Mechanisms of Gene Therapy for Tolerance: B7 Signaling Is Required for Peptide-IgG Gene-Transferred Tolerance Induction1
http://www.100md.com
免疫学杂志 2005年第14期
Abstract
LPS-activated B cells, transduced with IgG fusion proteins, are highly tolerogenic APCs. To analyze the mechanisms for this B cell-delivered gene therapy, we first followed the fate of CFSE-labeled B cell blasts. These cells primarily localized to the spleen, where a small population persisted for at least 1 mo after injection. By day 7 after injection, 95% of the transduced cells had divided at least once, presumably an effect of the in vitro LPS activation into the cycle, because resting cells did not divide. B cells from gld donors were not tolerogenic, initially suggesting a role for Fas ligand (FasL) in tolerance. Because transduced normal B cells expressed only low levels of FasL and did not kill Fas-expressing Jurkat or A20 B lymphoma cells in vitro, these data suggest that gld B cells are not tolerogenic due to unique characteristics of these B cells rather than the lack of functional FasL expression. The transduced B cell blasts displayed significant up-regulation of both B7 costimulatory molecules, and B7.2 up-regulation was maintained through day 7 in vivo. When B cells from B7 knockout donors were transduced to express Ig fusion proteins, they were not tolerogenic in two different mouse strains and Ag models. Moreover, anti-B7 Ab blocked tolerance induction in this model, a result consistent with a role for B7 in tolerance induction. We propose that tolerance may be induced in this model by B7-driven negative regulatory signaling, but tolerance is maintained by a lack of signal 2, because expression of B7 is eventually lost in vivo.
Introduction
Our laboratory has shown that IgG fusion proteins, delivered via retroviral gene therapy for B cell Ag presentation, are highly tolerogenic for the epitopes associated with the IgG (1, 2, 3, 4). In this system, retroviral constructs have been engineered with an immunodominant epitope (e.g., p12-26 of the cI repressor molecule) or full-length protein (e.g., OVA) in frame at the N terminus of a murine IgG1 H chain (4). B cells transduced with the fusion constructs efficiently induce both B cell and T cell tolerance in normal, immunocompetent animals (4). In the first demonstration of this, normal, immunocompetent BALB/c mice that were treated with B cell blasts expressing p12-26-IgG, were unresponsive to the peptide 12-26 at both cellular and humoral levels (4). Tolerance induction has subsequently been shown for targeted Ags in experimental models of autoimmune disease, including myelin basic protein (5), myelin oligodendrocyte glycoprotein (6), glutamic acid decarboxylase (5, 7), interphotoreceptor retinoid-binding protein (8), and clotting factor VIII (fVIII)4 for hemophilia (9).
B cells have been demonstrated to be potent tolerogenic APCs (10, 11, 12). Previous evidence from our laboratory underscores the importance of B cells in peptide-IgG gene-transferred tolerance induction, because transduced B cell-knockout (KO) (μMT) bone marrow failed to induce tolerance (13). Moreover, previous data suggest that the peptide-IgG fusion proteins are directly processed and presented by the transduced B cells, because MHC class II expression on the transduced cells was required for tolerance induction (13, 14).
To further clarify the mechanism of tolerance induction by peptide-IgG gene therapy, we have used CFSE to track the peptide-IgG-transduced B cell blasts in vivo. Furthermore, we have investigated the possible role of Fas ligand (FasL) and B7 expression by the B cells in tolerance induction. Our results suggest that B7 expression by the transduced B cells may be critical for tolerance induction via induction of regulatory T cells.
Materials and Methods
Virus-producer cell lines
The 12-26-IgG1 H chain cDNA was subcloned into the retroviral vector MBAE, as described previously (4). Peptide 12-26 is the immunodominant peptide from the cI p1-102 domain in BALB/c mice. MBAE contains long terminal repeats, + packaging signals, a neomycin resistance gene, and cloned human -actin promoter sequences. Similarly, a construct encoding OVA or the entire p1-102 sequence in frame with the IgG1 H chain was engineered. Virus-producer cell lines were prepared by lipofection of -2 packaging cells with the engineered constructs. The cell lines were found to be helper virus free and to contain 105 to 106 neomycin-resistant NIH 3T3 CFU/ml. The cell lines were cultured in RPMI 1640 (Life Technologies, Invitrogen) supplemented with 5% FBS, 2 mM L-glutamine, and 2-ME. Additionally, the cell lines were maintained in 0.6 mg/ml G418 (Geneticin; Life Technologies, Invitrogen). In one experiment, a bicistronic MSCV vector expressing a nominal Ag (domain C2 or A2 of human fVIII) (9) was also used to not only track the transduced B cells but also to validate their phenotype after in vivo transfer.
Retroviral transduction
Splenic BALB/c or C57BL/6 B cells were infected in vitro via coculture with irradiated viral-producing packaging cells or via the addition of a filtered viral supernatant, as described previously (5, 13). B cells were purified to 95% homogeneity with an anti-T cell Ab mixture (anti-Thy1, anti-CD4, and anti-CD8) plus complement (Low Tox M; Cedarlane Laboratories, Accurate Chemical and Scientific Corporation). Purified B cells were prestimulated with 10–25 μg/ml bacterial LPS (Escherichia coli 055:B5; Sigma) overnight and recultured with irradiated packaging cells (or viral supernatant) in the presence of 3–6 μg/ml polybrene and additional LPS for an additional 24 h.
