Gene Therapy to Inhibit Xenoantibody Production Using Lentiviral Vectors in Non-Human Primates

Gene Therapy (2007) 14, 49–57 & 2007 Nature Publishing Group All rights reserved 0969-7128/07 $30.00 www.nature.com/gt ORIGINAL ARTICLE Gene therapy to inhibit xenoantibody production using lentiviral vectors in non-human primates

JY Fischer-Lougheed1, AF Tarantal2, I Shulkin1, N Mitsuhashi1, DB Kohn3, CCI Lee2 and M Kearns-Jonker1 1Department of Cardiothoracic Surgery, The Saban Research Institute of Childrens Hospital Los Angeles and University of Southern California Keck School of Medicine, Los Angeles, CA, USA; 2California National Primate Research Center, University of California, Davis, CA, USA and 3Division of Research Immunology/Bone Marrow Transplantation, The Saban Research Institute of Childrens Hospital Los Angeles and University of Southern California Keck School of Medicine, Los Angeles, CA, USA

Xenoantibodies to the gala1,3 gal (gal) epitope impede the observed in several hematopoietic lineages in all monkeys. use of pig tissues for xenotransplantation, a procedure that Engraftment in animals receiving SIV-based a1,3GT con- may help overcome the shortage of human organ donors. structs was similar to that achieved using the HIV-1-derived Stable gal chimerism and tolerance to gal+ hearts could lentivector for the first 2 months post-transplantation, but be achieved in a1,3-galactosyltransferase (a1,3GT)À/À mice increased thereafter to reach higher levels by 5 months. using lentiviral vectors expressing porcine a1,3GT, the Upon immunization with porcine hepatocytes, the production enzyme that synthesizes the gal carbohydrate. In this study, of anti-gal immunoglobulin M xenoantibody was substantially we evaluated whether chimerism sufficient to inhibit anti-gal reduced in the gal+ BM recipients compared to controls. This xenoantibody responses can be achieved using lentivectors study is the first to report the application of gene therapy to in non-human primates. Rhesus macaques were trans- achieve low-level, long-term gal chimerism sufficient to inhibit planted with autologous, a1,3GT-transduced bone marrow production of anti-gal antibodies after immunization with (BM) following sublethal irradation. Simian immunodeficiency porcine cells in rhesus macaques. virus (SIV)- and human immunodeficiency virus (HIV)-1- Gene Therapy (2007) 14, 49–57. doi:10.1038/sj.gt.3302818; derived lentiviral constructs were compared. Chimerism was published online 3 August 2006

Keywords: xenotransplantation; lentiviral vector; chimerism; primate; rhesus monkeys

Introduction similar results using transfer of adenovirus-transduced, autologous splenic lymphocytes. More recently, we Xenotransplantation represents a potential solution to have shown that aGT-encoding lentivectors can achieve the shortage of human organs and tissues. Humans, sufficient chimerism in aGTÀ/À mice to result in per- however, reject porcine grafts owing to natural xeno- manent tolerance to gal+ hearts following sublethal antibodies, most of which are directed at the gal epitope irradiation.11 The objective of the present study was (Gala1–3Galb1–4GlcNAc-R).1 The gal carbohydrate, to apply non-myeloablative conditions and lentiviral synthesized by a1,3galactosyltransferase (aGT), is not vectors to achieve chimerism for the gal epitope in expressed in humans, apes or Old-World monkeys.2 non-human primates, and to study the effect of this Anti-gal antibodies are produced, however, in aGTÀ/À procedure on humoral immune responses to porcine knockout mice that lack this carbohydrate.3–5 Mixed-cell cells. This is a key step to clinical applications of a similar chimerism, achieved following transplantation of aGTÀ/À ex vivo gene therapy approach for humans. knockout mice with wild-type gal+ bone marrow (BM),6,7 leads to tolerance to gal+ heart grafts. The application of gene therapy as a means to achieve tolerance was first reported in aGTÀ/À knockout mice transplanted with Results BM, transduced with a murine retrovirus encoding the Juvenile rhesus monkeys (Macaca mulatta) were selected aGT enzyme.8 We reported that permanent tolerance for these studies based on their low, matched serum to gal+ hearts can be achieved in this animal model levels of pre-existing immunoglobulin M (IgM) and using lentiviral vectors,9 and Ogawa et al.10 reported immunoglobulin G (IgG) anti-gal antibodies. Anti-gal xenoantibody levels were determined by enzyme-linked Correspondence: Dr M Kearns-Jonker, Department of Cardiothor- immunosorbant assay (ELISA) using gal pentasacchar- acic Surgery, The Saban Research Institute, Childrens Hospital Los ide- and trisaccharide-coated plates. Autologous total Angeles and University of Southern California Keck School of BM and BM-derived CD34+ cells were transduced with a Medicine, 4661 Sunset Blvd., MS#137, Los Angeles, CA 90027, USA. E-mail: [email protected] lentivector-expressing porcine aGT, or a control lentivec- Received 15 January 2006; revised 28 April 2006; accepted 30 May tor. The cells were pooled and transplanted 24 h post- 2006; published online 3 August 2006 total body irradiation (TBI) using the intravenous route Inhibition of xenoantibody production using lentiviral vectors JY Fischer-Lougheed et al 50 into recipients that had received submyeloablative TBI, decline in the B-cell population in all treated monkeys. but no other immunosuppressive regimens before bone The data for four representative gal+ BMT-treated marrow transplantation (BMT). Table 1 summarizes animals and two corresponding controls are shown in the treatment of each BMT recipient. This method of Figure 4. The profiles of the other mononuclear cell transduction proved successful in achieving chimerism lineages examined were unaltered, with some variation at levels sufficient to induce tolerance to the gal carbo- from animal to animal. This observation suggests that B hydrate in our prior studies in aGTÀ/À knockout mice.9,11 cells in rhesus macaques were less resistant to irradiation An aliquot of transduced cells was kept for 4–6 days than other lineages. The decrease ranged from 48 to 92% in culture to test transduction efficiency.12–14 Cell-surface (n ¼ 7) in the peripheral blood. The depletion was similar gal expression on transduced cells, as monitored by the in the BM. By day 50 post-BMT, B-cell numbers returned 15–16 binding of lectin isolectin B4 (IB4) to gal carbohydrates, to values that were similar to pre-BMT levels. This was evaluated by flow cytometry and detected on up to 27% of rhesus monkey BM cells at 4–6 days post- transduction (Figure 1a). The presence of the porcine aGT transgene was demonstrated by quantitative poly- merase chain rection (qPCR) in the aGT-transduced cells that were transplanted (Figure 1b). Peripheral blood and BM samples were collected at 8 and 21 days post-transplantation, and regularly for a period of 1 year thereafter, to study the immune profile, cell reconstitution, and the pattern and persistence of engraftment in hematopoietic cell lineages.

