Gene Therapy (2001) 8, 362–368  2001 Nature Publishing Group All rights reserved 0969-7128/01 $15.00 www.nature.com/gt RESEARCH ARTICLE Receptor-mediated gene delivery to human peripheral blood mononuclear cells using anti-CD3 antibody coupled to polyethylenimine

MM O’Neill, CA Kennedy, RW Barton and RJ Tatake Boehringer Ingelheim Pharmaceuticals, Inc., 900 Ridgebury Road, Ridgefield, CT 06877, USA

Gene transfer to primary cells, especially to lymphoid cells, expression was detected until 96 h with a gradual diminution using a nonviral delivery system has been very challenging. in the signal after 48 h. Receptor-mediated gene delivery In the present studies, we have used a cationic polymer, was successfully used for freshly isolated, as well as pre- polyethylenimine (PEI) coupled to an anti-CD3 antibody for viously frozen lymphocyte samples. The transfections per- achieving receptor-mediated gene delivery to human periph- formed using ligands other than anti-CD3 were not as eral blood mononuclear cells (PBMC). Naive, unstimulated efficient as anti-CD3-PEI. These results suggest that in PBMC did not express transfected genes, whereas the addition to receptor-mediated endocytosis, signaling sub- transgenes were expressed efficiently in PHA activated sequent to engagement of the CD3 receptor with anti-CD3- PBMC. Transiently expressed gene products were detected PEI appears to be important for the efficacy of anti-CD3-PEI maximally at 24 and 48 h following transfection. Gene mediated gene delivery. Gene Therapy (2001) 8, 362–368.

Keywords: gene delivery; nonviral; receptor-mediated; lymphocytes

Introduction providing a co-stimulatory signal has increased the efficiency of retroviral transduction of lymphoid cells.9 Peripheral blood mononuclear cells (PBMC), consist of Pseudotyped retroviruses have also been used in order subpopulations of cells that are involved in specialized to increase the efficiency to transduce lymphoid cells.10–13 effector and regulatory functions in response to immuno- In these studies envelope protein of Moloney murine leu- logical stimuli. As a general rule, T lymphocytes are kemia virus has been replaced by the G glycoprotein of responsible for cellular immune response and humoral vesicular stomatitis virus to achieve vast host range,10,11 response is mediated by B lymphocytes. Due to their or with the glycoproteins of HIV for targeting specific, immunological specificity, genetically modified effector CD4-positive T cells.12,13 In comparison to retroviruses, 1 cells have been used for developing cell-based therapies. adenoviral vectors are used less frequently for transduc- Genetically modified lymphoid cells have been used to ing lymphoid cells,14,15 partly because lymphoid cells deliver therapeutic proteins in experimental rheumatoid inadequately express the CAR (receptor for adenovirus 2 3 arthritis, adenosine deaminase deficiency, and graft- and coxsackie virus) receptors that are essential for the 4,5 versus-host disease. Genetically modified immune cells attachment and entry of adenoviruses into cells.15 Besides 6 have also been shown to increase antitumor reactivity. retro- and adenoviral vectors, adeno-associated,16 and In addition to their potential for developing cell-based SV40-based17 viral vectors have been used in limited therapies, transiently transfected lymphoid cells are of studies to transduce lymphoid cells. Although the viral value to understand the role of certain genes that may be vectors continue to evolve to offer a better system, the differentially expressed in different lymphoid subpopula- major disadvantages are due to limitation of the length 7,8 tions in response to a stimulus. For this purpose, PBMC of DNA to be inserted, difficulties in consistent pro- provide an easily accessible cell population, which more duction of vectors and immunogenicity of viral proteins. closely represents a physiological system than established Nonviral gene delivery systems, on the other hand, cell lines. Unfortunately, PBMC have proven to be very appear to allow more flexibility with respect to the DNA difficult to transfect, using both viral as well as nonviral sequences to be inserted into cells but provide much gene delivery systems. lower efficiency of transfection. The attempts to facilitate Among viral vectors, retroviral vectors are commonly transfer of DNA into primary T or B lymphocytes using used for gene delivery to lymphoid cells because of their nonviral vectors have met with limited success.18,19 Sev- ability to transduce hematopoietic cells, although they eral rate-limiting steps, such as suboptimal attachment generally offer low transduction efficiency. In some cases, and internalization through the cell membrane, and/or inadequate release of