 invariant heterodimers termed CD3γɛ and CD3δɛ and a single invariant homodimer termed ζ or CD247 (Call et al. 2002). These 3 invariant dimers participate in assembly and surface expression of the whole TCR–CD3–CD247 complex, which we will refer to here as TCR. When a TCR interacts with its specific antigen through the variable heterodimers, intracellular signals are transduced to drive T-cell maturation or apoptosis in the thymus and T-cell activation, proliferation, and effector function or anergy–apoptosis in the periphery (Brownlie and Zamoyska 2013). The cytoplasmic tails of CD3ɛ and CD247 then expose their ITAM sequences (Love and Hayes 2010), which are phosphorylated by Lck, thus allowing the recruitment and Lck-mediated activation of ZAP-70. Activated ZAP-70 phosphorylates LAT, which recruits numerous signaling molecules to form a multiprotein complex termed the LAT signalosome. This signalosome, in turn, propagates signals to 3 major signaling pathways, namely MAPK, PI3K/AKT, and PLCγ-1, leading to the activation of transcription factors that are critical for gene expression and essential for T-cell growth and differentiation. Signals initiated from the TCR also result in actin reorganization and the activation of integrins by inside-out signaling (Kuhns and Davis 2012). Most αβTCR-bearing T cells recognize processed peptides associated with major histocompatibility complex (MHC) molecules, whereas γδTCR-bearing T cells can respond to a wide variety of antigens irrespective of their molecular nature including peptide and nonpeptide antigens, MHC-related proteins, phosphoantigens, and soluble proteins (Chien et al. 2014).

TCRIDs are low-prevalence autosomal recessive diseases characterized by impaired surface TCR expression, frequently associated with peripheral blood T lymphocytopenia, severe combined ID (SCID) and (or) autoimmune symptoms, but not associated with B or natural killer (NK) lymphocytopenia. The first TCRID was a human familial CD3 expression deficiency in a child with SCID and in his healthy sibling (Regueiro et al. 1986), later shown to be due to a complete CD3γ deficiency (Arnaiz-Villena et al. 1992). A second TCRID was soon reported (Thoenes et al. 1992) and shown to be due to partial CD3ɛ deficiency (Soudais et al. 1993). Several CD3, CD247, and TCR deficiencies have since been described (Figure 1, Table 1), which can be classified as complete or partial according to the absence or presence of residual levels of the affected protein. Although rare and sometimes based on single cases, TCRID offer rich information about the underpinnings of human TCR structure and function, which in turn impact our understanding of T-cell development and function. Therefore we, as have others (Casanova et al. 2014), believe they merit a detailed study.

Most patients show SCID or CID features such as recurrent respiratory infections, otitis, candidiasis, diarrhea, failure to thrive, and sometimes autoimmune phenomena in the first year of life. Thymus size can be small (CD3γ, δ, or ɛ deficiency (Arnaiz-Villena et al. 1991; de Saint Basile et al. 2004; Takada et al. 2005)) or normal (CD3γ or CD247 deficiency, unpublished). Chronic pyogenic infections, dysmorphic features, or bone abnormalities were not reported. Stem cell transplantation is indispensable when SCID is present, otherwise most patients die early in life as a consequence of viral infection (cytomegalovirus mainly, but also adenovirus). Autoimmune phenomena have been described in several patients, mainly when T cells are present. Most CD3γ-deficient patients showed autoimmune hemolytic anemia (AHA) or vitiligo (Arnaiz-Villena et al. 1992) and also autoimmune thyroiditis, thrombocytopenia, or autoimmune hepatitis (Gokturk et al. 2014). Two TCRα-deficient patients presented with vitiligo, AHA, or autoantibodies (Morgan et al. 2011). These results indicate that both positive selection and negative selection, which are TCR-dependent events, can be impaired in TCRID.

