Stable and functional lymphoid reconstitution of common cytokine receptor γ chain deficient mice by retroviral-mediated gene transfer

One of the most studied cytokine receptor deficiencies in man is X-linked severe combined immunodeficiency disease (SCIDX1), which results from defects in the common γ chain (γ_c). SCIDX1 accounts for 50%-60% of all cases of SCID (reviewed in 1) and is characterized by the complete absence of mature T and natural killer (NK) cells, whereas B cells are frequently present in increased numbers. The thymus and peripheral lymphoid organs are severely hypoplastic in patients with SCIDX1, suggesting an early block in T-cell differentiation. Following the co-localization of the gene encoding the γ chain of the interleukin-2 receptor (IL-2R; now denoted γ_c) to the SCIDX1 locus at Xq12-13.1, γ_c mutations were identified in a number of these patients,2 thereby demonstrating the essential roles of γ_c-dependent cytokines in human lymphoid development (reviewed in 3).

The γ_c chain was initially isolated as a functional component of the intermediate and high-affinity IL-2R.4Disruption of IL-2 mediated signaling, however, could not account for the SCIDX1 phenotype, because IL-2 deficiency was compatible with T-cell development.5 Further studies established that γ_c also participated in the receptors for IL-4, IL-7, IL-9, and IL-15 (reviewed in 6). The SCIDX1 phenotype, therefore, results from combined defects in these 5 cytokine systems. A T-cell developmental block similar to that seen in SCIDX1 can result from IL-7Rα-deficiency in man,7 suggesting that IL-7 is necessary for prothymocyte survival or expansion. In contrast, the NK-cell differentiation block results principally from defects in IL-15 signaling pathways, because this cytokine is required to promote NK-cell differentiation from bone marrow (BM) precursors.8Concerning human B-cell development, it appears that lymphoid precursors progress through γ_c-dependent stages, but γ_c-independent pathways can compensate in SCIDX1.

Without treatment, patients with SCIDX1 suffer from severe, recurrent infections, failure to thrive, and die within the first year of life. The recent results of Buckley et al9 clearly demonstrate that BM transplantation is the treatment of choice for SCIDX1, which is curative for those patients who have an HLA-identical donor. Haplo-identical transplants have also been performed, but they have a lower success rate and are plagued by poor B-cell reconstitution, thereby often necessitating long-term immunoglobulin replacement therapy.9,10 

Gene therapy remains an attractive alternative therapy for SCIDX1. In principle, γ_c-transduced hematopoietic precursors should demonstrate a marked selective advantage for lymphoid differentiation. This hypothesis has been supported by several in vitro studies, demonstrating that human γ_c gene transfer into Epstein-Barr virus–immortalized SCIDX1 B-cells lines could reconstitute IL-2R signaling.11-13 Retroviral transduction of SCIDX1 CD34⁺ BM precursor cells can permit NK- or T-cell differentiation,14 although these results were obtained in vitro under conditions that might not be attainable in vivo.

The existence of canine15 and murine16-18models of γ_c deficiency offer the possibility to test gene therapy as an alternative treatment for SCIDX1. γ_c-Deficient mice are characterized by severe reductions in B cells, NK cells, and gut-associated intraepithelial lymphocytes (IEL). In contrast, mature activated T cells develop and accumulate in γ_c^- mice, provoking inflammatory bowel disease and splenomegaly.19,20 Despite this abnormal T-cell development, γ_c^- mice are functionally immunodeficient: (1) γ_c^- lymphocytes fail to proliferate to mitogens or in mixed lymphocyte culture16,17; (2) γ_c^- mice fail to reject tumors21; and (3) γ_c^-mice fail to clear intracellular pathogens, such as Listeria monocytogenes and Toxoplasma gondii.22,23γ_c^- Mice, therefore, recapitulate many features of patients with SCIDX1. In this study, we have used γ_c^- mice to examine the feasibility of in vivo retroviral gene transfer to correct the immune system defects in this SCIDX1 mouse model.