An anti-gal IgG1 immune response is induced following transplantation of gal+ BM ELISA experiments were performed on gal pentasaccharide-coated plates to examine changes in the serum anti-gal antibody profiles after transplantation (Figure 2). The levels of anti-gal IgM antibodies did not significantly increase at days 8 and 21 post-BMT (Figure 2a). How- ever, transplantation of gal+-transduced BM cells did induce an anti-gal IgG response in the recipients, as early as 8 days post-BMT (Figure 2b). Anti-gal IgG xenoantibodies were not observed in the control BMT animals, which were transplanted with cells that did not express the gal epitope. In all gal+ BM recipients (gal+ BMT), the anti-gal IgG antibodies were of IgG1 subtype (Figure 2c). These antibodies were not capable of complement- dependent cytotoxicity to pig cells (Figure 2d). Genomic Figure 1 Expression of the aGT transgene in rhesus BM. (a) Flow DNA was extracted from the BM and peripheral blood cytometry histoplot demonstrating that rhesus monkey BM cells cells that were harvested post-transplantation. qPCR was expressed the gal carbohydrate following transduction with the aGT lentivector. The white area in the graph represents non- used to detect the presence of the aGT transgene. Figure transduced BM cells from animal no. 240, and the gray area depicts 3 shows that the transgene was detected in the marrow BM cells from animal no. 240 post-transduction. The bar indicates of the gal+ BM recipients at 8 days post-BMT. the region considered positive for lectin binding. The data are representative of similar results obtained from all gal+ BMT animals. B-cell numbers decline following sublethal irradiation (b) The presence of the aGT transgene was demonstrated by qPCR in autologous BM cells transduced with aGT-encoding lentivectors Hematopoietic reconstitution was found to be rapid in (nos. 240, 322, 296, 309 and 256; filled bars) and not in cells the animals post-TBI. Examination of peripheral blood transduced with control vector (nos. 318 and 796; dotted). mononuclear cell (PBMC) populations by flow cytometry Experiments were repeated twice, samples were run in duplicate using lineage-specific markers showed a consistent for each experiment and the data are shown as the mean7s.e.

Table 1 BMT recipients and treatment

Recipient Age and weight TBI Number of autologous cells transplanted Transduced nos. at transplantation CD34+ cells

240 8 months; 1.5 kg (female) 300 cGy (2 Â 150 cGy) 26 Â 106 aGT-transduced cells (17.3 Â 106 cells/kg) 1.4 Â 106 322 7.5 months; 2.0 kg (female) 350 cGy (single dose) 41 Â106 aGT-transduced cells (20 Â 106 cells/kg) 0.5 Â 106 296 9 months; 2.3 kg (female) 350 cGy (single dose) 70 Â 106 aGT-transduced cells (30 Â 106 cells/kg) 0.7 Â 106 309 12 months; 2.4 kg (female) 350 cGy (single dose) 96 Â 106 SIV-aGT-transduced cells (41 Â106 cells/kg) 1.7 Â 106 256 12 months; 2.7 kg (male) 350 cGy (single dose) 40 Â 106 SIV-aGT-transduced cells (15 Â 106 cells/kg) 1.1 Â106 796 10 months; 1.5 kg (female) 300 cGy (2 Â 150 cGy) 13.5 Â 106 control-transduced cells (9 Â 106 cells/kg) 0.9 Â 106 318 23 months; 2.5 kg (female) 350 cGy (single dose) 63 Â 106 control-transduced cells (25 Â 106 cells/kg) 0.6 Â 106

Abbreviation: BMT, bone marrow transplantation; TBI, total body irradiation.