vector–DNA complexes from the endosomes and transport of DNA to the nucleus, may be Correspondence: RJ Tatake partially responsible for the inefficiency of nonviral vec- Received 28 February 2000; accepted 30 November 2000 tors for gene delivery to lymphoid cells. To overcome Nonviral gene delivery to lymphocytes MM O’Neill et al 363 some of these hurdles, nonviral vectors have been modi- fied by covalent attachment of various ligands, which facilitate entry of DNA into cells via specific receptors.20–23 This approach was explored by Buschle et al24 to transfect primary lymphocytes using an anti-CD3 antibody conju- gated to poly-l-lysine to facilitate gene transfer via the CD3 receptor. Using anti-CD3-poly-l-lysine for gene transfer, they observed very efficient gene delivery to T lymphoblastoid cell lines, however, the efficiency of gene expression in prestimulated PBMC was low. Recently, Kircheis et al25 used a cationic polymer, poly- ethylenimine (PEI), to investigate combined effect of specificity of receptor-mediated gene delivery and high- transfection potential of PEI. PEI has been shown to be an efficient nonviral vector for gene delivery to a variety of cell types, both in vitro, as well as in vivo.26–32 The dis- Figure 1 Luciferase activity in human PBMC transfected with pGL3-con- tinctive properties of PEI, such as DNA-binding and con- trol DNA. Human PBMC were transfected with pGL3 control vector densation, along with its high buffering capacity is con- using electroporation (a) or anti-CD3-PEI (b). PBMC were prestimulated sidered to protect DNA from degradation.29 PEI acts as with PHA (1 ␮g/ml) for 48 h and supplemented with hrIL-2 (10 U/ml) a ‘proton sponge’ in the acidic environment of the lyso- for 24 h before transfection. Twenty-four hours after transfection, each transfected culture was split into two cultures. One culture was stimu- somes, and promote release of DNA from the lyso- + somes.33 Furthermore, PEI is also shown to promote lated with PMA (20 ng/ml) for 20 h (black bars, ), and the other was left unstimulated (gray bars, −). Cells were harvested and luciferase transgene localization to the nucleus in mammalian activity was determined as described in Materials and methods. The values 34,35 cells. When these properties of PEI were combined for relative light units (RLU) for the luciferase expression are given above with the specific mechanism of receptor-mediated gene each histogram since error bars are not visible due to relatively small delivery, ligand-conjugated PEI resulted in higher trans- standard deviation and the logarithmic scale. fection efficiency in various tumor cell lines.25 In the stud- ies reported here, we have used PEI coupled to various ligands specific for the receptors present on PBMC to was further augmented approximately 60% when trans- assess gene delivery to human PBMC. It was observed fected cells were treated with PMA (from RLU 45 406 ± that PBMC were efficiently transfected by anti-CD3-PEI, 851 to 72 361 ± 378). whereas anti-ICAM1-PEI and anti-MHC class I-PEI did Similar to various nonviral gene delivery vectors, non- not transfer DNA effectively. The results suggest that in targeted PEI was not effective in transfecting PBMC (data addition to ligand–receptor interaction and endocytosis not shown). Based on the specificity of anti-CD3-PEI to of the DNA/vector complexes, efficiency of expression target CD3-positive cells in the studies by Kirches et al,25 of a transgene is affected by signaling events following it is highly likely that the cell population expressing the receptor engagement. luciferase may be T lymphocytes. However, since PBMC were not fractionated to enrich T lymphocytes, the trans- fected cell population has been referred to as PBMC. The Results efficiency of transfection, in terms of the transgene expression, was variable in different PBMC samples. The Transfection of human PBMC with pGL3 plasmid DNA average luciferase expression without PMA stimulation Expression of the luciferase gene in human PBMC was of PBMC from nine different donors was 14 848 RLU/1 assessed following transfection of PBMC with pGL3 vec- × 105 cells/s (range 1236–45 406 RLU/1 × 105 cells/s). tor using various nonviral gene delivery carriers. The Stimulation of the transfected cells with PMA did not attempts to transfect PBMC using Superfect, Lipofectam- affect the variability in gene expression. Despite the ine or DEAE-dextran were unsuccessful (data not variability, luciferase activity was consistently seen in all shown). However, consistent gene expression was achi- PBMC samples transfected using anti-CD3-PEI. eved when PBMC were transfected either by electropor- ation or receptor-mediated gene delivery using anti-CD3- Effect of prestimulation of PBMC