Restoration of Lymphoid Populations in a Murine Model of X-Linked Severe Combined Immunodeficiency by a Gene-Therapy Approach

X-LINKED SEVERE COMBINED immunodeficiency (XSCID) is a life-threatening disease characterized by greatly diminished numbers of T and natural killer (NK) cells.1-3 Although B cells are present in normal numbers in XSCID patients, they are nonfunctional.1-3 XSCID is caused by mutations in the common cytokine receptor γ chain, γ_c,4 which is a component of the receptors for interleukin-2 (IL-2), IL-4, IL-7, IL-9, and IL-15.5-11The simultaneous inactivation of these 5 different cytokine systems therefore helps to explain the severity of the immunological defects in this disease, with defective IL-7 signaling likely explaining the defective T-cell development,12 and defective IL-15 signaling explaining the defective NK-cell development.12aXSCID is the most common form of SCID, accounting for approximately half of all cases of SCID.1-3 Although most children with XSCID can be successfully treated by haploidentical bone marrow transplantation, in many of these individuals, the donor T cells engraft, but B cells do not, resulting in clinically relevant hypogammaglobulinemia that necessitates chronic intravenous gammaglobulin therapy.13 14 Therefore, the development of gene therapy for individuals with this disease could represent a substantial therapeutic advance.

Previously, our group and others have generated γ_c-expressing retroviral vectors, which can infect mammalian cells in vitro.15-18 More recently, investigators have used retroviral transduction of bone marrow to direct γ_c expression in peripheral lymphocytes in normal dogs.19 Finally, γ_c has been transduced into bone marrow from XSCID hematopoietic stem cells with the maturation of CD34⁺ bone marrow cells into double positive (CD4⁺CD8⁺) and single positive (CD4⁺CD8⁻ and CD4⁻CD8⁺) lymphocytes in a hybrid human/mouse fetal thymic organ culture.20 As a logical next step towards the eventual development of gene therapy for this disease, we attempted to correct the immunological defect in mice in which γ_c had been deleted by homologous recombination. Such an approach has been successfully performed for mice lacking expression of Jak3.21 Jak3 deficiency is an autosomal recessive form of SCID that is clinically and immunologically similar to XSCID.22 23 

Like humans with XSCID, γ_c-deficient mice exhibit marked T-cell and NK-cell defects.24-26 Examination of γ_c-deficient mice has also shown the absence of “natural” NK1.1⁺ T cells as well as γδ T cells, including, for example, dendritic epidermal T cells. In addition, these mice lack both αβ and γδ intestinal intraepithelial lymphocytes (IELs).24,25 A major difference between humans and mice lacking γ_c expression is that B cells develop in humans with XSCID (although they are dysfunctional), but B-cell development is markedly diminished in the γ_c-deficient mice.24-26 This latter difference appears to result from the importance of IL-7 signaling for murine B-cell development, whereas this function in humans is either served by another cytokine or is redundant.12 Thus, while the murine syndrome is somewhat different from human XSCID, the γ_c-deficient mice nevertheless are valuable models for determining whether a gene therapeutic approach can result in correction of the defects in lymphoid development and T-cell function. We chose to use an ecotropic retrovirus directing expression of human γ_c to attempt gene therapy in the γ_c-deficient mice. Transduction of donor γ_c-deficient bone marrow with this virus followed by bone marrow transplantation into γ_c-deficient mice increased expression of T cells, B cells, and NK cells. This establishes the potential efficacy of gene therapy and also demonstrates the ability of human γ_c to cooperate with the relevant murine cytokines and their distinctive receptor components.