Mice with a targeted γ_c deletion16 were maintained in specific pathogen-free conditions in an isolator barrier facility (CNRS, Orleans, France) and were from the fourth generation backcross to the C57Bl/6 background. A novel alymphoid (T-, B-, NK-) mouse strain was generated by intercrossing recombinase activating gene (RAG)-2 deficient mice, and γ_c-deficient mice (RAG2/γ_cmice21) were used as recipients for the transplants. All mice were used at 4 to 12 weeks of age.

All cell lines were cultivated in DMEM supplemented with 10% fetal calf serum (FCS) and penicillin/streptomycin. The derivation of the retroviral packing cell line BOSC 23 and the transfection protocol used to generate infectious retrovirus has been described in detail.24 Transfectants expressing a soluble form of the murine stem cell factor and a hybridoma producing murine IL-6 were kindly provided by Genetics Institute and Dr. Van Snick, respectively. Cytokine-containing supernatants were used at optimal concentrations after titration on appropriate cytokine-dependent cell lines.

Production of infectious, replication-defective ecotrophic retrovirus using BOSC 23 cells was performed according to an established protocol.24 A retroviral expression plasmid (pMX25) using long terminal repeat (LTR) sequences from the Moloney murine leukemia virus was engineered to express the murine γ_c chain. A full-length γ_c complementary DNA (cDNA) was amplified with the use of reverse transcription-polymerase chain reaction (RT-PCR) and normal splenocyte RNA, subcloned into the EcoRI site of pMX, and fully sequenced. Subconfluent BOSC 23 cells were transfected with 10 μg of pMX-γ_c by the calcium phosphate method.24Retrovirus-containing supernatants were recovered 48 hours posttransfection, filtered through 0.45 μm filters, and used directly for infection of BM cells. Viral titer was determined using NIH3T3 cells as described.12 

Bone marrow hematopoietic precursors were isolated following intraperitoneal injection of 5-fluorouracil (5-FU at 150 mg/kg; 3 days prior to harvest). Cells flushed from femora and tibias of 5-FU-treated γ_c^- donor mice were cultured at 10⁶ cells/mL in X-Vivo-10 medium (BioWhittaker) supplemented with 5% FCS (Gibco BRL), murine stem cell factor (1/100 supernatant dilution), IL-6 (1/100 supernatant dilution), and Flt-3 ligand 100 ng/mL (kindly provided by Immunex Corp) with an equal volume of retroviral supernatant in the presence of 10 μg/mL of Polybrene (Sigma) on fibronectin-coated tissues culture plates at 37°C in 5% CO₂ fully humidified incubators. Purified whole fibronectin (Sigma; dissolved at 100 μg/mL in phosphate-buffered saline) was used to coat plates for 2 hours at room temperature. This infection protocol was repeated daily for 2 more days; supernatants were removed, and nonadherent cells recovered and replated in fresh virus-containing medium. A mock transduction protocol was performed in exactly the same fashion, except that BOSC 23 cells received the parental pMX vector. Recovered nonadherent BM cells (2 × 10⁶) were grafted into the tail vein of irradiated (0.3 Gy) RAG2/γ_c double mutant recipient animals. Total BM cells from primary recipient mice were used for secondary transfers.

Lymphoid organs were removed, and single-cell suspensions were prepared using a mesh filter. Hepatic lymphocytes are a rich source of NK cells and were analyzed as described.21 Splenocytes were cultured at 2 × 10⁵ cells/well in flat-bottom 96 well plates in DMEM supplemented with 10% FCS with or without concanavalin A (Con A; 2.5 μg/mL) and γ_c-dependent cytokines (IL-2 or IL-7 at 20 ng/mL) for 72 hours. Thymocytes were similarly cultured in U-bottom 96 well plates with or without phorbol myristate acetate (PMA) (10 ng/mL) and IL-4 (100 ng/mL). Cells were pulsed with 0.5 μCi of ³H-thymidine during the final 16 hours of culture.

Immunofluorescence analysis was performed as described.16,21 Antibodies against the following cell surface antigens were used for immunofluorescence analysis (all from Pharmingen) as FITC-, PE-, or TRICOLOR conjugates: CD4, CD8, TCRαβ, TCRγδ, B220, immunoglobulin M (IgM), IgD, and DX5. A combination of anti-TCRαβ and anti-B220 antibodies were used to sort splenocytes, using a FACStar⁺ cytometer prior to DNA isolation.