Gene Therapy Inhibition of xenoantibody production using lentiviral vectors JY Fischer-Lougheed et al 51

Figure 2 Anti-gal antibody response following gal+ BM transplantation. (a) Anti-gal IgM antibodies were not elicited in response to BMT as shown here by ELISA for three representative gal+ BMT and two control BMT animals. (b) Anti-gal IgG antibodies were induced upon transfer of gal+ BM cells. (c) The anti-gal IgG response was of the IgG1 subtype. Results are illustrated here for gal+ BMT no. 240. These data are representative of the IgG response demonstrated in ﬁve gal+ BMT recipients. (d) The anti-gal antibodies did not induce complement- dependent cytotoxicity to pig cells. Experiments are representative of 4–9 independent assays. Results are depicted as the mean7s.e.

and 22 days post-transplantation, then declined after 10 weeks to low levels that ranged between 1 and 8% of PBMCs over the course of the experiment. Similar levels of expression were detected in the BM. Chimerism was confirmed by qPCR (see Figure 3). Dual-color flow cytometry was used to identify which cell lineages expressed the gal carbohydrate in the aGT BMT chimeras (shown in Figure 5 for three representative gal+ BM recipients). Gal carbohydrate surface expression was observed on B cells (CD20+), T cells (CD3+) and on monocytes/macrophages (CD11b+, CD14+, CD56+), indicating multilineage chimerism. Lectin IB4 binding was observed on an average of 10–20% of cells within each lineage within the first 2 months post-transplantation. Chimerism was highest in CD20+ and CD11b+ populations at this time. At Figure 3 aGT transgene was detected in BM cells harvested 8 days post-BMT. qPCR was used to demonstrate the presence of the 5 months post-transplant, gal expression was detected + transgene in BM cells collected after transplantation. The assay was on 1–3% of lymphoid cells in all gal BMT animals and repeated twice and samples were run in duplicate for each gal+ by 1 year following BMT, persisting gal cell-surface BMT animal. The data are shown as the mean7s.e. expression (0.65%) was only identified within the CD11b+ cell population. transient drop in circulating lymphocytes was also Our experimental animals transduced with simian observed by others in rhesus monkeys following a 3 Gy immunodeficiency virus (SIV)-derived lentiviral aGT TBI.17 The sensitivity of B cells to submyeloablative vectors (nos. 309 and 256) demonstrated chimerism in irradiation has been reported in mice,18 and observed in several lineages at 5 months (160 days) post-BMT (Figure aGTÀ/À knockout mice;11 however, it has been suggested 6a). Chimerism was identified in these and all experi- that B cells producing natural anti-gal antibodies in mental animals within this study as the percentage of aGTÀ/À mice are irradiation resistant.6 gal+ cells obtained after subtracting background binding in peripheral blood or BM cells obtained from control, transduced and/or normal rhesus monkeys. At 5 Long-term, low-level chimerism can be achieved months, chimerism in this group was observed in 2.4 in rhesus monkeys and 4.7% of CD20+ cells, and 2.1 and 5.1% of CD11b+ cell Cell-surface gal-expressing cells were examined in the populations of recipient nos. 309 and 256, respectively, peripheral blood and BM of gal+ BM recipients by flow as illustrated by the flow cytometry acquisition dot- cytometry. The levels of total gal+ cells peaked between 8 plots (Figure 6b). Data in the literature suggest that

Gene Therapy Inhibition of xenoantibody production using lentiviral vectors JY Fischer-Lougheed et al 52

Figure 4 Flow cytometry analysis of peripheral blood cell reconstitution after transplantation. T-cell, B-cell and CD11b+ PBMC reconstitution following sublethal irradiation and BMT over a time course of 50 days is shown here for four lenti-gal BMTs and two control BMT animals. B cells were sensitive to sublethal TBI.

SIV-derived lentiviral vectors more efficiently transduce of IgM and IgG in the serum of these animals upon rhesus monkey cells compared with human immuno- immunization (not shown). deficiency virus (HIV)-1-derived vectors.19,20 At 1 year Examination of the levels of anti-pig IgM antibodies post-transplantation with SIV-based lentiviral vectors (Figure 7c), measured by ELISA on plates coated with in our study, 1–3% of gal+ cells were still detected by porcine aortic endothelial cells (PAECs), showed no flow cytometry in the spleen (Figure 6c). Our results are significant difference in the response of gal+ BMT consistent with reports that SIV-based lentiviral vectors animals and controls, suggesting that the reduced demonstrate improved efficiency of transduction of cytotoxicity21 in gal+ BMT monkeys was owing to the non-human primate cells, based on the percentage of attenuated production of antibodies directed at the gal transduced cells that we can identify over the long term carbohydrate epitope. The persistence of low levels of in our experimental animals. gal+-transduced cells, therefore, was sufficient to inhibit production of anti-gal antibodies. The immune response to non-gal xenoantigens remained intact. Immunization with porcine cells elicited a minimal anti- gal IgM antibody response in the gal+ BMT recipients + Seven monkeys (five gal BMT animals and two controls) Discussion were immunized with 60 million porcine hepatocytes and endothelial cells to assess whether chimerism was The use of rhesus monkeys to determine whether sufficient to inhibit xenoantibody responses at 2–4 gene therapy can induce tolerance to porcine xeno- months post-BMT. Figure 7a shows that the level of grafts represents an important pre-clinical model, as anti-gal IgM antibodies induced after immunization was these animals are most closely related to humans that considerably attenuated in the gal+ BM recipients when have pre-existing xenoantibodies that initiate xenograft compared with immunized controls. The immuniza- rejection.2,21,22 We show here that transplantation of tion did not elicit an increase in antibodies capable of autologous, transduced marrow expressing cell-surface complement-dependent cytotoxicity (Figure 7b) in the gal carbohydrate can attenuate the production of anti-gal gal+ BMT animals. There was no change in total amounts IgM antibodies, even after challenge with porcine

Gene Therapy Inhibition of xenoantibody production using lentiviral vectors JY Fischer-Lougheed et al 53