In contrast with all other reported TCRID due to complete protein deficiencies (Table 1), most CD3γ deficient individuals reported to date (7 out of 10) are presently alive and well without transplantation (Table 2), including some in their thirties. Late diagnosis, mild T lymphopenia, low TCR expression, and autoimmune features are common (see “TCRID” section). These limited immunological consequences are at odds with what is observed when the very homologous CD3δ chain is absent: early diagnosis, severe T lymphopenia, no TCR expression, and urgent transplantation. This clearly shows that CD3γ is not critical for survival in most cases, as compared with other TCR proteins, and particularly with CD3δ. The most likely molecular explanation is that CD3δ can replace CD3γ in the TCR, as has been reported (Zapata et al. 2004), whereas the reverse is not possible. As we review in the “Animal models” section, this is not the case in CD3γ Knock-Out (KO) mice. However, 3 out of 10 CD3γ-deficient patients, even though they had received transplants, did not survive to adulthood, thus suggesting that therapeutic decisions must be taken on a case by case basis.

A male or female infant showing T lymphopenia could be a result of TCRID, but also of mutations in IL7Rα, FOXN1, Coronin-1A, Zap70, TAP, MHC class II, transcription factors, PNP, ADA, or 22q11.2 (DiGeorge syndrome) (Al-Herz et al. 2014); however, only TCRID show TCR expression defects when T cells are present. Thus, a careful flow cytometry study is essential for the early differential diagnosis of TCRID, which is critical for patient survival, and should include both αβ and γδ T-cell subsets.

Samples are often received from abroad and travel long distances at changing temperatures before they reach the diagnostic lab. In such cases it is important to include age-matched controls that suffer similar shortcomings, so that comparisons can be meaningful.

TCRID may have no T cells and are thus properly called T⁻B⁺NK⁺. But often some T cells are present (T^±B⁺NK⁺), although their identification using CD3 as an extracellular marker can be difficult owing to their reduced TCR expression. Indeed, some authors have reported T⁻B⁺NK⁺ phenotypes despite the presence of abundant CD4⁺ and CD8⁺ lymphocytes expressing very low levels of surface CD3 (Roberts et al. 2007). As intracellular CD3 is not normally used for screening TCRID, we used extracellular CD4 and CD8 to identify T cells in TCRID patients. We have shown that essentially all CD4⁺ and CD8^bright cells within the lymphocyte gate are αβ T cells and, conversely, most double negative cells (DN, CD4⁻CD8⁻) in the same gate are γδ T cells (Muñoz-Ruiz et al. 2013). CD8^dull lymphocytes are mostly NK cells. For comparative purposes, we have considered CD3⁺ lymphocytes in Table 2 because they are widely reported, although it is likely an underestimation of the true number of T cells. Chimerism should be excluded using standard fingerprinting procedures, to avoid for instance counting the mother's T cells, as reported recently in a new case of CD3ɛ deficiency (Fuehrer et al. 2014).

Using these criteria, most patients with complete defects show selective peripheral blood T lymphopenia, either T⁻B⁺NK⁺ or T^±B⁺NK⁺ depending on the affected TCR chain, which is a first clue to the molecular basis of the pathology. T⁻B⁺NK⁺ has been associated with complete CD3ɛ or CD3δ defects, with less than 2% peripheral blood T cells. T^±B⁺NK⁺ has been reported in complete CD3γ, TCRα, or CD247 defects. All partial TCRID show a T^±B⁺NK⁺ or even a T⁺B⁺NK⁺ immunophenotype.

TCRID are best diagnosed if both αβ and γδ T cells are studied, as shown for complete TCRα deficiency (Morgan et al. 2011) and for partial CD3δ deficiency (Gil et al. 2011), both of which shared a Tαβ⁻γδ⁺B⁺NK⁺ immunophenotype despite their disparate molecular basis. Figure 2 (top) summarizes the effects of TCRID mutations in T-cell development.

We have listed the monoclonal antibodies used in our lab for surface staining of peripheral blood lymphocytes from possible TCRID patients (Table 3, extracellular).