For detection of the retrovirally transduced γ_ctransgene, total splenocyte or sorted lymphocyte DNA was extracted, using proteinase K in polymerase chain reaction (PCR) buffer containing 0.1% Tween 20. Following enzyme inactivation, a transgene specific PCR was performed, using exon 6 (5′-CTTCCTTGTTTGCACTGG-3′) and exon 8 (5′-GGGGAGGTTAGCGTCACTTAGGAC-3′) primers amplify a 400-base pair (bp) γ_c cDNA fragment. For RT-PCR, total peripheral blood lymphocyte RNA was converted into first strand cDNA using standard procedures, and amplification was performed using the primers described above.

Animals were immunized intraperitoneally with 100 μg of alum-precipitated nitrophenyl (NP)-conjugated bovine serum albumin (NP-BSA) and 10⁹ inactivated Bordatella pertussis. Pre-immune, day 7, and day 14 sera were collected. Levels of NP-specific IgM or IgG were determined by enzyme-linked immunosorbent assay (ELISA) as described.26 Briefly, ELISA plates were coated with 5 μg/mL NP-BSA overnight at 4°C and blocked with 5% BSA. Sera and standards were serially diluted and incubated for 1 hour at 37°C. Specific bound antibodies were revealed using isotype-specific anti-mouse antibodies directly conjugated to alkaline phosphatase followed by conversion of the chromogenic substrate p-NPP (Sigma). Total serum immunoglobulin isotype levels were determined using an ELISA kit (Pharmingen) according to the manufacturer's instructions. Statistical analysis was performed using the Studentt test.

Tissues (1-cm piece of small bowel) were fixed in Carnoy solution and embedded in paraffin. Sections were stained with methyl/pyronin or by the periodic acid-Schiff reaction. Intraepithelial lymphocytes were enumerated as described.16 

Bone marrow cells were harvested from 5-FU-treated γ_c^- mice and infected with retroviral particles (titer: 10⁶/mL) harboring the murine γ_c chain driven by the MoMLV LTRs.25 Three 1-day infection cycles were performed before transfer into alymphoid recipient animals (a new strain deficient in both the RAG2 and γ_c genes: RAG2/γ_c mutant mice). RAG2/γ_c rather than γ_c^- mice were used as hosts because the latter demonstrate abnormal T-cell development that has been shown to perturb steady-state hematopoiesis and provoke splenomegaly.19 RAG2/γ_c mice have no mature T, B, or NK cells but have otherwise normal hematopoietic parameters,21 thereby permitting donor lymphopoeisis to be determined in an unambiguous fashion. RAG2/γ_c mice (n = 7) that had been transplanted with retrovirally infected γ_c^- BM cells will be referred to as “γ_c-transduced” mice. Lymphoid reconstitution in an equivalent number of RAG2/γ_c mice that had been transplanted with mock infected γ_c^- BM cells gave an immunophenotype and function identical to nonmanipulated γ_c^- mice (data not shown). We have previously demonstrated lymphoid reconstitution in the RAG2/γ_c strain, using normal BM-derived hematopoietic precursors.21 

Peripheral lymphoid cells were analyzed in γ_c-transduced mice as well as γ_c, RAG2/γ_c, and control C57Bl/6 (γ_c⁺) mice. As previously reported,16-18 γ_c^- mice developed some mature T cells but very few B cells (Figure1A), whereas circulating lymphoid cells were not detected in the RAG2/γ_c mutants. In γ_c-transduced mice, mature T and B cells could be detected as soon as 4 weeks after the transplant, and, by 12 weeks postgraft, steady-state levels had been reached (Figure 1A). The kinetics and magnitude of lymphoid reconstitution observed in γ_c-transduced mice were comparable to that observed following transplantation of unmanipulated or mock-infected normal BM into RAG2/γ_c mice21 (Table1 and data not shown). Among individual γ_c-transduced mice, some variation between relative proportions of T and B cells could be documented, yet all animals showed peripheral lymphoid reconstitution (Table 1). These results contrast sharply with the immune phenotype in γ_c-deficient mice, whereby B lymphocyte numbers decline with age.19 