Figure 6 Chimerism in SIV-gal+ BMT animals. (a) PBMC lineages Figure 5 Cell-surface gal carbohydrate expression in PBMC expressing gal in SIV-gal+ BM recipients. Cell-surface gal expression lineages in gal+ BMT monkeys. Chimerism was examined by (lectin IB4 binding) at 160 days post-transplantation, as compared to dual-color flow cytometry following transplantation with autolo- a control animal is shown here. (b) Flow cytometry acquisition dot- gous, transduced BM. The levels of gal expression after background plots showing lectin IB4 binding CD11b+ and CD20+ PBMC levels from control BMT animals were subtracted are shown here. populations at the same time point in two SIV-gal+ BM recipients Symbols are: CD3 (full circles), CD11b (open squares), CD20 (full compared with a control animal. The percentages indicated in the triangles), CD14 (open circles) and CD56 (open triangles). Note: upper right quadrants were calculated excluding background CD14 and CD56 lineages are not shown for gal+ BMT no. 240, as binding. (c) Gal+ CD11b+ cells were still present in the spleen at these were not examined at the time this recipient was analyzed. day 335 post -transplantation with SIV-gal+ BM. hepatocytes. IgM antibodies to other porcine xenoanti- inhibitory effect on anti-gal-mediated marrow rejection.7 gens were produced at similar levels in the gal+ BM Our data in rhesus monkeys are consistent with the recipients as in controls. Specific tolerance to the gal concept that IgG1 anti-gal antibodies are not harmful to carbohydrate is considerably more difficult to achieve engraftment of gal+ cells. in monkeys compared with aGTÀ/À knockout mice Several types of vectors and conditions have been owing to the presence of memory cells and pre-existing assayed for gene transfer in non-human primate hema- xenoantibodies in non-human primates. High levels topoietic stem and progenitor cells.22–27 These studies of gal expression in the engrafted BM and low levels indicate that the maintenance of transgene expression of pre-existing cytotoxic xenoantibodies are minimal depends on the dose of irradiation, and hence the requirements for the success of the procedure. The space for engraftment, the repopulating capacity of transplantation of autologous, aGT-transduced cells the transduced progenitor cells and the immunogenicity induced an IgG1 antibody response in the gal+ recipients. of the transgene. In low-dose-irradiated rhesus These antibodies were not capable of complement- monkeys transplanted with high numbers of CD34+ cells dependent cytotoxicity to pig cells. In the aGTÀ/À transduced with a retrovirus expressing murine CD24, knockout mice, the observation that long-term chimer- transgene expression was detected in 10–15% of periph- ism could be achieved in the presence of high levels of eral blood leukocytes for over 4 months, and then anti-gal IgG1, predating transplantation of gal+ BM, led declined thereafter.28 These animals did not appear to the authors to suggest that IgG1 antibodies may have an develop an immune response to the product of the

Gene Therapy Inhibition of xenoantibody production using lentiviral vectors JY Fischer-Lougheed et al 54 function and safety. Testing in pre-clinical, non-human primate models is therefore an important pre-requisite for understanding the long-term consequences of gene therapy procedures in a clinically relevant model of transplantation. Our experimental procedure for transduction is novel in which we transplanted a pool of total BM transduced with low multiplicity of infection (MOI) and fractionated CD34+ cells transduced at high MOI (75–100). Transduc- tion at high MOI has been shown to over ride cellular factors that inhibit efficient transduction of rhesus cells with HIV-1-based vectors.12–14 High levels of chimerism were initially observed in all monkeys, then subse- quently dropped to low but detectable levels for a period of 1 year post-transplantation. This may have resulted from insufficient transgene integration into the pluripo- tent hematopoietic progenitor cells, vector silencing or lack of a selective advantage for the transduced cells. Other groups have studied CD34+ hematopoietic progenitor cell-mediated gene transfer using HIV-1- derived constructs to express enhanced green fluorescence protein (EGFP) in non-human primates, and have obtained sustained low transgene marking in recipient baboons,25 and rhesus macaques (0.5–8% EGFP in certain lineages up to 4 years post-transplantation; An et al.23 and Kung et al.27). Unlike the animals in those studies, our recipients were not subjected to myeloablative irradiation, which would enhance selection for transduced autologous marrow engraftment. Although tissue- specific tolerance has been achieved using lentiviral vectors following 10 Gy of irradiation, and following 500–650 cGy of irradiation using retroviral vectors, this paper represents the first demonstration that similar results can be achieved following 300–350 cGy of TBI. Given their potential as vehicles for gene therapy, lentiviral vectors are undergoing continuous remodeling to improve stable and safe transgene integration and Figure 7 Anti-gal antibody response is attenuated in Gal+ BMT expression in stem cells.33–37 We elected to use SIV- animals immunized with pig cells. (a) Anti-gal IgM antibody derived lentiviral vectors for this study based on reports response to immunization with porcine hepatocytes and endothelial + that they are more efficient for transduction of simian cells was considerably attenuated in gal BMT monkeys (n ¼ 5) 19,20,38–40 versus immunized controls (n ¼ 2), as shown here at 8 days post- stem cells. Our results using SIV-based aGT immunization. (b) Flow-cytometric cytoxicity assays did not show constructs are consistent with those findings. Under non- a significant increase in cytotoxic antibodies directed at pig cells myeloablative conditions, however, we were not able in the gal+ BM recipients collected 8 days after immunization. to achieve the levels of engraftment reported by Hanawa (c) ELISA performed on PAEC showing that the immunization et al.19 using an SIV-based EGFP-expressing vector + induced a similar production of anti-pig IgM antibodies in the gal (marking up to 14% of CD3+ and 17% CD20+ cells 7 BMT (n ¼ 5) and control BMT (n ¼ 2) animals. These experiments are representative of several independent experiments. All results are months after transfer). One significant difference is that expressed as the mean7s.e.m. their animals received larger doses of efficiently transduced CD34+ cells (7–20 Â 106 cells/kg) following myeloablative pre-conditioning. We elected to apply a non- myeloablative regimen for our studies, as this procedure introduced gene.29 The gene transfer efficiency we report is more appropriate for clinical applications in humans. here with lentivectors is similar. Lentivectors, however, Our group has recently shown that lentiviral vectors can offer several advantages as gene transfer vehicles. They be used to achieve chimerism and tolerance following efficiently transduce quiescent progenitors and non- 300 cGy of irradiation in the aGTÀ/À knockout mouse dividing cells, thus reducing the time in ex vivo culture model for xenotransplantation.11 This paper extends this and the risk of losing multipotency. Recent studies also study to a clinically applicable non-human primate model suggest that lentivectors integrate more randomly in the in which we have demonstrated that the level and recipient’s genome when compared with retroviral kinetics of chimerism that can be achieved in monkeys vectors.30 A careful analysis of vector integrating sites using these vectors is strikingly similar to the levels of has become the focus of studies in numerous laboratories chimerism in mice that we have recently reported to be owing to the fact that preferential integration into sufficient for tolerance induction. This raises the exciting specific sites has imposed a safety concern in vectors possibility that this regimen may be useful in the designed for clinical use.31,32 It is therefore important that development of procedures to induce transplantation vectors are designed and selected for both optimal tolerance in large animal models in the future.

