Class-, gene-, and group-specific HLA silencing by lentiviral shRNA delivery

J Mol Med (2006) 84: 425–437 DOI 10.1007/s00109-005-0024-2

ORIGINA L ARTI CLE

Constança Figueiredo . Axel Seltsam . Rainer Blasczyk

Class-, gene-, and group-specific HL A silencing by lentiviral shRNA delivery

Received: 9 June 2005 / Accepted: 2 November 2005 / Published online: 7 March 2006 # Springer-Verlag 2006

Abstract HLA incompatibility is the most relevant immunologic barrier to cell-based therapies. Improvement of histocompatibility is essential to achieving better survival of allogeneic cells in the foreign organism. RNA interference technology can be used to selectively and stably reduce cellular HLA class I expression. In the present study, we designed small interfering RNA (siRNA) molecules that target either β2-microglobulin (β2m) or HLA-A heavy chain transcripts and identified sensitive sites on the target RNAs using an in vitro transcription/ translation (IVTT) system. Transfection of siRNA into Blymphocyte cell lines (B-LCLs) resulted in specific reduction of HLA class I or HLA-A antigen expression by 79% at the mRNA and protein levels. An allele-specific HLA silencing rate of 65% was achieved in a B-LCL heterozygous for HLA-A*24,*68 allospecificities using HLA-A*68-specific siRNA. Lentiviral delivery of short hairpin RNA into HeLa and B-LCL cells resulted in selective and permanent silencing of HLA class I or HLAA by up to 90% even under inflammatory conditions. In cytotoxicity and proliferation assays, it was demonstrated that HLA class I knockdown was effective in preventing antibody-mediated cell lysis and CD8+ T cell response, while the residual HLA expression in HLA-silenced cells was protective against NK-cell-mediated lysis. The present data strongly suggest that silencing of HLA expression in a class-, gene-, and group-specific manner is an effective approach that may provide a new basis for developing new immunotherapies in the field of regenerative medicine. Keywords Lentiviral vector . RNAi . shRNA . Gene therapy . Immune tolerance . MHC Grant support: CF is supported by the Portuguese Foundation for Science and Technology (SFRH/9687/2002). C. Figueiredo . A. Seltsam . R. Blasczyk (*) Institute for Transfusion Medicine, Hanover Medical School, Carl-Neuberg-Str. 1, D-30625 Hanover, Germany e-mail: [email protected] Tel.: +49-511-5326700 Fax: +49-511-5322079

CONSTANC¸A FIGUEIREDO enrolled in a Ph.D. programme in Experimental Biology and Biomedicine at the University of Coimbra, Portugal. Since then, she has been working on her doctoral thesis in the Institute for Transfusion Medicine, Hanover Medical School, Germany. Her research interests focus on the development of new strategies to reduce cellular immunogenicity.

RAINER BLASCZYK received his M.D. from the Ruhr-University of Bochum, Germany. He is presently a University Professor of Transfusion Medicine and Head of the Institute for Transfusion Medicine at Hanover Medical School. His research interests include genomic diversity and its management and exploitation for regenerative and immunotherapeutic purposes.

Introduction The many applications for cellular therapy include cancer, autoimmune diseases, genetics disorders (hemophilia, thalassemia and chronic granulomatous disease), neurological disorders (Parkinson’s and Huntington’s disease), and repair of damaged organ tissues (cardiac, nervous, pancreatic and hepatic) [1]. Cells used in regenerative medicine can be derived from autologous or, more desirably, allogeneic sources. The advantages of allogeneic cells are their off-the-shelf availability and their ease of standardization, a vital prerequisite for the widespread use of stem cells in regenerative medicine [2–4]. However, histocom-

426

patibility is a major hurdle for the application of such products. Developing replacement tissues with little or no immunogenicity is therefore highly desirable. As HLA polymorphism is the most relevant immunologic barrier to organ transplantation, HLA incompatibilities will also be a major barrier to allogeneic-cell-based regenerative therapies [5–8]. Suppression of HLA expression could help to overcome this limitation by inducing immunologic tolerance and decreasing the risk of rejection, thus supporting the regenerative process. Even in gene therapy, the use of autologous cells modified to express foreign genes may result in an allogeneic profile due to the presentation of immunogenic peptides derived from transgene expression, which may ultimately lead to destruction of the modified cells [9, 10]. As is known from classical grafting applications such as allogeneic hematopoietic stem cell transplantation (HSCT) and solid organ transplantation, one major obstacle to therapeutic success is acute or chronic rejection, a process whereby complement-mediated lysis and T cell responses cause a loss of graft function [11–15]. The donor microcirculation is the principal target of action of complement, neutrophils, natural killer (NK) cells and macrophages. These effector systems are recruited by HLA class I and II alloantibodies, usually immunoglobulin-G (IgG) arising from earlier transplants, transfusions, and pregnancies, or produced during the rejection process by plasma cells after T cell stimulation. Anti-HLA class I antibodies are involved in acute rejection, whereas anti-HLA class II antibodies are of major importance in late rejection. HLAA, HLA-B, and HLA-DR matching is therefore essential to reducing the number of T and B cell determinants [16–18]. RNA interference (RNAi) has recently emerged as a prevailing genetic tool for silencing gene expression [19, 20]. The RNAi pathway triggers post-transcriptional degradation of homologous transcripts through a multistep mechanism involving double-stranded small interfering RNA (siRNA) [21]. Effective gene silencing has been achieved by transfecting siRNA and by using short hairpin RNA (shRNA) expression cassettes driven by Pol III promoters, such as U6 and H1 [22–24]. This report demonstrates that it is possible to introduce immunological unresponsiveness in regenerative medicine by selective and prolonged suppression of HLA or HLA subgroups. In this paper, we describe the employment of RNAi to silence HLA expression in a class-, gene-, and group-specific manner using synthetic siRNAs and lentiviral vectors encoding shRNAs. A cell-free method for rapid identification of effective siRNA sequences is also presented.