The full-length human γ_c cDNA was excised from pCRScript using BsaI and BamHI and cloned into the NcoI and BamHI sites of the pMMP vector to generate pMMP-γ_c; the construction of pMMP will be reported in detail elsewhere by Drs Jeng-Shin Lee and Richard C. Mulligan (Harvard Medical School, Boston, MA), who generously provided this vector to us. Briefly, the vector uses the myeloproliferative sarcoma virus (MPSV) long terminal repeats (LTRs) and glutamine tRNA primer binding site (PBSQ) instead of the wild-type proline PBS. Both modifications were introduced to improve gene expression.27 pMMP-γ_c was transfected together with pCDNA3.1 (which carries the Neomycin resistance gene; Invitrogen, Carlsbad, CA) using the calcium phosphate method into the Phoenix ecotropic packaging cell line (provided by Dr Garry Nolan, Stanford University, Stanford, CA; seehttp://cmgm.stanford.edu/micro/fac/nolan.html) and the 293 SPA amphotropic packaging cell line (provided by Dr H.L. Malech, National Institutes of Health [NIH]), and stable clones were selected in G418 (Life Technologies, Grand Island, NY). Supernatants from isolated, stable producer Phoenix and 293 SPA clones were screened for their ability to induce expression of human γ_c in NIH3T3 cells, as assessed by flow cytometric analysis.

The γ_c-deficient mice in this study have been previously described25,28 29 and were back-crossed for more than 10 generations to C57BL/6 mice. Wild-type C57BL/6 or γ_c-deficient donor mice (6 to 10 weeks old) were injected intravenously with 5-fluorouracil (150 mg/kg body weight) 48 hours before harvesting of bone marrow. Marrow was flushed from both lower limbs and prestimulated for 48 hours in bone marrow medium (Dulbecco’s modified Eagle’s medium [DMEM] containing 15% fetal bovine serum, 4 mmol/L L-glutamine plus 20 ng/mL murine IL-3, 50 ng/mL murine IL-6, and 100 ng/mL murine stem cell factor (cytokines were all from PeproTech, Rocky Hill, NJ). Bone marrow cells were then cocultured for 48 hours with the Phoenix producer cells in the same medium with 6 μg/mL polybrene (Sigma, St Louis, MO) at 32°C. Cells were then transferred to plates coated with RetroNectin (recombinant human fibronectin fragment CH-296, 20 μg/cm²; TaKaRa Biomedicals, Shiga, Japan) and infected for 24 hours at 32°C with viral supernatant. The viral supernatant was generated by culturing a human γ_c virus-producing Phoenix cell clone in bone marrow medium for 16 hours. After a further 24-hour recovery period in fresh bone marrow medium, cells were collected and approximately 1 × 10⁶ cells were injected intravenously into 6- to 8-week-old recipient C57BL/6 γ_c-deficient mice that had been lethally irradiated with 800 rads. Mock-transduced mice were handled identically except that bone marrow cells were cocultured with untransfected Phoenix cells that were not producing virus. Bone marrow cells from wild-type mice were just maintained in culture and were not exposed to Phoenix cells or virus.

EBV-transformed B cells from a patient with XSCID (provided by Dr J.M. Puck, NIH, IL-2RG database cDNA number 830del 4[3] containing a deletion/frameshift at Leu 272; seehttp://www.nhgri.nih.gov/DIR/LGT/SCID/scid_query_result.hts?ExonIntronNumber=6) were maintained in RPMI 1640 medium supplemented with 2 mmol/L L-glutamine, 10% fetal bovine serum, 100 U/mL penicillin, and 100 μg/mL streptomycin. The cells were transduced with the γ_c-expressing amphotropic retrovirus, which was collected after 12 to 16 hours culture of the confluent 293 SPA virus-producing clone. The infection was performed during 3 24-hour cycles of incubation on RetroNectin-coated tissue culture dishes at 32°C in the presence of 6 μg/mL of polybrene. The cells were allowed to recover in growth medium for 30 hours before fluorescence-activated cell sorting (FACS) analysis or stimulation with IL-2 or IL-4.

A total of 50 μL of peripheral blood was obtained by retroorbital bleed, and DNA was isolated using the Instagene Whole Blood kit (Bio-Rad, Hercules, CA). Primers used to amplify the human γ_c cDNA were as follows: forward primer (M10), 5′-CTCTTATTCCTGCAGCTGCC-3′; reverse primer (M19), 5′-CGTTCCAGCCAGAAATACAC-3′. Amplification of human γ_c was performed using 10 μL DNA per reaction for 40 cycles of 45 seconds at 94°C, 45 seconds at 58°C, and 90 seconds at 72°C.