Mice
BALB/c and C57BL/6 mice were purchased from the National Cancer Institute and from The Jackson Laboratory, respectively. FasL-deficient C57BL/6 gld and BALB/c gld mice were gifts from Dr. Wendy Davidson (Greenbaum Cancer Center, University of Maryland School of Medicine, Baltimore, MD), and fVIII KO mice on a C57BL/6 background (E16) were bred in our animal facility from a stock provided by Dr. Leon Hoyer (Holland Laboratory, American Red Cross). B7 KO mice (BALB/c background) were a kind gift from Dr. Arlene Sharpe (Brigham and Women’s Hospital, Boston, MA); B7 KO mice on an H-2b background were purchased from The Jackson Laboratory. All animals were used at 6–8 wk of age and housed in pathogen-free microisolator cages in our animal facility.
Tolerance induction and measurement of peptide-specific immunity
In a typical experiment, normal BALB/c recipients were given injections of gene-transferred LPS blasts (3 x 106 to 1 x 107), i.p. in a volume of 500 μl or i.v. in a volume of 100 μl serum-free medium. Five to 10 days after injection (or later), animals were immunized in a hind footpad with 25 μg of protein emulsified in CFA. Two weeks later, mice received an i.p. boost of 25 μg of protein in 500 μl of PBS. Cellular responses from the spleen were assayed 14 days after boosting. On day 3, cultures were pulsed with 1 μCi/well of [3H]thymidine (Amersham Biosciences) and incubated 16–20 h. Cells were then harvested on glass fiber filters and incorporated [3H] was detected via gas scintillation counting in a Packard Matrix 9600 reader.
CFSE staining
Cells for CFSE staining were resuspended in 1 ml of cold PBS and incubated with 5 μM CFSE (CFDA SE Cell Tracer; Molecular Probes) for 15 min at 37°C. Cells were washed extensively in serum-free medium before injection.
Abs and FACS analysis
The following Abs were purchased from PharMingen: -CD19-PE (1D3), -B7.1-PE (16-10A1), -B7.2-PE (GL1), -FasL-PE (MFL3), -B220-FITC (RA3-6B2), and -I-Ab-FITC (AF6-120.1). Cells were incubated with 2.4G2 supernatant for FcR blocking for 10 min at room temperature and stained with Ab at a 1/200 dilution for 40 min to 1 h at 4°C. Propidium iodide (PI; Sigma) was added at a 5 μg/ml final concentration, and cells were analyzed on a FACScan or FACSCalibur flow cytometer (Becton Dickinson). During acquisition, cells were gated on forward scatter (FSC-H) vs PI fluorescence (FL3) to exclude dead cells. Data were analyzed by Cell Quest software.
Apoptosis assays
For apoptosis induced in the Jurkat human T cell line, effector BALB/c LPS B cell blasts were obtained at 24 h of LPS stimulation and stained with 2.5 μM 1,1'-dioctadecyl-3,3,3',3'-tetramethylindodicarbocyanine perchlorate for 15 min at 37°C. BALB/c B cell blasts were incubated with Jurkat cells at a 10:1 E:T ratio. At 6 h, the cells were washed, stained with PI, and analyzed by flow cytometry. The percentage of apoptotic cells was determined from the fraction of PI-positive cells with the gated 1,1'-dioctadecyl-3,3,3',3'-tetramethylindodicarbocyanine perchlorate-negative cells.
For apoptosis induced in A20.2J murine B lymphoma cells (I-Ad), effector C57BL/6 (I-Ab) LPS B cell blasts were obtained at 24, 48, or 72 h of LPS stimulation. At 24 h of LPS stimulation, B cell blasts were transduced with retroviral supernatant for an additional 24 or 48 h in the presence of 6 μg/ml polybrene. Viral transduced and nontransduced C57BL/6 B cell blasts were incubated with A20.2J cells at E:T ratios of 3:1 or in other experiments up to 5:1 (data not shown). At 24 h, cells were washed, stained with -I-Ab-FITC and PI, and analyzed by flow cytometry. Alternatively, A20.2J cells were cultured with CFSE-labeled K562 human FasL effectors or with hamster anti-mouse Fas (Jo2; PharMingen) and used as positive control. The percentage of apoptotic cells was determined from the fraction of PI-positive cells with the gated I-Ab or CFSE-negative cells.
Results
Localization of tolerogenic B cells in vivo
Tolerance induced by peptide-IgG B cell blasts has been demonstrated to persist for as long as 6–10 mo (8) (Y. Kang, R. Agarwal, R. Caspi, and D. W. Scott, unpublished observations). However, the localization and persistence of the tolerogenic B cells in vivo was unknown. To track the B cells in vivo, two methods were used. In the first, CFSE was initially used to label the cells before injection. In later studies, we used a bicistronic vector with GFP to track the cells.
As demonstrated in Fig. 1A, CFSE-labeled peptide-IgG-transduced B cell blasts comprise a small but reproducible population within the recipient’s spleen. On day 3 after i.p. injection, the CFSE-labeled transduced B cells comprised 0.06% of the spleen (1% of the initial injected population). Intravenous injection increased the CFSE cell yield in the days immediately after injection compared with i.p. injection (e.g., 0.32% of recipient spleen on day 1 after injection vs 0.01%). However, by day 12, the two routes of injection had equalized, in that the CFSE-labeled transduced B cells comprised 0.04% of the recipient spleen (0.5% of injected cells), regardless of route of injection. Although an i.v. route of injection seemed to maximize CFSE cell yield in the spleen at early time points (before day 12), if animals were injected with half the number of cells above (5 x 106), the CFSE populations from i.v. vs i.p. injection were comparable by day 5 after injection (data not shown).