Study design and methods siRNA design The siRNA sequences for silencing HLA and β2-microglobulin (β2m) expression were designed either manually

considering the sequence variations of the HLA-A heavy chain or with the help of web-based algorithms (http:// www.qiagen.com and http://www.ambion.com). In vitro transcription/translation system HLA-A*2601 heavy chain (which belongs to the serological group HLA-A10) and β2m coding sequences, both followed by a V5/His tag coding sequence, were cloned separately into pTNT (Promega, Madison, WI), the expression vector containing a T7 promoter. The efficiency of the siRNAs was tested in a wheat germ extract in vitro transcription/translation (IVTT) system (Promega). For HLA-A-specific siRNA testing, 1.2 μg of siRNA targeting the HLA-A*2601 heavy chain and 650 ng of pTNT encoding HLA-A*2601 were added to each IVTT reaction and, for β2m-specific siRNA testing, 1 μg of synthetic siRNA and 500 ng of pTNT encoding β2m were used. Nonsense siRNAs were used as controls. After incubation of samples for 1 h at 30°C, expression of β2m and HLA-A heavy chains was analyzed by ELISA. ELISA for siRNA selection Briefly, mouse monoclonal antibodies recognizing the V5 tag (Serotec, Kindlington, UK) or the heavy chain of HLA class I molecules (Acris, Hiddenhausen, Germany) were used to immobilize β2m and HLA-A heavy chain, respectively, on a MaxiSorp plate (Nunc, Wiesbaden, Germany) at a concentration of 1 μg/μl. Twenty microliters of IVTT reaction product diluted with 150 μl of 5% milk was added to each well and incubated for 2 h. An HRPlabeled anti-β2m monoclonal antibody (DAKO, Carpinteria, CA) or an anti-His polyclonal antibody (Abcam, Cambridge, UK) was used for the detection of β2m or HLA-A heavy chain proteins. Cell culture Epstein–Barr-virus-immortalized B-lymphocyte cell lines (B-LCLs) from healthy blood donors, LCL721.221, HeLa, and K562 cells were cultured in RPMI-1640 medium (BioWhittaker/Cambrex, Hess. Oldendorf, Germany) supplemented with 2 mM glutamine, 10% (vol/vol) FCS (Bio Whittaker), 100 U/ml penicillin (C.C. Pro, Neustadt, Germany) and 100 μg/ml streptomycin (C.C. Pro) (complete medium). Human embryonic kidney (HEK) 293FT cells were grown in complete DMEM (Cambrex, Verviers, Belgium) supplemented with 0.1 mM nonessential amino acids (Gibco, Paisley, Scotland) and 500 μg/ml geneticin (Gibco). Primary T cells were cultured in X-Vivo (Cambrex) supplemented with 15% human serum (C.C. Pro). Blood for immortalization of B lymphocytes was obtained with informed consent as approved by the local ethics committee of the Hanover Medical School.

427 Table 1 siRNA/shRNA sequences targeting HLA-A heavy chain transcripts siRNA/shRNA

Locationa

siRNA/shRNA duplexb

siAhc01

259–279 (exon 2)

5' -CACACGGAAUGUGAAGGCCTT-3' 3' -TTGUGUGCCUUACACUUCCGG-5'

HLA-A group-specific A10/28

siAhc02

1003–1023 (exon 5–6)

5' -GAGCUCAGAUAGAAAAGGATT-3' 3' -TTCUCGAGUCUAUCUUUUCCU-5'

HLA-A gene-specific

siAhc03

1017–1037 (exon 6)

5' -AAGGAGGGAGCUACUCUCATT-3' 3' -TTUUCCUCCCUCGAUGAGAGU-5'

HLA-A gene-specific

siAhc04

836–856 (exon 4)

5' -AGAUACACCUGCCAUGUGCTT-3' 3' -TCUCUAUGUGGACGGUACACG-5'

HLA-A gene-specific

shAhc04

836–856 (exon 4)

5' -AGAUACACCUGCCAUGUGCUU 3' - UUUCUAUGUGGACGGUACACGAA

Specificity

C G A

HLA-A gene-specific

A a

The locations of the double-stranded small interfering RNA (siRNA) or short hairpin RNA (shRNA) sequences within the coding sequence of the HLA-A heavy chain are given b Nucleotide positions of siAh01 responsible for the HLA-A10/28 group-related alleles selectivity are indicated in bold

Transfection of synthetic siRNA into cultured cells 4

Twenty-four hours before transfection, 4×10 B-LCL 5 (ID1800083) and 1×10 HeLa cells were seeded in a 24well plate. One microgram of target-specific synthetic

siRNA (Tables 1 and 2) or nonsense siRNA as control were transfected by lipofection (RNAiFect Transfection Reagent Qiagen, Hilden, Germany) according to the manufacturer's instructions. After 3 h of incubation, transfection efficiency was evaluated by fluorescence microscopy and flow

Table 2 siRNA/shRNA sequences targeting β2-microglobulin transcripts siRNA/shRNA

Locationa

siRNA/shRNA duplex

siβ 2m01

260–280

5' -AAAGUGGAGCAUUCAGACUTT-3' 3' -TTUUUCACCUCGUAAGUCUGA-5'

siβ 2m02

76–96

5' -GAUGAGUAUGCCUGCCGUGTT-3' 3' -TTCUACUCAUACGGACGGCAC-5'

siβ 2m03

155–175

5' -UCCAUCCGACAUUGAAGUUTT-3' 3' -TTAGGUAGGCUGUAACUUCAA-5'

siβ 2m04

427–447

5' -UUCGAAGCUUGAAGGUAATT-3' 3' -TTAAGCUUCGAACUUCCAUU-5'

siβ 2m05

82–102

5' -GGUUUACUCACGUCAUCCATT-3' 3' -GTCCAAAUGAGUGCAGUAGGU-5'

siβ 2m06

400–420

5' -GUGGGAUCGAGACAUGUAATT-3' 3' -TTCACCCUAGCUCUGUACAUU-5'

shβ 2m06

400–420

5' -GUGGGAUCGAGACAUGUAATT 3' - UUCACCCUAGCUCUGUACAUUAA

C G A A a

The locations of the double-stranded small interfering RNA (siRNA) or short hairpin RNA (shRNA) sequences within the coding sequence of β2-microglobulin are given

428

cytometry using FITC-labeled nonsense siRNA duplex (Qiagen). Transfection was repeated every 48 h. The cells were incubated for a maximum of 6 days. Lentiviral constructs A lentiviral vector was used for stable shRNA expression. Short hairpin RNA expression cassettes (Tables 1 and 2) were cloned in the pENTR/U6 entry vector (BLOCK-iT Lentiviral RNAi Expression System, Invitrogen, Karlsruhe, Germany) and transferred into the lentiviral destination vector (pLenti6/BLOCK-iT-DEST) using Gateway Technology. To create a HLA-A11 single-antigen expressing K562 cell line (K562-A11), the coding sequence of HLA-A*1101 was cloned in the lentiviral construct pRRL-cPPT-PGK-GFP-W (kindly provided by D. Trono, Lausanne, Switzerland).

antibodies with HLA-A9 (serological group of HLA-A*23 and HLA-A*24 alleles) or HLA-A28 (serological group of HLA-A*68 and HLA-A*69 alleles) specificities (One Lambda, Canoga Park, CA) to detect HLA-A allelespecific expression. Anti-HLA-A, -B, and -C (Serotec) or anti-β2m (DAKO) FITC-labeled antibodies were used to evaluate HLA class I expression. HLA-A11 expression on K562-A11 cells was evaluated by staining with biotinlabeled monoclonal antibody with HLA-A11 specificity (One Lambda) and streptavidin-PE (Molecular Probes, Leiden, The Netherlands). Heavy chain or β2m levels were measured in a FACScan flow cytometer (BD Bioscience Imunocytometry Systems, San Jose, CA), and the data were processed using CellQuest software (BD Bioscience Imunocytometry Systems). Western blot Cell extracts were prepared in a lysis buffer (50 nM HEPES, 420 mM KCl, 1 mM EDTA, and 0.1% NP-40).