We start with just UCHT-1 (anti-CD3ɛ), 11F2 (anti-γδTCR), and any anti-CD4 and anti-CD8α antibodies, and in a different staining IP26 (anti-αβTCR) and 11F2 (anti-γδTCR). Within a subset (for instance, CD4⁺ lymphocytes), TCR Mean Fluorescence Intensity is evaluated with any of those CD3- or TCR-specific antibodies in comparison with that of an age-matched healthy control (Figure 3, top center). The expression of CD4 itself can serve as an internal control (Figure 3, top left). We normally use a range of CD3-specific monoclonal antibodies, such as those listed in Table 3 (extracellular), to confirm any TCR expression defect.

Family members are not appropriate controls, because in some cases they also show (partial) TCR expression defects that may underestimate the patient's defect (see Figure 3 and Muñoz-Ruiz et al. 2013, top center and “Surface TCR expression in carriers’ lymphocytes”). If no TCR expression defect is observed, TCRID can be excluded. If a TCR expression defect is ascertained, both αβ and γδ T cells should be analyzed. If αβ T cells (defined as CD4⁺ and (or) CD8^bright) show impaired TCR expression and γδTCR⁺ show normal surface TCR expression, TCRα deficiency is likely. If αβ and γδ T cells share a surface TCR expression defect, a CD3 or CD247 defect may be suspected and further studies are required to identify the culprit protein (see “Protein identification in patients’ lymphocytes”), as indicated in the diagnostic flow chart (Figure 3, bottom). The following TCR expression hierarchy may be useful for the diagnosis of TCRID due to the lack of the indicated chains: CD3ɛ=CD3δ=TCRα≥CD247>CD3γ (our experience). Also, studying NK cells can be informative, as CD247 deficiency affects the expression of NK surface receptors such as CD16 or NKp30 (Reyburn, unpublished data). It is important to note that in contrast to αβ T cells, the range of surface TCR levels in γδ T cells in different individuals is quite heterogeneous (Garcillán et al. 2014), a fact that must be taken into account when analyzing the overlays of γδ T cells.

Lastly, it is advisable to grow T cells using IL-2 plus allogeneic feeder cells or immortalize them with Herpesvirus saimiri (HVS) or Human T-cell leukemia virus type 1 (HTLV-1), (see “Functional studies” and “Cellular models”), so that the detected TCR expression defects can be ascertained and further analyzed.

The detection of TCRID mutation carriers using flow cytometry is difficult because surface TCR expression is often within the normal range. Again, overlays of extracellular CD3ɛ staining can be used to compare several family members, as shown in Figure 3 (top center), for CD247 (Marin and Garcillán, unpublished data). Using similar strategies, we have shown that CD3γ^+/−, CD3δ^+/− (and even CD3δ^+/leaky) carriers can be identified, although the difference is not so clear and can be confounded by normal surface TCR variation (Muñoz-Ruiz et al. 2013). Of course, a genetic analysis is required for confirmation.

In the past, immunoprecipitation (Perez-Aciego et al. 1991) or Western blot techniques (Gil et al. 2011) had to be used for the identification of protein defects. But these approaches require large amounts of T cells, which are difficult to obtain when samples are scarce (as is often the case with infant patients), and even more difficult in cases with severe T-cell lymphopenia. Generation of T-cell lines can be a solution, but they take precious time, thereby delaying diagnosis. CD3 chain-specific antibodies for intracellular staining have become available lately, and in our hands simplified the diagnosis of TCRID, which can now be done rapidly. Table 3 (intracellular) lists such antibodies and Figure 3 (top right) shows a recent example of their use. Once the culprit protein is defined, genetic validation is indispensable to establish the molecular basis of the protein defect.

In summary, excellent reagents are now available to diagnose TCRID defects in a record time, even when very few T cells are present.