Finally, to establish that these results were not due to intrinsic defects of these KO mice, transduced B cells were treated with anti-B7.1 (CD80) plus anti-B7.2 (CD86) at the time of transfer to recipients; transduced B cells treated with anti-B220 were used as a control for opsonization. As shown in Fig. 7C, blocking B7 expression interfered with tolerance induction, whereas blocking B220 on these B cells did not.
Other B cell stimulators, such as anti-CD40 and CpG, also up-regulate both B7 molecules, similar to that induced by LPS. Our laboratory has previously shown that, whereas LPS- or anti-CD40-stimulated peptide-IgG B cells induce tolerance, CpG-stimulated peptide-IgG B cells do not (14) (Y. Su, T.-C. Lei, and D. W. Scott, manuscript in preparation). Interestingly, in vivo, although LPS-stimulated transduced B cells maintain B7.2 up-regulation through day 10 after injection, the CpG-stimulated cells return to resting levels of B7.2 by day 3 (Fig. 8). In contrast, levels of B7.1 fall similarly in both groups, declining in both by day 3 after injection.
Discussion
We have shown previously that B cells that have been retrovirally transduced to express IgG fusion proteins are highly tolerogenic in vivo (4, 5, 6, 13, 17), in keeping with the established tolerogenicity of B cell Ag presentation (10, 11, 12). Moreover, we know that MHC class II but not FcR expression is necessary for this tolerogenic effect (13, 14, 18, 19). To further examine the mechanisms of this gene therapy for tolerance, we elected to follow the fate and phenotype of the transduced B cell after in vivo transfer, using CFSE labeling. After i.p. injection, the peptide-IgG B cells were present in the largest numbers in the peritoneum and spleen, as well as in limited numbers in the peripheral lymph nodes. The cells were undetectable in peripheral blood, heart, lung, liver, and kidney. Cells in the peritoneum likely were trapped after injection. Although some cells in this peritoneal pool probably died, a significant number of living cells remained at day 12 after injection, and cells may continue to leave the peritoneum for the secondary lymphoid organs. Within the spleen, microscopy revealed CFSE+ peptide-IgG B cells to be present in the white pulp, although it is not clear whether they are predominantly in the marginal zone (data not shown). The splenic white pulp would provide an environment in which the peptide-IgG B cells could contact and tolerize Ag-specific T cells.
CFSE+ peptide-IgG B cells divide in vivo, seemingly a consequence of the in vitro LPS stimulation before injection. Because CFSE-labeled, LPS-stimulated nontransduced cells divide as well, but resting B cells do not, we conclude that homeostatic regulation is not responsible for the continued division. In the spleen, the numbers of CFSE+ peptide-IgG B cells peaked at day 3 after injection, but a small population remained at days 30 and 60 after injection. Although CFSE+ peptide-IgG B cells were no longer evident in the spleen at day 120 after injection, T cell tolerance was seen in animals immunized at that time (Y. Kang and D. W. Scott, unpublished observations; R. Agarwal and R. Caspi, manuscript in preparation). This T cell hyporesponsiveness seen at 4 mo was not surprising, because studies on the kinetics of tolerance induction have shown that although B cell tolerance wanes within 7 wk, T cell tolerance can persist for nearly 6 mo (20). Unfortunately, CFSE has limited use in determining the life span of the peptide-IgG B cells because their division dilutes the fluorescent signal. Nonetheless, it is interesting that these activated B cells do persist for >1 mo, because the life span of B cells has long been controversial (20).
LPS-stimulated B cells have been reported to express up-regulated levels of FasL and to kill Fas-sensitive target cells (15, 16, 17). However, in our system, the peptide-IgG B cell blasts only expressed low FasL, and did they kill Fas-sensitive Jurkat or A20 lymphoma cells. We reported previously that peptide-IgG-transduced B cells from gld mice did not induce tolerance (5), and we reproduce those results in this study with another Ag (Fig. 4). However, data from the Caspi laboratory suggests that such cells can mediate tolerance induction in a uveitis model, notably in another mouse strain (R. Agarwal and R. Caspi, manuscript in preparation). We suggest that the phenotype, state of activation, and behavior of gld B cells may be aberrant in gld mice: this is a more plausible explanation to explain the differences. For example, gld B show some features of activated B cells, lack expression of CD21 and CD23, and are often oligoclonal. Indeed, regulation of B7 after LPS stimulation may also be different in B cells from gld mice, because we have recently found that B7.2 is rapidly modulated in vivo on gld B cells compared with controls (Y. Su and D. W. Scott, data not shown). Thus, although the role of Fas in our system remains unclear, recent data support the premise that FasL is not required in our system of peptide-IgG gene therapy and, moreover, that deletion of target T cells is not occurring (M. T. Litzinger, Ph.D. thesis, George Washington University, and unpublished results).