Fresh cells from thymus, spleen, and bone marrow were stained and analyzed on a FACSort (Becton Dickinson, San Jose, CA) using CellQuest software. Splenocyte cell suspensions were treated with ACK lysing buffer to remove red blood cells. Some splenocytes were also removed for 48-hour stimulation in 6-well plates coated with 10 μg/mL of anti-mouse CD3ε (145/2C11 monoclonal antibody [MoAb]). Cells were then stained with the following antibodies (PharMingen, San Diego, CA): fluoroscein isothiocyanate (FITC)-conjugated anti-mouse IgM, Cy-chrome–conjugated anti-mouse CD45R/B220, FITC-conjugated anti-mouse CD25 (anti–IL-2Rα, PC61), phycoerythrin (PE)-conjugated anti-mouse CD62L (L-selectin), Cy-chrome–conjugated anti-mouse CD4 (L3T4), allophycocyanin (APC)-conjugated anti-mouse CD8α (Ly-2) (53-6.7), APC-conjugated anti-mouse CD3ε (145-2C11), PE-conjugated anti-mouse NK-1.1, Cy-chrome–conjugated anti-mouse T-cell receptor (TCR) β (H57-597), and PE-conjugated TUGh4 MoAb (against human γ_c). To eliminate possible binding to Fc receptors, anti-mouse CD16/CD32 MoAb (2.4G2) was added during staining. EBV-transformed B cells that were not transduced or transduced with the γ_c retrovirus were stained with PE-conjugated mouse anti-human γ_c MoAb AG184 or a PE-conjugated mouse isotype-matched control MoAb (PharMingen).

Serum was obtained from mice at the time of sacrifice by retroorbital bleed. Total serum IgA, IgG, and IgM levels were measured by sandwich enzyme-linked immunosorbent assay (ELISA) using Ig quantitation kits according to the manufacturer’s instructions (Bethyl Laboratories, Montgomery, TX). Briefly, 96-well, flat-bottom ELISA plates (Immunon 4TM; Dynex, Chantilly, VA) were coated with a polyclonal goat anti-mouse Ig Fc-specific antibody. Serum samples or a standard reference serum were serially diluted in phosphate-buffered saline (PBS) with 1% (wt/vol) bovine serum albumin (Sigma) and added to the plates. Bound Igs were detected using horseradish peroxidase (HRP)-labeled goat anti-mouse Ig Fc-specific antibodies followed by addition of the HRP-substrate o-phenylenediamine (1 mg/mL; Sigma) and H₂O₂ (0.03%) in phosphate-citrate buffer (pH 5.0). Serum Ig levels were determined by comparing the absorbance values (450 nm) of the samples with the values from the linear portion of the curve generated with the reference serum.

Splenocytes prepared as above were plated in triplicate in a 96-well plate at 2 × 10⁵ cells/well in RPMI 1640 medium containing 10% fetal bovine serum, 2 mmol/L L-glutamine, and antibiotics. The cells were stimulated for 48 hours at 37°C with and without human IL-2 (2 nmol/L, kindly provided by Hoffmann LaRoche, Nutley, NJ). Cells were then pulsed with 1 μCi/well of methyl-³H-thymidine (NEN, Boston, MA) for 8 hours. ³H-thymidine uptake was then determined using a beta plate counter.

293T cells (a human embryonic kidney cell line expressing SV40 large T antigen) were transfected by the calcium phosphate method (transfection kit from 5′-3′, Inc, Boulder, CO). One million cells were plated into 100-mm tissue culture dishes 24 hours before transfection with 2 μg of murine IL-7Rα, 2 μg of Jak3, 0.5 μg each of Stat5a and Stat5b, and 2 μg of murine or human γ_c expression vectors. Each of these cDNAs was cloned in pME18S, a eukaryotic expression vector in which transcription is driven by the SRα promoter.30 Twenty-four hours after transfection, cells from each dish were split between 2 × 100 mm dishes and 24 hours later, cells were either stimulated or not stimulated with murine or human IL-7 for 10 minutes at 37°C, washed with cold PBS, and nuclear extracts were prepared.

The murine IL-7Rα cDNA was generated by PCR-mediated amplification of mRNA from C57BL/6 lymphocytes and from 70Z/3 cells using primers based on the published murine IL-7Rα sequence.31 DNA sequencing of our cDNAs in each case showed 5 nucleotide alterations corresponding to 4 amino acid changes, as compared with the previously reported murine IL-7Rα sequence.31 Our sequence has been submitted to GenBank (Accession no. AF078906).

