Report

Gene Therapy of Human Severe Combined Immunodeficiency (SCID)-X1 Disease

See allHide authors and affiliations

Science  28 Apr 2000:
Vol. 288, Issue 5466, pp. 669-672
DOI: 10.1126/science.288.5466.669

Abstract

Severe combined immunodeficiency–X1 (SCID-X1) is an X-linked inherited disorder characterized by an early block in T and natural killer (NK) lymphocyte differentiation. This block is caused by mutations of the gene encoding the γc cytokine receptor subunit of interleukin-2, -4, -7, -9, and -15 receptors, which participates in the delivery of growth, survival, and differentiation signals to early lymphoid progenitors. After preclinical studies, a gene therapy trial for SCID-X1 was initiated, based on the use of complementary DNA containing a defective γc Moloney retrovirus–derived vector and ex vivo infection of CD34+ cells. After a 10-month follow-up period, γc transgene–expressing T and NK cells were detected in two patients. T, B, and NK cell counts and function, including antigen-specific responses, were comparable to those of age-matched controls. Thus, gene therapy was able to provide full correction of disease phenotype and, hence, clinical benefit.

In considering diseases that might be ameliorated by gene therapy, a setting in which a selective advantage is conferred by transgene expression, in association with long-lived transduced cells such as T lymphocytes, may prove critical. SCID-X1 offers a reliable model for gene therapy because it is a lethal condition that is, in many cases, curable by allogeneic bone marrow transplantation (1–4). It is caused by γc cytokine receptor deficiency that leads to an early block in T and NK lymphocyte differentiation (1–3). In vitro experiments of γc gene transfer have shown that γc expression can be restored (5–7), as well as T and NK cell development (8–9), while the immunodeficiency of γc mice can be corrected by ex vivo γc gene transfer into hematopoietic precursor cells (10,11). Long-term expression of human γc has also been achieved by retroviral infection of canine bone marrow (12). It has been anticipated that γc gene transfer should confer a selective advantage to transduced lymphoid progenitor cells because, upon interaction with interleukin-7 (IL-7) and IL-15, the γc cytokine receptor subunit transmits survival and proliferative signals to T and NK lymphocyte progenitors, respectively (2,3). This hypothesis received further support from the observation that a spontaneously occurring γc gene reverse mutation in a T cell precursor in one patient led to a partial, but sustained, correction of the T cell deficiency, including at least 1000 distinct T cell clones (13, 14). Spontaneous correction of the immunodeficiency has otherwise not been observed in several hundred γc-deficient SCID patients nor in γcmice (2–4).

Two patients, aged 11 months (P1) and 8 months (P2), with SCID-X1 met the eligibility criteria for an ex vivo γc gene therapy trial. SCID-X1 diagnosis was based on blood lymphocyte phenotype determination and findings of γc gene mutations resulting either in a tail-less receptor expressed at the membrane (P1) (R289 X) or in a protein truncated from the transmembrane domain that was not expressed at cell surface (P2) (a frameshift causing deletion of exon 6) (15). After marrow harvesting and CD34+ cell separation, 9.8 × 106 and 4.8 × 106 CD34+cells per kilogram of body weight from P1 and P2, respectively, were preactivated, then infected daily for 3 days with the MFG γc vector–containing supernatant (16). CD34+ cells (19 × 106 and 17 × 106/kg, respectively) were infused without prior chemoablation into P1 and P2, ∼20 to 40% and 36% of which expressed the γc transgene as shown by either semiquantitative PCR analysis (P1) or immunofluorescence (P2). As early as day +15 after infusion, cells carrying the γc transgene were detectable by PCR analysis (17) among peripheral blood mononuclear cells. The fraction of positive peripheral blood mononuclear cells increased with time (Fig. 1). T lymphocyte counts increased from day +30 in P1 (who had a low number of autologous T cells before therapy), whereas γc-expressing T cells became detectable in the blood of P2 at day +60 (Fig. 2). Subsequently, T cell counts, including CD4+ and CD8+ subsets, increased to 1700/μl from day +120 to +150 and reached values of ∼2800/μl after 8 months (Fig. 2). Transgenic γc protein expression could not be studied on P1 cells given the presence of the endogenous tail-less protein. However, semi-quantitative PCR performed at day +150 showed that a high proportion of T cells carry and express the γc transgene (Fig. 1, A and B). Similar results were observed at day +275. Southern blot analysis of provirus integration in peripheral T cells from both patients revealed a smear indicating that multiple T cell precursors had been infected by the retroviral vector (18).