We know that B7 expression on LPS-stimulated B cells is initially high. Indeed, we reported previously that anti-CTLA-4 treatment could modulate tolerance induction in our model (13), as in others (22, 23, 24, 25, 26, 27, 28). Thus, it was important to determine whether B7.1 or B7.2 was initially required for tolerance in our system. We found that transduced B cells from B7 KO donors in two model systems were not tolerogenic, suggesting an initial role for B7, particularly B7.2, in tolerance induction. Moreover, anti-B7, but not anti-B220, could block tolerance induction in this model (Fig. 7C). Intriguingly, CpG-stimulated peptide-IgG B cells, which do not mediate tolerance, exhibited faster down-regulation of B7.2 in vivo compared with LPS B cell blasts (Fig. 8). B7 has been hypothesized to function in tolerance, variously by altering the cytokine profile of responding T cells (22, 23, 24, 25), by interacting with the negative regulator CTLA-4 (26, 27) and/or by regulating CD4+CD25+(CTLA-4+) regulatory T cells (26, 28). In our gene therapy system, recipients of peptide-IgG-transduced B cells show a reduction in IL-2, IFN-, and IL-4 (i.e., both Th1 and Th2 cytokines), suggesting a mechanism other than immune deviation (4, 13, 18).
With regard to negative signaling via CTLA-4, when the B7-high, peptide-IgG B cell blasts are injected, Ag-specific T cells in the animals would be naive and thus would not be expected to express any appreciable CTLA-4. Moreover, the peptide-IgG B cells do not appear to induce significant CTLA-4 up-regulation, because OVA-IgG B cells do not induce activation in DO11.10 OVA-specific T cells during the first 72 h in vivo, as measured by proliferation, CD69 up-regulation, CD25 up-regulation, and CTLA-4 expression (M. T. Litzinger and D. W. Scott, unpublished data). Nonetheless, CTLA-4 could be expressed at low but functionally significant levels. Although previous data from our laboratory showed that anti-CTLA-4 treatment reversed tolerance only in previously primed animals (13), the use of Ab treatment to block signaling may be less efficient than KO models.
B7 has been hypothesized to play a critical role in regulatory T cell homeostasis, as evidenced by the fact that B7 KO mice show a deficit in CD4+CD25+ T cells (28, 29, 30). Although the CD4+CD25+ population in our recipient animals does not appear to expand initially, the possibility remains that B7 expression regulates CD4+CD25+ cells to induce Ag-specific tolerance. However, because B7 expression is lost by day 10 after injection, we propose that tolerance may be maintained by a lack of the costimulatory second signal required for productive T cell activation. Alternatively, it is possible that basal B7 expression is sufficient to sustain a population of regulatory T cells and thus maintain tolerance (28). Indeed, recent data suggest that CD25+ regulatory T cells are required for and/or induced by our B cell gene therapy protocol (7, 9). Additional studies on the role of CD4+CD25+ regulatory T cells in this tolerogenic gene therapy system are under investigation.
Acknowledgments
We thank Dr. Arlene Sharpe (Brigham and Women’s Hospital, Boston, MA) for providing B7.1 and B7.2 single and double knockout mice and Dr. Wendy Davidson for gld mice and valuable discussions. We also thank Dr. Mark Williams for assistance with apoptosis assays, Dr. Moustapha El-Amine for advice, Dr. Biying Xu for critical reading of this manuscript, and Damaris Lopez for technical support.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by National Institutes of Health Grants R01 AI035622, HL061883, and T32 HL007698.
2 Current address: Laboratory of Tumor Immunology and Biology, National Cancer Institute, Building 10–Magnuson Cancer Center, Room 8B12, 9000 Rockville Pike, Bethesda, MD 20892. E-mail address: litzingm@mail.nih.gov
3 Address correspondence and reprint requests to Dr. David W. Scott, Department of Surgery, BioPark1, 800 West Baltimore Street, #317, University of Maryland School of Medicine, Baltimore, MD 21201. E-mail address: dscot005@umaryland.edu
4 Abbreviations used in this paper: fVIII, factor VIII; KO, knockout; FasL, Fas ligand; PI, propidium iodide.
Received for publication December 16, 2003. Accepted for publication May 11, 2005.
References
Scott, D. W., M. Venkataraman, J. J. Jandinski. 1979. Multiple pathways of B lymphocyte tolerance. Immunol. Rev. 43: 241-280.
Zambidis, E. T., D. W. Scott. 1996. Epitope-specific tolerance induction with an engineered immunoglobulin. Proc. Natl. Acad. Sci. USA 93: 5019-5024.
Zambidis, E. T., R. K. Barth, D. W. Scott. 1997. Both resting and activated B lymphocytes expressing engineered peptide-Ig molecules serve as highly efficient tolerogenic vehicles in immunocompetent adult recipients. J. Immunol. 158: 2174-2182.
Zambidis, E. T., A. Kurup, D. W. Scott. 1997. Genetically transferred central and peripheral immune tolerance via retroviral-mediated expression of immunogenic epitopes in hematopoietic progenitors or peripheral B lymphocytes. Mol. Med. 3: 212-224.
Melo, M. E. F., J. Qian, M. El-Amine, R. K. Agarwal, N. Soukhareva, Y. Kang, D. W. Scott. 2002. Gene transfer of Ig-fusion proteins into B cells prevents and treats autoimmune diseases. J. Immunol. 168: 4788-4795.
Xu, B., D. W. Scott. 2004. A novel retroviral gene therapy approach to inhibit specific antibody production and suppress experimental autoimmune encephalomyelitis induced by MOG and MBP. Clin. Immunol. 111: 47-52.