Whole-cell or nuclear extracts were prepared as previously described.32 The β-casein, PRRIII, and FcγRI probes were labeled with ³²P-deoxycytidine triphosphate (dCTP) and the Klenow fragment of DNA polymerase, as previously described.33 Samples (15 μg of whole-cell extract or 7 μg of nuclear extract) were preincubated with 2 μg of poly dI-dC for 20 to 30 minutes and then the probe was added and reaction mixtures were incubated on ice for an additional 20 minutes. Samples were then run on native 6% polyacrylamide gels in 0.5 × tris/borate/EDTA (TBE) buffer.

Whole intestines were dissected from control or bone marrow transplanted mice and fixed in 10% buffered formalin. Ten-micrometer paraffin sections were stained with hematoxylin and eosin and evaluated by standard light microscopy. Intestinal IEL were identified as small cells with densely staining nuclei that are located above the epithelial basement membrane and below the plane of the epithelial cell nuclei. Goblet cells, the nuclei of which are in the same location as IEL, were distinguished from IEL by the presence of mucin vacuoles. Sections from different mice were evaluated blindly and 2 experiments were performed with similar results.

As a step toward evaluating possible gene therapy for XSCID, we assessed whether a human γ_c–expressing retrovirus could correct the immunological defects in γ_c-deficient mice. Previously, it was established that murine γ_c can functionally cooperate with human IL-2Rβ in mediating responses to human IL-2.34,35 Conversely, human γ_c can functionally cooperate with murine IL-4Rα in mediating responses to murine IL-4.7 However, the ability of human γ_c to cooperate with the other relevant cytokine receptor chains in response to murine γ_c-dependent cytokines has not been reported. Because the T-cell defect in γ_c-deficient mice is believed to result from defective IL-7 signaling, we investigated the ability of human γ_cto cooperate with murine IL-7Rα in response to murine IL-7. We transfected murine IL-7Rα together with human or murine γ_c into 293T cells along with Jak3, Stat5a, and Stat5b to create an in vitro reconstitution system similar to one previously developed for IL-2 signaling.36 Cells were then analyzed for the ability of murine IL-7 to induce Stat5 DNA binding activity as evaluated by EMSAs. As shown in Fig 1, like murine γ_c (lane 4), human γ_c cooperated with murine IL-7Rα to mediate a response to murine IL-7 (lane 6). This suggested that gene therapy using human γ_c to correct γ_c deficiency in mice was a rational experimental approach. Such an approach would potentially also determine the ability of human γ_c to cooperate with murine γ_c-dependent cytokines in vivo and if successful would allow the identical vector to be evaluated in murine and human cells.

Therefore, we constructed a retroviral expression vector in which the γ_c cDNA was inserted into pMMP, a retroviral expression vector in which expression is under the control of the MPSV LTR. A schematic is shown in Fig 2A. We transfected Phoenix ecotropic retroviral packaging cells to make a stable clone producing the virus (see Materials and Methods).

Bone marrow cells were harvested and treated as described in Materials and Methods. These cells were infected by coculture with virus-producing Phoenix cells, followed by continued exposure to viral supernatant on plates coated with RetroNectin. As a control, bone marrow cells were also cocultured with untransfected Phoenix cells (mock transduction). Approximately 1 × 10⁶cells were injected intravenously into γ_c-deficient irradiated recipient mice, and the mice were returned to a standard animal facility. Six to 10 weeks later, we analyzed γ_c-deficient mice and wild-type mice as controls, as well as γ_c-deficient mice that had received bone marrow from wild-type mice or γ_c-transduced or mock-transduced bone marrow from γ_c-deficient mice. As shown in Fig 2B, human γ_c DNA was detected in peripheral blood cells from the γ_c-deficient mice that received bone marrow transduced with the human γ_c.