on the efficiency of PEI conjugates. The results from a representative experi- transfections using anti-CD3-PEI ment comparing relative efficiencies of electroporation In order to determine optimal conditions for gene deliv- and anti-CD3-PEI-mediated gene delivery are summar- ery to PBMC using anti-CD3-PEI, experiments were per- ized in Figure 1. A moderate level of gene expression was formed using unstimulated PBMC and PBMC stimulated observed when cells were transfected by electroporation. with PHA. As shown in Figure 2a, naive, unstimulated The transgene expression was augmented when trans- PBMC did not express the transgene, whereas, transgene fected cells were stimulated with a transcriptional acti- expression was achieved by prestimulation of PBMC vator, PMA (20 ng/ml). In contrast, significantly higher with PHA. Optimal transfection efficiency was achieved luciferase activity was seen in the cells transfected using when transfections were carried out 48–72 h after PHA anti-CD3-PEI in the absence of stimulation with PMA. stimulation. Addition of exogenous IL-2 to PHA blasts Since CD3-receptor crosslinking by anti-CD3 antibody before transfection resulted in increased gene expression, induces the T cell activation signaling cascade, we optimally, when added 24 h before transfection (Figure wanted to determine if cells transfected using anti-CD3- 2b). Interestingly, addition of IL-2 to the cells 72 h before PEI were capable of responding to another activation sig- gene delivery (IL-2 added along with PHA on day 0) nal, eg PMA. As shown in Figure 1, luciferase activity resulted in gene expression even lower than the cells that

Gene Therapy Nonviral gene delivery to lymphocytes MM O’Neill et al 364

Figure 2 Effect of prestimulation of PBMC with PHA and human rIL-2 on the efficiency of transfection. (a) PBMC were transfected with pGL3- control DNA (10 ␮g) using anti-CD3-PEI at various times after PHA stimulation. PHA blasts were either untreated (open circles) or treated (filled circles) with human rIL-2 (10 units/ml) 24 h before transfection. Due to relatively low standard deviation and the logarithmic scale, error bars are not visible. (b) PHA blasts were supplemented with human rIL- Figure 3 Kinetics of gene expression in PBMC transfected using anti- 2 (10 units/ml) for different periods before transfection. Transfections were ␮ carried out 72 h after PHA stimulation. The values for RLU for the lucifer- CD3-PEI. Optimally prestimulated PBMC were transfected with 10 g ase expression are given above each histogram since error bars are not pEGFP (a) or pGL3-control vector (b). The GFP expression was quantit- visible due to relatively small standard deviation and the logarithmic scale. ated daily for 4 consecutive days using FACS analyzer (Becton In both experiments the transfected cells were not further activated with Dickinson). Dotted histograms represent mock-transfected cells, whereas PMA. The luciferase gene expression was assessed 24 h after transfection. solid histograms represent cells transfected with pEGFP. The data from histograms are summarized in Table 1. The values for RLU for the lucifer- ase expression are given above each histogram since error bars are not did not receive IL-2 before transfection. These data sug- visible due to relatively small standard deviation and the logarithmic scale. gest that the efficiency of gene expression is influenced by different culture conditions, including presence of either endogenously produced (following stimulation with Dickinson, Mountain View, CA, USA) (Figure 3a and PHA) or exogenously supplemented growth factors such Table 1). The transfected cells were monitored for lucifer- as IL-2. ase activity after 24, 48, 72 and 96 h following transfection (Figure 3b). Expression of both luciferase and GFP was Kinetics of transgene expression following anti-CD3-PEI- detected up to 96 h after transfection, although gene mediated gene delivery expression began to gradually decrease after 48 h. Optimally prestimulated PBMC were transfected separ- Although the percentage of GFP positive cells is likely to ately with pGL3 or pEGFP vectors to assess duration of be variable in different PBMC samples, as high as 56–43% gene expression. The percentage of cells expressing GFP of cells expressing GFP were detected up to 72 h with was determined by flow cytometry using standard pro- significant fluorescence intensity. The decrease in the per- cedures and analysis with a FACScan cytometer (Becton centage and the intensity of GFP-positive cells by 96 h

Gene Therapy Nonviral gene delivery to lymphocytes MM O’Neill et al 365 Table 1 Expression of GFP in anti-CD3-PEI transfected PBMC

Time (h) Mock-transfected GFP-transfected after transfection % positive MCF % positive MCF