To verify the presence and expression of the retroviral γ_c transgene in γ_c-transduced animals, genomic DNA and RNA were prepared from either total peripheral blood cells, total splenocytes, or from sorted splenic T and B cells. A PCR was then performed, using primers specific for exons 6 and 8 of the γ_c gene (Figure 1B-D). These primer pairs amplify a 0.8-kilobase (kb) fragment from the wild-type γ_c locus and a 0.4-kb fragment from the γ_c cDNA transgene or endogenous transcript, but they do not amplify the targeted γ_c locus lacking exon 6.16 

A PCR product (0.4 kb) corresponding to the size expected for γ_c cDNA transgene was detected, using DNA derived from peripheral blood of all γ_c-transduced mice (data not shown). Genomic retroviral integration could be demonstrated in total splenocyte DNA from 7 of 7 γ_c-transduced animals at 7 to 47 weeks postgraft (Figure 1B; data not shown) and was also detected in sorted splenic lymphocytes from 2 of 2 animals tested at 23 weeks postgraft. No amplification products were found in mock-transduced animals, and, in C57Bl/6 mice, only the expected fragment corresponding to the wild-type γ_c locus was detected (Figure 1B; data not shown). Moreover, γ_c-transduced animals (5 of 5 tested) expressed transgene-specific γ_c transcripts of the expected size in peripheral blood cells (Figure 1C). Finally, γ_c expression could be detected on the surface of γ_c-transduced cells (Figure 1E). We conclude that retroviral infection of γ_c^- BM precursors can generate peripheral lymphoid cells with stable integration and expression of the transduced γ_c transgene.

Splenic-cell populations were further characterized in γ_c-transduced animals. The total number of splenocytes was significantly increased in γ_c-transduced animals compared with RAG2/γ_c recipients or to the γ_c^- donor mice (Table 1). A normal ratio of CD4- to CD8-expressing cells was observed in 7 of 7 γ_c-transduced animals (Figure2A); these cells expressed normal levels of TCRαβ (data not shown). Splenocytes from these mice demonstrated a normal pattern of IgM and IgD expression, indicating the presence of both newly generated and recirculating mature B cells (Figure 2B; these cells co-expressed CD19, data not shown). In contrast, CD3⁺TCRγδ T cells were detected in 3 of 7 γ_c-transduced animals (data not shown), while DX5⁺ TCRαβ^-NK cells were present in 3 of 7 γ_c-transduced mice (Figure 2C). One γ_c-transduced mouse harbored both TCRγδ cells and NK cells (data not shown). Why γδ T cells and NK cells were found in only a fraction of γ_c-transduced mice remains unclear but may be related to the level or timing of transgene expression by the retroviral promoter.

To determine whether peripheral T lymphocytes present in γ_c-transduced animals expressed functional γ_c-containing receptor complexes, splenocytes were stimulated with Con A alone or in the presence of the γ_c-dependent cytokines IL-2 or IL-7 (Table2). We have previously demonstrated that IL-2 and IL-7 do not promote the proliferation of γ_c-deficient cells.16 In contrast, IL-2 or IL-7 stimulated the Con A mitogenic response of splenocytes from 3 of 3 γ_c-transduced animals (Table 2). Thymocytes from γ_c-transduced mice also responded to the combination of IL-4 plus PMA, unlike γ_c^- thymocytes (Table2). These results demonstrate that γ_c gene transfer can restore expression of functional IL-2, IL-4, and IL-7 receptors on T cells from γ_c-transduced animals.

To assess B-cell function in γ_c-transduced mice, we first analyzed steady-state circulating plasma immunoglobulin levels in animals >8 weeks posttransplant. γ_c^- mice exhibit abnormal serum immunoglobulin levels, reflecting defective B-cell differentiation.16,17 In particular, γ_c^- mice have extremely reduced levels of IgG1 and a 2 log reduction in the concentrations of serum IgG2a (Figure3A). Following γ_c gene transfer, we found a normal concentration of all serum immunoglobulin levels tested in γ_c-transduced animals (Figure 3A). Levels of IgG1 and IgG2a were not significantly different from control C57Bl/6 mice (P = .21 for IgG1 and P = .35 for IgG2a). Sera from RAG2/γ_c mice were negative for all immunoglobulin subclasses.