Because of the greatly diminished numbers of T, B, and NK cells in γ_c-deficient mice,24-26 we evaluated the extent to which cellularity of these lineages was normalized by gene therapy. As shown in Fig 3A, the number of thymocytes was markedly increased in the mice receiving the γ_c-transduced, but not in mice receiving the mock-transduced, bone marrow. Similarly, although γ_c-deficient splenocytes exhibit a substantially decreased percent of CD3⁺ T cells with an elevated CD4:CD8 ratio as compared with wild-type mice (Fig 3B and C, second vfirst dot plots), both the percent of CD3⁺ T cells (Fig 3B) and the CD4:CD8 ratio (Fig 3C) were substantially corrected in the mice receiving γ_c-transduced bone marrow (fourth dot plot in each panel), but not in those receiving mock-transduced bone marrow (fifth dot plot in each panel). There was also partial correction of NK cells in the spleen (Fig 3D). Analysis of spleen (Fig 3B) and bone marrow (Fig 3E) in the γ_c-transduced mice also showed a substantial correction of the number of B cells. The numbers of reconstituted T cells, B cells, and NK cells in the spleen are shown in Table 1. Because γ_c-deficient mice exhibit a peripheral expansion of CD4⁺ T cells,25 37 it is only after irradiation that this population is diminished (in Table 1, compare mock-transduced to nontransduced mice). Thus, the reconstitution in γ_c-transduced mice is much better than that seen in the mock-transduced recipients and approximately equivalent to that seen in mice receiving wild-type bone marrow.

Whereas splenocytes from γ_c-deficient mice transplanted with mock-transduced bone marrow did not exhibit native and IL-2–induced lytic activity against the NK-sensitive target cell, YAC-1, splenocytes from γ_c-transduced mice had detectable cytolytic activity, suggesting that the NK cells reconstituted in these mice were functional (data not shown). The reconstituted B cells were also functional as evaluated by the augmentation of IgA, IgG, and IgM Ig levels associated with reconstitution (Table 2). In addition to correction of lymphoid populations in the thymus, spleen, and bone marrow, transduction of γ_c allowed the reconstitution of intraepithelial lymphocytes in the intestine (Fig 4). Consistent with the reconstitution of these lymphocyte populations, we confirmed expression of human γ_c in thymocytes, splenocytes, and bone marrow cells, as evaluated by flow cytometry using TUGh4 MoAb to human γ_c(PharMingen) (compare Fig 5A, C, and E to Fig 5B, D, and F). Note that in the thymus, it is evident that the entire peak is shifted, suggesting that most or all of the thymocytes express human γ_c. In the spleen, γ_c was expressed on CD3⁺B220⁻, CD3⁻B220⁺, and CD3⁻B220⁻ cells. In the bone marrow, both B220⁺ and B220⁻ populations expressed γ_c (data not shown). Based on forward versus side scatter, the lymphoid cells in the bone marrow expressed more γ_c than did the myeloid cells (data not shown).

As previously reported, splenic T cells from γ_c-deficient mice have an activated memory phenotype, with an increase in CD62L^low cells.28 However, after retroviral transduction with the γ_c retrovirus, the wild-type naive pattern was restored (Fig6A). Moreover, anti-CD3 stimulation potently induced IL-2Rα (CD25) expression on the γ_c-transduced, but not mock-transduced CD4⁺ T cells (Fig 6B). As shown in Fig 7, we also evaluated the ability of the splenocytes from variously treated mice to respond to IL-2. Splenocytes from wild-type mice exhibited the greatest relative increase in proliferation in response to IL-2 and cells from γ_c-deficient mice receiving mock-transduced γ_c-deficient bone marrow did not proliferate in response to IL-2. Interestingly, splenocytes from mice receiving either γ_c-transduced or wild-type bone marrow proliferated to a similar extent. The absolute increase in proliferation was as high as seen with wild-type mice in 5 of 8 γ_c-transduced mice examined, but the relative increase was lower due to the very low basal proliferation in wild-type mice. The reason for the differences in basal proliferation levels is unclear. Note that splenocytes from γ_c-deficient mice (Fig 7B) have a similarly high basal proliferation in response to IL-2 as compared with splenocytes from γ_c-deficient mice receiving any type of bone marrow transplant (Fig 7C, D, and E). Finally, as shown in Fig 8, splenocytes from mice that received γ_c-transduced bone marrow could respond to IL-2 effectively, as judged by Stat5 DNA binding activity.