24 13.32 79.99 56.05 368.48 48 16.8 71.14 73.27 273.43 72 17.06 64.21 62.27 243.23 96 14.06 63.24 57.35 161.88

Optimally prestimulated PBMC were transfected with pEGFP. The GFP expression was quantitated daily for 4 consecutive days using FACS analyzer (Becton Dickinson). The data from histograms shown in Figure 3a are summarized in Table 1. Figure 4 Effect of costimulation with immobilized anti-CD3 antibody on gene expression by PBMC. Optimally prestimulated PBMC were trans- ␮ could be due to overproliferation of GFP nonexpressing fected with pGL3 control vector (5 g DNA). For anti-CD3 antibody immobilization, 24-well plates were coated with 1.0 ml of IOT3b (anti- cells or GFP-expressing cells that did not transfer the CD3 antibody, clone UCHT1, B-D Coulter, Fullerton, CA, USA) at transgene into the daughter cells after cell division. 1 ␮g/ml for 4 h at 37°C. The wells were washed five times with 2.0 ml of PBS before addition of cells. Transfected cells were either cultured in Gene transfer to PBMC using ICAM1 and MHC class I the wells coated with anti-CD3 antibody (black bars, +) or uncoated wells surface receptors (gray bars, −). The luciferase activity was detected at 24 h after transfec- Since efficient gene delivery could be obtained using tion without an additional stimulation with PMA. A and B represent two separate transfections performed using optimal nitrogen (PEI) to phos- receptor-mediated endocytosis via the CD3 surface recep- phate (DNA) ratios that were predetermined for each transfecting reagent. tor, we sought to address if gene delivery could be achi- Error bars are not visible for some data points due to relatively small eved via other receptors expressed on PHA-activated standard deviation. PBMC. We used PEI conjugated to anti-ICAM1 or anti- HLA class I (W6/32) monoclonal antibodies, to target ICAM1 or HLA class I receptors on the cell surface, lines has been achieved by targeting the CD3, CD4 or the respectively. Neither of these antibodies coupled PEI con- transferrin receptors using anti-CD3-poly-l-lysine,24 anti- jugates were as efficient as anti-CD3-PEI (data not CD3-PEI,25 transferin-polycation,36 transferin-PEI25 or shown). Since only anti-CD3-PEI was effective in trans- anti-CD4-poly-l-lysine.37 Among these vectors, anti-CD3- fecting PBMC, we asked if receptor–ligand interaction poly-l-lysine was used for gene transfer to previously and internalization are sufficient for gene expression, and activated PBMC, which resulted in transgene expression is activation following engagement of the CD3 receptor in about 5% cells.24 In light of relatively low efficiency of with anti-CD3 antibody important for gene expression. transfection of PBMC shown in these studies, we have To address this, PBMC transfected with anti-ICAM1-PEI achieved a significant improvement in transfection of or anti-HLA class I-PEI were cultured in the presence of PBMC using anti-CD3-PEI. The increase in transfection immobilized anti-CD3 antibody. Expression of the trans- efficiency using anti-CD3-PEI in comparison with anti- gene was significantly augmented when the cells trans- CD3-poly-l-lysine could be primarily attributed to dis- fected using anti-ICAM1-PEI were cultured in the pres- tinctive properties of PEI, such as its buffering capacity ence of immobilized anti-CD3 antibody (Figure 4). The for the condensed DNA in the acidic environment of transgene expression was as good as or better than that endosomes and promoting release of DNA from the obtained in transfection of the same PBMC using anti- endosomes.26 In these studies by Kircheis et al,25 30–1000- CD3-PEI. The gene expression in anti-CD3-PEI trans- fold enhancement in transgene expression was observed fected cells was further augmented when these cells were in three out of five tumor cell lines when these cells were cultured in the presence of anti-CD3-PEI. In comparison, transfected with PEI conjugated to the ligands against anti-HLA class I-PEI and nontargeted PEI-transfected specific surface receptors on various tumor cell lines. In PBMC did not show appreciable increase in luciferase the present studies, in addition to showing efficacy of expression in the presence of immobilized anti-CD3 anti- anti-CD3-PEI as a gene delivery vector to human PBMC, body. The gene expression in these transfections, even in we have shown that previously frozen lymphocytes can the presence of immobilized anti-CD3 antibody, was 3.5– also be transfected using anti-CD3-PEI and assessed 12 times less than that achieved using anti-CD3-PEI in transgene (GFP) expression in individual cells by FACS the absence of immobilized anti-CD3. analysis. In the present studies, we observed disparity in trans- Discussion fection efficiency of PBMC using