Figure 1

γc transgene integration and expression. Primers used to detect both PCR and RT PCR products amplify a 904–base pair stretch encompassing the 3′ end of the γc sequence and downstream vector sequence (5). (A) Semiquantitative PCR analysis of leukocyte subset DNA from P1 and P2. Blood samples were drawn at day +150. T cells (CD3+), B cells (CD19+), monocytes (CD14+), granulocytes (CD15+), and NK cells (CD56+) as well as CD34+ from a bone marrow sample obtained at day +150 from P2 were isolated by a FACStar plus cell sorter (Becton Dickinson) after staining with appropriate mAbs (19). Purity was >99%. Sorted cells were analyzed for the frequency of vector-containing cells (17). Actin DNA was amplified in parallel. Samples from peripheral blood mononuclear cells (PBMC) obtained before treatment are shown as negative controls. A standard curve was constructed by diluting cells containing one copy of the MFG γc vector (5) with noninfected cells. All specimens were tested at three dilutions: 1:1, 1:20, and 1:200. (B) Semiquantitative RT-PCR analysis of leukocyte-subset RNA from P1. The same blood sample as in (A) was used. Actin cDNA was amplified in parallel as a control of RNA content. The standard curve was constructed as in (A) (17). No signal was detected in the absence of reverse transcriptase (not shown). Each specimen was diluted to 1:1, 1:500, and 1:5000.

Figure 2

Longitudinal study of lymphocyte subsets from patient 1 (P1) and patient 2 (P2). Absolute counts of T cells (CD3+, CD8+, and CD4+), B cells (CD19+), and NK cells (CD16+, CD56+) are shown as a function of time. Day 0 is the date of treatment. The scale for NK cells is on the right- hand side of each panel.

Immunofluorescence studies showed that γc was expressed on the membrane of T cells in P2. The magnitude of expression was similar to that of control cells (Fig. 3A), as found in previous in vitro gene transfer experiments (5, 8, 9). These results indicate that sufficient transgene expression had been achieved and that γc membrane expression is likely to be regulated by the availability of the other cytokine receptor subunits with which γc associates (3). Both αβ and γδ T cell receptor (TCR)–expressing T cells were detected (Fig. 3B). Polyclonality and Vβ TCR diversity were demonstrated by using antibodies specific for TCR Vβ (19) and the immunoscope method (18, 20). In both patients, naı̈ve CD45RA+ T cells were detected, accounting for a majority of the T cell subset (Fig. 3B). In both patients, T cells proliferated from day +105 in the presence of phytohemagglutinin (PHA) and antibodies to CD3 (anti-CD3). The extent of proliferation was the same as that of age-matched controls (Fig. 4A). After primary vaccination, in vitro T cell proliferative responses to tetanus toxoid (P1 and P2: 18,000 and 12,000 cpm, respectively) and polioviruses (P2: 38,000 cpm) were observed within normal range (21). P1 T cells were also found to proliferate in the presence of protein pure derivitive (PPD) (12,000 cpm) as a likely consequence of bacillus Calmette-Guerin (BCG) persistence after immunization at 2 months of age in this immunocompromised child. Five months after cessation of intravenous immunoglobulin (Ig) therapy, antibodies to tetanus and diphtheria toxoids as well as to polioviruses were found in the serum of both patients, together with detectable concentrations of IgG and IgM (Fig. 4B). A normal level of IgA was also detected in the serum of P1. As determined by semi-quantitative PCR and reverse transcriptase–PCR analysis, it was observed that in both cases, a low fraction of B cells carry and express the γc transgene (Fig. 1). It is therefore unknown whether antibody responses are provided by untransduced or the few transduced B cells. Residual persistence (< 1%) of administered intravenous immunoglobulins (last given 5 months before measurement of antibody response) could, in part, also contribute. The γc-expressing NK cells were detected in the blood of P2 by day 30 (Figs. 1, 2, and3A). These cells efficiently killed K562 cells in vitro (18). NK cells became detectable in the blood of P1 only from day +150.

Figure 3

γc protein expression and lymphocyte subsets. (A) γc protein detection at the surface of lymphocyte subsets from a control and from P2 obtained at day +150. γc expression on B cells from P2 after treatment was undetectable (not shown). The y axis depicts the relative cell number, and thex axis shows the logarithm of arbitrary immunofluorescence units. Thin lines are isotype controls; thick lines, staining by the anti-γc. Similar results were observed on blood samples obtained at days 275 (P1) and 240 (P2). (B) The percentage of CD45RO+ and CD45RA+ among CD4 and CD8 T cells from P1 and P2 obtained at day +275 and 240, respectively, as well as the percentage of T cells expressing either an αβ TCR or a γδ TCR.