Flow cytometric analysis Cells transfected with siRNA or transduced for shRNA production were stained with FITC-labeled monoclonal

60 40

b

siNS

siAhc04

siAhc03

mock

siAhc01

0

siAhc02

20

beta2-microglobulin silencing 100 80 60 40 20

siNS

siβ2m06

siβ2m05

siβ2m04

siβ2m03

0

siβ2m02

Cells were resuspended at 4×10 per milliliter of virus supernatant and plated in a six-well plate. A mock transduction in an equivalent volume of medium was used as the control. The HeLa cells were repetitively transduced by spinoculation at 800×g for 2 h at 32°C in the presence of 6 μg/ml Polybrene (Sigma-Aldrich, Steinheim, Germany) and incubated overnight at 37°C. On the following day, the cells were washed and cultured in fresh culture medium. HLA expression was evaluated by flow cytometry on days 6, 15, 30, and 90. Lamin A/C expression was assessed by Western blot on days 6, 30, and 90.

80

siβ2m01

5

HLA-A*2601 silencing 100

mock

Transduction

a % of production

For HLA silencing, lentiviral particles were produced by transfecting 6 ×106 HEK293FT with 9 μg plasmid mix (pLP1 containing a gag/pol fusion protein, pLP2 encoding the viral reverse transcriptase, and pLP/VSVG encoding the VSV-G protein) and 3 μg pLenti6 encoding the specific shRNAs. The Lenti6-GW/U6-laminshRNA vector (Invitrogen) encoding shRNA, which targets lamin A/C, and vectors encoding nonsense shRNAs were used as controls. For expression of HLA-A*1101 in K562 cells, 10 μg of pRRL-cPPT-PGK-GFP-W encoding HLA-A*1101, 9 μg of the packaging plasmid PAX2, and 3 μg of the envelope plasmid pMD2G-VSVG were used to co-transfect 6×106 HEK293FT producer cells. After 16 h, the cells were washed and incubated with complete DMEM supplemented with 1% sodium pyruvate (C.C. Pro). Virus supernatants were collected 48 h after transfection.

% of production

Lentivirus production

Fig. 1 Screening of siRNA efficiency using an in vitro transcription/translation (IVTT) system. a HLA-A*2601 heavy chain (belonging to serological group HLA-A10) silencing was evaluated by IVTT using three HLA-A gene-specific small interfering RNAs (siRNA) (siAhc02, siAhc03, and siAhc04) and one HLA-A10/28 group-specific siRNA (siAhc01). b β2-microglobulin (β2m) silencing was assessed using six siRNAs specific for β2m (siβ2m01 to siβ2m06). In both assays, the negative control experiments were performed without siRNAs (mock) and with nonsense (NS) siRNAs. Expression plamids (pTNT) containing the β2m and HLA-A coding sequence served as templates for transcription. The percentage of relative target production in the siRNA-mediated inhibition reactions refers to the production in the mock control

429

The protein samples were separated on sodium dodecyl sulfate 10–20% gradient polyacrylamide gels (Invitrogen), electroblotted to polyvinylidene difluoride membranes (Invitrogen), and exposed to antibodies against lamin A/ C (BD Biosciences) or actin (Sigma). Secondary antibodies conjugated with horseradish peroxidase (DAKO) and Roti–Lumin chemiluminescence substrate (ROTH, Karlsruhe, Germany) were used for detection.

AGATACACCTGCCATGTG-3'), respectively. One hundred nanograms of RNA volumes were used in each reaction. The thermal cycling conditions used on an ABI Prism 7700 (Applied Biosystems) were 50°C for 15 min and 95°C for 10 min followed by 40 cycles at 95°C for 15 s and 60°C for 1 min. The constitutively expressed β-actin gene was amplified as the reference standard for normalization of mRNA levels.

Real-time RT-PCR

Complement-dependent cytotoxicity assay

Total cellular RNA was isolated from HeLa and B-LCLs (Rneasy Mini Kit, Qiagen, Hilden, Germany) 48 h after siRNA transfection or 3 months after transduction. Primers were designed to amplify transcripts of β2m (5'GCTGTGCTCGCGCTACTC-3' and 5'-ATGAAACCCA GACACATAGC-3') or HLA-A heavy chain (5'-CTCG TGGAGACCAGGCCT-3' and 5'-CTCAGGGTGAGGG GCTTG-3'). One-step real-time RT-PCR (RT-PCR Master Mix, Applied Biosystems, Darmstadt, Germany) was performed using MGB-TaqMan probes for β2m (5'-AGCA GAGAATGGAAAGT-3') and HLA-A heavy chain (5'-

A microcytotoxicity test (MCT) was performed on nontransduced or shRNA transduced HeLa and B-LCL cells (both typed as homozygous for HLA-A28) according to the standard lymphocytotoxicity test (LCT) protocol for serological HLA typing and cross-matching. An anti-HLAA2,28 IgM monoclonal (One Lambda) or two local antisera (H-561-1 and RK-858) specific for HLA-A28 were used. Briefly, 2×103 cells were incubated with 1 μl of each antiserum in a microtest chamber for 30 min at room temperature. Then, 6 μl of rabbit complement (Biotest AG, Dreieich, Germany) were added and incubated for 60 min at

Fig. 2 Gene- and class-I-specific HLA silencing mediated by siRNAs. Synthetic small interfering RNAs (siRNAs) specific for the HLA-A heavy chain (HLA-gene-specific silencing) or β2-microglobulin (HLA-class-I-specific silencing) were used for transfection in HLA-A28 homozygous B-lymphoblastoid cell lines (B-LCL; ID: 1800083). Relative quantification of HLA-A heavy chain (a) or β2m (b) mRNA transcripts were performed by real-time PCR (RTPCR). The target mRNA expression was normalized to β-actin expression in the cells. c Flow cytometry was performed to analyze

HLA-A expression in B-LCL after 6 days of culture using an FITClabeled HLA-A2/28-specific antibody; the results for non-transfected B-LCL (blue) and for B-LCL transfected with siAhc02 (red), siAhc04 (pink) or nonsense siRNA (green) are shown. d Flow cytometry analysis of β2m surface expression in non-transfected BLCL (blue) and in BCL transfected with β2m03 siRNA (pink), β2m05 siRNA (red), or nonsense siRNA (green) using an anti-β2m specific monoclonal antibody. siNS, nonsense siRNA

430 Fig. 3 Group-specific HLA-A silencing. A small interfering RNA (siRNA) specific for the sequences of HLA-10/28 grouprelated alleles was used for transfection in B-LCL cells heterozygous for HLA-A24,28. Flow cytometry analysis of HLA-A28 (a) or HLA-A24 (b) expression on non-transfected B-LCLs (blue) and on B-LCLs transfected with siAhc01 (red) or nonsense siRNA (green)

room temperature. Three microliters of ethidium bromide plus acridine orange solution was used to detect the number of lysed cells under a fluorescence microscope. The percentage of lysed cells was scored as previously described [25].