anti-CD3-PEI, anti- ICAM1-PEI and anti-HLA class I-PEI. These observations In the present studies, we have shown that efficient gene suggest that for efficient transfection of primary cells, like delivery to PHA-activated human PBMC was achieved PBMC, receptor binding and internalization may be via the CD3 receptor using anti-CD3-PEI as a gene deliv- receptor-specific. These results also suggest that higher ery vector, whereas gene delivery using anti-ICAM1-PEI efficiency of anti-CD3-PEI-mediated transfection of and anti-HLA class I-PEI was not efficient. Receptor- PBMC, as compared with anti-ICAM1-PEI or anti-MHC mediated gene delivery has been shown to be effective class I-PEI, could be partly attributed to activation of for a variety of cell types, including lymphoid cells.20–23 transfected cells following CD3/anti-CD3 interaction. Increased transfection efficiency for lymphoblastoid cell This is supported by the observations that efficiency of

Gene Therapy Nonviral gene delivery to lymphocytes MM O’Neill et al 366 anti-ICAM-1-PEI-, as well as anti-CD3-PEI-mediated TNF␣ in comparison with no production of TNF␣ (below transfection of PBMC was augmented in the presence of detectable level) by nontransfected cells. Thus, both immobilized anti-CD3 (Figure 4). It remains to be determ- TNF␣ and FasL could be responsible for some apoptotic ined if activation of transfected cells following CD3/anti- events immediately following anti-CD3-PEI-mediated CD3 interaction promotes internalization of the bound transfection of PBMC. However, beyond 24 h after trans- ligand/PEI/DNA complex or augments expression of a fection, fewer apoptotic cells were seen in the transfected transgene in the internalized complex, or it is a combined cultures. The overall gene expression in PBMC using effect of both. Increased membrane fluidity due to anti-CD3-PEI, thus appears to be a balance between the CD3/anti-CD3 crosslinking could promote better inter- beneficial effects for promoting gene expression and nalization of a DNA/vector complex bound to another unfavorable effects, such as induction of apoptosis, that receptor. Moreover, transcriptional activation following occur due to triggering the CD3 receptor by anti-CD3 anti-CD3 crosslinking could potentially enhance respon- antibody conjugated to PEI. siveness of the internalized reporter gene. We observed Thus considering the risks and benefits associated with marginal increase in the transfection efficiency using non- anti-CD3-PEI-mediated gene delivery, it appears that targeted PEI in the presence of immobilized anti-CD3 development of such gene delivery protocols using non- antibody. A similar effect of immobilized anti-CD3 anti- viral vectors, where high percentage of primary cells can body along with anti-CD28 antibody has been reported be transfected in a specific manner, is of value. This when primary T lymphocytes were transduced with method of transfecting human primary lymphocytes can retroviral vectors.9 Therefore, it is possible that costimu- be used as another tool for analyzing gene expression lation with immobilized anti-CD3 antibody can enhance and dissecting signaling pathways with an ability to gene delivery by any vector and may not be restricted to introduce exogenous genes. This approach can be used receptor-mediated gene delivery. However, the differ- for assessing effects of various pharmacological agents or ences in the efficiency to transfect PBMC by anti-ICAM1- physiological stimuli on expression of certain genes or PEI, anti-HLA class I-PEI and nontargeted PEI in the the promoters of activation-inducible genes, following presence of immobilized anti-CD3 antibody, suggest that their introduction in a primary lymphoid cell population. the unique properties of different receptor–ligand inter- Most of the experiments reported in these studies were actions can also affect the efficiency of a transgene performed using cryopreserved PBMC samples. Thus an expression. added advantage of using anti-CD3-PEI for gene delivery Thus anti-CD3-PEI conjugates appear to offer multiple is the ability to use either freshly isolated or cryopre- advantages derived from both PEI, as well as anti-CD3 served lymphocytes, that can be efficiently transfected antibody for efficient gene delivery to primary CD3-posi- within 48–72 h of thawing. Finally, as shown by genetic tive lymphoid cells. However, there are other conse- modification of T cells using viral vectors,1–5 a desired quences of the CD3/anti-CD3 antibody interaction, lymphocyte subpopulation can be genetically engineered which need to be considered. For example, activation- using targeted nonviral gene delivery methods to carry induced expression of a transgene in PBMC could poten- therapeutic proteins to the target