A mutation database has been established that contains most described mutations in the genes encoding for TCR chains (Piirilä et al. 2006). Mutations causing TCRID include missense, nonsense, and splice-site mutations as well as less common genomic deletions and insertions (Table 4). Figure 4 illustrates their presumed protein products that, when truncated, are often unstable.

Most mutations drastically affect the expression of the protein (complete defects), whereas in some leaky splice-site mutations, small amounts of normal transcripts were produced, allowing for the expression of some normal protein and thus causing a partial phenotype. This is the case for CD3δ and CD3ɛ TCRID.

Potentially truncated proteins have also been reported and were likely unstable, as they were undetectable in T cells. Indeed, the in-frame deletion of exon 7 predicted a mutant CD3ɛ protein that was not found even with very sensitive techniques (Thoenes et al. 1992). Also, a predicted mutant CD3δ protein selectively lacking the extracellular domain encoded by exon 2 was undetectable in peripheral blood mononuclear cells (PBMC) from the patient or carriers, although it could be shown in transfected non-T cells (Gil et al. 2011). This may be mutation specific, as shown recently in a new splicing mutation in CD3E intron 2, which was predicted to delete exon 2 and thus the start codon. As a consequence, no protein was detected by Western blot either in carrier PBMC or in transfected non-T cells (Fuehrer et al. 2014). In TCRα deficiency, a c.*1G>A mutation resulted in an aberrant transcript joining exon 2 to the normally untranslated exon 4. In the predicted translation product, the 35 C-terminal amino acids would be replaced by 56 amino acids encoded by exon 4. This would result in partial loss of the connecting peptide domain and abolition of the transmembrane and cytoplasmic domains of the TCRα chain. In PBMC, no TCRα or TCRβ proteins were detected by Western blotting with specific antibodies, but minute amounts of both chains were detectable by immunofluorescence microscopy using a different set of antibodies. The authors concluded that the residual predicted truncated TCRα failed to bind normally to TCRβ, because in the latter experiment the TCR chains did not colocalize. This was later confirmed in transfected HeLa cells (Morgan et al. 2011).

Nonsense mutations are clearly deleterious and frequently associate with severe TCRID immunophenotypes, except for complete CD3γ and partial CD3ɛ deficiencies. In CD3δ deficiency, a microarray analysis in thymocytes revealed more than 2-fold lower specific transcripts, but no CD3δ protein was detected by Western blot analysis of the same lysates or by thymus immunohistochemistry (Dadi et al. 2003).

Somatic mosaicism caused by reversions can ease otherwise severe clinical phenotypes. Reversions have been reported in several disorders, including dermatological, metabolic, and immunological disorders such as ADA deficiency, X-linked SCID, and Wiskott–Aldrich syndrome (Hirschhorn 2003; Wada and Candotti 2008).

Reversion of some T cells to carrier TCR expression levels has been shown to be the consequence of CD247 germ-line mutations when studied (Figure 4). However, TCR signaling after stimulation with phytohemagglutinin (PHA) or anti-CD3 of revertant T cells can be impaired (Rieux-Laucat et al. 2006) or restored to carrier levels (Marin, unpublished data) for reasons yet to be defined. Interestingly, no reversions have been reported in other TCRID, suggesting that CD247 is more prone to somatic mutations than the CD3 or TRAC genes.

CD3GDE haplotype analysis using polymorphic markers can be useful (i) to rule out CD3 mutations, if segregation does not fit the pedigree; (ii) to establish the existence of founder alleles even when families are unaware of such connections, as we have reported for CD3G in Turkey (Recio et al. 2007) and for CD3D in Ecuador (Gil et al. 2011); and (iii) for carrier detection and (or) prenatal diagnosis, because recombination within the CD3 gene complex is rare.

Screening tests specific for some founder CD3G and CD3D mutations using TaqI and BsaAI (Table 5) have been developed for several Turkish and Ecuadorian families, respectively, which help to reach quick diagnoses and segregation data for genetic counselling (Recio et al. 2007; Gil et al. 2011). Therefore, in Turkish or Ecuadorian patients with an immunophenotype suggestive of CD3γ or CD3δ deficiency, regardless of the accompanying spectrum of clinical symptoms, screening for such mutations should be considered in their molecular diagnosis.