The cells were then analyzed by FACS to detect the level of HLA class I antigen expression. NK cell cytotoxicity assay

Stimulation assay with interferon-γ Native or shRNA-transduced HeLa and B-LCL cells were stimulated for 48 h with recombinant human IFN-γ (R&D Systems, Mineapolis, MN) at a concentration of 50 ng/ml.

Fig. 4 Lentiviral-mediated HLA silencing. Flow cytometry results obtained with HeLa (a and b) and B-LCL (c and d) cells expressing short-hairpin RNAs (shRNAs) specific for the HLA-A heavy chain (a and c) or β2-microglobulin (β2m) (b and d) are indicated in red curves. Cells in (a) and (c) were stained with a FITC-labeled monoclonal antibody having HLA-A28 specificity to detect HLA-A allele-specific expression. In (b) and (d), an anti-β2m FITClabeled antibody was used to evaluate β2m expression. Nontransduced cells (blue curve) and those transduced with vectors encoding nonsense shRNAs (orange curve) or with shRNAs specific for lamin A/C (green curve) were used as positive and negative controls, respectively. e and f In Western blot analysis, β-actin levels did not change in HeLa and B-LCL cells expressing the different shRNAs. HeLa/B-LCL Native cells, shAhc cells expressing shRNAs specific for the HLA-A heavy chain, shβ2m cells expressing shRNAs specific for β2m, shLamin A/C cells expressing shRNAs specific for lamin A/C, shNS cells expressing nonspecific shRNAs

Natural killer cells (NK cells) were isolated from peripheral blood of healthy individuals positive for HLA-A*03 or HLA-A*11 (human NK cell isolation kit II, Myltenyi Biotec, Bergisch Gladbach, Germany). Target cells were labeled with CFSE (Molecular Probes) as previously

431

described [26]. Native K562 cells, K562-A11 cells, and K562-A11-shβ2m06 cells which had been transduced with shβ2m06-encoding vectors were used as target cells and incubated in separate assays with freshly isolated NK cells or with pre-exposed NK cells. Pre-exposure was done by incubation with K562-A11 at a 1:1 E/T ratio in the presence of 50 U/ml IL-2 (Serotec) for 48 h. In an autologous setting, native B-LCLs and B-LCLs transduced with shβ2m06-encoding vectors were used as target cells and incubated with freshly isolated NK cells. In this LCL-assay, HLA-deficient LCL721.221 cells served as target cells for the positive control. In a final step, target cells were incubated with NK cells at a 50:1 E/T ratio in the presence of 50 U/ml IL-2 (Serotec) for 24 h to determine cytotoxicity. Target cell lysis was assessed by PI staining (Sigma).

a Cell Lysis Score

8

Results Evaluation of effective siRNAs To determine whether HLA expression can be inhibited in a gene- and group-specific manner, siRNAs were designed to target either conserved or polymorphic regions of the HLAA heavy chain. Three siRNAs were HLA-A gene-specific and one was specific for HLA-A alleles belonging to the A10 and A28 groups (Table 1). In addition, positions within the coding region of β2m were selected as targets for six siRNAs to induce HLA class I-specific silencing (Table 2). The potential of both siRNA sets was initially assessed by testing their efficiency to inhibit expression of

5 4 3 2

HeLa shNS HeLa shLamin A/C

HeLa shAhc

HeLa shβ2m

b 9

H-561-1 RK-858

8 7 6 5 4 3 2 1

Cell Lysis Score

B-LCL

B-LCL shNS

B-LCL shLamin A/C

B-LCL shAhc

HeLa

HeLa shNS

HeLa shLamin A/C

HeLa shAhc

B-LCL

B-LCL shNS

B-LCL shLamin A/C

B-LCL shAhc

B-LCL shβ2m

9 8 7 6 5 4 3 2 1

d Cell Lysis score

Allogeneic lymphocytes isolated from peripheral blood were pre-stimulated by incubating them with irradiated native B-LCLs at a ratio of 1:1 E /T the in presence of 50 U/ ml of IL-2 for 14 days. The CD8+ T cells were then isolated with CD8+ T cell isolation kit II (Miltenyi Biotec) and labeled with CFSE (Molecular Probes). Irradiated target cells (native B-LCLs or B-LCLs expressing a nonsense shRNA and shβ2m, respectively) were incubated with CFSE-labeled CD8+ T effector cells at a ratio of 1:1 E/T for 6 days. CFSE intensity as an indicator for T cell proliferation was measured by flow cytometry. Culture supernatants were harvested to measure IFN-γ secretion using ELISA technique (Cytimmune, Maryland, DC).

H-561-1 RK-858

6

HeLa

c T cell assays

7

1

Cell Lysis Score

Fig. 5 Complement-dependent microcytotoxicity assay. A comple- " ment-dependent microcytotoxicity assay was performed with HeLa and B-LCL cells (all homozygous for HLA-A28). Non-transduced cells and cells transduced with lentiviruses encoding nonsense (shNS), HLA-A heavy chain-specific (shAhc), lamin A/C-specific (shLamin A/C) or β2-microglobulin-specific (shβ2m) short-hairpin RNAs were stained with anti-HLA-A28 patient sera (H-561-1, RK858) (a and b) or a monoclonal IgM antibody having HLA-A2,28 specificity (c and d). The percentages of complement-lysed cells are scored on a scale of 1 to 8 (score 1, 0–10%; 2, 11–20%; 4, 21–40%; 6, 41–80%; 8, 81–100%). The mean values obtained from eight experiments are shown

HeLa shβ2m

9 8 7 6 5 4 3 2 1 0

B-LCL shβ2m

the target proteins in an IVTT system. Addition of the genespecific siRNAs (siAhc02, siAhc03, and siAhc04) to the IVTT reactions caused up to 60% reduction in HLAA*2601 production. The HLA-A10/28-specific (siAhc01) siRNA inhibited heavy chain translation by 39% (Fig. 1a).