sites in a specific tially be affected if PBMC were already activated during manner. gene transfer using anti-CD3-PEI. The results shown in Figure 1 indicate that expression of a transfected gene whose expression is driven by a constitutive promoter Materials and methods can be further augmented by another stimulus, such as PMA, even though PBMC were apparently activated by Isolation and culturing of mononuclear cells from human PEI-conjugated anti-CD3 antibody. In addition to these peripheral blood lymphocytes studies, we have successfully used anti-CD3-PEI to trans- Heparinized venous blood was obtained from healthy fect PBMC, with or without additional activation and volunteers. Peripheral blood mononuclear cells (PBMC) have assessed functional activity of a transgene con- were isolated by Ficoll–Hypaque (density = 1.077 g/ml, sisting of an activation inducible promoter (unpublished Amersham Pharmacia Biotech, Piscataway, NJ, USA) observations). Thus activation via the CD3 receptor density gradient centrifugation at 400 g for 20 min at during transfection does not seem to affect subsequent room temperature. The cells at the interface were washed activation by a different stimulus. 2× with Dulbecco’s phosphate buffered saline (DPBS) Another outcome of the CD3/anti-CD3 interaction (Gibco BRL), and resuspended at 1 × 106 cells/ml in com- during anti-CD3-PEI-mediated gene delivery could lead plete medium (RPMI-1640 + 10% FBS, 25 mm Hepes, to apoptosis of transfected cells. We have often observed 50 ␮g/ml gentamycin, 1× non-essential amino acids, 1× − some apoptotic cells in transfected cultures within 24 h, sodium pyruvate, 5 × 10 5 m 2 ME, and 1× glutamine (Life which could partly be due to activation induced cell Technologies Gibco/BRL, Rockville, MD, USA)). PBMC death (AICD) as a result of Fas/Fas ligand (FasL) interac- were stimulated with phytohemaglutinin (PHA-P) at tion in anti-CD3-PEI transfected PBMC. In addition to 1 ␮g/ml (Sigma, St Louis, MO, USA) and were split 1:2 Fas/FasL-mediated AICD, TNF␣ produced by anti-CD3 as and when needed. In some cultures, 10 units/ml of activated T cells may also be responsible for the observed human recombinant interleukin-2 (rIL-2) (Sigma) was apoptosis. Of importance is the observation that in cyto- added 24 h before transfection. PHA-stimulated PBMC kine-induced killer cells (CIK), nonviral gene delivery are referred to as PHA blasts. vectors, which were not coupled to anti-CD3 antibody, induced apoptosis of the transfected cells by production Transfection of PBMC by electroporation of TNF␣.38,39 In the present investigations, we have PHA blasts were resuspended at 2 × 106 cells in 20 ␮l assessed production of TNF␣ in a set of transfections. complete medium containing 10 units of human rIL-2. Anti-CD3-PEI-transfected cells produced 50.83 pg/ml of The cells were incubated with 5 ␮g of pGL3-control plas-

Gene Therapy Nonviral gene delivery to lymphocytes MM O’Neill et al 367 mid DNA for 20 min at room temperature. The cell/DNA Acknowledgments mixture was placed between the electrodes of a microchamber (Life Technologies) and electroporated on We would like to thank Carol Stearns and Kyung Yu for ice at 60 ␮F, low ohms and 135 volts (900 volts/cm) using the FACS analysis, Ann Graham for estimating TNF␣ in BRL CellPorator. The cells were incubated for an transfected cultures and Raju Patel for technical assist- additional 15 min at room temperature, suspended in ance in transfections. We thank Dr Ernst Wagner, Marga- 2.5 ml of complete medium and cultured at 37°C in a six- retha Kursa and Robert Holzhauser (Boehringer well plate (Corning Costar, Acton, MA, USA). The cells Ingelheim, Vienna) for supplying the transfecting were harvested for assessment of the transgene reagents anti-CD3-PEI, anti-HLA class I (W6/32)-PEI and expression at 24, 48 or 72 h after transfection. anti-ICAM1-PEI and Dr Richard Schneiderman for criti- cally reading the manuscript. Transfection of PBMC using anti-CD3-PEI PHA blasts used for transfection were either sup- plemented with 10 units/ml of human rIL-2 before trans- References fection or were left without exogenous supplementation of human rIL-2. DNA/anti-CD3-PEI complexes were pre- 1 Hege KM, Roberts MR. T-cell gene therapy. Curr Opin Biotechnol pared as described earlier.25 Briefly, 10 ␮g DNA (pGL3- 1996; 7: 629–634. control DNA or pEGFP DNA) was diluted in 250 ␮l 2 Chernajovsky Y et al. Pathogenic lymphoid cells engineered to Hepes buffered saline (HBS, 150 mm NaCl, 20 mm Hepes express TGF beta 1 ameliorate disease in a collagen-induced pH 7.3). Six ␮g anti-CD3-PEI (refers to the concentration arthritis model. Gene Therapy 1997; 4: 553–559. of PEI in the conjugates) diluted in 250 ␮l HBS was added 3 Mullen CA et al. Molecular