In view of our successful transduction of murine bone marrow, we also prepared an amphotropic version of the pMMP-human γ_cvirus. As expected based on previous studies,16-18 this virus could successfully direct the expression of human γ_c in an EBV-transformed B-cell line derived from a patient with XSCID and moreover could confer to these cells the ability to mediate IL-2–induced and IL-4–induced STAT protein activation (data not shown).

XSCID is a severe inherited disease that is uniformly fatal when bone marrow transplantation is not successful. In this study, we used γ_c-deficient mice as a model for human XSCID and tested whether a retroviral-mediated gene therapeutic approach could correct in vivo the defects in these mice. Whereas no correction was detected in mice that received mock-transduced γ_c-deficient bone marrow, substantial correction of lymphoid development of T cells, NK cells, and B cells was observed in mice receiving γ_c-transduced bone marrow. The T cells not only increased in number, but their cell surface phenotypic profiles also normalized. Specifically, the CD4:CD8 ratio largely was corrected, and resting T cells no longer exhibited a memory-activated phenotype (increased CD62L expression). Moreover, T-cell function was partially restored as demonstrated by responsiveness to IL-2. In addition, intestinal intraepithelial lymphocytes were reconstituted. B-cell function also improved as evaluated by the increase in serum Ig levels. Importantly, the retrovirally transduced bone marrow was almost as effective as wild-type bone marrow. As such, these studies demonstrate functional reconstitution in a murine in vivo model of XSCID. These studies complement a study in which mice deficient in Jak3 were treated by a similar gene therapy type of approach.21 Thus, for both of the two major forms of T⁻B⁺NK⁻ SCID, gene therapy has been tested and demonstrated to be effective in murine models. Although we have demonstrated that reconstitution occurred, the latest time point at which we analyzed mice in this study was 63 days postbone marrow transplantation. Thus, it is unknown whether life-long reconstitution was achieved, and it is possible that the number of transduced clones might decrease with time. Nevertheless, it is relevant that approximately 80% of lymphocyte engraftment typically occurs within 2 months after bone marrow transplantation.38Moreover, it is reasonable to hypothesize that γ_c-transduced cells might exhibit a growth advantage and/or augmented survival given their ability to respond to γ_c-dependent cytokines, which are well-known to induce both proliferative and antiapoptotic signals.39 

The fact that the retroviral transduction of human γ_ceffectively restored T- and NK-cell development, as well as IL-2 signaling in mice, indicates that human γ_c successfully cooperated with murine IL-2Rβ (which is essential for IL-2–dependent proliferation and IL-15–dependent development of NK cells) and with murine IL-7Rα (which is essential for IL-7–dependent development of T cells and B cells in mice) in response to the endogenous murine cytokines. The effectiveness of human γ_c virus in murine stem cell transduction also suggested that an amphotropic version of the virus used in this study might have potential use for human gene therapy. As a preliminary step in this direction, we confirmed that such a virus could successfully transduce an EBV-transformed B-cell line from a patient with XSCID and confer de novo responsiveness to IL-2 and IL-4.

In XSCID, haploidentical bone marrow transplantation is highly successful even though many individuals do not successfully engraft B cells. Thus, in an investigational gene therapy clinical protocol, one would want to assure that conventional therapy was still available even if gene therapy were not successful. In this regard, it is appealing to contemplate transduction of cord blood stem cells in the perinatal period, allowing ample time for conventional bone marrow transplantation should gene therapy be unsuccessful. As haploidentical bone marrow transplantation is not uniformly successful, gene therapy is obviously a desirable eventual approach for the treatment of XSCID.

We thank Drs Jeng-Shin Lee and Richard C. Mulligan, Harvard Medical School, for providing the pMMP vector; Dr Garry Nolan, Stanford University, for providing Phoenix cells; Dr Harry Malech and Gilda Linton, NIAID, for providing 293 SPA cells; Dr Jennifer Puck, NHGRI, for providing the human XSCID EBV line mentioned in the results; and Dr Atsushi Miyajima for the pME18S expression vector; and Drs Cynthia Dunbar and Harry Malech for valuable conversations/critical comments.