Song, L., J. Wang, R. Wang, M. Yu, Y. Sun, G. Han, H. Chang, J. Qian, D. W. Scott, B. Shen. 2004. B-cell delivery of GAD-Ig fusion construct induces specific tolerance and prevents diabetes in NOD mice: role of CD4 regulatory cells and TGF-. Gene Ther. 11: 1487-1496.(Mary T. Litzinger2,*,, Ya)
LPS-activated B cells, transduced with IgG fusion proteins, are highly tolerogenic APCs. To analyze the mechanisms for this B cell-delivered gene therapy, we first followed the fate of CFSE-labeled B cell blasts. These cells primarily localized to the spleen, where a small population persisted for at least 1 mo after injection. By day 7 after injection, 95% of the transduced cells had divided at least once, presumably an effect of the in vitro LPS activation into the cycle, because resting cells did not divide. B cells from gld donors were not tolerogenic, initially suggesting a role for Fas ligand (FasL) in tolerance. Because transduced normal B cells expressed only low levels of FasL and did not kill Fas-expressing Jurkat or A20 B lymphoma cells in vitro, these data suggest that gld B cells are not tolerogenic due to unique characteristics of these B cells rather than the lack of functional FasL expression. The transduced B cell blasts displayed significant up-regulation of both B7 costimulatory molecules, and B7.2 up-regulation was maintained through day 7 in vivo. When B cells from B7 knockout donors were transduced to express Ig fusion proteins, they were not tolerogenic in two different mouse strains and Ag models. Moreover, anti-B7 Ab blocked tolerance induction in this model, a result consistent with a role for B7 in tolerance induction. We propose that tolerance may be induced in this model by B7-driven negative regulatory signaling, but tolerance is maintained by a lack of signal 2, because expression of B7 is eventually lost in vivo.
Introduction
Our laboratory has shown that IgG fusion proteins, delivered via retroviral gene therapy for B cell Ag presentation, are highly tolerogenic for the epitopes associated with the IgG (1, 2, 3, 4). In this system, retroviral constructs have been engineered with an immunodominant epitope (e.g., p12-26 of the cI repressor molecule) or full-length protein (e.g., OVA) in frame at the N terminus of a murine IgG1 H chain (4). B cells transduced with the fusion constructs efficiently induce both B cell and T cell tolerance in normal, immunocompetent animals (4). In the first demonstration of this, normal, immunocompetent BALB/c mice that were treated with B cell blasts expressing p12-26-IgG, were unresponsive to the peptide 12-26 at both cellular and humoral levels (4). Tolerance induction has subsequently been shown for targeted Ags in experimental models of autoimmune disease, including myelin basic protein (5), myelin oligodendrocyte glycoprotein (6), glutamic acid decarboxylase (5, 7), interphotoreceptor retinoid-binding protein (8), and clotting factor VIII (fVIII)4 for hemophilia (9).
B cells have been demonstrated to be potent tolerogenic APCs (10, 11, 12). Previous evidence from our laboratory underscores the importance of B cells in peptide-IgG gene-transferred tolerance induction, because transduced B cell-knockout (KO) (μMT) bone marrow failed to induce tolerance (13). Moreover, previous data suggest that the peptide-IgG fusion proteins are directly processed and presented by the transduced B cells, because MHC class II expression on the transduced cells was required for tolerance induction (13, 14).
To further clarify the mechanism of tolerance induction by peptide-IgG gene therapy, we have used CFSE to track the peptide-IgG-transduced B cell blasts in vivo. Furthermore, we have investigated the possible role of Fas ligand (FasL) and B7 expression by the B cells in tolerance induction. Our results suggest that B7 expression by the transduced B cells may be critical for tolerance induction via induction of regulatory T cells.
Materials and Methods
Virus-producer cell lines
The 12-26-IgG1 H chain cDNA was subcloned into the retroviral vector MBAE, as described previously (4). Peptide 12-26 is the immunodominant peptide from the cI p1-102 domain in BALB/c mice. MBAE contains long terminal repeats, + packaging signals, a neomycin resistance gene, and cloned human -actin promoter sequences. Similarly, a construct encoding OVA or the entire p1-102 sequence in frame with the IgG1 H chain was engineered. Virus-producer cell lines were prepared by lipofection of -2 packaging cells with the engineered constructs. The cell lines were found to be helper virus free and to contain 105 to 106 neomycin-resistant NIH 3T3 CFU/ml. The cell lines were cultured in RPMI 1640 (Life Technologies, Invitrogen) supplemented with 5% FBS, 2 mM L-glutamine, and 2-ME. Additionally, the cell lines were maintained in 0.6 mg/ml G418 (Geneticin; Life Technologies, Invitrogen). In one experiment, a bicistronic MSCV vector expressing a nominal Ag (domain C2 or A2 of human fVIII) (9) was also used to not only track the transduced B cells but also to validate their phenotype after in vivo transfer.
Retroviral transduction
Splenic BALB/c or C57BL/6 B cells were infected in vitro via coculture with irradiated viral-producing packaging cells or via the addition of a filtered viral supernatant, as described previously (5, 13). B cells were purified to 95% homogeneity with an anti-T cell Ab mixture (anti-Thy1, anti-CD4, and anti-CD8) plus complement (Low Tox M; Cedarlane Laboratories, Accurate Chemical and Scientific Corporation). Purified B cells were prestimulated with 10–25 μg/ml bacterial LPS (Escherichia coli 055:B5; Sigma) overnight and recultured with irradiated packaging cells (or viral supernatant) in the presence of 3–6 μg/ml polybrene and additional LPS for an additional 24 h.