analysis of T lymphocyte-directed to previously diluted DNA and mixed vigorously. The gene therapy for adenosine deaminase deficiency: long-term expression in vivo of genes introduced with a retroviral vector. DNA/anti-CD3-PEI complexes were allowed to form at Hum Gene Ther 1996; 7: 1123–1129. room temperature for 25–30 min. The complexes were 4 Bonini C et al. HSV-TK gene transfer into donor lymphocytes × 5 then gently mixed with the cells (5 10 cells in 1.5 ml for control of allogeneic graft-versus-leukemia. Science 1997; 276: complete medium in a 24-well plate). After a 4 h incu- 1719–1724. bation at 37°C, the medium containing the DNA/anti- 5 Gallot G et al. Human HLA-specific T-cell clones with stable CD3-PEI mixture was replaced with 2.0 ml of complete expression of a suicide gene: a possible tool to drive and control medium containing human rIL-2 (10 units/ml). The cells a graft-versus-host–graft-versus-leukemia reaction? Blood 1996; were transferred to a humidified incubator at 37°C con- 88: 1098–1103. taining 5% CO for 24, 48, 72 or 96 h before assessment 6 Clay TM et al. Efficient transfer of a tumor antigen-reactive TCR 2 to human peripheral blood lymphocytes confers anti-tumor of the transgene expression. reactivity. J Immunol 1999; 163: 507–513. 7 Novak TJ, Yoshimura FK, Rothenberg EV. In vitro transfection Luciferase assay of fresh thymocytes and T cells shows subset-specific expression Luciferase activity was analyzed using the luciferase of viral promoters. Mol Cell Biol 1992; 12: 1515–1527. assay system (Promega, Madison, WI, USA). Briefly, the 8 Hughes CC, Pober JS. Transcriptional regulation of the interleu- cells were pelleted at 14 000 r.p.m. in a microcentrifuge kin-2 gene in normal human peripheral blood T cells. Conver- for 5 min. The cells were lysed for 10 min in 100 ␮lof1× gence of costimulatory signals and differences from transformed lysis buffer and centrifuged for 5 min at 14 000 r.p.m. to T cells. J Biol Chem 1996; 271: 5369–5377. remove nuclei. Luciferase activity in the supernatant was 9 Quinn ER, Lum LG, Trevor KT. T cell activation modulates quantitated using duplicate 20 ␮l samples on a Wallac retrovirus-mediated gene expression. Hum Gene Ther 1998; 9: 1457–1467. Microbeta luminometer (model 1450–024), in which the 10 Sharma S, Cantwell M, Kipps TJ, Friedmann T. Efficient infec- substrate (Luciferase assay system kit, Promega) is tion of a human T-cell line and of human primary peripheral injected into each well before counting. blood leukocytes with a pseudotyped retrovirus vector. Proc Natl Acad Sci USA 1996; 93: 11842–11847. Reagents 11 Gallardo HF, Tan C, Ory D, Sadelain M. Recombinant retro- pGL3-control plasmid DNA (referred to as pGL3) was viruses pseudotyped with the vesicular stomatitis virus G glyco- purchased from Promega and pEGFP plasmid DNA was protein mediate both stable gene transfer and pseudotransduc- purchased from Clontech, Palo Alto, CA, USA. Plasmids tion in human peripheral blood lymphocytes. Blood 1997; 90: were prepared in the laboratory using standard molecu- 952–957. 12 Schnierle BS et al. Pseudotyping of murine leukemia virus with lar biology techniques. Transfecting reagents anti-CD3- the envelope glycoproteins of HIV generates a retroviral vector PEI, anti-HLA class I (W6/32)-PEI and anti-ICAM1-PEI with specificity of infection for CD4-expressing cells. Proc Natl were supplied by Dr Ernst Wagner (Boehringer Acad Sci USA 1997; 94: 8640–8645. Ingelheim, Vienna, Austria). Anti-CD3-PEI conjugates 13 Lodge R et al. MuLV-based vectors pseudotyped with truncated were prepared using Ortho-mune-OKT3 antibody (Ortho HIV glycoproteins mediate specific gene transfer in CD4+ per- Diagnostic Systems, Raritan, NJ, USA).25 Anti-HLA class ipheral blood lymphocytes. Gene Therapy 1998; 5: 655–664. I (W6/32) antibody (hybridoma was purchased from 14 Abe J et al. In vivo antitumor effect of cytotoxic T lymphocytes ATCC) and anti-ICAM1 (BIRR1)40 were generated and engineered to produce interferon-gamma by adenovirus- purified at Boehringer Ingelheim, R&D, Ridgefield, CT, mediated genetic transduction. Biochem Biophy Res Comm 1996; USA. Transfections using Superfect (Qiagen, Valencia, 218: 164–170. 15 Leon RP et al. Adenoviral-mediated gene transfer in lympho- CA, USA), Lipofectamine (LifeTechnologies Gibco/BRL) cytes. Proc Natl Acad Sci USA 1998; 95: 13159–13164. and DEAE-Dextran (Promega) were performed according 16 Chen JD et al. Intra- and extracellular immunization against to the suppliers’ protocols. Phytohemagglutinin-P (PHA- HIV-1 infection with lymphocytes transduced with an AAV vec- P), human rIL-2 and phorbol 12-myristate 13-acetate tor expressing a human anti-gp120 antibody. Hum Gene Ther (PMA) were purchased from Sigma. 1996; 7: 1515–1525.