B-cell immunoglobulin responses were evaluated, following immunization with the T cell-dependent antigen NP-BSA. As shown in Figure 3B, normal titers of circulating anti-NP-BSA-specific IgM antibodies could be detected in γ_c-transduced animals by day 7 postimmunization and persisted at day 14. Moreover, switched IgG antibodies could be detected in these mice at day 14. In contrast, immunization of γ_c^- mice, using either T-dependent or T-independent antigens, fails to elicit any antigen-specific immunoglobulin (Figure 3B; Vosshenrich et al, submitted). Taken together, these data demonstrate that γ_c-transduced animals harbor mature B cells and T cells capable of functional immune responses.

Distinct morphological structures located in between the intestinal crypts (cryptopatches) contain cells with an immature lymphoid phenotype (c-kit⁺, IL-7Rα⁺) that appear to play a role in the generation of some IEL T-cell subsets.27,28 Cryptopatch generation requires the IL-7Rα chain28 or the γ_c chain (our unpublished observations). The development of all IEL subsets is γ_cdependent as well.16 

Gut-associated lymphoid cells developed in γ_c-transduced mice. Cryptopatches were detected in 7 of 7 γ_c-transduced animals (Figure 2D). Their morphological structure was comparable to that of control C57Bl/6 mice (data not shown). In addition, IELs could be detected in the γ_c-transduced animals (Figure 2D, Table 1), whereas IELs have never been observed in γ_c^- mice.29 Together, these observations suggest that γ_c gene transfer into hematopoietic precursors capable of generating intestinal lymphoid cells had occurred.

Two γ_c-transduced animals were analyzed at 40 or 47 weeks after the initial transplant. In these animals, we were able to detect normal percentages of circulating mature B and T cells (Figure 4). The retrovirally transduced γ_c transgene was detected in splenocyte DNA from these mice, and development of gut-associated lymphoid cells had occurred (data not shown). No pathological effects of gene transfer were observed in γ_c-transduced animals analyzed during the course of this study.

To determine whether stable integration of the γ_ctransgene could be achieved in γ_c^-hematopoietic stem cells, we transplanted 10⁶ BM cells from 3 different γ_c-transduced mice (at 8 to 12 weeks postgraft) into secondary RAG2/γ_c recipient mice. The primary γ_c-transduced mice demonstrated normal circulating B and T cells at this time, and, on secondary transfer, the γ_c-transduced BM inoculum could again give rise to normal percentages of mature B and T cells (a representative example is shown in Figure 4). A total of 8 mice were analyzed up to 23 weeks following secondary transfers, and the kinetics and stability of peripheral lymphoid reconstitution were similar between primary and secondary recipients (data not shown).

In this study we have shown that ex vivo γ_c gene transfer into hematopoietic precursor cells from γ_c-deficient mice can lead to a correction of the immune deficiency without obvious side effects. Following transfer of the retrovirally transduced γ_c^- BM cells to alymphoid recipients, we found that engrafted animals generated peripheral lymphoid cells that stably integrated and expressed the γ_c transgene. Both mature B cells and T cells were present in the recipients: γ_c-transduced lymphocytes responded to γ_c-dependent cytokines and were capable of generating a cooperative (B-T cell) immune response following antigen immunization. Finally, γ_c-transduced mice showed development of gut-associated lymphoid cells, the presence of which is entirely dependent on expression of the γ_cchain.16 Taken together, these results demonstrate the feasibility of γ_c gene transfer ex vivo and fulfill an important prerequisite for further clinical studies of γ_cgene therapy for patients with SCIDX1.