Figure 4

Functional characteristics of transduced lymphocyte subsets. (A) Longitudinal follow-up of PHA (□, ▪)– and anti-CD3 (○, •)–induced proliferation of lymphocytes from P1 (open symbols) and P2 (filled symbols) (8). Background [3H]thymidine uptake was less than 400 cpm. Positive control values are >50 × 103 cpm. (B) Serum immunoglobulin analysis was determined by nephelometry and serum antibody by enzyme-linked immunosorbent assay after immunization (see above). Diphtheria toxoid (Dipht. tox.) was also used for immunization. The last intravenous Ig injections were given at day +90 in both patients. Tet. Tox., tetanus toxoid. Isohemagglutinins to blood group A have now been detected in both patients' sera.

As a likely consequence of development and sustained function of the immune system, clinical improvement was observed in both patients. In P2, protracted diarrhea as well as extensive graft-versus-host disease (GVHD)–like skin lesions disappeared. Both patients left protective isolation at days 90 and 95 and are now at home 11 and 10 months, respectively, after gene transfer without any treatment. Both enjoy normal growth and psychomotor development. No side effects have been noted. A similar result has since been achieved in a third patient 4 months after gene transfer (22). These results demonstrate that in these patients, a selective advantage was conferred to T and NK lymphocyte progenitors, enabling full-blown development of mature and functioning T and NK lymphocytes (23).

These overall positive results contrast with the failure of previous attempts to perform ex vivo gene therapy in adenosine deaminase (ADA)–deficient patients (24–27). Concomitant administration of ADA enzyme to these patients is likely to have counterbalanced the potential growth advantage of the transduced cells in this setting (23). Also, advances in the methodology of gene transfer into CD34+ cells, i.e., the use of a fibronectin fragment (28) as well as of a cytokine combination enabling potent CD34 cell proliferation, contributed to the success of γc gene therapy.

Because γc gene transfer was achieved without any additional myeloablative or immunosuppressive therapy, these results pave the way for a possible extension of this therapeutic approach to other genetic diseases characterized by defective cell-subset generation, such as other forms of SCID (29). The kinetics of T cell development in γc gene transfer is similar to that observed in SCID patient recipients of haploidentical stem cell transplantation (4), suggesting that early progenitor cells have been infected by the MFG γc virus and effectively transduced. The hypothesis that transduced autologous T cells in P1 account for the development of the T cell compartment is unlikely because (i) the infected CD34+ cell population was contaminated by less than 0.1% CD3+ T cells; (ii) a thymic gland (27 mm by 25 mm by 25 mm at day +275) became detectable by ultrasound echography, indicative of thymopoiesis, whereas most T cells at day +275 exhibit a naı̈ve CD45RA+ phenotype; and (iii) the T cell repertoire was polyclonal and diverse. In both patients, it was shown that at day +150, a fraction of bone marrow CD34+ cells harbored and expressed the γc transgene (Fig. 1, P2). It was not possible to determine whether more primitive cells, i.e., CD34+CD38 cells, were transduced because of insufficient bone marrow sample. In the mouse, a common lymphoid progenitor (CLP) gives rise to the different lymphocyte populations (30). If a human counterpart of CLP exists, it would be the best candidate from among the earliest cells that were transduced ex vivo from these patients. Identification of integration sites in the various cell lineages could help determine the permissive differentiation stage. The question of the persistence of T and NK cell generation has yet to be addressed. If infected cells have no self-renewal capacity and have a short life-span, new generation of T and NK cells should cease. However, the fact that a thymic gland is still detectable 9 months after γc gene transfer suggests that thymopoiesis is still ongoing. Follow-up of the SCID-X1 patient in whom a spontaneous reversion mutation occurred in a T cell precursor (13, 14) indicates that gene transfer could be sufficient to provide a functional memory T cell pool for a number of years. This optimistic view will require careful sequential appraisal. Kohn et al. have previously shown that transgenes placed under the control of the long-terminal repeat (LTR) viral promoter can be silenced in quiescent T cells (31). Although the identification of silencing sequences in the MFG LTR makes this a strong possibility (31), down-regulation of γc expression has not been observed so far in these two patients, in γc-deficient mice treated by ex vivo γc gene transfer (11), or in cell lines maintained in culture over 1 year (5).

Follow-up will be required to assess the long-term effects of ex vivo γc gene transfer in CD34+ cells of SCID-X1 patients. To date, this methodology has resulted in the sustained correction (up to 10 months) of the SCID-X1 phenotype in two patients, including a patient in whom the mutated protein is expressed at the cell surface. It is presumed that the effect results from a strong positive selective pressure provided to the corrected lymphoid progenitors.

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed at INSERM Unit 429, Hôpital Necker–Enfants Malades, 149 rue de Sèvres, 75743 Paris Cedex 15, France. E-mail: fischer{at}necker.fr

REFERENCES AND NOTES

View Abstract

Navigate This Article