432

In the β2m IVTT system, two of the six β2m-specific siRNAs reduced target protein expression by 60% (siβ2m03 and siβ2m05), whereas the others were incapable of reducing β2m production (Fig. 1b). The introduction of nonsense siRNAs into the IVTT reactions did not affect the levels of target protein production.

hibited the expression of the HLA-A28 molecule (Fig. 3b). The four mismatches between the siAhc01 target sequence and HLA-A24 appear to be responsible for the specificity of this siRNA (Table 1). Co-transfection experiments with two different siRNA duplexes did not reveal any additional effects on HLA-suppression (data not shown).

Transient HLA silencing

Stable HLA silencing

The siRNAs shown to be most effective in the IVTT system were used to treat B-LCLs and HeLa cells. Transfection of 1 μg HLA-A gene-specific siRNAs into an HLA-A28 homozygous B-LCL reduced mRNA expression by 72% (siAhc04) and 76% (siAhc02) (Fig. 2a) and suppressed HLA-A protein production by 59% (siAhc04) and 70% (siAhc02) (Fig. 2b). The β2m-specific siRNAs, siβ2m03 and siβ2m05, decreased β2m mRNA expression by 60% and 79%, respectively (Fig. 2c). At the protein level, transfection of these two siRNAs reduced β2m expression by 60% (siβ2m03) and 80% (siβ2m05), respectively (Fig. 2d). Transfection of the group-specific siRNA (siAhc01) into a B-LCL heterozygous for HLA-A28 and HLA-A24 (ID: 1800160) decreased HLA-A28 protein expression by 65% (Fig. 3a). However, no change was observed at the HLAA24 protein level, indicating that siAhc01 specifically in-

To overcome the transient effect of synthetic siRNAs, we constructed VSV-G pseudotyped lentiviruses to deliver HLA-A gene-specific or β2m-specific shRNAs into HeLa cells (typed as homozygous for HLA-A28). The siAhc04 and siβ2m05 siRNA sequences, which proved to be the most effective in the transient HLA silencing experiments, were used to design the short hairpin expression cassettes for lentiviral expression. Lentiviral constructs were transduced with a multiplicity of infection (MOI) of 3. HeLa cells stably expressing HLA-A-specific shRNA reduced HLA-A surface expression by up to 70% after 3 months in culture compared to the expression rate in non-transduced cells (Fig. 4a). HeLa cells transduced with lentiviruses encoding β2m-specific shRNAs suppressed β2m protein expression by 90% (Fig. 4b). B-LCL cells expressing HLA-A gene-specific shRNAs showed a reduction of

a HLA class I (MFI)

60 50 40

unstimulated

30

IFN-gamma stimulated

20 10 0

HeLa

HeLa shNS

HeLa HeLa shLamin shβ2m A/C

b 900

HLA class I (MFI)

Fig. 6 Silencing of HLA class I expression under inflammatory conditions. Non-transduced Hela or B-LCL cells as well as cells transduced with vectors encoding nonsense short-hairpin RNAs (shNS), lamin A/Cspecific (shLamin A/C), or β2-microglobulin-specific short-hairpin RNAs (shβ2m) were stimulated with IFN-γ for 48 h. HLA class I expression was measured by flow cytometry. The results are presented as geometric mean fluorescence intensity (MFI) values±SEM (n=5)

800

unstimulated

700

IFN-gamma stimulated

600 500 400 300 200 100 0

B-LCL

B-LCL shNS

B-LCL shLamin A/C

B-LCL shβ2m

433

HLA-A expression by up to 80% (Fig. 4c). B-LCLs expressing HLA class I-specific shRNAs showed a reduction of β2m expression by up to 70% (Fig. 4d). HLA-A expression remained unaffected in HeLa and BLCL cells stably expressing nonsense shRNAs or shRNAs directed against lamin A/C (Fig. 4a–d). Silencing of lamin A/C in HeLa and B-LCL cells transduced with the lamin A/ C-specific vector was determined by Western blot after 3 months (data not shown). Lack of inhibition of β-actin expression in HeLa and B-LCL cells transduced with HLA-A heavy-chain- or β2m-specific shRNAs (Fig. 4e and f) as well as lack of inhibition of HLA-B expression in HeLa cells transduced with HLA-A heavy chain specific shRNAs (data not shown) documented the specificity of the shRNA sequences used.

cells and cells expressing lamin A/C-specific or nonspecific shRNA was around 80–100% (mean score of 6–8) when polyclonal sera were used and 100% (mean score of 8) when incubated with monoclonal antibodies. However, for HLA-suppressed HeLa and B-LCL cells stably transduced with HLA-A gene-specific shRNAs, the cell lysis was only 11 to 40% (mean score of 2–4) when polyclonal sera were used and 40% (mean score of 4) when incubated with monoclonal antibodies. For HeLa and BLCL cells stably transduced with β2m-specific shRNAs (indirectly HLA-class-I-specific) the cell lysis was only up to 20% (mean score of up to 2) when polyclonal sera and the monoclonal antibody were used.

Microcytotoxicity assay

The capacity of resistance to up-regulation of HLA class I antigen under inflammatory conditions was tested by IFN-γ stimulation. As shown in Fig. 6, even under inflammatory conditions, HeLa (Fig. 6a) or B-LCL (Fig. 6b) cells transduced with shβ2m06 were incapable of up-regulating HLA class I expression, whereas non-transduced HeLa cells, B-LCL cells, or cells transduced with control vectors (HeLa/B-LCL-shNS and HeLa/B-LCL-shLamin A/C) showed increased levels of HLA class I antigen expression. These results demonstrate the capacity of RNAi-mediated

As HLA antigens of transplanted cells or tissues are targets for humoral rejection, the effect of HLA silencing was determined in a complement-dependent MCT. HeLa and B-LCL cells (both homozygous for HLA-A28), nontransduced or expressing different types of shRNAs, were either incubated with polyclonal or monoclonal HLA antisera specific for HLA-A2,28. The results are summarized in Fig. 5. Cell lysis in non-transduced HeLa and B-LCL

Fig. 7 NK cell cytotoxicity assay. Freshly isolated NK cells were incubated with a native K562 cells, b K562 cells expressing HLA-A11 (K562-A11), c K562-A11-shβ2m06 cells expressing shβ2m06, g LCL721.221, h native B-LCL cells, and i B-LCL cells expressing shβ2m06, whereby cytotoxicity assays with B-LCL cells were performed in an autologous setting. NK cells preexposed to K562-A11 cells for 48 h were incubated with d native K562 cells, e K562A11 cells, and f K562-A11shβ2m06 cells. Target cells were labeled with CFSE. Only the target cells are shown in the dot plot. Cellular cytotoxicity was evaluated by PI staining