Mice
BALB/c and C57BL/6 mice were purchased from the National Cancer Institute and from The Jackson Laboratory, respectively. FasL-deficient C57BL/6 gld and BALB/c gld mice were gifts from Dr. Wendy Davidson (Greenbaum Cancer Center, University of Maryland School of Medicine, Baltimore, MD), and fVIII KO mice on a C57BL/6 background (E16) were bred in our animal facility from a stock provided by Dr. Leon Hoyer (Holland Laboratory, American Red Cross). B7 KO mice (BALB/c background) were a kind gift from Dr. Arlene Sharpe (Brigham and Women’s Hospital, Boston, MA); B7 KO mice on an H-2b background were purchased from The Jackson Laboratory. All animals were used at 6–8 wk of age and housed in pathogen-free microisolator cages in our animal facility.
Tolerance induction and measurement of peptide-specific immunity
In a typical experiment, normal BALB/c recipients were given injections of gene-transferred LPS blasts (3 x 106 to 1 x 107), i.p. in a volume of 500 μl or i.v. in a volume of 100 μl serum-free medium. Five to 10 days after injection (or later), animals were immunized in a hind footpad with 25 μg of protein emulsified in CFA. Two weeks later, mice received an i.p. boost of 25 μg of protein in 500 μl of PBS. Cellular responses from the spleen were assayed 14 days after boosting. On day 3, cultures were pulsed with 1 μCi/well of [3H]thymidine (Amersham Biosciences) and incubated 16–20 h. Cells were then harvested on glass fiber filters and incorporated [3H] was detected via gas scintillation counting in a Packard Matrix 9600 reader.
CFSE staining
Cells for CFSE staining were resuspended in 1 ml of cold PBS and incubated with 5 μM CFSE (CFDA SE Cell Tracer; Molecular Probes) for 15 min at 37°C. Cells were washed extensively in serum-free medium before injection.
Abs and FACS analysis
The following Abs were purchased from PharMingen: -CD19-PE (1D3), -B7.1-PE (16-10A1), -B7.2-PE (GL1), -FasL-PE (MFL3), -B220-FITC (RA3-6B2), and -I-Ab-FITC (AF6-120.1). Cells were incubated with 2.4G2 supernatant for FcR blocking for 10 min at room temperature and stained with Ab at a 1/200 dilution for 40 min to 1 h at 4°C. Propidium iodide (PI; Sigma) was added at a 5 μg/ml final concentration, and cells were analyzed on a FACScan or FACSCalibur flow cytometer (Becton Dickinson). During acquisition, cells were gated on forward scatter (FSC-H) vs PI fluorescence (FL3) to exclude dead cells. Data were analyzed by Cell Quest software.
Apoptosis assays
For apoptosis induced in the Jurkat human T cell line, effector BALB/c LPS B cell blasts were obtained at 24 h of LPS stimulation and stained with 2.5 μM 1,1'-dioctadecyl-3,3,3',3'-tetramethylindodicarbocyanine perchlorate for 15 min at 37°C. BALB/c B cell blasts were incubated with Jurkat cells at a 10:1 E:T ratio. At 6 h, the cells were washed, stained with PI, and analyzed by flow cytometry. The percentage of apoptotic cells was determined from the fraction of PI-positive cells with the gated 1,1'-dioctadecyl-3,3,3',3'-tetramethylindodicarbocyanine perchlorate-negative cells.
For apoptosis induced in A20.2J murine B lymphoma cells (I-Ad), effector C57BL/6 (I-Ab) LPS B cell blasts were obtained at 24, 48, or 72 h of LPS stimulation. At 24 h of LPS stimulation, B cell blasts were transduced with retroviral supernatant for an additional 24 or 48 h in the presence of 6 μg/ml polybrene. Viral transduced and nontransduced C57BL/6 B cell blasts were incubated with A20.2J cells at E:T ratios of 3:1 or in other experiments up to 5:1 (data not shown). At 24 h, cells were washed, stained with -I-Ab-FITC and PI, and analyzed by flow cytometry. Alternatively, A20.2J cells were cultured with CFSE-labeled K562 human FasL effectors or with hamster anti-mouse Fas (Jo2; PharMingen) and used as positive control. The percentage of apoptotic cells was determined from the fraction of PI-positive cells with the gated I-Ab or CFSE-negative cells.
Results
Localization of tolerogenic B cells in vivo
Tolerance induced by peptide-IgG B cell blasts has been demonstrated to persist for as long as 6–10 mo (8) (Y. Kang, R. Agarwal, R. Caspi, and D. W. Scott, unpublished observations). However, the localization and persistence of the tolerogenic B cells in vivo was unknown. To track the B cells in vivo, two methods were used. In the first, CFSE was initially used to label the cells before injection. In later studies, we used a bicistronic vector with GFP to track the cells.
As demonstrated in Fig. 1A, CFSE-labeled peptide-IgG-transduced B cell blasts comprise a small but reproducible population within the recipient’s spleen. On day 3 after i.p. injection, the CFSE-labeled transduced B cells comprised 0.06% of the spleen (1% of the initial injected population). Intravenous injection increased the CFSE cell yield in the days immediately after injection compared with i.p. injection (e.g., 0.32% of recipient spleen on day 1 after injection vs 0.01%). However, by day 12, the two routes of injection had equalized, in that the CFSE-labeled transduced B cells comprised 0.04% of the recipient spleen (0.5% of injected cells), regardless of route of injection. Although an i.v. route of injection seemed to maximize CFSE cell yield in the spleen at early time points (before day 12), if animals were injected with half the number of cells above (5 x 106), the CFSE populations from i.v. vs i.p. injection were comparable by day 5 after injection (data not shown).