Gene Therapy Inhibition of xenoantibody production using lentiviral vectors JY Fischer-Lougheed et al 55 Materials and methods into two doses over 24 h) or a single dose of 350 cGy on the day of transplantation (see Table 1). No additional Animals immunosuppressive treatment was given. Monkeys All animal procedures conformed to the requirements of nos. 240 and 796 were challenged at 3 and 5 months the Animal Welfare Act, and protocols were approved post-BMT with a subcutaneous injection of 60 million before implementation by the Institutional Animal Care hepatocytes and endothelial cells (kind gift from Ronald and Use Committee at the University of California, Faris at MultiCell Technologies Inc., Warwick, RI, USA). Davis. Sixteen juvenile rhesus monkeys (M. mulatta) Monkeys nos. 322, 296 and 318 were challenged with ranging in age from 7.5 to 23 months and weights of porcine cells at 2 and 3 months post-BMT, and monkeys 1.5–2.7 kg were screened for pre-existing anti-gal nos. 309 and 256 were challenged at 4 months post-BMT. xenoantibodies. Seven animals with matched levels of The porcine cell preparation used for immunization has xenoantibodies were selected for BMT for these studies been shown to induce a robust anti-pig antibody (five experimental and two controls). Their peripheral response in rhesus macaque controls.42 Serum from blood cells were gal negative, as demonstrated by flow these animals obtained pre- and post-immunization with cytometry with fluorescence isothiocyanate (FITC)-lectin an identical number of cells provided additional control IB4 (from Bandieraea simplicifolia, Sigma, St Louis, MO, samples for this study. USA) that binds to gal epitopes.16 Serum, blood and BM samples from four rhesus monkeys that did not receive Hematopoietic cell reconstitution and molecular a control or gal+ BM transplant provided additional chimerism controls for this study. BM and peripheral blood samples were collected Cell isolation, CD34+ cell enrichment and transduction regularly. Complete blood counts were analyzed each Autologous BM was harvested and mononuclear cells time a blood sample was collected. Hematopoietic were isolated by Ficoll–Histopaque (Sigma, USA) density cell reconstitution and gal expression were analyzed by centrifugation. CD34+ cells were enriched from half of dual-color flow cytometry, performed on a FACScalibur the BM, using mouse anti-human-CD34 antibody 12.8 cytometer using CellQuest software (Becton Dickinson, (kind gift from RG Andrews, Fred Hutchinson CRC, San Jose, CA, USA). Labels were FITC-lectin IB4 (Sigma, Seattle, WA, USA) and anti-mouse IgM-micromagnetic USA), and PE-conjugated mouse antibodies anti-CD3, anti-CD11b, anti-CD20, anti-CD14, anti-CD56 and anti- beads (Milltenyi Biotec, Auburn, CA, USA). Purity À was assessed by flow cytometry using R-phycoerythrin CD16 (BD Pharmingen, USA). Human AB,Rh serum (PE)-conjugated anti-CD34 antibody (BD Pharmingen, was used at 2% to block nonspecific binding to Fc San Diego, CA, USA). The remainder of the BM aspirate receptors. The values reported for detecting transduced (total BM) was transduced without CD34+ cell enrich- cells (chimerism) were determined conservatively by ment. For two animals (nos. 296 and 318), BM aspirates setting the gates to exclude nonspecific binding to control had been collected over several weeks before transplan- animals (this includes animals transplanted with BM tation and cryopreserved, in order to increase the transduced with the control vector). Background labeling amount of available cells for transfer. These cells were was determined using cells from control animals for each experiment and was subtracted from the lectin binding placed in culture 24 h before transduction. Cells were + seeded on RetroNectin-coated (Takara, Japan) plates and in gal BMT animals to determine the levels of chimer- transduced with lentivector aGT (pCSO-rre-cppt-MCU3- ism. Owing to the very conservative manner in which we have calculated chimerism, our quantitative assessment a1,3gal-WPRE) or a control lentivector (aGT in reverse + orientation).9 The cells were plated at 4 Â 106 cells/well of gal cells provided may be an under-representation of and transduced in 0.5 ml of X-VIVO15 (BioWhittaker, labeled cells in these animals. Walkersville, MD, USA) supplemented with 100 ng/ml each of human cytokines interleukin (IL)-3, stem cell qPCR analysis factor, Flt3-ligand and TPO (BioSource, Camarillo, CA, Real-time qPCR was used to detect the porcine aGT USA) and IL-6 (Stem Cell Technologies, Canada). The transgene in DNA isolated from transduced BM samples MOI ranged from 75 to 100 for CD34+ cells, and 8–15 kept 4–6 days in culture, as well as from BM and for total BM. Cells were exposed to the lentiviral vector peripheral blood samples of BMT recipients at various twice in 16 h. The same procedure was used to prepare time points post-transplantation. qPCR was performed and transduce the cells with a SIV-derived aGT using the ABI PRISM 7700 Sequence Detector and lentivector (pSIV-SIN-MNDU3-a1,3gal), pseudotyped analyzed using SDS1.9.1 software (Applied Biosystems with vesicular stomatitis virus-glycoprotein. The latter Inc., Foster City, CA, USA). For each sample, the data vector was constructed at the CHLA vector core using an were normalized to the signal obtained for the rhesus SIV-parent vector derived from SIVmac251 and obtained housekeeping glyceraldehyde-3-phosphate dehydrogen- from FL Cosset (LVRTG, INSERM U412, Ecole Normale ase (GAPDH) gene (ratio method as detailed in the Supe´rieure de Lyon, France).38,41 The MOI range for SIV Quantitect SYBR Green Handbook). A control standard vector transduction was 3 for total BM and 15–25 for the curve for GAPDH was generated using serial dilutions CD34+ fractions. (copies/ml) of a control GAPDH DNA. Briefly, a GAPDH gene fragment was cloned into the pCR2.1 cloning vector BMT and immunization (Invitrogen, Carlsbad, CA, USA) and linearized, the Twenty-four hours after BM aspiration, the transduced DNA concentration was determined by spectophotome- cells (CD34+ cells and total BM) were pooled, and try, confirmed by gel electrophoresis and the number of injected intravenously into recipients that had previously copies for the standard was calculated as described in the received submyeloablative irradiation of a 300 cGy (split SYBR Green Handbook protocols (Qiagen, Valencia, CA,