Influence of IFN-γ on β2m silencing

434

cytotoxicity assays with B-LCLs as targets, only freshly isolated NK cells were used as effector cells. NK-cellmediated cell lysis of up to 95% was observed in HLAdeficient LCL721.221 cells (Fig. 7g), whereas almost no lysis was detectable when non-transduced B-LCLs were used as target cells (Fig. 7h). Only a minimal NK cell cytotoxicity of 10% was observed when HLA-class-Isuppressed B-LCLs were used (Fig. 7i). T cell response T cell proliferation and IFN-γ secretion assays were used to assess CD8+ T cell response against allogeneic cells. Primed CD8+ T cells strongly proliferated in the presence of either non-transduced B-LCLs or B-LCLs transduced for nonsense shRNA expression (three cycles of cell division). CD8+ T cells showed only one cycle of cell division when exposed to HLA-class-I-silenced B-LCL cells (Fig. 8a). In the presence of allogeneic B-LCL cells presenting normal levels of HLA class I expression, CD8+ T cells produced IFN-γ levels of up to 8596 pg/ ml compared to levels of 459,3 pg/ ml and 31 pg/ ml when incubated with HLA-class-I-suppressed and HLA-class-Ideficient cells, respectively (Fig. 8b).

Discussion

Fig. 8 T cells assays. a CD8+ T cell proliferation assay. Prestimulated effector T cells were labeled with CFSE and incubated with different target cells: native B-LCLs (green), B-LCLs expressing shLaminA/C (pink), B-LCLs expressing shβ2m06 (red), or HLA-class-I-deficient LCL721.221 cells (blue). b Interferon-γ secretion assay. Interferon-γ secretion was quantified by ELISA on culture supernatants

HLA class I silencing both under normal and inflammatory conditions. NK cell cytotoxicity NK cell cytotoxicity assays were performed by incubating NK cells with native K562 cells (NK cell lysis susceptible), K562-A11 cells, or K562-A11-shβ2m06 cells (90% reduction in HLA-A11 expression, data not shown). Freshly isolated NK cells lysed HLA-deficient K562 and HLAclass-I-silenced K562-A11-shβ2m06 cells by up to 96 and 92%, respectively (Fig. 7a,c), whereas 41% lysis was detectable in K562-A11 cells (Fig. 7b). When pre-exposed NK cells were used, they lysed up to 90% of HLAdeficient K562 and HLA-class-I-silenced K562-A11shβ2m06 cells (Fig. 7d,f). In contrast, expression of HLA-A*11 in K562-A11 almost completely inhibited cell lysis by pre-exposed NK cells (Fig 7e). In the NK cell

The emergence of cell-based regenerative medicine as potential therapy for substitution of malignant or injured tissues is intimately correlated with the necessity to inhibit the host immune response to the modified autologous or transdifferentiated allogeneic cells [27]. HLA silencing can prevent the immune system from recognizing immunogenic peptides in the genetically modified autologous transplant or optimize the matching of recipient and allogeneic donor cells [28]. The principle of the new approach is that not the recipient's immune system but the transplanted donor cells are modified to induce immunologic tolerance in the recipient. Thus, rejection of transplanted cells or tissue can be prevented without the risk of the hazardous side effects associated with a general impairment of the immune system. This strategy would allow for minimizing post-transplant long-term immunosuppressive therapy. Recently, it was shown that silencing HLA antigens by transfection of RNAi expression cassettes could inhibit Tcell-mediated immune recognition [29, 30]. This work presents a new strategy for long-term knockdown of HLA using shRNA-lentiviral delivery. HLA class I suppression was carried out in a class- and gene-specific way as well as on the HLA allele level. Moreover, we showed that suppression of HLA class I on the cell surface is even maintained under inflammatory conditions. Finally, it was demonstrated that HLA class I knockdown was effective in preventing antibody-mediated cell lysis and CD8+ T cell response and that the residual HLA expression in HLA-silenced cells was effective in inhibiting NK-cell-mediated lysis.

435

The development of alloantibody responses to transplanted tissue is one of the most relevant factors contributing to graft injury and rejection. Hyperacute rejection is the classical example of antibody-mediated rejection. Pre-transplant HLA compatibility testing is routinely done using a complement-dependent lymphocytotoxicity test whereby donor lymphocytes are stained with the recipient's serum and rabbit complement. The finding of lysed donor lymphocytes, representing a positive crossmatch, should be interpreted as a contraindication for transplantation (e.g., renal transplantation) or, at least, as a risk factor for graft failure (HSCT) [31]. In this study, we used such a microcytotoxicity test to assess the effect of HLA silencing on complement-mediated cell lysis. It was shown that HLA class I silencing effectively protects cells from antibodymediated lysis. Moreover, HeLa cells transduced with HLA-class-I-specific shRNAs obtained a negative crossmatch in the MCT with patient sera, suggesting that the cells had acquired resistance to antibody-mediated, complement-dependent cytotoxicity. CD8+ T-cell-mediated rejection presents another significant barrier to allogeneic transplantation. Proliferation and interferon-γ secretion assays indicated a markedly reduced CD8+ T cell response against HLA class I suppressed cells when compared to native cells. These results support the hypothesis that HLA silencing might have a relevant impact on decreasing the risk of graft failure. Interestingly, our in vitro data suggest that complete silencing of non-permissive HLA antigens may not be necessary to induce acceptance of otherwise rejected tissues. The clinical applicability of HLA-silencing also depends on the susceptibility of transplanted cells to NK-cellmediated cytotoxicity. According to the ‘missing self’ hypothesis, it is the function of NK cells to recognize and eliminate cells that fail to express certain self HLA class I molecules [32]. NK cell specificity is determined by a balance of signals generated by various stimulatory and inhibitory receptors [33]. Several inhibitory NK cell receptors have a specificity for HLA class I allotypes. Our NK cell assays strongly suggest that the residual HLA antigen expression on HLA-silenced cells is protective against autologous NK cell attack. However, in an artificial system using as targets K562 cells transduced to express a single HLA class I specificity, we observed a partial inhibition with freshly isolated NK cells and an almost complete inhibition when pre-exposed NK cells were used as effector cells. This loss of activity of pre-exposed NK cells was unexpected with regard to the polyclonal nature of NK cells. As NK cells showed lytic activity against native HLA-class-I-deficient K562 cells independently from their pre-treatment, the inhibition of the preexposed NK cells might be due to a modulation of the NK cell receptor repertoire during pre-exposure to HLA-A*11expressing K562-A11 target cells. This is supported by recent reports suggesting that expression of NK cell receptors can be influenced by certain cytokines and that chronic exposure to tumor cells alters NK cell receptor signaling, facilitating the evasion of tumor cells from NK cell lysis [34, 35]. Alternatively, the polyclonal pattern of