Finally, to establish that these results were not due to intrinsic defects of these KO mice, transduced B cells were treated with anti-B7.1 (CD80) plus anti-B7.2 (CD86) at the time of transfer to recipients; transduced B cells treated with anti-B220 were used as a control for opsonization. As shown in Fig. 7C, blocking B7 expression interfered with tolerance induction, whereas blocking B220 on these B cells did not.
Other B cell stimulators, such as anti-CD40 and CpG, also up-regulate both B7 molecules, similar to that induced by LPS. Our laboratory has previously shown that, whereas LPS- or anti-CD40-stimulated peptide-IgG B cells induce tolerance, CpG-stimulated peptide-IgG B cells do not (14) (Y. Su, T.-C. Lei, and D. W. Scott, manuscript in preparation). Interestingly, in vivo, although LPS-stimulated transduced B cells maintain B7.2 up-regulation through day 10 after injection, the CpG-stimulated cells return to resting levels of B7.2 by day 3 (Fig. 8). In contrast, levels of B7.1 fall similarly in both groups, declining in both by day 3 after injection.
Discussion
We have shown previously that B cells that have been retrovirally transduced to express IgG fusion proteins are highly tolerogenic in vivo (4, 5, 6, 13, 17), in keeping with the established tolerogenicity of B cell Ag presentation (10, 11, 12). Moreover, we know that MHC class II but not FcR expression is necessary for this tolerogenic effect (13, 14, 18, 19). To further examine the mechanisms of this gene therapy for tolerance, we elected to follow the fate and phenotype of the transduced B cell after in vivo transfer, using CFSE labeling. After i.p. injection, the peptide-IgG B cells were present in the largest numbers in the peritoneum and spleen, as well as in limited numbers in the peripheral lymph nodes. The cells were undetectable in peripheral blood, heart, lung, liver, and kidney. Cells in the peritoneum likely were trapped after injection. Although some cells in this peritoneal pool probably died, a significant number of living cells remained at day 12 after injection, and cells may continue to leave the peritoneum for the secondary lymphoid organs. Within the spleen, microscopy revealed CFSE+ peptide-IgG B cells to be present in the white pulp, although it is not clear whether they are predominantly in the marginal zone (data not shown). The splenic white pulp would provide an environment in which the peptide-IgG B cells could contact and tolerize Ag-specific T cells.
CFSE+ peptide-IgG B cells divide in vivo, seemingly a consequence of the in vitro LPS stimulation before injection. Because CFSE-labeled, LPS-stimulated nontransduced cells divide as well, but resting B cells do not, we conclude that homeostatic regulation is not responsible for the continued division. In the spleen, the numbers of CFSE+ peptide-IgG B cells peaked at day 3 after injection, but a small population remained at days 30 and 60 after injection. Although CFSE+ peptide-IgG B cells were no longer evident in the spleen at day 120 after injection, T cell tolerance was seen in animals immunized at that time (Y. Kang and D. W. Scott, unpublished observations; R. Agarwal and R. Caspi, manuscript in preparation). This T cell hyporesponsiveness seen at 4 mo was not surprising, because studies on the kinetics of tolerance induction have shown that although B cell tolerance wanes within 7 wk, T cell tolerance can persist for nearly 6 mo (20). Unfortunately, CFSE has limited use in determining the life span of the peptide-IgG B cells because their division dilutes the fluorescent signal. Nonetheless, it is interesting that these activated B cells do persist for >1 mo, because the life span of B cells has long been controversial (20).
LPS-stimulated B cells have been reported to express up-regulated levels of FasL and to kill Fas-sensitive target cells (15, 16, 17). However, in our system, the peptide-IgG B cell blasts only expressed low FasL, and did they kill Fas-sensitive Jurkat or A20 lymphoma cells. We reported previously that peptide-IgG-transduced B cells from gld mice did not induce tolerance (5), and we reproduce those results in this study with another Ag (Fig. 4). However, data from the Caspi laboratory suggests that such cells can mediate tolerance induction in a uveitis model, notably in another mouse strain (R. Agarwal and R. Caspi, manuscript in preparation). We suggest that the phenotype, state of activation, and behavior of gld B cells may be aberrant in gld mice: this is a more plausible explanation to explain the differences. For example, gld B show some features of activated B cells, lack expression of CD21 and CD23, and are often oligoclonal. Indeed, regulation of B7 after LPS stimulation may also be different in B cells from gld mice, because we have recently found that B7.2 is rapidly modulated in vivo on gld B cells compared with controls (Y. Su and D. W. Scott, data not shown). Thus, although the role of Fas in our system remains unclear, recent data support the premise that FasL is not required in our system of peptide-IgG gene therapy and, moreover, that deletion of target T cells is not occurring (M. T. Litzinger, Ph.D. thesis, George Washington University, and unpublished results).