The retroviral infection protocol used in our model was not aimed at optimizing gene transfer conditions. Nevertheless, cytokine stimulation of BM progenitor cells followed by 3 cycles of infection (using viral supernatants and fibronectin-coated plates) was found to work efficiently as previously documented.30-32 Importantly, this protocol closely matches current conditions established for treatment of patients with SCIDX1 by BM transplantation or ex vivo gene transfer, including noncytoablative host conditioning.9 The observations that distinct T-cell subsets, B cells, and IEL-cell populations were stably reconstituted in γ_c-transduced mice up to 47 weeks postprimary transplant as well as on secondary marrow transplantation (up to 23 weeks after transfer) strongly suggest that early hematopoietic progenitor cells with self-renewal capacity were successfully infected. It should be noted that both primary and secondary RAG2/γ_c recipients were not subjected to lethal irradiation, so that competition between endogenous hematopoietic cells and transduced donor cells could well have occurred.

A human autosomal SCID syndrome due to JAK-3 deficiency has been described that is immunologically identical to SCIDX1,33,34thereby highlighting the requirement for JAK-3 activation following γ_c-receptor signaling. Recently, a mouse model of JAK-3 deficiency was successfully treated by ex vivo gene transfer with the generation of peripheral T and B cells that were functionally responsive to γ_c-dependent cytokines.35 Thus, ex vivo retroviral gene transfer appears capable of correcting immune deficiencies secondary to γ_c or JAK-3 defects.

It is important to note that, in these experiments, the transduced genes were constitutively expressed under the control of the viral LTR. Dysregulated γ_c or JAK-3 signaling could in theory lead to autonomous cell activation caused by receptor or kinase overexpression. However, in the case of γ_c-transduced mice, no evidence of myelo- or lymphoproliferation or autoimmune manifestations was observed in the long-term reconstituted animals. The absence of side effects could reflect the low level of γ_c expression detected in the treated animals. Still, the lymphoid system of γ_c-transduced mice appeared functional by a number of criteria (in vitro proliferative responses, specific immunoglobulin production following immunization), suggesting that high-level γ_c expression may not be a requisite for immune reconstitution. Because γ_c is constitutively expressed in multiple hematopoeitic lineages (reviewed in 36), a potential protein excess might not have negative consequences. These results suggest that (1) regulated transgene expression may not be required for the generation of functional lymphoid cells from γ_c^-hematopoietic precursors and that (2) retroviral gene transfer to correct deficiencies in the γ_c-signaling pathways may have low-potential toxicities, which is an important consideration for any new therapeutic approaches.

Although secondary extinction of retroviral transgenes has been reported,37-39 extinction of the γ_c transgene was not observed in this study, including mice receiving secondary transplants and analyzed 23 weeks later. Whether extinction is a randomly occurring event or requires a particular chromatin conformation around the proviral integration site is not known. It is, therefore, difficult to predict from this animal model of SCIDX1 what the outcome of gene therapy in patients with SCIDX1 will be with regard to long-term expression from retroviral vectors. Natural selection of γ_c-expressing cells may be an important factor in preserving transgene expression in vivo following gene transfer. It is worthwhile noting that γ_c-transduced SCIDX1 Epstein-Barr virus–transformed B-cell lines were maintained for more than 1 year in culture without evidence of loss of the γ_c transgene expression.12 

Spontaneous in vivo reversion of inherited mutations has been observed in patients with ADA and SCIDX1 (reviewed in 1). In the latter case, the γ_c reversion resulted in a significant and long-lasting correction of the T-cell deficiency. This result strongly suggests that a γ_c⁺ lymphoid precursor carries a major survival and/or growth advantage over γ_c^- cells in vivo and are in agreement with the original observations of skewed X-inactivation patterns in SCIDX1 obligate female carriers made by Puck et al.40 These observations provide a logical basis to hypothesize that a selective advantage would be conferred to corrected progenitor cells in patients with SCIDX1. The results obtained in γ_c^- and JAK3-deficient mice support the notion that human SCID syndromes represent a rather favorable model for treatment by ex vivo gene transfer.

We are grateful to Dr. Francois Huetz for providing immunization reagents, Fabian Gross for fibronectin, Françoise Selz for cell sorting, and Michele Malassis for technical help.