Gene Therapy Inhibition of xenoantibody production using lentiviral vectors JY Fischer-Lougheed et al 56 USA). Samples were run in duplicate and included a AI52079-02 (to MKJ) and RR00169 to the California negative control (deionized water), a standard curve for National Primate Research Center. GAPDH and monkey cell DNA prepared from the peripheral blood and BM taken at various times after BMT. Both GAPDH and aGT primers were used to References amplify DNA within each of the experimental samples. The oligonucleotide primers (Operon Biotechnologies 1 Sandrin MS, Vaughan HA, Dabkowski PL, McKenzie IF. Inc., Alameda, CA, USA) used for the qPCR assay were: Anti-pig IgM antibodies in human serum react predominantly GAPDH reverse, 50-TGCACCACCAACTGCTTAGC-30 with Gala(1–3)Gal epitopes. Proc Natl Acad Sci USA 1993; 30: and forward, 50-GGCGTGGACTGTGGTCATGAG-30; 11391–11395. porcine aGT reverse, 50-CCAAAACACAACCATTACAG 2 Galili U, Shohet SB, Kobrin E, Stults CL, Macher BA. Man, apes, TTGAG-30 and forward (within the MCU3 region of the and Old World monkeys differ from other mammals in the lentivector), 50-GTTCGCTTCTCGCTTCTGTT-30. Quanti- expression of alpha-galactosyl epitopes on nucleated cells. J Biol tative amplifications were performed using 1 mg of DNA Chem 1988; 263: 17755–17762. and the Quantitect SYBR Green PCR kit (Qiagen, USA) 3 Tearle RG, Tange MJ, Zannettino ZL, Katerelos M, Shinkel TA, Van Denderen BJ et al. The alpha-1,3-galactosyltransferase according to the manufacturer’s protocol. The PCR knockout mouse. Implications for xenotransplantation. Transplanta- conditions were: 951C 15 min; 45 cycles (951C15s;561C 1 tion 1996; 61: 13–19. 30 s; 72 C 30 s). Data are presented as relative amounts, 4 Thall A, Murphy H, Lowe JB. Alpha1,3-galatosyltransferase standardized to GAPDH. deficient mice produce cytotoxic natural antibodies. Transplant Proc 1996; 28: 561–562. Enzyme-linked immunosorbant assays 5 LaTemple DC, Galili U. Adult and neonatal anti-gal response in Anti-gal IgM and IgG serum levels were measured by knock-out mice for a-1,3-galactosyltransferase. Xenotransplanta- ELISA, as described previously,9 using purified gal tion 1998; 5: 191–196. carbohydrates in tri- and pentasaccharide forms (V-Labs, 6 Yang YG, deGoma E, Ohdan H, Bracy JL, Xu Y, Iacomini J et al. Covington, LA, USA) or PAECs as targets. Peroxidase- Tolerization of anti-gala1–3gal natural antibody-forming B conjugated anti-human IgM and IgG antibodies were cells by induction of mixed chimerism. J Exp Med 1998; 187: obtained from Jackson ImmunoResearch Labs and 1335–1342. 7 Ohdan H, Swenson KG, Kitamura H, Yang YG, Sykes M. Sigma, respectively. IgG subtyping was performed using Tolerization of Gala1,3Gal-reactive B cells in pre-sensitized a1,3- peroxidase-conjugated anti-human IgG antibodies (The 43 galactosyltransferase-deficient mice by nonmyeloablative induc- Binding Site Inc., San Diego, CA, USA). Total IgM and tion of mixed chimerism. Xenotransplantation 2001; 8: 227–238. IgG serum levels were measured using human IgM 8 Bracy JL, Chase CM, Russell PS, Mauiyyedi S, Colvin RB, and IgG ELISA quantitation kits (Bethyl Laboratories, Iacomini J. Induction of molecular chimerism by gene therapy Montgomery, TX, USA), according to the manufacturer’s prevents antibody-mediated heart transplant rejection. Gene instructions. The peroxidase substrate was SureBlue Therapy 2001; 8: 1738–1744. TMB (KPL, Gaithersburg, MD, USA). 9 Kearns-Jonker M, Fischer-Lougheed J, Shulkin I, Kleihauer A, Mitsuhashi N, Kohn DB et al. Use of lentiviral vectors to induce Antibody-mediated cytotoxicity long-term tolerance to gal+ heart grafts. Transplantation 2004; 77: Levels of xenoantibodies capable of complement-depen- 1748–1754. dent cytotoxicity were measured by a flow-cytometric 10 Ogawa H, Yin DP, Galili U. Induction of immune tolerance to toxicity assay.44 Minipig kidney cells (ATCC, Manasses, a transplantation carbohydrate antigen by gene therapy with autologous lymphocytes transduced with adenovirus containing VA, USA) or PAEC were incubated at room temperature the corresponding glycosyltransferase gene. Gene Therapy 2004; 45–60 min with heat-inactivated serum samples diluted 11: 292–301. 10-fold, and then 60 min with Low-Tox M Rabbit 11 Mitsuhashi N, Fischer-Lougheed J, Shulkin I, Kleihauer A, Kohn Complement (Cedarlane, Hornby, ON, Canada). Cell DB, Weinberg K et al. Tolerance induction by lentiviral gene death was evaluated by propidium iodide uptake, therapy with a non-myeloablative regimen. Blood 2006; 107: measured using a FACSCalibur cytometer (Becton 2286–2293. Dickinson, USA). 12 Owens CM, Yang PC, Go¨ttlinger H, Sodroski J. Human and simian immunodeficiency virus capsid proteins are major viral Statistical analysis determinants of early, postentry replication blocks in simian The results were analyzed by Student’s t-test. Po0.0507 cells. J Virol 2003; 77: 726–731. was assessed as significant. All data are shown as 13 Kootstra NA, Munk C, Tonnu N, Landau NR, Verma IM. mean7s.e. Abrogation of postentry restriction of HIV-1-based lentiviral vector transduction in simian cells. Proc Natl Acad Sci USA 2003; 100: 1298–1303. 14 Stremlau M, Owens CM, Perron MJ, Kiessling M, Autissier P, Acknowledgements Sodroski J. The cytoplasmic body component TRIM5alpha restricts HIV-1 infection in Old World Monkeys. Nature 2004; We thank Karen Pepper and Denise Petersen of the 427: 848–853. CHLA vector core facility for preparing the vectors, 15 Strokan V, Molne J, Svalander CT, Breimer ME. Heterogeneous Ronald Faris at MultiCell Technologies Inc., Warwick, expression of Gal alpha1-3Gal xenoantigen in pig kidney: a RI, USA for providing the porcine hepatocytes used lectin and immunogold electron microscopic study. Transplanta- for immunization and the staff of the University of tion 1998; 66: 1495–1503. California National Primate Research Center for assis- 16 Murphy LA, Goldstein IJ. Five alpha-D-galactopyranosyl-bind- tance and care with the animals. This work was ing isolectins from Bandeiraea simplicifolia seeds. J Biol Chem 1977; supported by NIH Grants 1R21 AI49922-01 and 7RO1 252: 4739–4742.