reactivity could have been reduced to a limited number of receptors in the artificial K562-A11 system during preexposing the NK cells to their targets. This might be due to a consumptive process leading to the elimination of NK cells reactive to other ligands than HLA-A11. Subsequently, cell lysis might have been reduced when NK cells were re-exposed to the HLA-A11-positive target cells. It would be interesting to further investigate whether or not modified cells silenced for particular HLA molecules or expressing recombinant HLA can be used for modulation of NK cell activity. In the present study, both the dimeric structure of the HLA class I molecule, comprising the HLA heavy chain plus β2m, and the polymorphic nature of the HLA heavy chain were used to develop strategies for class-, gene- and group-specific HLA silencing. Target sequences within the coding region of β2m were selected to design siRNAs to silence all HLA class I molecules on the cell surface (HLA class I-specific silencing), while conserved regions within the mRNA sequence of the HLA-A heavy chain served as targets for gene-specific HLA-A silencing. Sequences covering polymorphic transcript sites specific for a defined group of HLA-A alleles were used as targets for groupspecific HLA-A suppression. Due to the high degree of HLA heterozygosity, the application of a group-specific silencing strategy might even enable allele-specific HLA suppression in the majority of patients exhibiting heterozygosity at the respective HLA locus. The feasibility of silencing a single HLA specificity was demonstrated in our experiment where a group-specific siRNA (HLA-A10/28) specifically caused suppression of a single HLA-A allele (HLA-A28) in a lymphocyte cell line heterozygous for HLA-A24 and HLA-A28. In this case, the allele specificity of the siRNA was based on four polymorphic sites within the HLA-A target sequence that distinguish alleles belonging to the HLA-A10/28 group from those that are not. Other investigators have already shown that allele-specific silencing of protein expression based on the presence of one or more mismatches is possible [36, 37]. From the therapeutic viewpoint, allele-specific silencing might become an important tool in the treatment or prevention of dominant inherited diseases [38]. Allele- or group-specific silencing is also an attractive possibility in cellular therapeutics as the suppression of a specific non-permissive allele might contribute to better engraftment. An efficient delivery system is crucial for the development of RNAi-based therapy [39]. A variety of strategies to express interfering RNAs with the use of virus vectorbased cassettes have been explored, including retroviral and lentiviral vectors [40–43]. In this study, a lentiviral vector system was used due to its capability to transduce dividing and non-dividing cells and to mediate stable protein suppression by integrating the RNAi cassette into the genome. Persistent silencing of HLA-A or HLA class I molecules was achieved by transducing HeLa cells with lentiviral vectors coding for shRNAs that target HLA heavy chains or β2m, respectively. In addition, the fact that the expression of the non-targeted HLA-molecules and the housekeeping protein lamin A/C remained unaf-

436

fected over the whole observation period reflects the specificity of the expressed shRNAs. Transplants might be more susceptible to post-transplant virus and bacteria infections by downregulation of β2m, as inhibition of virus-derived peptide, microbial lipids, or lipopeptides presentation to Cluster of Differentiation 1 (CD1)-restricted T cells will be indirectly affected [44]. Experiments using inducible expression systems with regulatory promoters will have to show whether RNAimediated suppression of HLA can temporarily be reversed if therapeutically required. The selection of the most effective siRNA targets is crucial to optimize mRNA cleavage by RISC and to successfully silence the protein expression. Efficacy and specificity of siRNA sequences strongly depend on the position of the target sites [45]. Although several prediction algorithms exist, a number of siRNAs typically have to be tested before finding an effective one. However, testing the action of different synthetic siRNAs by cell transfection is laborious and time-consuming. Using the IVTT system, we have developed a simple and rapid screening method for selecting efficient siRNA duplexes. A good correlation was observed between the results in the IVTT and the silencing results obtained with eukaryotic cells constitutively expressing HLA. In the IVTT, the silencing effect is determined as the degree of translation inhibition. As the read-out of the results is feasible in the ELISA format, this approach allows for high-throughput screening of the best siRNAs at the protein level. In conclusion, our data strongly support the idea that HLA expression can be effectively silenced in a class-, gene-, and group-specific manner. The possibility to deliver the siRNAs by viral transduction, which offers the advantages of stable transcript reduction and organ selective applicability, provides the basis for exciting new approaches to cell therapy in the field of regenerative medicine.

References 1. Strom TB, Field LJ, Ruediger M (2002) Allogeneic stem cells, clinical transplantation and the origins of regenerative medicine. Curr Opin Immunol 14:601–605 2. Ferrari G, Cusella-De Angelis G, Coletta M, Paolucci E, Stornaiuolo A, Cossu G, Mavilio F (1998) Muscle regeneration by bone-marrow derived myogenic progenitors. Science 279:1528–1530 3. Jakson KA, Mi, T, Goodell MA (1999) Hematopoietic potential of stem cells isolated from murine skeletal muscle. Proc Natl Acad Sci USA 96:14482–14486 4. Oddorico JS, Kaufman DS, Thomson JS (2001) Multilineage differentiation from human embryonic stem cell lines. Stem Cells 19:193–2004 5. Blume KG (1993) A general overview of the status of bone marrow transplantation for hematologic diseases. Leukemia 7:1078–1079 6. Hansen JA, Petersdorf E, Martin PJ, Anasetti C (1997) Hematopoietic stem cell transplants from unrelated donors. Immunol Rev 157:141–151 7. Gallon LG, Leventhal JR, Kaufman DB (2002) Pretransplant evaluation of renal transplant candidates. Semin Nephrol 22:515–525