Gene Therapy Nonviral gene delivery to lymphocytes MM O’Neill et al 368 17 Strayer DS, Kondo R, Milano J, Duan LX. Use of SV40-based transfer into cells in culture and in vivo: polyethylenimine. Proc vectors to transduce foreign genes to normal human peripheral Natl Acad Sci USA 1995; 92: 7297–7301. blood mononuclear cells. Gene Therapy 1997; 4: 219–225. 30 Coll JL et al. In vivo delivery to tumors of DNA complexed with 18 McMahon SB, Norvell A, Levine KJ, Monroe JG. Transient trans- linear polyethylenimine. Hum Gene Ther 1999; 10: 1659–1666. fection of murine B lymphocyte blasts as a method for examin- 31 Goula D et al. Polyethylenimine-based intravenous delivery of ing gene regulation in primary B cells. J Immunol Meth 1995; 179: transgenes to mouse lung. Gene Therapy 1998; 5: 1291–1295. 251–259. 32 Ringenbach L et al. Polyethylenimine-mediated transfection of 19 Cron RQ, Schubert LA, Lewis DB, Hughes CC. Consistent tran- human monocytes with the IFN-gamma gene: an approach for sient transfection of DNA into non-transformed human and cancer adoptive immunotherapy. Gene Therapy 1998; 5: 1508– murine T-lymphocytes. J Immunol Meth 1997; 205: 145–150. 1516. 20 Cotten M, Wagner E, Birnstiel ML. Receptor-mediated transport 33 Boussif O, Zanta MA, Behr JP. Optimized galenics improve in of DNA into eukaryotic cells. Meth Enzymol 1993; 217: 618–644. vitro gene transfer with cationic molecules up to 1000-fold. Gene 21 Batra RK et al. Receptor-mediated gene delivery employing lec- Therapy 1996; 3: 1074–1080. tin-binding specificity. Gene Therapy 1994; 1: 255–260. 34 Pollard H et al. Polyethylenimine but not cationic lipids pro- motes transgene delivery to the nucleus in mammalian cells. J 22 Guy J, Drabek D, Antoniou M. Delivery of DNA into mam- Biol Chem 1998; 273: 7507–7511. malian cells by receptor-mediated endocytosis and gene ther- 35 Godbey WT, Wu KK, Mikos AG. Tracking the intracellular path apy. Mol Biotechnol 1995; 3: 237–248. of poly(ethylenimine)/DNA complexes for gene delivery. Proc 23 Cristiano RJ. Targeted, nonviral gene delivery for cancer gene Natl Acad Sci USA 1999; 96: 5177–5181. therapy. Front Bioscience 1998; 3: D1161-D1170. 36 Wagner E, Cotten M, Foisner R, Birnstiel ML. Transferrin-polyc- 24 Buschle M et al. Receptor-mediated gene transfer into human T ation–DNA complexes: the effect of polycations on the structure lymphocytes via binding of DNA/CD3 antibody particles to the of the complex and DNA delivery to cells. Proc Natl Acad Sci CD3 T cell receptor complex. Hum Gene Ther 1995; 6: 753–761. USA 1991; 88: 4255–4259. 25 Kircheis R et al. Coupling of cell-binding ligands to polyethylen- 37 Puls RL, Minchin RF. Gene transfer and expression of a nonviral imine for targeted gene delivery. Gene Therapy 1997; 4: 409–418. polycation-based vector in CD4+ cells. Gene Therapy 1999; 6: 26 Abdallah B et al. A powerful nonviral vector for in vivo gene 1774–1778. transfer into the adult mammalian brain: polyethylenimine. 38 Ebert O et al. Lymphocyte apoptosis: induction by gene transfer Hum Gene Ther 1996; 7: 1947–1954. techniques. Gene Therapy 1997; 4: 296–302. 27 Baker A et al. Polyethylenimine (PEI) is a simple, inexpensive 39 Ebert O et al. TNF-alpha secretion and apoptosis of lymphocytes and effective reagent for condensing and linking plasmid DNA mediated by gene transfer. Cytokines Cell Mol Ther 1999; 5: to adenovirus for gene delivery. Gene Therapy 1997; 4: 773–782. 165–173. 28 Boletta A et al. Nonviral gene delivery to the rat kidney with 40 Smith CW et al. Recognition of an endothelial determinant for polyethylenimine. Hum Gene Ther 1997; 8: 1243–1251. CD18-dependent human neutrophil adherence and transendo- 29 Boussif O et al. A versatile vector for gene and oligonucleotide thelial migration. J Clin Invest 1988; 82: 1746–1756.

Gene Therapy