Gene Therapy Inhibition of xenoantibody production using lentiviral vectors JY Fischer-Lougheed et al 57 17 Laukkanen MO, Kuramoto K, Calmels B, Takatoku M, von Kalle primate hematopoietic stem and progenitor cells. PLOS Biol C, Donahue RE et al. Low-dose total body irradiation causes 2004; 2: e423. clonal fluctuation of primate hematopoietic stem and progenitor 31 Engel BC, Kohn DB, Podsakoff GM. Update on gene therapy cells. Blood 2005; 105: 1010–1015. of inherited immune deficiencies. Curr Opin Mol Ther 2003; 5: 18 Goud SN. Effects of sublethal radiation on bone marrow cells: 503–507. induction of apoptosis and inhibition of antibody formation. 32 Montini E, Cesana D, Schmidt M, Sanvito F, Ponzoni M, Toxicology 1999; 135: 69–76. Bartholomae C et al. Hematopoietic stem cell gene transfer in a 19 Hanawa H, Hematti P, Keyvanfar K, Metzger ME, Krouse A, tumor-prone mouse model uncovers low genotoxicity of Donahue RE et al. Efficient gene transfer into rhesus repopulat- lentiviral vector integration. Nat Biotech 2006; 24: 687–696. ing hematopoietic stem cells using a simian immunodeficiency 33 Yam PY, Li S, Wu J, Hu J, Zaia JA, Yee J-K. Design of HIV vectors virus-based lentiviral vector system. Blood 2004; 103: 4062–4069. for efficient gene delivery into human hematopoietic cells. Mol 20 Yoshioka T, Ageyama N, Shibata H, Yasu T, Misawa Y, Takeuchi Ther 2002; 5: 479–484. K et al. Repair of infarcted myocardium mediated by trans- 34 Sandrin V, Boson B, Salmon P, Gay W, Ne`gre D, Le Grand R et al. planted bone marrow-derived CD34+ stem cells in a nonhuman Lentiviral vectors pseudotyped with a modified RD114 envelope primate model. Stem Cells 2005; 23: 355–364. glycoprotein show increased stability in sera and augmented 21 Kawauchi M, Nakajima J, Endoh M, Oka T, Takamoto S. transduction of primary lymphocytes and CD34+ cells derived Ontogeny of anti-pig xenoantibody and hyperacute rejection. from human and nonhuman primates. Blood 2002; 100: 823–832. Transplantation 2000; 70: 686–688. 35 Hanawa H, Kelly PF, Nathwani AC, Persons DA, Vandergriff JA, 22 Donahue RE, Dunbar CE. Update on the use of nonhuman Hargrove P et al. Comparison of various envelope proteins for primate models for preclinical testing of gene therapy their ability to pseudotype lentiviral vectors and transduce approaches targeting hematopoietic cells. Hum Gene Ther 2001; primitve hematopoietic cells from human blood. Mol Ther 2002; 12: 607–617. 5: 242–251. 23 An DS, Kung SK, Bonifacino A, Wersto RP, Metzger ME, 36 Logan AC, Lutzko C, Kohn DB. Advances in lentiviral vector Agricola BA et al. Lentivirus vector-mediated hematopoietic design for gene-modification of hematopoietic stem cells, review. stem cell gene transfer of common gamma-chain cytokine Curr Opin Biotechnol 2002; 13: 429–436. receptor in rhesus macaques. J Virol 2001; 75: 3547–3555. 