Knock-Out (KO) mouse models (Table 6) have been used to replicate human congenital TCRID, because mouse and human TCR and CD3 chains share significant sequence and functional homology (Dave 2009). Indeed, a block in intrathymic T-cell development and reduced TCR expression was reported in all cases, although sharp differences were found for CD3γ and CD3δ (Figure 2).

CD3δ^−/−mice and humans shared a severe selective αβ T lymphocytopenia, but this was due to a block at different checkpoints (DP in mice vs DN in humans, Dave 2009). However, although CD3δ-deficient patients also lacked γδ T cells, CD3δ^−/−mice showed no obvious defect in γδ T-cell development or TCR expression (Dave et al. 1997). The mechanism became clear when stoichiometry studies showed that in humans, but not in mice, the γδTCR incorporates CD3δ (Hayes and Love 2006; Siegers et al. 2007).

CD3γ^−/−mice showed very few peripheral T cells (around 10% of normal values) with low surface TCR and abnormal stoichiometry (αβ/δɛδɛζζ) owing to a severe block at the DN stage (Haks et al. 1998). By contrast, CD3γ deficiency in humans allows the selection of substantial numbers of peripheral T cells (Recio et al. 2007), which in the case of αβ T lymphocytes expressed relatively high levels of functional αβTCR complexes (Pacheco-Castro et al. 1998), albeit with an abnormal stoichiometry (αβTCR/CD3δɛδɛζζ) and an impaired association to CD247 (Zapata et al. 2004). Therefore, the lack of CD3γ or CD3δ during αβ T-cell development gives rise to inverted phenotypes in mice versus humans, often precluding a meaningful comparison (Siegers et al. 2007).

CD3δ and CD3γ are the most homologous CD3 proteins. Yet, although patients lacking CD3δ are devoid of mature T cells, CD3γ-deficient individuals do have seemingly functional αβ and γδ T lymphocytes and generally less severe clinical features (Table 2). These differential phenotypes indicate that in humans CD3δ replaces CD3γ in the pre-TCR and TCR for stable expression on and signalling from the cell surface, allowing thymic development of fairly functional T cells. In support of this notion, expression of a human, but not a mouse, CD3δ transgene rescued αβ and γδ T-cell development in mice that were deficient for both CD3δ and CD3γ (Fernández-Malavé et al. 2006; Siegers et al. 2007). Because pre-TCR and γδTCR complexes in mice do not seem to incorporate CD3δ (Malissen et al. 1999), these data further suggest that CD3δ has evolved distinct features in humans that allows it to substitute CD3γ in both species.

CD3ɛ^−/−mice showed abrogated T-cell development with a severe block at the DN stage (DeJarnette et al. 1998) (Figure 2), supporting a shared critical role on thymocyte development versus humans (de Saint Basile et al. 2004).

As observed in humans, CD247^−/−mice (Liu et al. 1993) showed a partial block at the DP stage with weak transition to SP T cells and very low surface TCR in both thymocytes and peripheral T cells, which are nonfunctional (Liu et al. 1993).

T-cell development in TCRα^−/−mice is blocked at the DP stage, indicating that TCR gene rearrangements are critical for αβ T-cell differentiation, whereas γδ T cells are unscathed (Mombaerts et al. 1992).

The complex structure of the TCR and the severe T-cell lymphopenia normally associated to TCRID (Mazariegos et al. 2013) has hampered the dissection of the precise role of each TCR chain for TCR assembly and T-cell selection and function. KO mice, TCRID patients, mutagenized T cells lacking certain chains, TCR-transfected non-T cells and, more recently, Knocked-Down (KD) T cells tell different yet complementary tales in each species, as we review here for cellular models.