8. Hodgson DM, Behfar A, Zingman LV, Kane GC, Perez-Terzic C, Alekseev AE, Puceat M, Terzic A (2004) Stable benefit of embryonic stem cells therapy in myocardial infarction. Am J Physiol Heart Circ Physiol 287:471–479 9. Chae DW, Nosaka Y, Strom TB, Malinski W (1996) The distribution of IL-15 receptor α chains on human PBMCs and the effect of immuno-suppressive drugs on receptor expression. J Immunol 157:2813–2819 10. Marrack P, Bender J, Hildeman D, Jordan M, Mitchel T, Murakami I, Sakamoto A, Schaefer BC, Swanson B, Kappler J (2000) Homeostasis of alpha beta TCR+ T cells. Nat Immunol 1:107–111 11. Armitage JO (1994) Bone marrow transplantation. N Engl J Med 330:827–838 12. Thomas ED (1995) History, current results, and research in marrow transplantation. Perspect Biol Med 38:230–237 13. Madrigal JA, Scott I, Arguello R, Szydlo R, Little A-M, Goldman JM (1997) Factors influencing the outcome of bone marrow transplants using unrelated donors. Immunol Rev 157:153–166 14. Barret AJ, Rezvani K, Solomon S, Dickinson AM, Wang XN, Stark G, Cullup H, Jarvis M, Middelton PG, Chao N (2003) New developments in allotransplant immunology. Hematology 350–371 15. Rocha P, Plumb T, Crowley S, Coffman T (2003) Effector mechanisms in transplant rejection. Immunol Rev 196:51–64 16. Opelz G (1992) Strength of HLA-A, HLA-B, and HLA-DR mismatches in relation to short and long-term kidney graft survival. Transpl Int 5:621–624 17. McKenna RM, Takemoto SK, Terasaki PI (2000) Anti-HLA antibodies after solid organ transplantation. Transplantation 69:319–326 18. Vongwiwatana A, Tasanarong A, Hidalgo LG, Halloran PF (2003) The role of B cells and alloantibody in the host response to human organ allografts. Immunol Rev 196:197–218 19. Elbashir SM, Lendeckel W, Tuschl T (2001) RNA interference is mediated by 21 and 22 nucleotide RNAs. Genes Dev 15:188–200 20. McManus MT, Sharp PA (2002) Gene silencing in mammals by small interfering RNAs. Nat Rev 3:737–747 21. Hammond SM, Bernstein E, Beach D, Hannon GJ (2000) An RNA-directed nuclease mediate post-transcriptional gene silencing in Drosophila cells. Nature 404:293–296 22. Elbashir SM, Harborth J, Lendeckel W, Yalcin A, Weber K, Tuschl T (2001) Duplexes of 21-nucleotide RNAs mediate RNA interference in cultured mammalian cells. Nature 411:494–498 23. Brummelkamp TR, Bernards R, Agami R (2002) A system for stable expression of short interfering RNAs in mammalian cells. Science 296:550–553 24. Paddison PJ, Caudy AA, Bernstein E, Hannon GJ, Conklin DS (2002) Short hairpin RNAs (shRNAs) induce sequence-specific silencing in mammalian cells. Genes Dev 16:948–958 25. Terasaki PI, Bernoco D, Park MS, Ozturk G, Iwaki Y (1978) Microdoplet testing for HLA-A, B, C and D antigens. Am J Clin Pathol 69:103–120 26. Jedema I, van der Werff NM, Barge RM, Wilemze R, Falkenburg JH (2005) New CFSE-based assay to determine susceptibility to lysis by cytotoxic T cell of leukemic precursor cells within a heterogeneous target cell population. Blood 103:2677–2682 27. Passier R (2003) Potential of human embryonic stem cells in regenerative medicine. Horm Res 60(3):11–14 28. Ottinger HD, Rebmann V, Pfeiffer KA, Beelen DW, Kremens B, Runde V, Schaefer UW, Grosse-Wilde H (2002) Positive serum crossmatch as predictor for graft failure in HLAmismatched allogeneic blood stem cell transplantation. Transplantation 73:1280–1285 29. Gonzalez S, Castanotto D, Li H, Olivares S, Jensen MC, Forman SJ, Rossi JJ, Cooper LJN (2005) Amplification of RNAi-targeting HLA mRNAs. Mol Ther 11:811–818

437 30. Figueiredo C, Seltsam A, Blasczyk R (2004) HLA silencing by RNAi-mediated transcript targeting. Genes Immun 56:(Suppl) 31. Abdelgany A, Wood M, Beeson D (2003) Allele-specific silencing of a pathogenic mutant acetylcholine receptor subunit by RNA interference. Hum Mol Genet 12:2637–2644 32. Lyunggren HG, Karre K (1990) In search of the ‘missing self’: MHC molecules and NK cell recognition. Immunol Today 14:52–58 33. Brumbaugh KM, Binstadt BA, Leibson PJ (1998) Signal transduction during NK cell activation: balancing opposing forces. Curr Top Microbiol Immunol 230:103–122 34. Farag SS, Fehniger TA, Ruggeri L, Velardi A, Caligiuri MA (2002) Natural killer receptors: new biology and insights into graft-versus-leukemia effect. Blood 100:1937–1947 35. Draghi M, Yawata N, Gleimer M, Yawata M, Valiante VM, Parham P (2005) Single cell analysis of the human NKL cell response to missing self and its inhibition by class I. Blood 105:2028–2035 36. Gonzalez-Alegre P, Miller VM, Davidson BL, Paulson HL (2003) Toward therapy for DYT1 dystonia: allele specific silencing of mutant TorsinA. Ann Neurol 53:781–787 37. Ding H, Schwarz DS, Keene A, Affar B, Fenton L, Xia X, Shi Y, Zamore PD, Xu Z (2003) Selective silencing by RNAi of a dominant allele that causes amyotrophic lateral sclerosis. Aging Cell 2:209–217

38. Miller VM, Gouvion CM, Davidson BL, Paulson HL (2004) Targeting Alzheimer's disease gene with RNA interference: an efficient strategy for silencing mutant alleles. Nucleic Acids Res 32:661–668 39. Van Den Haute C, Eggermont K, Nuttin B, Debyser Z, Baekelandt V (2003) Lentiviral vector mediated delivery of short hairpin RNA result in persistent knockdown of gene expression in mouse brain. Hum Gene Ther 14:1799–1807 40. Devroe E, Silver PA (2002) Retrovirus-delivered siRNA. BMC Biotechnol 2:15–19 41. Abbas-Terki T, Blanco-Bose W, Délgon N, Pralong W, Aebisher P (2002) Lentiviral-mediated RNA interference. Hum Gene Ther 13:2197–2201 42. Scherr M, Battmer K, Blomer U, Schiedlmeier B, Ganser A, Grez M, Eder M (2002) Lentiviral gene transfer into peripheral blood-derived CD34+ NOD/SCID-repopulating cells. Blood 99:709–712 43. Qin X-F, An DS, Chen I.SY, Baltimore D (2003) Inhibiting HIV-1 infection in human T cells by lentiviral-mediated delivery of small interfering RNA against CCR5. Proc Natl Acad Sci USA 100:183–188 44. Beckman EM, Porcelli SA, Morita CT, Behar SM, Furlong ST, Brenner MB (1994) Recognition of lipid antigen by CD1restricted alpha beta+ T cells. Nature 372:691–694 45. Miller VM, Xia H, Marrs GL, Gouvion CM, Lee G, Davidson BL, Paulson HL (2003) Allele-specific silencing of dominant disease genes. Proc Natl Acad Sci USA 100:7195–7200