37 Delenda C. Lentiviral vectors: optimization of packaging, 24 Hanazono Y, Terao K, Shibata H, Nagashima T, Ageyama N, transduction and gene expression. J Gene Med 2004; 6: S125–S138. Asano T et al. Introduction of the green fluorescent protein gene 38 Ne`gre D, Cosset FL. Vectors derived from simian immunodefi- into hematopoietic stem cells results in prolonged discrepancy of ciency virus (SIV). Biochimie 2002; 84: 1161–1171. in vivo transduction levels between bone marrow progenitors 39 Hanazono Y, Asano T, Ueda Y, Ozawa K. Genetic manipulation and peripheral blood cells in nonhuman primates. J Gene Med of primate embryonic and hematopoietic stem cells with simian 2002; 4: 470–477. lentivirus vectors. Trends Cardiovasc Med 2003; 13: 106–110. 25 Horn PA, Morris JC, Bukovsky AA, Andrews RG, Naldini L, 40 Nagata M, Takahashi M, Muramatsu S, Ueda Y, Hanazono Y, Kurre P et al. Lentivirus-mediated gene transfer into hema- Takeuchi K et al. Efficient gene transfer of a simian immuno- topoietic repopulating cells in baboons. Gene Therapy 2002; 9: deficiency viral vector into cardiomyocytes derived from 1464–1471. primate embryonic stem cells. J Gene Med 2003; 5: 921–928. 26 Hu J, Kelly P, Bonifacino A, Agricola B, Donahue R, Vanin E et al. 41 Mangeot PE, Negre D, Dubois B, Winter AJ, Leissner P, Direct comparison of RD114-pseudotyped versus amphotropic- Mehtali M et al. Development of minimal lentivirus vectors pseudotyped retroviral vectors for transduction of rhesus derived from simian immunodeficiency virus (SIVmac251) and macaque long-term repopulating cells. Mol Ther 2003; 8: 611–617. their use for gene transfer into human dendritic cells. J Virol 27 Kung SK, An DS, Bonifacino A, Metzger ME, Ringpis GE, Mao 2000; 74: 8307–8315. SH et al. Induction of transgene-specific immunological toler- 42 Zahorsky-Reeves J, Gregory C, Cramer DV, Patanwala I, Kyles ance in myeloablated nonhuman primates using lentivirally A, Borie D et al. Similarities in the immunoglobulin response and + transduced CD34 progenitor cell. Mol Ther 2003; 8: 981–991. VH gene usage in rhesus monkeys and humans exposed to 28 Rosenzweig M, MacVittie TJ, Harper D, Hempel D, Glickman porcine hepatocytes. BMC Immunol 2006; 7:3. RL, Johnson RP et al. Efficient and durable gene marking of 43 Baquerizo A, Mhoyan A, Kearns-Jonker M, Arnaout WS, hematopoietic progenitor cells in nonhuman primates after Shackleton C, Busuttil RW et al. Characterization of human nonablative conditioning. Blood 1999; 94: 2271–2286. xenoreactive antibodies in liver failure patients exposed to 29 Rosenzweig M, Connole M, Glickman R, Yue S-P S, Noren B, pig hepatocytes after bioartificial liver treatment: an ex vivo DeMariaet MA et al. Induction of cytotoxic T lymphocyte and model of pig to human xenotransplantation. Transplantation antibody responses to enhanced green fluorescence protein 1999; 67: 5–18. following transplantation of transduced CD34+ hematopoietic 44 Wetzsteon PJ, Head MA, Fletcher LM, Lye WC, Norman DJ. cells. Blood 2001; 97: 1951–1959. Cytotoxic flow-cytometric crossmatches (Flow-Tox): A compar- 30 Hematti P, Hong B-K, Ferguson C, Adler R, Hanawa H, Sellers S ison with conventional cytotoxicity crossmatch techniques. Hum et al. Distinct genomic integration of MLV and SIV vectors in Immunol 1992; 35: 93–99.

Gene Therapy