We know that B7 expression on LPS-stimulated B cells is initially high. Indeed, we reported previously that anti-CTLA-4 treatment could modulate tolerance induction in our model (13), as in others (22, 23, 24, 25, 26, 27, 28). Thus, it was important to determine whether B7.1 or B7.2 was initially required for tolerance in our system. We found that transduced B cells from B7 KO donors in two model systems were not tolerogenic, suggesting an initial role for B7, particularly B7.2, in tolerance induction. Moreover, anti-B7, but not anti-B220, could block tolerance induction in this model (Fig. 7C). Intriguingly, CpG-stimulated peptide-IgG B cells, which do not mediate tolerance, exhibited faster down-regulation of B7.2 in vivo compared with LPS B cell blasts (Fig. 8). B7 has been hypothesized to function in tolerance, variously by altering the cytokine profile of responding T cells (22, 23, 24, 25), by interacting with the negative regulator CTLA-4 (26, 27) and/or by regulating CD4+CD25+(CTLA-4+) regulatory T cells (26, 28). In our gene therapy system, recipients of peptide-IgG-transduced B cells show a reduction in IL-2, IFN-, and IL-4 (i.e., both Th1 and Th2 cytokines), suggesting a mechanism other than immune deviation (4, 13, 18).
With regard to negative signaling via CTLA-4, when the B7-high, peptide-IgG B cell blasts are injected, Ag-specific T cells in the animals would be naive and thus would not be expected to express any appreciable CTLA-4. Moreover, the peptide-IgG B cells do not appear to induce significant CTLA-4 up-regulation, because OVA-IgG B cells do not induce activation in DO11.10 OVA-specific T cells during the first 72 h in vivo, as measured by proliferation, CD69 up-regulation, CD25 up-regulation, and CTLA-4 expression (M. T. Litzinger and D. W. Scott, unpublished data). Nonetheless, CTLA-4 could be expressed at low but functionally significant levels. Although previous data from our laboratory showed that anti-CTLA-4 treatment reversed tolerance only in previously primed animals (13), the use of Ab treatment to block signaling may be less efficient than KO models.
B7 has been hypothesized to play a critical role in regulatory T cell homeostasis, as evidenced by the fact that B7 KO mice show a deficit in CD4+CD25+ T cells (28, 29, 30). Although the CD4+CD25+ population in our recipient animals does not appear to expand initially, the possibility remains that B7 expression regulates CD4+CD25+ cells to induce Ag-specific tolerance. However, because B7 expression is lost by day 10 after injection, we propose that tolerance may be maintained by a lack of the costimulatory second signal required for productive T cell activation. Alternatively, it is possible that basal B7 expression is sufficient to sustain a population of regulatory T cells and thus maintain tolerance (28). Indeed, recent data suggest that CD25+ regulatory T cells are required for and/or induced by our B cell gene therapy protocol (7, 9). Additional studies on the role of CD4+CD25+ regulatory T cells in this tolerogenic gene therapy system are under investigation.
Acknowledgments
We thank Dr. Arlene Sharpe (Brigham and Women’s Hospital, Boston, MA) for providing B7.1 and B7.2 single and double knockout mice and Dr. Wendy Davidson for gld mice and valuable discussions. We also thank Dr. Mark Williams for assistance with apoptosis assays, Dr. Moustapha El-Amine for advice, Dr. Biying Xu for critical reading of this manuscript, and Damaris Lopez for technical support.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by National Institutes of Health Grants R01 AI035622, HL061883, and T32 HL007698.
2 Current address: Laboratory of Tumor Immunology and Biology, National Cancer Institute, Building 10–Magnuson Cancer Center, Room 8B12, 9000 Rockville Pike, Bethesda, MD 20892. E-mail address: litzingm@mail.nih.gov
3 Address correspondence and reprint requests to Dr. David W. Scott, Department of Surgery, BioPark1, 800 West Baltimore Street, #317, University of Maryland School of Medicine, Baltimore, MD 21201. E-mail address: dscot005@umaryland.edu
4 Abbreviations used in this paper: fVIII, factor VIII; KO, knockout; FasL, Fas ligand; PI, propidium iodide.
Received for publication December 16, 2003. Accepted for publication May 11, 2005.
References
Scott, D. W., M. Venkataraman, J. J. Jandinski. 1979. Multiple pathways of B lymphocyte tolerance. Immunol. Rev. 43: 241-280.
Zambidis, E. T., D. W. Scott. 1996. Epitope-specific tolerance induction with an engineered immunoglobulin. Proc. Natl. Acad. Sci. USA 93: 5019-5024.
Zambidis, E. T., R. K. Barth, D. W. Scott. 1997. Both resting and activated B lymphocytes expressing engineered peptide-Ig molecules serve as highly efficient tolerogenic vehicles in immunocompetent adult recipients. J. Immunol. 158: 2174-2182.
Zambidis, E. T., A. Kurup, D. W. Scott. 1997. Genetically transferred central and peripheral immune tolerance via retroviral-mediated expression of immunogenic epitopes in hematopoietic progenitors or peripheral B lymphocytes. Mol. Med. 3: 212-224.
Melo, M. E. F., J. Qian, M. El-Amine, R. K. Agarwal, N. Soukhareva, Y. Kang, D. W. Scott. 2002. Gene transfer of Ig-fusion proteins into B cells prevents and treats autoimmune diseases. J. Immunol. 168: 4788-4795.
Xu, B., D. W. Scott. 2004. A novel retroviral gene therapy approach to inhibit specific antibody production and suppress experimental autoimmune encephalomyelitis induced by MOG and MBP. Clin. Immunol. 111: 47-52.
Song, L., J. Wang, R. Wang, M. Yu, Y. Sun, G. Han, H. Chang, J. Qian, D. W. Scott, B. Shen. 2004. B-cell delivery of GAD-Ig fusion construct induces specific tolerance and prevents diabetes in NOD mice: role of CD4 regulatory cells and TGF-. Gene Ther. 11: 1487-1496.(Mary T. Litzinger2,*,, Ya)